Advances in Industrial and Labor Relations, Volume 13

Citation preview

CONTENTS LIST OF CONTRIBUTORS

vii

INTRODUCTION David Lewin and Bruce E. Kaufman

ix

EUROPEAN RESPONSE TO HIGH RATES OF UNEMPLOYMENT: MONETARY POLICY OR DEREGULATION OF PART-TIME WORK? Stephen M. Hills and Teresa Schoellner

1

THE EFFECTS OF USING ALTERNATIVE TYPES OF STAFFING ARRANGEMENTS Cynthia L. Gramm and John F. Schnell

27

ADOPTION AND USE OF DISPUTE RESOLUTION PROCEDURES IN THE NONUNION WORKPLACE Alexander J. S. Colvin

69

THE ORGANISATIONAL OMBUDS: IMPLICATIONS FOR VOICE, CONFLICT RESOLUTION AND FAIRNESS AT WORK Sue Fernie and David Metcalf

97

BUILDING AND SUSTAINING LABOR-MANAGEMENT PARTNERSHIPS: RECENT EXPERIENCES IN THE U.S. Susan C. Eaton, Saul A. Rubinstein and Robert B. McKersie

137

UNION AND EMPLOYER TACTICS IN ONTARIO ORGANISING CAMPAIGNS Felice F. Martinello and Charlotte Yates

157

v

vi

LABOR’S “LAST STAND” IN NATIONAL POLITICS? THE CAMPAIGN FOR STRIKER REPLACEMENT LEGISLATION, 1990–1994 John Logan

191

THE TRANSFER OF IDEAS IN INDUSTRIAL RELATIONS: DUNLOP AND OXFORD IN THE DEVELOPMENT OF AUSTRALIAN INDUSTRIAL RELATIONS THOUGHT, 1960–1985 Diana Kelly

245

LIST OF CONTRIBUTORS Alexander J. S. Colvin

Department of Labor Studies and Industrial Relations, Pennsylvania State University, USA

Susan C. Eaton (Deceased)

Kennedy School of Government, Harvard University, USA

Sue Fernie

London School of Economics and Political Science, UK

Cynthia L. Gramm

College of Administrative Science, University of Alabama – Huntsville, USA

Stephen M. Hills

Fisher College of Business, Ohio State University, USA

Bruce E. Kaufman

Department of Economics and Beebe Institute, Georgia State University, USA

Diana Kelly

University of Wollongang, Australia

David Lewin

UCLA Anderson School of Management, University of California at Los Angeles, USA

John Logan

London School of Economics and Political Science, UK

Felice Martinello

Department of Economics, Brock University, Canada

Robert B. McKersie

Sloan School of Management, Massachusetts Institute of Technology, USA

David Metcalf

London School of Economics and Political Science, UK

Saul A. Rubinstein

School of Management and Labor Relations, Rutgers University, USA vii

viii

John F. Schnell

College of Administrative Science, University of Alabama – Huntsville, USA

Teresa Schoellner

PhD, Department of Economics

Charlotte Yates

Labour Studies and Department of Political Science, McMaster University, Canada

INTRODUCTION Volume 13 of Advances in Industrial and Labor Relations (AILR) contains eight papers that deal with a variety of industrial relations topics. The first chapter, by Stephen Hills and Teresa Schoellner, analyzes the influence of the creation in 1999 of the European Monetary Union (EMU) on the deregulation of part-time work in each of the 13 EMU member nations. The second chapter, by Cynthia Gramm and John Schnell, provides a systematic review and analysis of the effects of flexible staffing arrangements (e.g. temporary employment) on individuals working under such arrangements, regular employees in the same organizations, organizational performance, and macroeconomic outcomes. The third chapter, by Alexander Colvin, analyzes the adoption, structure, and function of peer review and arbitration type dispute resolution procedures, respectively, for nonunion employees in two different business units of the TRW company. The fourth chapter, by Sue Fernie and David Metcalf, nicely complements the Colvin paper by examining the uses and implications for employee voice, conflict resolution, and fairness at work of ombuds arrangements in four British enterprises. The fifth chapter, by Susan Eaton, Saul Rubinstein, and Robert McKersie, analyzes recent U.S. experience with 50 labor-management partnerships by identifying key factors that influence the formation and sustainability of such partnerships, including the development of network ties across traditional boundaries, and by providing theoretical and empirical perspectives on this experience as a whole. The sixth chapter, by Felice Martinello and Charlotte Yates, applies cluster analysis to a sample of Ontario, Canada union organizing campaigns, finds several distinctive (though interdependent) union and employer campaign strategies, and estimates the effects of these strategies on campaign outcomes. The seventh chapter, by John Logan, provides the first detailed scholarly analysis of the U.S. AFL-CIO’s 1990–1994 campaign to achieve legislation proscribing the use of permanent striker replacements, and also analyzes this campaign’s significance for several other on-going labor problems. The eighth chapter, by Diana Kelly, analyzes the influences of U.S. scholar John Dunlop and of the British Oxford School, respectively, on the development during 1960–1985 of Australian industrial relations thought, and shows how and why Dunlop had substantially greater influence than the Oxford School in this regard. On a note of recognition, the papers in this volume by Colvin, Martinello and Yates, and Logan were winners of the 2003 AILR/Industrial Relations Research ix

x

Association (IRRA) “best papers” competition. On an extremely somber note, Susan Eaton, co-author with Saul Rubinstein and Robert McKersie of a paper in this volume, died of an incurable illness in late 2003 at a very young age. It may come as some solace to Susan’s family and friends that the editors of AILR were able to tell Susan that her paper had been accepted for publication. It is not by happenstance that the research settings for the papers contained in Volume 13 of AILR include North America, Europe and the Pacific Rim. AILR has long encouraged manuscript submissions from researchers worldwide, and seeks to publish articles that expand theoretical and empirical industrial relations knowledge beyond that obtained from U.S. settings and data sources. Taken as a set, the eight papers contained herein clearly reflect achievement of this objective. David Lewin Bruce E. Kaufman Editors

EUROPEAN RESPONSE TO HIGH RATES OF UNEMPLOYMENT: MONETARY POLICY OR DEREGULATION OF PART-TIME WORK? Stephen M. Hills and Teresa Schoellner INTRODUCTION Decreased regulation of part-time work is one way a country responds to high rates of unemployment. Proponents of deregulation argue that a more flexible labor market is required to allow labor markets to clear. A more traditional response to high unemployment is change in monetary policy, where interest rates are lowered to stimulate the economy and increase rates of employment. Both policies have been tried in Europe, a good place to study the effects of the two policy responses, both because European unemployment has been high and because the trade off between monetary policy and the deregulation of part time work has varied from country to country. The establishment of the European monetary union (EMU) in 1999 created a natural experiment in which any one country’s ability to adjust its monetary policy was curtailed, creating pressure for deregulatory policies to come into play (Aaronovitch & Grahl, 1997; Pisani-Ferry, 1998). The establishment of the EMU lowered the relative cost of easing restrictions on part-time work and may have caused changes in labor regulation in certain

Advances in Industrial and Labor Relations Advances in Industrial and Labor Relations, Volume 13, 1–26 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0742-6186/doi:10.1016/S0742-6186(04)13001-1

1

2

STEPHEN M. HILLS AND TERESA SCHOELLNER

European countries. In fact, changes in the monetary policy of each of the eleven signatory countries to the EMU became, in terms of international political relationships, far more costly on January 1, 1999, when the Maastricht treaty was implemented.1 On that date, all eleven countries aligned their interest rates and agreed to exchange rate fluctuations within a narrow band, placing monetary policy under the authority of a single central bank. Nevertheless, each maintained autonomy over its separate regulatory environment. We, therefore, expected the regulatory environment to change as some countries turned to their next best options for dealing with unemployment. What had formerly been the politically less costly option – an independent monetary policy – was no longer viable. In this chapter we describe the European march toward monetary union in terms of interest rates, inflation rates, and unemployment. Secondly, we estimate a fixed effects model of change in each country’s unemployment, hypothesizing that the EMU would result in differential rates of unemployment that reflect the variety of institutional regulations across Europe. In other words, a common monetary policy would make the unemployment costs of different labor regulations more obvious, country by country, producing rates of unemployment that were higher in some and lower in others. A common monetary policy might increase pressure on countries with tighter regulations over part time work to ease them, but countries might also seek a common standard of regulation and encourage adjustment through migration from countries with high unemployment to countries with low unemployment. Since the latter would mean giving up even more economic control to the EU, we expected pressures to increase to ease regulations on part time work. Before examining the record on deregulation, however, we looked at the trade off between unemployment and inflation, using rates of inflation as a way of inferring which countries had, in the past, relied more heavily on monetary policy to control unemployment. Those with historically high rates of inflation might be most likely to try and retain control over unemployment by easing restrictions on part time work. The fixed effects model is modified to reflect changes in both unemployment and inflation as countries sought to align their monetary policies under the dictates of the Maastricht treaty. We use the fixed effects model to predict what countries experienced unexpectedly high unemployment rates after full implementation of the EMU. We also use it to identify countries where the trade off between unemployment and inflation was particularly difficult in the years immediately preceding monetary union. Finally, we turned to the regulatory record. We asked whether policy makers in the countries most affected by the EMU had already begun to ease restrictions on part-time work, even before 1999.2 We speculated on the type of policy changes that might continue to occur, given the country’s cultural and political environment. We asked whether deregulation of part time work prior to 1999 had

European Response to High Rates of Unemployment

3

helped some countries adjust to a loss of control over monetary policy subsequent to the establishment of the EMU.

PART-TIME WORK AND THE EUROPEAN INSTITUTIONAL LANDSCAPE Rates of part-time work have varied widely from country to country in Europe, even before Maastricht. Rates can be obtained from the statistical offices of the separate countries, but definitions are often not consistent. The Organization for Economic Cooperation and Development (OECD), however, uses a consistent definition for part-time work: less than 30 hours of work per week. Using OECD data, we found part-time work varying from 5% of employment in Spain to 27% in the Netherlands in 1988. (Comparable data for the U.S. showed a rate of 13%.) For women the range was even greater, from 12% in Spain to 51% in the Netherlands (see Table 1). Not only do rates of part-time employment vary widely at one point in time; they are changing significantly over time. In Denmark, rates of part-time work fell in the past ten years, particularly for women. A decline also occurred in Sweden. But both Italy and Luxembourg have seen dramatic increases in part-time work over the same time period. In Luxembourg, part-time rates for women went from Table 1. Rates of Part-time Employment, Potential EMU Countries (Total Labor Force, Women Only, 1988 and 1998). Country

Austria Belgium Denmark Finland France Germany Ireland Italy Luxembourg Netherlands Portugal Spain Sweden U.K.

Both Sexes, % Labor Force

Women, % Labor Force

1988

1998

1988

1998

NA 13.2 19 7.4 12.3 11.4 9.3 8.3 7.1 26.9 6.7 5 16 20.5

11.5 16.3 17 9.6 14.8 16.6 18 11.2 12.8 30 9.9 7.7 13.5 23

NA 27.7 31.4 10.5 21.8 26.4 19.4 17.3 18.3 51.2 12.3 12.1 27.6 40.8

22.8 32.2 25.4 13 25 32.4 31.2 22.4 29.6 54.8 15.8 16.6 22 41.2

4

STEPHEN M. HILLS AND TERESA SCHOELLNER

18% in 1988 to 30% just a decade later. Rates of part-time work are a complex result of legislative regulation, collective bargaining provisions, cultural norms, and economic conditions, all of which can change over time. In the late 1990s, member states of the OECD were tracking changes in the cultural and regulatory environment of part-time work through the framework of the OECD Jobs Strategy. The OECD’s Economic and Development Review Committee recommended in 1995–1996 a number of specific changes for each of the member countries. Reports in 1997 and 1999 documented the changes that occurred. The 1999 report demonstrated that the reason for falling rates of parttime employment in Denmark and Sweden was, in part, deteriorating economic conditions from 1993 to 1997. In Sweden, both full time and part-time employment opportunities fell for women during this time period. In Denmark, however, full time opportunities increased for women at the same time that part-time work decreased (OECD, 1999, Table 4.4, p. 86). Some of the increases in part-time employment can also be traced to changing economic conditions. In Luxembourg, decreasing full-time opportunities for women accompanied an increase in part-time work. The same was true, though to a lesser extent, for France and Germany. Ireland’s dramatic increase in part-time work for women was accompanied by an equally dramatic 7% increase in total employment from 1993 to 1997 (OECD, 1999, Table 4.4, p. 86). To some degree, increasing part-time work may be a result of the long standing decline in work hours throughout European countries. This contrasts with rising hours worked in the U.S. during the 1990s. The European decline in work hours is largely a result of voluntary agreement between employers and employees (or their union representatives), but government health and safety regulations also reinforced the trend (OECD, 1999, p. 122). Changes in the rate of part time work clearly are influenced by a variety of factors. Our task is to determine if unexpected changes in the level of unemployment put pressure on European governments to ease regulations on part time work. Is there evidence that regulatory reform was used as a substitute for monetary policy in the countries associated with monetary union? To answer this question, we first examine Europe’s peculiar kind of unemployment.

EUROPEAN UNEMPLOYMENT: STRUCTURAL OR A RESULT OF THE DRIVE TO MONETARY UNION? The rise in European unemployment for the last twenty years has begged explanation. The U.S. fall in unemployment over the same period of time – in the year 2001 near or below what had been thought the “natural rate” – caused soul

European Response to High Rates of Unemployment

5

searching about the flexibility of European labor institutions. Real wage rigidity, minimum wages, unemployment insurance systems, governmental regulations on discharge, and other collectively bargained regulations have been candidates for blame. (Scarpetta, 1996; Vinals & Jimeno, 1998). The rise in unemployment made monetary union more difficult since individual countries had to relinquish control over monetary policy, one of the main policy instruments for dealing with unemployment. If high rates of unemployment were structural, however, relinquishing monetary policy would perhaps not be so hard, since monetary

Fig. 1. Unemployment Rates.

6

STEPHEN M. HILLS AND TERESA SCHOELLNER

policy is not the best tool for encouraging structural change anyway. Changes in bargaining rules or legislative regulations might be more appropriate. All European countries have experienced rising rates of unemployment over the last twenty years. In Fig. 1 are displayed the unemployment rates for the eleven EMU countries plus Great Britain, Denmark and Sweden, all of which ultimately opted out of the union but could have joined if they wished. The general rise since 1970 can be seen plus a growing dispersion in rates for all but Spain. Most countries had an upward trend in unemployment, but Spain did not. Spanish unemployment rates rose to 21% in 1985, fell to 16% in 1990, peaked at 24% in 1994 and then fell dramatically thereafter, registering 14% in 2000. Irish rates peaked at 18% in 1987, fluctuated around that rate until 1993, and fell to 3.6% in 2000. Only the Netherlands had a rate lower than the Irish in the year 2000, registering just 2.5%. Finnish rates of unemployment peaked at 17% in 1994 and fell to 9.2% in 2000. While Finland, Ireland, and Spain showed considerable convergence around the new European norm for unemployment, other countries did not show a similar movement toward convergence.

EUROPEAN MONETARY POLICY The Maastricht Treaty of 1992 established the framework for monetary union. The treaty was signed by the then Heads of State of the European Community and established the criteria for member states to enter into monetary union. The criteria placed limits on the amount of government deficit and total debt that a country could incur and asked member states to achieve inflation rates and longterm interest rates close to the average for the three countries with the lowest inflation. In 1994, Austria, Finland and Sweden voted to enter the European Community and were thereby considered eligible for the EMU. Sweden would ultimately opt out of the EMU, as would the UK and Denmark, leaving eleven countries who agreed to the EMU as of January 1, 1999. Greece, which had been a member of the European Community since 1981 was not accepted for the EMU in 1999 but it did enter the European Monetary System, aligning its exchange rates with other European countries that agreed to the EMU. Given the coordination in monetary and fiscal policy necessary to meet Maastricht requirements, we would expect convergence in European interest rates. Figure 2 shows the full range of short term interest rates that have reflected independent monetary policy in European countries for the last thirty years. One can see the effects of Maastricht on interest rates in the 1990s. Sharp convergence has taken place. With the exception of the UK whose interest rates were at 6.6%

European Response to High Rates of Unemployment

7

Fig. 2. Interest Rates (Includes Only Short-term Interest Rate Data from Datastream).

in 2000, interest rates in other European countries have been just above 4%: 4.6% in Denmark, 4.2% in Sweden, and the forced convergence of the eleven EMU countries at 4.25%. Historically high unemployment rates in Europe may, indeed, be the result of labor market “inflexibility.” Alan Krueger addressed the impact of European integration on the “labor compact,” namely the “implicit bargain among labor, capital and government encompassing pay, work effort, social protection, and labor – such market rules as union organization, safety standards, and termination laws” (Krueger, 2000, p. 1). But he concluded that the threat to separate labor

8

STEPHEN M. HILLS AND TERESA SCHOELLNER

compacts in individual European countries has likely been exaggerated. Imperfect mobility of labor, capital and services will prevent the convergence of labor compacts according to Krueger, but countries will have to bear costs for their unique regulatory environments. An integrated monetary policy will make the costs of different regulatory environments more transparent. One can see this by examining the trade off between unemployment and inflation in the countries meeting the stringent monetary and fiscal requirements of Maastricht. In Fig. 3 we have plotted inflation rates against unemployment rates for fourteen European countries – for the eleven who eventually founded the EMU and the three others who considered it but opted out, at least in 1999. For many of the countries, in the late 1980s and through the 1990s, higher rates of unemployment were accepted in exchange for the lower rates of inflation that would ultimately be the criteria for monetary union. The trade off appears to have been

Fig. 3. Inflation vs. Unemployment Rate.

European Response to High Rates of Unemployment

9

Fig. 3. (Continued )

made in the early 1980’s for most countries. Starting at that point, a steady increase in unemployment was matched by an equally steady fall in inflation for the next fifteen years. Finland, Sweden and Portugal lagged in their decisions. Nevertheless, for almost all 14, conscious policy choice appears to have been an important factor in rising unemployment rates. The different institutional arrangements (including regulations regarding part-time work) undoubtedly influenced how high unemployment would rise, but one should not discount the effect that a planned monetary union has had, even well before 1999.

MEASURING THE EFFECT OF THE EMU THROUGH UNEXPECTEDLY HIGH UNEMPLOYMENT We predicted that the EMU’s effects might be seen, in part, through unusually high unemployment in some countries. But in which countries? The EMU’s effects should be most visible in those countries where an independent monetary policy had previously reduced unemployment to a rate lower than its institutions would permit under a common, multi-country monetary policy. We sought to show this effect by constructing a model to compare predicted and actual rates of unemployment at the time of monetary union. For these estimations we used a fixed effects model in which real interest rates, and each country’s lagged unemployment rate were predictors for the country’s annual rate of unemployment. A pooled cross-sectional, time series design was used for the data that began in 1970 and continued through 1999. Data were obtained from Data Stream, a British based on-line source for economic and financial data over time. The single source helped to insure consistent definitions for the variables over time. The results of our estimations are shown in Table 2. The fit for the model was good – with an overall R squared of 0.95%; much of the fit was attributable to

10

STEPHEN M. HILLS AND TERESA SCHOELLNER

Table 2. Fixed Effects Model, 1998 (Pooled Annual Data, 1970–1998). Country

Country Specific Intercept

Group Intercept

Beta 1 × Inflation

Beta 3 × Lagged Unemployment

Predicted Value

Actual Value

Residual

Austria Belgium Denmark Finland France Germany Ireland Italy Luxembourg Netherlands Portugal Spain Sweden U.K.

−0.136 0.346 1.043 0.166 0.457 0.155 −1.205 0.214 −0.467 −0.152 0.176 2.516 −0.260 Base

0.255 0.255 0.255 0.255 0.255 0.255 0.255 0.255 0.255 0.255 0.255 0.255 0.255

0.001 0.002 0.002 0.002 0.002 0.002 0.004 0.003 0.002 0.003 0.004 0.003 0.000

6.456 8.557 7.020 11.503 11.312 8.857 9.421 10.758 3.037 5.038 6.156 18.933 7.302

6.577 9.161 8.321 11.927 12.026 9.270 8.476 11.230 2.828 5.145 6.592 21.707 7.297

7.20 9.51 6.38 11.37 11.81 10.88 7.60 11.94

−0.623 −0.349 1.941 0.557 0.216 −1.610 0.876 −0.710

4.15 5.04 18.82 6.52

0.995 1.552 2.887 0.777

the lagged unemployment figure. All countries that initially tried to meet EMU requirements, including those that ultimately opted out of monetary union, were included in the estimates. The UK, with its independent monetary policy, was used as the base.3 Table 3. Residuals from Fixed Effects Model. Residual

Germany Italy Austria Luxembourg Belgium France Ireland Finland Sweden Denmark Netherlands Spain Portugal

1999

1998

−0.116 0.313 0.426 0.577 0.594 0.736 0.814 1.078 1.125 1.183 1.281 1.712 2.096

−1.15111 −0.1962 −0.16463 0.537426 0.024033 0.579613 1.265345 1.082641 1.567412 1.530381 1.545402 0.570743 3.01023

European Response to High Rates of Unemployment

11

Residuals are shown in Table 3 (residuals are calculated as predicted values minus actual values, thus a negative means unexpectedly high actual unemployment). When predicted values from the fixed effects model were compared with actual values for 1998 or 1999, Germany, Italy and Austria showed noticeably higher rates of unemployment than we would have expected. In Fig. 4 are shown the eleven EMU countries by level of unemployment in 1998. The fixed effects model shows that three countries with already high rates of unemployment in 1998 (Germany, Italy and Austria) were disproportionately affected by EMU in 1999. Residuals ranged from −0.6 to −1.6. Not all the high unemployment countries were affected, however. Spain, which had the highest rate of unemployment in 1998, did not have unexpectedly high unemployment that might be attributed to EMU – in fact, our estimates show a rate of unemployment considerably lower than expected.

Fig. 4. Unemployment in 11 EMU Countries, (a) 1998 and (b) 1999.

12

STEPHEN M. HILLS AND TERESA SCHOELLNER

Table 4. Fixed Effects Model with Cross Term for Unemployment 1992–1999 (Pooled Annual Data, 1970–1999). Variable

df

B-Value

Std. Error

t-Ratio

Approx. Prob.

Intercept LAGURATE INFL RGDP Austria Belgium Finland France Germany Ireland Italy Luxembourg Netherlands Portugal Spain Sweden U.K. Austria D Belgium D Finland D France D Germany D Ireland D Italy D Luxembourg D Netherlands D Portugal D Spain D Sweden D U.K. D

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

0.886 0.913 0.128 – −0.250 0.494 −0.011 0.673 0.0508 −0.890 0.0212 −0.686 −0.172 0.587 5.16 −0.452 0.113 0.253 0.329 0.186 0.188 0.478 −1.66 0.587 0.228 −0.451 0.379 1.48 0.397 −0.810

0.254 0.022 0.055 – 0.316 0.331 0.306 0.342 0.3104 0.727 0.2991 0.339 0.301 0.397 1.34 0.324 0.299 0.449 0.447 0.491 0.455 0.455 0.45 0.452 0.431 0.444 0.465 0.67 0.457 0.445

3.487 41.686 2.341 −3.155 −0.791 1.494 −0.037 1.969 0.164 −1.224 0.071 −2.024 −0.571 1.477 3.867 −1.395 0.377 0.564 0.735 0.378 0.413 1.050 −3.695 1.298 0.529 −1.017 0.815 2.191 0.868 −1.821

0.001 0.000 0.020 0.002 0.430 0.136 0.970 0.050 0.870 0.222 0.944 0.044 0.568 0.141 0.000 0.164 0.706 0.573 0.463 0.706 0.680 0.295 0.000 0.195 0.597 0.310 0.416 0.029 0.386 0.070

SSE MSE SBC Reg. Rsq. Durbin-Watson

353.3129 1.021135 1221.529 0.9539 0.8465

DFE Root MSE AIC Total Rsq.

346 1.010512 1103.641 0.9539

Note: F-test on including the crossterm dummies. F-test: Numerator df: 11, denominator df: 346; F-statistic: 2.5119, Prob > F: 0.0047.

European Response to High Rates of Unemployment

13

MEASURING THE EFFECT OF THE EMU THROUGH CHANGES IN UNEMPLOYMENT, 1992–1999 A problem exists in measuring the effects of the EMU only in 1998 and 1999. Well prior to 1999, when the Maastricht treaty was implemented, countries were aligning their monetary and fiscal policies. The common monetary policy was a process, rather than a single event. To provide a more complete picture of the effects of the EMU, we re-estimated the fixed effects model, including a cross term to measure the increases in unemployment that occurred while a common monetary policy was being instituted between 1992 and 1999. The results (Table 4) show that in all countries except Ireland, the Netherlands, or the UK, unemployment was increasing. In Spain, Germany, and Italy unemployment increased the most. The dummy variables for all countries were jointly significant, though not close to significant separately.

MEASURING THE EFFECT OF THE EMU THROUGH RATES OF INFLATION A second way of gauging the effect of the EMU on individual countries is to examine past monetary policies that were used to lower the rate of unemployment. Assuming: (1) that inflation is always and everywhere a monetary phenomenon;

Fig. 5. Inflation Rates, Potential EMU Countries, 1970–1999.

14

STEPHEN M. HILLS AND TERESA SCHOELLNER

and (2) that monetary authorities rationally respond to political pressure to reduce unemployment, we need only look for countries that had “relatively” high rates of inflation prior to Maastricht. These same countries should be the ones with the lowest tolerance for high levels of unemployment and would also have the most trouble adapting to the low inflation policies of the European Central Bank. In Fig. 5, inflation rates are displayed over time for the eleven countries of the EMU. Based on our rather straightforward criteria, Portugal, Spain, and Italy should have the most difficult time adjusting to the EMU, followed by Finland, Sweden and Belgium.

MEASURING THE EFFECT OF THE EMU THROUGH CHANGES IN RATES OF INFLATION FROM 1992–1999 To measure how the Maastricht process changed rates of inflation from 1992–1999, we re-estimated the fixed effects model, including a cross term to measure the rates of inflation that occurred while a common monetary policy was being instituted. Results are shown in Table 5. Finland, Spain, and Belgium had the highest rates of inflation after 1992. Cross terms were jointly significant, but not individually significant so results must be interpreted somewhat cautiously. To summarize our results: In Germany, Italy, and Austria unemployment was unexpectedly high during the implementation period for the Maastricht treaty (1992–1998). But in virtually all 11 countries (with the exception of the Netherlands) unemployment increased, compared to years prior to 1992. In Belgium, Finland and Spain, rates of inflation were highest during the seven years of adjustment to Maastricht’s requirements. Of those three, we suspected that Spain would have the most difficulty adjusting to Maastricht due to its historical willingness to accept inflation in its monetary policy. Finally, only Germany experienced a rate of unemployment that was higher than predicted after the implementation of Maastricht was complete in 1999. To reiterate: our strategy here was to estimate the effects of EMU on unemployment and on monetary policy (including inflation). In countries where effects of EMU were strong, we hypothesized that political pressure may have led countries to reduce labor market regulation, given that an independent monetary policy was no longer an option. We turn now to the regulatory record. Our focus will be on Austria, Germany, Italy, and Spain, all countries that experienced unexpectedly high unemployment, either in the run-up to the EMU or after the EMU was fully implemented. We will also pay special attention to the Netherlands, the exception to rising unemployment in years prior to the EMU. We expect to find a complex, though interesting,

European Response to High Rates of Unemployment

15

Table 5. Fixed Effects Model with Cross Term for Inflation 1992–1999 (Pooled Annual Data, 1970–1999). Variable

df

B-Value

Std. Error

t-Ratio

Approx. Prob.

Intercept LAGURATE INFL RGDP Austria Belgium Finland France Germany Ireland Italy Luxembourg Netherlands Portugal Spain Sweden U.K. Austria I Belgium I Finland I France I Germany I Ireland I Italy I Luxembourg I Netherlands I Portugal I Spain I Sweden I U.K. I

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

1.03 0.895 0.195 – −0.327 0.445 −0.213 0.624 0.0225 −1.763 0.00142 −0.800 −0.236 0.501 5.24 −0.626 0.0958 16.9 30.5 77.0 31.5 25.1 −0.112 20.2 11.8 −14.8 17.0 56.7 48.1 −31.1

0.24 0.020 0.051 – 0.296 0.310 0.290 0.316 0.2907 0.648 0.28560 0.318 0.291 0.361 1.11 0.294 0.2870 19.2 23.0 35.1 27.5 18.5 0.037 11.5 18.8 18.9 10.8 14.3 17.7 15.9

4.284 45.508 3.859 −3.98 −1.105 1.435 −0.735 1.977 0.077 −2.721 0.005 −2.514 −0.808 1.388 4.722 −2.129 0.334 0.879 1.326 2.19 1.147 1.359 −3.012 1.755 0.628 −0.78 1.582 3.961 2.718 −1.954

0.0001 0.0001 0.0001 0.0001 0.2701 0.1522 0.4629 0.0489 0.9383 0.0068 0.996 0.0124 0.4195 0.1661 0.0001 0.034 0.7386 0.38 0.1856 0.0292 0.2523 0.175 0.0028 0.0802 0.5306 0.4362 0.1145 0.0001 0.0069 0.0515

SSE MSE SBC Reg. Rsq. Durbin-Watson

334.0426 0.965441 1200.441 0.9564 0.8662

DFE Root MSE AIC Total Rsq.

346 0.982569 1082.553 0.9564

Note: F-test on including the crossterm dummies. F-test: Numerator df: 11, denominator df: 346; F-statistic: 3.5730, Prob > F: 0.0001.

16

STEPHEN M. HILLS AND TERESA SCHOELLNER

picture of interaction among political forces, a country’s culture, and its regulatory stance.

POLICY PRESSURES IN COUNTRIES NEGATIVELY AFFECTED BY EMU In the run up period prior to monetary union, European countries considered decreased regulation of the labor market as an alternative to monetary policy. To demonstrate this, we turn to a set of “European Employment Strategy Papers” required by the Employment Title of the Treaty of Amsterdam. The treaty was signed by members of the European Union in 1997. Since then, each country of the EU has drawn up a National Employment Action Plan that is available publicly on the “Europa” web site. Action plans are created annually. In the action plans, proposals for legislative changes are described, and changes that have already been made in the regulatory environment are noted. We have studied the plans for each of the countries identified in our statistical analysis as significantly affected by the Maastricht agreement, paying special attention to plans that would change the regulation of part time work. Countries of interest were Germany, Austria, Italy and Spain. Germany’s 1999 National Employment Action Plan began its section on labor market adaptability by pointing out an important fact: 90% of workers required to pay social security contributions are covered by collective bargaining contracts. Therefore, flexibility depends greatly on the stance taken by German unions, which have traditionally been very inflexible. The job crisis, aggravated by German unification in 1990 caused the Confederation of German Trade Unions (the DGB) to change its policy, however. Gradually, the DGB accepted more flexible ways of calculating work hours, including provisions for part-time work and sabbaticals, particularly for older workers who make the transition to retirement. The DGB also wanted to extend collective bargaining protection to “non-standard jobs,” most importantly to temporary workers (Sarfati, 1998, p. 21). By the end of 1998, about 40% of workers covered by collective contracts had negotiated agreements that contained provisions for part-time employment for older workers (German Employment Action Plan, 1999, p. 57). The Law on Development of Part-time Work of Older Workers came into effect on January 1, 2000 and further strengthened the legal parity of part-time and full-time workers. Future collective agreements were expected to gradually extend the scope of part-time work (German Employment Action Plan, 2000, p. 9). Despite gradual change, the unexpectedly high unemployment in Germany was undoubtedly a reflection of its fairly inflexible

European Response to High Rates of Unemployment

17

institutions. In the years following implementation of the EMU, we should expect increasing pressure for German institutions to be made more flexible as a way of compensating for the country’s loss of an independent monetary policy that could attack unemployment more directly. Rates of part-time work are neither high nor low in Austria. Yet it is one of the countries identified by our statistical analysis as experiencing disproportionately high unemployment in 1999, as countries rushed to meet the requirements for monetary union. The Austrian government in 1998 was focussed on promoting job creation and on alleviating the financial strains of its social insurance funds. A package of legislative measures took effect on January 1, 1998 (Sarfati, 1998, p. 13). As in France, the debate between Austrian employers’ associations and trade union confederations centered mostly on reducing weekly hours. For its part, the Austrian government offered to help compensate for loss of pay and to promote training and educational leaves (Sarfati, 1998, p. 13). The Austrians did not, as of 1998, seem inclined to use regulatory reform as an offset to the loss of an independent monetary policy. Despite the unexpectedly high unemployment, Austrian policy response was ambivalent. Historically, Italy has suffered consistently high unemployment, particularly in the south. At 11.2% in 1998, Italy’s rate of part-time employment was relatively low, with only three EMU countries registering a lower rate. Italy has had wide gender gaps in both employment and unemployment rates; its overall employment rate is only 51%. In the 1990s, the Italian government was concerned with persisting inflation, accompanied by recession, public deficits, and high unemployment. Italy’s 1999 National Employment Action Plan took a holistic approach to these interrelated problems, arguing for a high level pact involving all the “social partners,” namely employers’ associations, trade unions, and the Italian government. In 1992, a pact was reached that abolished the Italian automatic pay indexing system known as the “scala mobile,” thereby allowing the introduction of effective incomes policies. The second such pact, agreed to in 1997, was in part aimed at increasing labor market flexibility. The agreement legalized private temporary employment agencies and promoted part-time work by reducing employers’ social security taxes. Other provisions funded community work in areas of high unemployment, discouraged black market transactions, and subsidized vocational training (Sarfati, 1998, p. 17). The 1999 action plan set a goal to raise part-time employment to the European average of 17% by the year 2003 (Italian National Employment Action Plan, 1999, p. 41). Part time work was not to be targeted at women, the long term unemployed, or young people. Rather, part-time work was intended to carry with it benefits and regulatory protections equal to full time work. In short, the low level of part-time work in Italy have been part cultural and part structural. The structural initiatives

18

STEPHEN M. HILLS AND TERESA SCHOELLNER

of 1997 were intended to raise part time employment by about 50%, by no means a small change. It is unclear how much more change could have occurred without important shifts in cultural values. In Spain, where unemployment in the 1990s was the highest in the EU, economic growth has from 1996 to 1999 exceeded significantly the average growth of other member countries (European Commission, 2000, p. 28). Yet Spain had the lowest employment rate of any country in the EU, and gender differences were large. The employment rate for women in 1999 was only 35%, compared with a 51% rate for women in all EU countries. Spanish women had an unemployment rate twice the rate for women in the EU and double the rate for Spanish men. About one fourth of Spanish women were unemployed in 1999 (European Commission, 2000, p. 29). Given the high rates of unemployment in Spain, rates of part time work were surprisingly low. At 7.7%, the rate was the lowest of any EMU country. Only two countries had rates of part time employment for women that were lower than for Spanish women – Finland and Portugal (see Table 1). The National Employment Action Plans for Spain in 1999 and 2000 revealed a strong emphasis on the positive effects of economic growth on reducing unemployment. Several references were made to positive rates of employment growth in Spain and their link to the overall rise in GDP. The action plan emphasized the beneficial effects of the reduction in budgetary deficits and wage restraint in collective bargaining. But little emphasis was placed on labor market flexibility or on encouraging a higher rate of part-time work. The only reference to part-time work was in the 1999 action plan where reference was made to legislation passed in November, 1998 encouraging “stable part-time employment.” Labor market reform in Spain dates back to 1984 when a Spanish Workers’ Statute was amended. Regulations on fixed term contracts were lifted, dismissals were made easier, and labor mobility was encouraged (Sarfati, 1998, p. 16). Labor market reform, as defined in 1997, meant creating better protection for part-time workers and limiting the use of temporary contracts. These two provisions grew out of a year-long negotiation between the Spanish employer’s association and Spain’s two main trade union confederations (Sarfati, 1998, p. 39). In 1997, the Spanish government again relaxed regulations on dismissals and on severance pay, but with the aim of encouraging open ended (as opposed to fixed, temporary term) contracts. In sum, Spain’s public policy was aimed mainly at macro-economic stimulation of the growth in GDP. To be sure, legislation restricting the use of temporary employment contracts was modified but little direct policy emphasis was placed on encouraging part-time work. Our empirical results for Spain were mixed. On the one hand, Spain showed unexpectedly high unemployment between 1992 and 1999. But in 1999 alone, we

European Response to High Rates of Unemployment

19

showed an unexpectedly low rate of unemployment. In the late 1990s, Spain was growing its way out of unemployment, with the help of the free trade associated with the European Union. On the other hand, Spain’s independent monetary policy and its associated higher inflation (relative to its neighbors) had to yield to an EU common monetary policy that would raise Spain’s unemployment. The two opposing effects may have been reflected in our mixed results. Our results for inflation identified three countries which had particularly high rates of inflation in the years during which the EMU was established – Belgium, Finland, and Spain. In each country, a special set of circumstances helps explain these results. In Belgium, debate over work hours, part-time employment and part-time early retirement was part of a failed tripartite negotiation initiated by the government in 1996. Labor unions would not accept the government’s proposed wage restraint unless job creation was part of the employment side of the agreement. Employers wanted wage restraint and labor flexibility without guarantees for job creation. After negotiations broke down, the government unilaterally imposed an incomes policy and passed legislation aimed at meeting EMU fiscal requirements (Sarfati, 1998, p. 15). The need for an incomes policy probably means that monetary policy prior to 1996 had an inflationary bias. In Finland, a “Rainbow Coalition” elected in 1995 promised to halve unemployment by 1999. Unemployment, which had peaked in 1994 at 16.6%, fell steadily to 9.2% in 2000. The new government followed a coalition government that had not taken much responsibility for employment promotion in the recessionary years of the first part of the decade. The Finnish push to lower unemployment probably increased rates of inflation from 1992 to 1999. The third country where inflation was evident just prior to the EMU was Spain, whose emphasis on (inflationary) economic growth we have already discussed.

COUNTRIES WITH RATES OF UNEMPLOYMENT UNAFFECTED BY THE EMU Perhaps the best example of a country whose unemployment was not affected by the EMU was the Netherlands. From 1992 to 1999, our data showed that the Dutch were the only ones who did not show an increase in unemployment associated with the run-up to the EMU. In the Netherlands, low rates of unemployment have historically been coupled with extraordinarily high rates of part-time employment. Since the recessionary periods of the 1970s and early 1980s, the Dutch government promoted part-time employment and urged labor and management to improve the quality of part-time jobs. Access to social security coverage was the same, whether part-time or full-time. Dutch legislation also simplified procedures for temporary

20

STEPHEN M. HILLS AND TERESA SCHOELLNER

help agencies. The result was the highest incidence of part-time work among industrialized countries (Sarfati, 1998, p. 12). With such high rates of part-time work, one might ask whether those holding part-time jobs did so voluntarily or involuntarily. The European Union fields its own Labour Force Survey (LFS), which helps to answer this question. In the EU’s LFS, workers are asked if they are employed in part-time jobs voluntarily or involuntarily. Among men, where somewhat less than 10% held part-time jobs in 1998, less than 2% of all part time employed men responded that they held their part-time jobs involuntarily. The figures were even more dramatic for women. Among women, where about 55% held part-time jobs in 1998, less than 2% of part time employed women responded that they held their part-time jobs involuntarily. Clearly, part-time jobs were overwhelmingly viewed as good jobs in the Netherlands (European Commission, 1999, Tables 35 & 36, p. 39 and Table 1).

DISCUSSION AND CONCLUSIONS Perhaps the key question underlying the study of industrial relations is how control over employment relationships is, or should be, shared. In the case of the EMU, we see several forces at work, each vying for its own control. Eleven European nation states have opted to cede control over monetary policy to a central bank. Two others (Sweden and Denmark) have aligned their interest rates with the eleven, even though they have not entered the union. We estimated the effects of giving up control over monetary policy, with the expectation that some countries would experience an increase in unemployment. We hypothesized that countries with abnormally high unemployment would, in turn, show pressure to substitute regulatory control for monetary control, and we found a wide variety of possibilities, even when looking only at regulations concerning part-time work. In some cases, governments were quite aggressive in changing legislation that affected the amount of part-time work. In other cases, governments worked indirectly through negotiations with social partners to change the nature and amount of part-time work. In some countries, negotiations and changes in legislation were long standing. Others began in the late 1990s. Our statistical analysis showed that pressures for more labor market flexibility would be expected in Austria, Germany, Italy and Spain. Italy had begun to change its stance toward part-time work, but it still had one of the lowest rates of participation in Europe Furthermore, change in the provisions for part time work in Italy had to be bargained through the national labor federations. In 1997, changes were bargained, but before that, concern was largely over the inflationary effects

European Response to High Rates of Unemployment

21

of the “scala mobile.” Spain successfully countered increased pressure from unemployment with economic growth, but if growth were to slacken, a change in policy toward part time work could be expected. In Germany, the DGB agreed to changes in regulation of part-time work, but the changes were directed at special groups – those readying for retirement in particular. Pressure could, perhaps, be expected for broader changes affecting all part-time employment. In Austria we should expect discussion to turn from weekly work hours to the use of part-time and temporary labor as the unemployment costs of Austria’s regulatory environment become more clear. From the limited information available, it appeared that Austrian discussion on these issues in the late 1990s was just beginning. The country most aggressive in its encouragement of part-time work was the Netherlands. The EU’s Labour Force Survey helped us establish that part-time work in the Dutch context was not seen as “involuntary.” But the Dutch also enjoyed one of the lowest unemployment rates in Europe in 1999. In this context, Dutch workers appeared to be using part-time work to accommodate a variety of patterns of work and non-work activity. The Dutch government sought to insure that part-time work did not carry with it lower benefits than full time work, at least where government-provided social insurance is concerned. The flexibility of part-time work arrangements probably helped insulate the Dutch economy from the negative effects on unemployment that the introduction of the EMU produced. What does the policy record show since 1999 in regard to monetary policy and deregulation of part time work? Arguments for flexibility in labor market institutions were more compelling when European labor markets were loose in comparison with the tight labor markets in the United States in the late 1990s. The distribution of unemployment from country to country in Europe was more controversial when policy makers believed that the overall rate of unemployment could be lower if only there were more flexibility. Subsequent to September 11, 2001, rates of unemployment worldwide increased significantly, thereby changing the policy debate considerably. Overall monetary policy in the EU and the U.S. became the focus of attention as central banks sought to engineer an economic recovery. The deregulation of part time work fell rapidly down the list of policy concerns. As evidence, one only needs to examine the most recent National Action Plans listed on the Employment and Social Affairs section of Europa’s web site: (http://europa.eu.int/comm/employment social/news/2001/ jun/napsincl2001 en.html). New policy themes are the following: policies to combat social exclusion, and policies to impact significantly the level of poverty within the EU. The variety of policy approaches to labor market regulation have been great in Europe. Nevertheless, with the establishment of the EMU, pressure in the late 1990s increased throughout Europe to make labor markets more “flexible.” (Our

22

STEPHEN M. HILLS AND TERESA SCHOELLNER

proxy for flexibility is a policy change that eases restrictions on part time work.) In each of the countries most affected by the EMU, we found evidence that restrictions on part time work were being lifted. Was this good policy? In the Netherlands, with low unemployment and strong economic growth, it appeared to be. Only small numbers of Dutch citizens responded that their part time employment was involuntary. In Germany, Italy, and Spain, where the unemployment effects of the EMU were large, relaxation of the rules on part time work occurred slowly – through bargained agreements. Rates of part time work in these countries wee much lower than in the Netherlands. The slow, negotiated change may mean that unemployment continues to be problematic in these countries, but the negotiating partners who change the rules governing part time work may also insure that part time jobs are ultimately considered good jobs and therefore the policy that encourages them, good policy.

NOTES 1. The eleven countries are Austria, Belgium, Finland, France, Germany, Ireland, Italy, Luxembourg, the Netherlands, Portugal, and Spain. Denmark, Sweden and the U.K. opted not to enter the EMU. 2. Part-time work, because of the availability of data, is used as a proxy for other types of institutional reform – the removal of restrictions on the use of temporary work, for example. 3. Alternative specifications, including models scaled to real GDP, and models using the Parks method for analyzing autoregressive time series in a panel excluding the fixed effects were also considered. To test for sensitivity, the model was re-estimated with Denmark as the base. The pattern of results was the same as when the U.K. was used as the base.

ACKNOWLEDGMENTS The authors thank Dustin Burt, Ryan Leszczynski, and Frederic Petipre for their help with this project and OSU’s Center for International Business, Education, and Research (CIBER) for partial funding of the project. The OSU CIBER is supported by a grant from the U.S. Department of Education. Views expressed in this research do not necessarily reflect the views of the U.S. Department of Education.

REFERENCES Aaronovitch, S., & Grahl, J. (1997). Building on Maastricht. In: P. Gowan & P. Anderson (Eds), The Question of Europe (pp. 178–202). London, NY: Verso. European Commission (1999). Employment in Europe 1999. Brussels: European Commission.

European Response to High Rates of Unemployment

23

European Commission (2000). Guidelines for member states’ employment policies for the year 2000. Brussels: Employment and Social Affairs, European Commission. Krueger, A. B. (2000). From Bismarck to Maastricht: The march to European Union and the labor compact. NBER Working Paper 7456. Cambridge, MA: National Bureau of Economic Research (January). Organisation for Economic Cooperation and Development (OECD) (1999). Implementing the OECD jobs strategy: Assessing performance and policy. Paris: OECD. Pisani-Ferry, J. (1998). Monetary Union with variable geometry. In: J. Frieden, D. Gros & E. Jones (Eds), The New Political Economy of the European Monetary Union (pp. 125–161). Lanham, MD: Rowman and Littlefield. Sarfati, H. (1998). Labour market flexibility vs. jobs: The challenge for social dialogue in Europe. Paper presented at the International Industrial Relations Association Meetings, Bologna, Italy (September 22–26). Scarpetta, S. (1996). Assessing the role of labour market policies and institutional setting on unemployment: A cross-country study. OECD Economic Studies, 26, 43–98. Vinals, J., & Jimeno, J. F. (1998). Monetary Union and European unemployment. In: J. Frieden, D. Gros & E. Jones (Eds), The New Political Economy of the European Monetary Union (pp. 13–52). Lanham, MD: Rowman and Littlefield.

APPENDIX: HISTORY OF THE EUROPEAN MONETARY UNION http://cec.caixacatalunya.es/euro/int2eng.html Web site of Caixa Catalunya, a financial institution with a network of 800 branches in Spain. Managing assets of more than 2.5 billion pesetas, it is the third largest savings bank in Spain. 1957 The Treaty of the European Economic Community promoted the coordination of national economic and monetary policies. The member states of the European Community committed themselves to taking exchange rates into consideration as a matter of common interest. 1959 Creation of the convertibility of all the currencies in various European states. 1970 The Werner Plan was proposed to promote the step-by-step approach to economic and monetary union, with a single currency by 1980. However, the major instability of currency in the early seventies, added to divergent economic policies as a result of the first petrol crisis, prevented this materializing.

24

STEPHEN M. HILLS AND TERESA SCHOELLNER

1972 In consequence of the evolution of the Bretton-Woods system, the Council of Ministers of the European Community decided to limit fluctuation between the six currencies of the European Community to a margin of ±2.25%. The U.K., Ireland and Denmark agreed to become part of the “monetary snake,” in anticipation of their joining the European Community in 1973. The pound sterling, the Irish pound, the French franc and the Italian lira withdrew at a very early date. 1979 The European Monetary System (EMS) was founded on Franco-German initiative. Its main features were exchange rate, intervention and credit mechanisms. With the exception of the U.K., the eight member states joined the exchange rate mechanism of the European Monetary System, with set but flexible exchange rates. 1987 The European Single Act for the creation of a Single European Market came into effect on 1 July. The objective of monetary union was laid down for the first time in the Treaty of the European Community. 1989 At the meeting in June, in Madrid, the Council of Europe passed the Delors report outlining a three-stage plan for the founding of the European Economic and Monetary Union. The peseta joined the exchange rate mechanism with fluctuation extended to ±6%. 1990 On 1 July, the Economic and Monetary Union embarked on its first phase. In October, the U.K. joined the exchange rate mechanism with extended fluctuation. 1991 In December, the Heads of State and Government of the European Community decided, in Maastricht, to create the Economic and Monetary Union no later than 1999. 1992 In April, the Portuguese escudo joined the exchange rate mechanism with a broad margin of fluctuation, of 6%. In September, the pound sterling and Italian lira withdrew from the exchange rate mechanism.

European Response to High Rates of Unemployment

25

1993 As a result of the monetary crisis in the EMS, in August the ministers of finance and the presidents of European Union central banks decided to extend the fluctuation of this system to 15% (exception: the German mark -DEM- and the Dutch guilder -NLG-, which continued at 2.25%). 1994 1 January saw the start of the second phase of the European Monetary Union. The European Monetary Institute (EMI) was set up as a forerunner of the European Central Bank (ECB). 1995 Austria joined the exchange rate mechanism at the start of the year. As a result of the weakness of the American dollar, in February and March many European currencies came under pressure from the German block. March saw the devaluation of the peseta and the escudo in the European Monetary System. European Council confirms in Madrid the date of the beginning of EMU’s 3rd phase: 1.1.99 and it changes the name of unique currency, which goes to euro. Also, they set up three stages A, B and C for European Monetary Union: (a) Preparation of EMU, from 1.5.98 to 31.12.98 (b) Beginning of EMU, from 1.1.99 to 31.12.01 (c) Currency duality, from 1.1.02 to 30.6.02 at most. 1996 European Council in Dublin approves the new structure of change process, SME2 or MTC2 and decides the new juridical environment of euro. They present the first design of euro bills. 1997 European Council in Amsterdam approves MTC2, the process of change rates which has to substitute the actual. They present the first euro coins design. The CE autumn rapport about convergence is positive for every country, except for Greece. European Council in Luxembourg in December approves the creation of an informal organ to coordinate economic politics of the 11 countries in the euro area, called Euro-11. Spanish government presents in December the “National Plan for Transition to Euro.” Sweden, Denmark and U.K. announce that they will not accede to euro in the first phase.

26

STEPHEN M. HILLS AND TERESA SCHOELLNER

1998 Greece enters to EMS in March. Special European Council in Brussels at the beginning of May decides the 11 countries which will accede to euro in the first phase, also the ECB managers and the process of determination of euro conversion rates. On the first of July the ECBS initiates its activity and its main organ, the ECB, with central office in Frankfurt. In December Spanish Parliament approves Euro Introduction Law. 1999 (January 1) Introduction of the Euro and monetary union

THE EFFECTS OF USING ALTERNATIVE TYPES OF STAFFING ARRANGEMENTS Cynthia L. Gramm and John F. Schnell INTRODUCTION Traditionally, hiring indefinite duration contract employees has been the dominant method used by U.S. organizations to staff their labor needs. Indefinite duration contract employees, hereafter referred to as “regular” employees, have three defining characteristics: (1) they are hired directly as employees of the organization whose work they perform; (2) the duration of the employment relationship is unspecified, with a mutual expectation that it will continue as long as it is mutually satisfactory; and (3) the employment relationship provides ongoing – as opposed to intermittent – work. When their demand for labor increases, organizations staffed exclusively by regular employees can respond by having their employees work overtime or by hiring additional regular employees. Conversely, when their demand for labor decreases, such organizations can either maintain “inventories” of excess regular employees or reduce labor inputs by laying-off or reducing the work hours of regular employees. “Flexible staffing arrangements” (FSAs), defined by the absence of one or more of the defining characteristics of “regular” employment, contribute to an organization’s flexibility to adjust labor inputs on an as-needed basis (Abraham, 1988). FSAs, together with regular employment, comprise a set of options available to an organization for staffing its work.1 Table 1 defines several common forms of FSAs.

Advances in Industrial and Labor Relations Advances in Industrial and Labor Relations, Volume 13, 27–68 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0742-6186/doi:10.1016/S0742-6186(04)13002-3

27

28

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

Table 1. Definitions of Different Types of Flexible Staffing Arrangements. Type of Staffing Arrangement

Definition

Direct-hire temporaries or short-term hires

Individuals employed directly by an organization under a fixed term contract, for the duration of a project, or to fill positions that are explicitly identified as temporary. Individuals in a pool of workers who are called in to work for an organization on an as-needed basis. Individuals employed by a temporary help agency who perform work for client organizations of the agency. Individuals provided by a labor market intermediary firm (i.e. contractor firm or consulting agency) who perform work for client organizations of the intermediary firm. (Such individuals may or may not be regular employees of the intermediary firm.) Depending on the terms of the contract between the client organization and the intermediary firm, the work may be performed either on the premises of the client organization or on the intermediary’s premises. The client organization’s procurement of a contract with the intermediary firm to provide this form of staffing is referred to as subcontracting or contracting out. Work force employed by an employee leasing organization, which “leases” the work force to a client organization. Self-employed individuals who provide labor or services to client organizations. Includes self-employed consultants.

On-call workers Agency temporariesa Contract firm workers/contractors/job shoppers/agency consultantsa

Leased work forcesa Independent contractors a Individuals

working in these types of FSA may be regular employees of the temporary or leasing agency, the contractor firm, or the consulting agency. These types of staffing arrangements are sometimes referred to as “intermediated” and the contractor firms, temporary help agencies, consulting agencies, and employee leasing firms that employ such workers are referred to as “labor market intermediary firms.”

Many other labels – “contingent” (Freedman, 1985), “market-mediated” (Abraham, 1990), “nonstandard” (Kalleberg et al., 1997), “nontraditional” (Ferber & Waldfogel, 1998), “atypical” (Delsen, 1995), and “alternative” (Polivka, 1996; U.S. Bureau of Labor Statistics, 1995) – have been coined to describe FSAs. Of these, “contingent work” is probably the most widely used. Freedman (1985, p. 35) originally characterized contingent work as an array of staffing options available to the employer that provided for “conditional and transitory employment arrangements as initiated by a need for labor,” which is consistent with our conceptualization of flexible staffing arrangements. We prefer the label “flexible staffing arrangements,” rather than the label “contingent work” for these staffing options because “contingent work” subsequently has come to have so many diverse meanings that its use is confusing. Many define contingent work to include all or

The Effects of Using Alternative Types of Staffing Arrangements

29

Table 2. Definitions of Contingent Workers. BLS definition 1

BLS definition 2

BLS definition 3 Kalleberg, et al., definition

a U.S.

Wage and salary workers who expect their jobs will last for an additional year or less and who had worked at their jobs for 1 year or less. Excludes self-employed and independent contractors.a Workers including the self-employed and independent contractors who expect that their jobs will last for an additional year or less and who had worked at their jobs or were self-employed for 1 year or less.a Workers including the self-employed and independent contractors who do not expect their jobs to last.a All on-call/day laborers as well as workers who: “(1) report their job as temporary; (2) report they cannot work for the employer as long as they wish; (3) are not sure about criteria “1” or “2”; or (4) expect their job to last for only one year or less.”b

Bureau of Labor Statistics (1995). et al. (1997).

b Kalleberg

virtually all work arrangements except work as full-time regular employees. Others characterize contingent work as jobs with a variety of negative attributes, including low prospects for long-term employment, unpredictable hours, and lack of access to benefits. Such broad definitions result in groupings that are too heterogeneous to be theoretically or empirically useful. Even the U.S. Bureau of Labor Statistics’ (BLS) (1995) recent efforts to define contingent work in terms of a single trait – insecure employment – have resulted in multiple definitions that some critics assert are too narrow (Kalleberg et al., 1997). Table 2 summarizes the BLS definitions as well as an alternative definition offered by Kalleberg et al. (1997).2 Following the BLS, in the next section of this paper we will use the term “contingent” to describe a job trait – insecure employment – that may be correlated with working in many types of FSAs. The perception that U.S. organizations increasingly are staffing their work using arrangements other than regular employment has focused scholarly and media attention on FSA usage. Available evidence confirms that U.S. employment in at least one type of FSA, agency temporaries, has grown rapidly and comprises an increasing share of total employment (Golden, 1996; Laird & Williams, 1996; Segal & Sullivan, 1995, 1997a).3 Specifically, from 1972–1995, temporary help supply (THS) industry employment grew at an annual rate of over 11%, compared to a 2% annual growth rate for aggregate nonfarm employment (Segal & Sullivan, 1997a). In the short-run, THS employment growth appears to be influenced by both supply-side factors, such as the entry of married women into the labor force, and demand side factors, such as economic expansion and foreign competition (Golden & Appelbaum, 1992; Golden, 1996; Laird & Williams, 1996). Laird & Williams

30

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

(1996), however, fail to find evidence of any long-run equilibrium relationships between THS employment and such supply- and demand-side factors. Rather, changes in the legal environment may have fueled long-term growth. Careful analyses by Autor (2000a) suggest that THS employment increases significantly in response to a state’s adoption of the implied contract exception to the employmentat-will doctrine, with such changes accounting for as much as 20% of the growth in THS employment from 1973 to 1995.4 Lee (1996) argues that other changes in employment law may have fueled growth in FSA usage, but these hypotheses are yet to be tested. Autor (2000a) also provides evidence of more rapid growth in THS employment in states where union density declined more slowly, although the direction of causality in this relationship is unclear. The unavailability of longitudinal data on employment in other types of FSAs makes it impossible to verify whether they exhibit similar growth trends. Nevertheless, the limited evidence available from surveys of employers is suggestive of growth and expected growth in the use of other types of FSAs (Abraham, 1990; Abraham & Taylor, 1996; Houseman, 2001b; Lautsch, 1998). The growth in the use of FSAs has been controversial. Much of the controversy stems from two contrasting views of the effects of using FSAs, a positive view, which emphasizes the perceived beneficial effects of using FSAs for both organizations and individuals, and a negative view, which emphasizes the perceived adverse effects of using FSAs. Adherents to the positive view argue that using FSAs can contribute to organizational efficiency by facilitating a firm’s ability to adjust its work force in response to fluctuating market conditions and by more efficiently providing screening, selection, and training capabilities that enhance worker-job matches (e.g. Lentz, 1996). They also emphasize that FSAs provide training and work opportunities for individuals whose inexperience, lack of training, need for a flexible schedule, or need for short-term or supplemental income may make regular employment an undesirable or unattainable alternative (e.g. Lentz, 1996). If true, the alleged positive microeconomic effects of growth in the use of FSAs imply positive macroeconomic effects, such as productivity growth. Adherents to the positive view tend to advocate legal changes that accommodate the use of FSAs (e.g. Lentz, 1996). Adherents of the negative view assert that FSAs provide less favorable terms and conditions of employment than regular employment, including lower pay, fewer employee benefits, less desirable working conditions, fewer training and advancement opportunities, and less secure employment. The alleged substandard nature of FSA jobs is thought to be attributable, at least in part, to fact that individuals working in many FSAs either are not covered by, or often will fail to meet, eligibility requirements for protections and benefits provided by U.S.

The Effects of Using Alternative Types of Staffing Arrangements

31

labor and employment laws. If these perceptions are true, at the macroeconomic level, a growing share of the labor force working in FSAs implies a growing share of the labor force working in insecure jobs with “substandard” compensation and working conditions and limited advancement opportunities. Moreover, to the extent that organizations use FSAs to support efficiency wage policies and long-term employment commitments for regular employees, a growing share of the labor force working in FSAs may imply a growing gap between the quality of work and the standard of living experienced by regular employees and FSA incumbents. Adherents to the negative view tend to advocate the adoption of public policies designed to restrict the use of FSAs, to prevent organizations from offering individuals working in FSAs less favorable compensation packages than their regular employees, and/or to extend legal protections granted to regular employees to workers in FSAs (e.g. Carnevale et al., 1998). Both views of flexible staffing arrangements pose a number of empirical questions about the effects of using FSAs. A review of research on such effects can inform policy discussions as well as pinpointing future research needs. Our review will focus on U.S. empirical research published since 1990. The next section will review the relatively extensive body of research on the effects of using FSAs on individuals working in such arrangements. The third section provides a review of the much sparser body of research examining the effects of using FSAs on regular employees in the same organization, on organizational performance, and on macroeconomic performance.

THE EFFECTS OF WORKING IN FSAS ON INCUMBENT WORKERS The dominant prediction associated with the negative view of flexible staffing arrangements is that they provide less desirable terms and conditions in employment than regular employment. This proposition has motivated a sizable body of quantitative empirical research investigating the effects of working in an FSA on incumbents. Most such studies seek to compare employment outcomes for workers in one or more types of FSA to outcomes for some reference group. However, because the categorization of staffing arrangements varies across studies, the reference group does not always correspond directly to our definition of regular employees. Table 3 identifies the data source, categories of FSAs, and reference group used in each of the studies that we discuss below. Pay. The most extensive body of research investigating the effects of working in FSAs on incumbents focuses on the effects of such work on pay and other measures of earnings. Data from recent U.S. household surveys – the Contingent

32

Table 3. Categorization of Staffing Arrangements in Studies of the Effects of Working in FSAs on Incumbents. Data Source

FSA Categories

Reference Category

Allen & Sienko (1998)

Sample of 197 employees of a U.S. telecommunications company who were given the Hackman & Oldham (1998) Job Diagnostic Survey. 1994 U.S. B. L. S. Occupational Compensation Survey of Temporary Help Supply Services – sample of 1033 THS establishments. Telephone interviews of a stratified random sample of temporary help supply establishments in six metropolitan areas. 1990 Biennial Survey of Illinois Registered Nurses 1997 CAEAS

Short-term hires

Regular employees

Agency temporaries

None

Agency temporaries

None

Agency temporaries

Direct hire employeesa

Agency temporaries, independent contractors, contract firm employees, on-call/day-laborers Short-term hires

Standard employeesa

Autor (2000)

Autor, Levy, & Murname (1999)

Bellemore (1998) Belman & Golden (2000)

Chattopadhyay & George (2001) Cohany (1998)

Christensen (1998)

DiNatale (2001)

Ellingson, Gruys, & Sackett (1998)

Surveys of workers in two computer hardware manufacturing organizations. 1997 CAEAS

Survey and interviews of representatives of 21 firms that are members of Boston College’s Work-Family Roundtable. 1999 CAEAS

Survey of 174 agency temporaries employed by a single temporary help supply firm.

Regular employees

Agency temporaries, independent contractors, on-call workers, contract firm employees Agency temporaries, short-term hires, independent contractors/consultants

Workers not in any of the FSA categoriesa

Agency temporaries, independent contractors, on-call workers, contract firm employees Agency temporaries

Workers not in any of the FSA categoriesa

Regular employees

None

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

Author(s)

Data from the Displaced Worker Supplements to the February 1994 & 1996 Current Population Surveys matched to CAEAS in the subsequent years (1995 & 1997).

Feldman, Doerpinghaus & Turnley (1995)

Survey of 186 agency temporaries employed by 7 temporary help supply firms. 1995 CAEAS

Hipple & Stewart (1996)

Houseman (2001b)

Houseman & Polivka (1998)

James (1998) Jarmon, Paulson, & Rebne (1998)

Kalleberg et al. (1997)

Upjohn Institute’s survey of a national stratified random sample of private sector establishments. 1995 CAEAS

Employees & nonemployee contractors in a single firm. Survey of managers in a convenience sample of six client organizations. Responding managers had responsibility of supervising one or more contract company employees. 1995 CAEAS

Kalleberg, Reskin & Hudson (2000)

1995 CAEAS

Kalleberg & Reynolds (2000)

Second National Organizations Survey (NOS-II)

Three categories: (1) independent contractors, consultants, & other free-lance workers; (2) other self-employed workers; and (3) agency temporaries, workers in “temporary jobs,” on-call, & contract workers. Agency temporaries

Regular employees

None

Agency temporaries, independent contractors, on-call workers, contract firm employees Agency temporaries, short-term hires, on-call workers, contract company employees Agency temporaries, on-call workers, contract company employees, short-term hires, & independent contractorsb Non-employee contractors

Workers not in any of the FSA categoriesa

Regular employees

Contract company employees

Regular employees

Agency temporaries, on-call workers, day laborers, and independent contractorsb Agency temporaries, contract workers, on-call/day labor workers, and independent contractorsb Agency temporaries (on-site and off-site), on-call workers, contract company employees (on-site and off-site)

Full-time regular employees

Full-time regular employees

Full-time regular employees

The Effects of Using Alternative Types of Staffing Arrangements

Farber (1999)

Full-time regular employees

Full-time regular employees

33

34

Table 3. (Continued ) Author(s)

Data Source

FSA Categories

Reference Category

Lapidus (1993)

May 1985 Current Population Survey

Temporary help agency employees

Lautsch (1998)

Agency temporaries (worker survey); direct hire temporaries and temporaries employed through intermediaries – agency temporaries or contract company workers (establishment survey) Short-term hires Short-term hires Temporary help agency employees

Polivka, Cohany & Hipple (2000)

Survey of workers at Polaroids’ Digital Products Division & a 1997 establishment survey conducted by the University of Massachusetts Center for Survey Research. 1992 survey of U.S. establishments 1992 survey of U.S. establishments Survey of employees of 20 THS firms in a southeastern city Survey of supervisors and of engineers and engineering technicians who were either employees or contractors working for a large aerospace firm. 1995 & 1997 CAEAS

Non-temporary help agency employees Regular employees

Schur (2003)

(1) 1995, 1997, 1999, & 2001 CAEAS

Lautsch (2000) Lautsch (2003) McClurg (1999)

Rebitzer (1998) Segal & Sullivan (1995)

(2) Survey of Income and Program Participation (3) A nationally representative household survey of people with and without disabilities Telephone survey of workers at petrochemical facilities Current Population Survey

Independent contractors.

Employees.

Agency temporaries, on-call workers, short-term hires, contract company workers, and independent contractorsb (1) Agency temporaries, on-call workers & day laborers, short-term hires, contract company workers, and independent contractors (2) Agency temporaries

Full-time regular employees

Varies depending on the data set used.

(3) Temporary workers

Contract company employees

Direct hire employees

Employment in the Personnel Supply Industry (SIC 736)

Non-SIC 736 employment

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

Pearce (1993)

None None No reference category

Segal & Sullivan (1997b)

Segal & Sullivan (1998)

Wiens-Tuers (1998)

a

1983–1993 outgoing rotations of the Current Population Survey. 1984–94 time series data on sample of individuals taken from the administrative files of the Unemployment Insurance system in state of Washington. 1984–94 time series data on sample of individuals taken from the administrative files of the Unemployment Insurance system in state of Washington 1995 CAEAS

Employment in the Personnel Supply Industry (SIC 736) Temporary services industry employmentd

Non-SIC 736 employment

Temporary services industry employmentd

Employment outside of temporary services industry.

Agency temporaries, contract company workers, day laborers, leased employees, on-call employees, independent contractors

“Standard” employeesc

Employment outside of temporary services industry and unemployment

The author does not indicate whether this category includes short-term hires as well as regular employees. The authors do not distinguish between those FSAs who work part-time and those who work full-time. c Includes all those who do not fall into any of the categories of FSAs listed in the second column and who are not self-employed. Thus, this category presumably includes short-term hires as well as regular employees. d Employed by firms in the Temporary Services industry, which includes leased workers as well as agency temporaries. b

The Effects of Using Alternative Types of Staffing Arrangements

Segal & Sullivan (1997a)

35

36

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

and Alternative Employment Arrangements Supplements (CAEAS) to the Current Population Survey (CPS) – lend themselves to addressing this empirical question.5 Descriptive evidence from the 1995, 1997, and 1999 CAEAS indicates that in comparison to workers with traditional arrangements, median weekly earnings are lower for on-call workers and agency temporaries but higher for independent contractors and contract company workers (Cohaney, 1998; DiNatale, 2001; Hipple & Stewart, 1996). Several studies use the CAEAS data to estimate multiple regression models of the effect of FSA work on wages or other measures of earnings. Polivka et al. (2000) provide evidence that full-time regular employees have higher real hourly earnings than both agency temporaries and short-term hires, but lower real hourly earnings than independent contractors and, in 1997, than contract company workers.6 Kalleberg et al. (1997) estimate separate wage models for men and women. In their base model for both men and women, agency temporaries, on-call workers, independent contractors, and contract company workers all earn significantly lower wage rates than full-time regular workers (Kalleberg et al., 1997).7 When additional controls are added, the pattern of results changes.8 Specifically, whereas wages are significantly lower for female on-call workers than for female fulltime regular employees, there are no apparent wage differentials between female agency temporaries and female full-time regular employees (Kalleberg et al., 1997). Conversely, male full-time regular employees’ wages are higher than the wages of male agency temporaries but are no different than the wages of male on-call workers (Kalleberg et al., 1997). For both women and men, wage rates are significantly higher for independent contractors and for contract company employees than for full-time regular employees (Kalleberg et al., 1997). These studies do not isolate the effects of part-time work from the effects of working in FSAs. As Lee and Johnson (1991) have pointed out, work schedule (i.e. full- or part-time) and type of staffing arrangement are conceptually distinct features of work. Moreover, there is evidence that part-time workers tend to have lower rates of pay (as well as less access to employer-provided benefits than fulltime workers (e.g. Ferber & Waldfogel, 1998, 2000).9 As a result, the failure to clearly isolate the effect of work schedule from the effect of the type of staffing arrangement in empirical analyses may result in statistical bias when estimating the effect of working in an FSA on wages (or benefits). For example, in wage models in which the reference category is full-time regular employees and in which the model does not control for whether FSA employment is full-time or part-time, the coefficients associated with most measures of FSA employment will be biased, thereby overstating or understating the effect of FSA employment.10 Belman and Golden (2000) explore whether the estimated effects of FSA-work on wages are robust to the inclusion of controls for work schedule. They use the

The Effects of Using Alternative Types of Staffing Arrangements

37

1997 CAEAS data to estimate three alternative models of ln(hourly earnings). Their base model (Model 1) augments a standard human capital equation by including indicator variables of four types of FSAs as right-hand-side variables: agency temporary, independent contractor, contract company employees, and on-call/daylabor. Their Model 2 adds a dummy variable for part-time status. Finally, their third model adds controls for major industry and occupational classifications. In general, their findings indicate that hourly earnings are lower for agency temporaries and higher for independent contractors and contract company employees than for the “standard workers;” the hourly earnings differential between on-call/day-labor workers and “standard workers” is not statistically significant.11 A comparison of coefficients on the FSA variables in Belman and Golden’s (2000) Models 1 and 2 provides evidence that failing to isolate work schedule effects from FSA effects can bias the estimation of FSA effects: Controlling for part-time status increases the magnitude of the hourly earnings advantage of independent contractors. In contrast, controlling for work schedule has minimal effect on the estimates of hourly earnings gap between “standard workers” and either agency temporaries, contact company employees, or on-call/day-labor workers. Controlling for the type of staffing arrangement, the part-time work schedule effect is negative, significant and sizable. When industry and occupational controls are added, the hourly earnings differential between contract company employees and “standard workers” is no longer significant and the part-time effect is reduced but still considerable. Belman and Golden’s (2000) results confirm the need to isolate work schedule from FSA effects. Their model assumes, however, that the work schedule effect is constant for each type of staffing arrangement. Estimation of an alternative model that allows the work schedule effect to vary by type of staffing arrangement would be a useful extension of their work. Using data from the 1985 CPS, Lapidus (1993) estimates a wage equation using samples of nonunionized males and females who are not employed by THS firms and then uses its results to obtain a predicted mean wage for samples of males and females employed by temporary help agencies.12 While she finds no difference between the predicted mean wage and the actual mean wage for female temporaries, for male temporaries, the predicted mean wage is $2.21 less than their actual wage. Segal and Sullivan (1995) use the CPS outgoing rotation data for 1983–93 to estimate the wage effects of employment in the Personnel Supply Services (PSS) industry (SIC 736).13 Other things equal, they estimate that, whereas pinkcollar and blue-collar PSS workers earn 10.2% and 34.3% less, respectively, than comparable workers outside that sector, hourly white-collar PSS workers earn 2.4% more than their counterparts in other industries (Segal & Sullivan, 1995).14 In a subsequent paper, Segal and Sullivan (1997a) demonstrate that the percentage wage gap favoring regular employees declines from 21.8 to 7.7% after controlling

38

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

for worker and job characteristics.15 Their estimates suggest a lower wage gap for white-collar than for either blue-collar or pink-collar workers (Segal & Sullivan, 1997a). Because cross-sectional estimates of wage differentials among different types of staffing arrangements may reflect differences in unobserved productivity-related characteristics among individuals employed in different staffing arrangements, investigating how an individual’s wage varies with changes in the staffing arrangement in which she is employed may provide a clearer picture of the effect that working in an FSA has on wages. Three recent papers by Segal and Sullivan (1995, 1997a, 1998) explore whether wage differences apparent in cross-sectional studies are robust when time series data on individuals are used to estimate the wage effects of temporary work. In the first paper, Segal and Sullivan (1995) follow up their cross-sectional estimates of wage differentials between PSS and non-PSS workers, which are described above, by estimating the impact of changing the type of staffing arrangement in which a worker is employed on a worker’s one-year wage growth. Compared to the base case, in which the worker had non-PSS employment in both time periods, for pink-collar workers moving from non-PSS to PSS employment is associated with 4.7% less wage growth, whereas moving from PSS to nonPSS employment status is associated with 9.6% higher wage growth. The wage growth model results for pink-collar workers are of a magnitude consistent with the 10.2% gap estimated in the cross-sectional wage-level models summarized above. For blue-collar workers, changing from non-PSS to PSS employment is associated with 12.2% less wage growth and changing from PSS to non-PSS employment is associated with 14.2% more wage growth than the base case (Segal & Sullivan, 1995). Both of these estimates are of considerably lower magnitude than the 34.3% estimated effect of PSS employment in Segal and Sullivan’s (1995) cross-sectional wage-level models, which suggests that some of the estimated differential in the cross-sectional wage-level models may be attributable to differences between PSS and non-PSS employees in unobserved productivity-related traits. For white-collar workers, changing from non-PSS to PSS employment is associated with 5.75% less wage growth relative to the base case, but wage growth is not significantly different between the base case and changing from PSS to non-PSS employment (Segal & Sullivan, 1995). For all three groups of workers, the wage growth of workers who are PSS employees at both points in time did not differ significantly from the wage growth of individuals who are non-PSS employees at both points in time (Segal & Sullivan, 1995). In their second paper, Segal and Sullivan (1997a) estimate models with workerspecific fixed effects that may influence earnings. These fixed effects control for unobservable characteristics, such as intelligence, education quality, or work ethic,

The Effects of Using Alternative Types of Staffing Arrangements

39

that are constant over time, that are likely to influence wage rates, and that may vary systematically between PSS and non-PSS workers. The wage gap between non-PSS and PSS workers in the fixed effect model is 3.1%, which is substantially lower than the 7.7% gap that they observe in models without individual fixed effects (Segal and Sullivan, 1997a).16 Observing that the disadvantage of estimating models with fixed effects is that unobserved heterogeneity may bias the estimated effects toward zero, Segal and Sullivan (1997a) also estimate specifications correcting for such bias. These estimates suggest that wage rates for PSS workers are about 5% lower than those for non-PSS workers (Segal & Sullivan, 1997a). In a third study using Washington State Unemployment Insurance data in which the career histories of a sample of individual workers are tracked for 11 years (1984–1994), Segal and Sullivan (1998) are able to estimate the wage differential associated with THS employment controlling for worker-specific fixed effects, which capture the effects of individual characteristics that are constant over time, and worker-specific time trends, which capture the effects of worker characteristics that vary over time at a near constant rate (e.g. age, experience). With such controls, they estimate a 10% differential between temporary wages and “reasonable indicators” of the non-temporary wage opportunities that the THS workers might have had (Segal & Sullivan, 1998, p. 13). Observed wage differences across staffing arrangements in the above studies, which at best control only for broad occupational categories, may be attributable to wage differences among occupations. Studies of wage differences across staffing arrangements within a single occupation facilitate distinguishing the effect of staffing arrangements on wages from occupational effects. In one such study using data from the 1990 Biennial Survey of Illinois Registered Nurses, wages are significantly higher for nurses employed through temporary agencies than for those who are direct-hire employees (Bellemore, 1998). Interestingly, the return to tenure is greater for agency temporary nurses whereas the return to experience is greater for direct-hire employee nurses (Bellemore, 1998). Lautsch (1998) estimates wage differentials using data on a single firm’s work force, which includes agency temporaries and regular employees working in the same occupations and receiving identical training and supervision. Controlling for human capital characteristics, race, gender, and job characteristics, she finds that the wage rates of agency temporaries are not significantly different from those of regular employees. Studies of employing organizations regarding how their compensation practices vary across staffing arrangements among workers in comparable positions provide an alternative source of information to data in which the individual worker is the unit of analysis. Houseman (2001b) notes that the majority of respondents to an employer survey report that their on-call workers and short-term hires receive

40

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

hourly pay rates that are about the same as the hourly pay rates of their full-time regular employees in similar positions. In contrast, the majority of employers state that their hourly billed rate for agency temporaries was higher than the hourly pay rate for full-time regular employees in similar positions (Houseman, 2001b). Evidence from case studies suggests that job-level wage differentials between FSAs and regular employees in an organization vary substantially with the organizational context in which the FSAs are used (Erickcek et al., 2002; Lautsch, 2002). Lautsch’s (2002) analyses of four cases suggests that the within-firm wage gap between FSAs and regular employees in similar job classifications is lowest when FSAs are used to facilitate numerical flexibility and when the technology prevents managers from shifting the skill mix of tasks between regular employees and FSAs and that the within-firm wage gap between the two types of workers is highest when FSAs are used to lower costs, regardless of technology. Benefits. For a variety of reasons, employers may not want to offer all of its workforce an identical benefits package.17 Under the U.S. Federal Tax Code’s nondiscrimination (ND) rules, however, benefits qualify for favorable tax treatment (tax-exemption or tax deferral, depending on the type of benefit) only if the firm can demonstrate that the benefit plan does not discriminate in favor of highly compensated employees. The nondiscrimination rules permit firms to exclude workers who are not their employees (e.g. agency temporaries, contract company workers, and independent contractors) as well as seasonal employees, which would include some direct-hire-temporaries and on-call workers, from nondiscrimination calculations, thereby creating an incentive for firms to use staffing arrangements that are excluded from nondiscrimination calculations for workers who are likely to have atypical tastes for benefits relative to their co-workers or who are likely to be outliers in the firm’s pay distribution (Carrington et al., 2002). Additionally, laws that mandate employer-provided benefits (e.g. workers’ compensation laws, the Family and Medical Leave Act) only require organizations to provide such benefits to their employees, which excludes independent contractors from coverage.18 Thus, even in cases where comparable workers in different staffing arrangements receive similar pay rates, there is concern that FSA workers are likely to lack the employer-provided benefits that are commonly part of the regular employee’s compensation package. Segal and Sullivan (1997a) report that PSS workers are less likely than nonPSS workers to have health insurance that is paid for (wholly or in part) by their employer, controlling for worker and job characteristics. Multiple regression analyses of the 1995 CAEAS data suggest that on-call workers are 2.2 times less likely and agency temporaries are 7.1 times less likely to have health insurance from any source than “standard employees” (Wiens-Tuers, 1998). The adverse

The Effects of Using Alternative Types of Staffing Arrangements

41

effect that working in these two types of FSAs has on access to health insurance increases if the worker is black or Hispanic (Wiens-Tuers, 1998). Several studies provide evidence that the odds of having employer-provided health insurance and the odds of having an employer-provided pension plan are lower for each type of FSA studied (i.e. agency temporaries, on-call work, wage and salaried independent contractors, and contract company employment) than they are for full-time regular workers (Kalleberg et al., 1997, 2000; Polivka et al., 2000). Moreover, agency temporaries, on-call workers, short-term hires, and contract company employees are all less likely than full-time regular employees to have health insurance from any source (Polivka et al., 2000). Establishmentlevel evidence indicates that although the vast majority of U.S. establishments provide full-time regular employees with paid vacations and holidays, paid sick leave, pension benefits, and health insurance benefits, only a small percent of establishments provide those benefits for their on-call workers or short-term hires (Houseman, 2001b). Although establishments are less likely to provide health insurance and pensions to short-term hires than to regular employees, the likelihood of providing these benefits for short-term hires is higher when the organization provides the benefits to its regular employees and when the organization has consistent performance objectives for workers in both groups (Lautsch, 2003). Because many FSAs may work part-time and part-time workers are less likely than full-time workers to have benefits, the estimated differences in access to benefits between FSAs and full-time regular employees may be attributable to the part-time status of many FSA workers rather than to the type of staffing arrangement. Belman and Golden (2000) distinguish FSA-effects on benefits from the effects of work schedule. They use the 1997 CAEAS data to investigate the effect of working in four types of FSAs – agency temporary, independent contractor, contract company employees, and on-call/day-labor – on having employer-provided pension coverage, on having health insurance through the employer, and on having health insurance from any source. They estimate three alternative models for each dependent variable, which parallel their three alternative ln(hourly earnings) models described in the previous subsection. The coefficients associated with each of their FSA measures are consistently negative and significant, suggesting that FSA workers are less likely than “standard” workers to have employer-provided pensions, employer-provided health insurance, or health insurance from any source. Their findings also indicate that part-timers are significantly less likely than full-time workers to have any of these benefits. Moreover, controlling for work schedule reduces the magnitude of the adverse effects of working as independent contractors or as on-call/day-laborers on access to benefits. Adding industry and occupational controls produces minimal change in these estimated effects.

42

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

Bellemore (1998) finds that agency-temporary nurses have fewer benefits but higher wages than regular-employee nurses and, moreover, that nurses who value wages highly relative to benefits, especially pensions, are more likely to work for temporary agencies. This suggests that the wage share of total compensation may differ by type of staffing arrangement and that workers may sort themselves among different staffing arrangements based upon their degree of preference for cash compensation relative to benefits. Investigating whether these relationships can be replicated for other occupational groups, especially occupations that are less female-dominated, would be worthwhile. Among establishments employing both regular employees and short-term hires, Lautsch (2000) finds that the likelihood that the establishment provides healthcare benefits to short-term hires is positively related to the existence of a human resources management (HRM) department, to the importance of job ladders and to the age of the establishment, and is inversely related to the importance of seniority rules and the log of the number of regular employees in the establishment. The likelihood that the establishment provides pensions for short-term hires is positively related to the provision of pensions for the establishment’s regular employees, the presence of an HRM department, an efficiency wage policy, whether the establishment is part of a larger organization, establishment age, the importance of job ladders, and the ratio of short-term hires to regular employees; and is inversely related to the importance of seniority rules, the log of the number of regular employees, and the importance of “considering worker needs” (Lautsch, 2000). Evidence from organizational case studies suggests that union contracts and legal mandates may also influence whether organizations provide benefits for their FSA workers (Erickcek et al., 2002). While most studies of the effects of working in FSAs on access to benefits have focused on pension and health insurance benefits, there is also some evidence suggesting that organizations are more likely to provide other types of benefits for regular employees than for FSA workers. In a sample of 21 “family-sensitive” U.S. firms, the incidence of availability of six family-friendly benefits (i.e. flextime, seminars on work family issues, child care location and referral, parent support groups, child care information assistance, elder care/adult dependent care information assistance) is approximately twice as high for regular employees as for short-term hires who, in turn, are more likely to receive these benefits than either agency temporaries and contractors/consultants performing work for these firms (Christensen, 1998). In a second study, Lautsch (2003) finds that establishments are less likely to provide child care for short-term hires than for regular employees. There is some evidence suggesting that the decision to provide benefits for FSA workers varies with a variety of organizational attributes. Lautsch (2000) finds that the likelihood that an establishment will provide day care benefits for

The Effects of Using Alternative Types of Staffing Arrangements

43

short-term hires is higher in establishments that provide day care benefits to the company’s regular employees and in establishments that are part of larger organizations. Providing these benefits also is positively related to the number of high performance work practices used by the establishment (i.e. job rotation, total quality management, quality circles, teams), to the extent to which absenteeism is a serious problem for the company, and to the age of the establishment; and is negatively related to the log of the number of regular employees in the establishment (Lautsch, 2000). Standard of living. Differences across types of staffing arrangements in the terms and conditions of employment are likely to be associated with differences in the standard of living. Wiens-Tuers (1998) finds that both temporaries and on-call workers are less likely than regular employees to live in a home owned by a member of the individual’s household, suggesting that household economic well-being is higher for the latter group. Independent contractors were even more likely than regular employees to live in a home owned by the household (Wiens-Tuers, 1998). In a second study, Lautsch (2003) finds that the likelihood that short-term hires have child care benefits is higher when regular employees have day care benefits, when flexibility considerations motivate the use of short-term hires, and when the organization has consistent performance objectives for both groups. Among female workers, compared to full-time regular employees, the odds of receiving a poverty level wage are higher in two types of FSAs – on-call work and agency temporary work, and lower for independent contractors (Kalleberg et al., 1997). For both men and women, the odds of being paid a wage in the highest quintile of the wage distribution is significantly greater for independent contractors and for contract company employees than for those full-time regular employees (Kalleberg et al., 1997). Kalleberg et al. (2000) find that individuals with full-time regular employment who are not contingent workers are less likely to earn a low wage (defined as a wage rate in the lowest quintile of the 1995 CAEAS sample) than on-call workers, agency temporaries, or independent contractors; among females, the odds of earning a low wage is also higher for full-time, noncontingent regular employees than for contract company employees.19 Employment security. Much of the controversy surrounding the use of FSAs stems from the assumption that work in flexible staffing arrangements is more likely to be insecure or “contingent” than work as a regular employee. Testing this assumption requires individual-level data on both the type of staffing arrangement in which individuals are employed and whether their employment is contingent, with contingent defined such that its definition is not tautological with the definitions of the types of staffing arrangements. Table 4 reports CAEAS estimates for 1995 of contingent workers as a percent of employment by type of staffing arrangement, using the three alternative BLS definitions of contingent workers

44

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

Table 4. Contingent Workers as a Percent of Employment by Type of Staffing Arrangement. Definition of Contingent Worker

BLS Definition 1 BLS Definition 2 BLS Definition 3

Contingent Workers as % of Total Workers Employed by Type of Staffing Arrangement All Staffing Arrangements

Agency Temporaries

On-Call Workers

Contract Firm Employees

Independent Contractors

Nonalternative Employment Arrangements

2.2 2.8 4.9

39.4 48.0 66.5

17.6 18.0 35.2

7.7 11.7 19.8

Not applicable 3.8 3.8

1.6 1.8 3.6

Note: Percent of total employed estimates are taken from Polivka (1996) using the BLS’ 1995 CAEAS data.

summarized in Table 2. While none of these definitions may be ideal, one can draw some tentative conclusions when the pattern of results is consistent across different definitions of contingent work. In general, the descriptive statistics reported in Table 4 challenge the assumption that the work provided by FSAs is necessarily contingent. Interestingly, contingent workers comprise a minority of the workers employed as on-call workers, as independent contractors, and as contract company employees and comprise a majority of agency temporaries only when the broadest definition of contingency is used. Moreover, contingent work is much more common in some types of FSAs – agency temporary work and on-call work – than in others. Consistent with the proposition that FSAs provide less employment security than regular employment, workers in “nonalternative staffing arrangements” are, for the most part, less likely to be contingent than agency temporaries, on-call workers, workers provided by contract firms, or independent contractors. Because the “nonalternative staffing arrangements” category includes short-term hires as well as regular employees, the estimate of the share of workers in this category who are contingent is likely to overstate the percent of regular employees who are contingent.20 Subsequent research using the CAEAS data explores whether the relationships apparent in these descriptive statistics are robust to regression analyses controlling for other factors likely to influence whether an individual has a contingent job. Specifically, in a model using the third and most inclusive BLS contingent work measure as a dependent variable and controlling for human capital, personal characteristics, industry, and occupation, Wiens-Tuers (1998) finds that relative to “standard” employees, on-call employees are 8 times more likely to be contingent, temporaries are 50 times more likely to be contingent, and day laborers are 155 times more likely to be contingent. Moreover, being black or Hispanic increases the odds of a temporary being in contingent work (Wiens-Tuers, 1998).

The Effects of Using Alternative Types of Staffing Arrangements

45

Houseman and Polivka (1998) use multivariate analyses of the CAEAS data to investigate differences across types of staffing arrangements in employment stability. Their key findings are that: (1) on-call workers, direct hire temporaries, and contract workers are more likely than full-time regular employees to switch jobs, to become unemployed, and to leave the labor force within a year; (2) agency temporaries are more likely than full-time regular employees to switch jobs and to become unemployed within a year; and (3) there are no significant differences between independent contractors and full-time regular employees in the likelihood of switching employers, becoming unemployed, and leaving the labor force within a year (Houseman & Polivka, 1998). Taken together, these findings suggest work in three types of flexible staffing arrangements – agency temporaries, on-call work, and contract company employees – provide less employment stability than full-time regular employment; conversely, there appears to be little difference between full-time regular employees and independent contractors in the degree of employment instability. Although Houseman and Polivka (1998) control for a variety of demographic traits in their analyses, they caution that other unobserved personal characteristics may account for the observed differences between full-time regular employees and on-call workers, direct hire temporaries, and contract workers. Moreover, their findings do not isolate the effect of FSA vs. regular employment from the effect of full-time vs. part-time work. Training. A related concern is that work in FSAs provides few opportunities for human capital development, thereby limiting opportunities for workers in FSAs to obtain future employment as regular employees. Most research investigating the training consequences of FSA work focuses on agency temporaries. Recent surveys suggest that a majority of THS establishments provide training to their employees (Autor, 2000b; Autor et al., 1999). The incidence of providing training, however, appears to vary by occupation, with workers in clerical occupations being the most likely to receive training (Autor et al., 1999). The goal of training for industrial and blue collar THS workers is more likely to be to help them qualify for an assignment, whereas for other occupations (i.e. managerial and professional, clerical, and technical workers) the goal of training is more likely to be the advancement of existing skills (Autor et al., 1999). Training provided by THS firms is necessarily general (Autor, 2000b) and typically is provided prior to the employee’s placement (Krueger, 1993). These stylized facts run counter to the common perception of THS firms as providers of jobs with limited training opportunities, as well as to the human capital model, which predicts that workers pay for general training. To explain this, Autor (2000b) proposes a model in which THS firms provide free, up-front general skills training to induce self-selection of high-ability employees, where ability is defined

46

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

as the motivation and ability to learn job skills. Autor’s (2000b) model assumes that training is more productive and valuable to high ability workers, that workers have some prior (albeit incomplete) private information about their ability, and that providing skills training enables a THS firm to acquire information about worker ability that in the short-run is unavailable to other firms. The model predicts that workers who perceive their ability as high will seek work with THS firms offering a package of initially lower wages and up-front general training in expectation of future wage gains in permanent employment, while workers who perceive their ability as low will avoid such firms because of the lower initial wages and their limited expected gains from training (Autor, 2000b). The THS firms offering the low initial wage/up-front training package profit from their training investment via the short-run informational advantage about worker ability that they derive from providing training (Autor, 2000b). Autor’s model predicts that training THS establishments will pay their employees lower wages and charge their clients higher mark-ups than non-training THS establishments. As hypothesized, Autor (2000b) finds lower wages at THS establishments providing training than at those which do not, other things equal. Using a smaller sample, Autor et al. (1999) observe higher mark-ups at training than nontraining THS establishments but comparable wages at both types of establishments despite higher quality workers at THS establishments providing training. This result, although weaker, is still consistent with the notion that it is the training THS firm, rather than the worker, which profits from the higher labor market value of its workers during their stint as temporaries. Second, consistent with the prediction that training will be rationed by ability, hiring selectivity is significantly higher at THS establishments providing computer training, and workers with better employment credentials are more likely to have access to training in THS firms (Autor et al., 1999). Third, although direct evidence of the hypothesis that individuals who have worked at training THS establishments will have steeper wage profiles than those who have worked at non-training THS establishments is unavailable, estimates that permanent job placement rates are 60–80% higher for temporaries at training establishments than for their counterparts at non-training establishments provide indirect support for this hypothesis (Autor, 2000b; Autor et al., 1999). Finally, market competition produces higher THS wages, especially at training THS establishments (Autor, 2000b). Evidence exploring whether workers in other types of FSAs have less access to employer-provided training than regular employees is limited. In one case study of a single firm, although comparable amounts of training were provided to regular employees and workers in FSAs, regular employees both expressed higher training needs and reported their jobs as more skill-enhancing than workers in

The Effects of Using Alternative Types of Staffing Arrangements

47

FSAs (Lautsch, 1998). Additional case study evidence suggests that FSA workers are likely to receive the same training as regular employees in the same job category and firm when technology prevents managers from shifting the skill mix of tasks between regular employees and FSAs (Lautsch, 2002). When technology enables managers to shift the skill mix of tasks among the two types of workers, however, the training deficit of the FSA workers relative to their regular employee counterparts is higher when FSAs are used to generate cost savings than when they are use to achieve numerical flexibility (Lautsch, 2002). Further research is warranted on how organizations allocate training opportunities among individuals working in different types of staffing arrangements and with different ability levels, on what factors influence those allocation decisions, and on the implications of such allocation decisions for workers’ career paths and wage profiles. Three types of data would be particularly helpful in addressing such questions: (1) case studies of a single organization in which the individual workers, whose staffing arrangements, access to training, and career outcomes may vary, are the unit of analysis; (2) samples of organizations in which one observes each organization’s policies and practices with respect to training workers in different types of staffing arrangements; and (3) longitudinal data on individuals. Other unanswered questions include: Are the training practices of non-THS employers and the outcomes of those practices comparable to those of THS employers? Given the widespread provision of basic general skills training by THS establishments, do THS employers provide employment and training opportunities to workers who otherwise would have limited access to such opportunities? How do the wage profiles and career paths of FSA workers with experience at training firms compare to those of FSA workers with experience at non-training firms? How does the practice of using FSAs to screen workers for positions as regular employees influence training opportunities and career paths for such workers? Advancement opportunities. Given available evidence that many FSA jobs provide employment conditions that appear to be inferior to those in regular employment, it becomes important to assess whether jobs in such FSAs are “deadends” or bridges to jobs with better terms and conditions of employment. Evidence from surveys of firms and from case studies suggests that some firms use FSAs to screen workers for regular employment (Abraham, 1988; Abraham, 1990; Erickcek et al., 2002; Houseman, 2001b). Case studies suggest, moreover, that absent such screening arrangements some employers would be unwilling to consider workers perceived as risky hires (e.g. those with criminal records or poor work histories) for regular employment (Erickcek et al., 2002). Thus, FSA work may provide some of the least marketable members of the labor force with opportunities they would not otherwise have to be considered for regular employment.

48

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

Few studies examine the role of FSA work as transitional employment. Using matched CPS data (DWS and CAEAS), Farber (1999) estimates the extent to which “alternative employment arrangements” are responses to job loss. He finds that workers who lose their jobs are more likely subsequently to be employed in temporary or part-time work (involuntarily) than are non-losers and that such jobs are part of a transitional process back to full-time regular employment. Based on analyses of multiple data bases, Schur (2003) concludes that FSAs offer employment opportunities for disabled workers whose health might prevent them from holding regular jobs and, for some disabled workers, may facilitate transitions to regular employment. Segal and Sullivan’s (1995) analyses of U.S. CPS data suggest that less than one-third of temporary workers were still temporaries a year later and more than half were in permanent positions. Observing that whether employment as an agency temporary is dead-end employment or a bridge to a job in a more desirable form of staffing arrangement depends upon whether employment spells as agency temporaries are long or short, Segal and Sullivan (1997b) use Washington State Unemployment Insurance (UI) data in which career histories on a sample of individual workers are tracked for 11 years (1984–1994). Their analyses suggest that for most individuals, spells of work as an agency temporary tend to be brief – the average duration is only about two quarters with 75% of temporary employment being accounted for by spells of a year or less, and, for a majority (58%), such spells end with a transition to permanent work (Segal & Sullivan, 1997b). Nevertheless, for a sizable minority, spells of temporary work do not end with a transition to permanent work. Because the Washington UI data do not have any demographic or occupational information, the authors are not able to examine whether such traits influence the duration of temporary employment spells or the likelihood that a spell of work as an agency temporary will lead to regular employment. This would be a fruitful topic for future investigation. The limited evidence available to date suggests that FSA work provides bridges to regular employment or transitional employment between spells of regular employment for many workers. Nevertheless, a sizable minority of FSA workers appears to be “stuck” in FSA jobs. This raises an empirical question that warrants further investigation: what factors facilitate a successful and speedy transition to regular employment? Although common sense suggests that individual traits are likely to influence such transitions, there is some evidence that organizational factors also play a role. Using a sample of establishments, Lautsch (1998) finds that the percent of temporaries later hired as regular employees by the establishment is positively related to the importance to management of flexibility relative to cost considerations and inversely related to the competitiveness of the environment and to the presence of a union. Analyses of four case studies suggest that the withinfirm mobility from FSA jobs to regular employment is highest when FSAs are

The Effects of Using Alternative Types of Staffing Arrangements

49

used to facilitate numerical flexibility and when the technology prevents managers from shifting the skill mix of tasks between regular employees and FSAs and is lowest when FSAs are used to achieve cost savings (Lautsch, 2002). Union representation. U.S. labor law is geared to providing union representation and collective bargaining rights to individuals who are employees. Not surprisingly, descriptive data suggest that full-time, regular employees are more likely to have union representation than employees in most flexible staffing arrangements (Kalleberg et al., 1997, p. 49). On-call workers are an exception to this generalization; they are slightly more likely than full-time regular employees to be union members or covered by a collective bargaining agreement (Kalleberg et al., 1997, p. 49). Noting a variety of potential benefits to unions from organizing individuals in FSAs, Stone and Gallagher (1997) hypothesize that both on-call workers and agency temporaries would probably be more receptive to organizing efforts than independent contractors, with on-call workers being the easiest of the three types of workers to organize. They note further that both on-call workers and agency temporaries might be good candidates for associate union membership whereas independent contractors might be receptive to occupational unionism (Stone & Gallagher, 1997). Their conceptual analysis suggests a need for careful empirical analyses of the impact that the type of staffing arrangement in which an individual works has on his or her attitude toward unions and desire for union representation. Other job attributes. The type of staffing arrangement in which an individual works also may have implications for the worker’s physical well-being. U.S. rules for assigning legal liability for the costs of job-related accidents deter client organizations from providing employees of contract firms working at the client site with safety-related training and supervision (Rebitzer, 1998). Arguing that FSAs typically are less experienced and skilled; less familiar with equipment, people, and norms at a work site; and receive less safety-related training than regular employees, Rousseau and Libuser (1997) hypothesize that FSAs are at greater risk of on-the-job accidents, injuries, and deaths than regular employees. Rousseau and Libuser’s (1997) review of the largely descriptive evidence from two high-risk industries – coal mining and petrochemicals – provides preliminary support for this hypothesis. Rebitzer’s (1998) analysis of the determinants of accidents among contract and direct-hire employees working in maintenance jobs in the petrochemical industry provide even stronger support: other things equal, the incidence of accidents was significantly and substantially higher among contractor employees than among direct-hire employees. Rebitzer’s (1998) results also suggest that closer safety supervision of contract employees by the client organization – proxied by a requirement that contract employees report onthe-job accidents, injuries, and safety problems to the client organization as

50

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

well as to the contracting organization – significantly reduces the incidence of contract employee accidents. Consistent with the view that safety training by contractors may be less than adequate, Rebitzer (1998) finds no evidence of a relationship between contractor-provided safety training and the incidence of contract employee accidents at client organization sites. Alternatively, individuals who take jobs as contract employees may be inherently more difficult to train than those who take jobs as direct-hires (Rebitzer, 1998). Chattopadhyay and George (2001) report evidence that short-term hires perceive their jobs as having a lower level of prestige than regular employees. Additionally, both the degree and nature of control exerted over workers may vary by type of staffing arrangement. Gottfried (1991) describes the control mechanisms developed by four temporary help firms and hypothesizes that agency temporaries will be subjected to different methods of control than regular employees. Although Gottfried (1991) lacks the control group of regular employees needed to test his hypothesis, James’ (1998) finding that workers who are employees are subject to less managerial control than non-employee contractors provides support for the general proposition that control mechanisms differ for the two types of workers. Attitudes. The assumption that the work-related attitudes and behavior of workers in FSAs will differ from those of regular employees permeates the literature on FSAs. On one hand, FSAs may attract workers from different populations and with correspondingly different attitudinal and behavioral tendencies than regular employment. On the other hand, the nature of work in FSAs may provide different psychological experiences for workers than regular employment, resulting in attitudinal and behavioral differences between workers in FSAs and regular employees. Lee and Johnson (1991) find that temporary employees had significantly higher organizational commitment and job satisfaction than regular employees in the same organization. Tests of individual hypotheses in Lee and Johnson’s model generally yield stronger results for temporary than for permanent employees. Allen and Sienko (1998) report that regular employees evince higher satisfaction with job security than short-term hires in similar jobs in the same organization, but observe no differences in the mean levels of pay satisfaction reported by the two types of employees. The attitudes of FSA workers, like those of regular employees, vary with other attributes of the individual and the job. In support of their general hypothesis that organizational commitment and job satisfaction are influenced by both the individual’s work schedule (full- or part-time) and whether it is the individual’s preferred work schedule, Lee and Johnson (1991) find that full-time temporary employees have higher organizational commitment than part-time temporary employees when both work a preferred schedule, but have lower organizational

The Effects of Using Alternative Types of Staffing Arrangements

51

commitment and lower job satisfaction than part-time temporary employees when both work an unpreferred schedule. Using data from a survey of agency temporaries, Feldman et al. (1995) find that such temps report higher satisfaction with several aspects of their work when they are temps by choice, are placed in positions consistent with prior experience and education, and are not trying to convert to more permanent employment. Similarly, Ellingson et al. (1998) find that pay satisfaction and satisfaction with temporary work are higher among voluntary than among involuntary agency temporaries. Taken together, these findings suggest that the work-related attitudes of individual’s working in FSAs may be strongly influenced by whether an individual is working in his preferred type of staffing arrangement and whether other characteristics of the job (e.g. work schedule) are compatible with the individual’s human capital investments and preferences. Organizations that staff the same jobs with a mixture of similarly-qualified FSAs and regular employees provide ideal settings for exploring the effect that working in an FSA has on worker attitudes or behavior. Findings from two studies of this type fail to support the conventional wisdom that regular employees will exhibit greater organizational commitment and more extra-role behaviors than workers in FSAs (Lautsch, 1998; Pearce, 1993). Gallagher and Parks (2001) observe that the various foci (i.e. employment, occupational, job, and organizational commitment) of work commitment may have less relevance or different meanings for FSA workers than for regular employees (Gallagher & Parks, 2001). Although the concept of employment commitment is likely to be relevant for workers in FSAs, as well as for regular employees, one might expect the level of employment commitment to be lower for most FSAs than for regular employees, especially among FSA workers who prefer FSA jobs over regular employment. Depending on the type of FSA, the concepts of occupational commitment, job commitment, and organizational commitment may be either inapplicable or have different meanings for FSA workers (Gallagher & Parks, 2001). For example, for workers employed by labor market intermediaries, it might be appropriate to adapt the concept of dual (organizational) commitment (to the employing organization and the client organization) that has successfully been applied to unionized workers. Thus, it may be necessary to rethink or modify these concepts of commitment when applying them to FSA workers. Consistent with this view, McClurg (1999) finds that the Steers (1977) model of organizational commitment, which has been widely used in non-THS settings, performed poorly when tested using a sample of employees of THS firms. Similarly, the concept of job satisfaction and its various components may be irrelevant or have different meanings for various types of FSA workers than for regular employee. For example, application of the concept of job satisfaction to employees of labor market intermediaries requires careful consideration of how

52

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

to define the job. Specifically, is the “job” defined by the worker’s “job” as an employee of the labor market intermediary or by the tasks and attributes of the work the worker performs for a particular client organization? Which definition of job is used is likely to have an impact on both measurement and results. Worker performance. In one study comparing the work performance of similarly situated FSAs and regular employees, managers of high-tech organizations who supervised both contractors and regular employees in professional jobs perceived overall contractor performance to be at least as good as that of regular employees (Jarmon et al., 1998). While this suggests that using FSAs is unlikely to have either positive or negative effects on organizational performance, it would be useful to test whether this result applies to other professions, to workers in nonprofessional occupations, and to workers in other types of FSAs. Moreover, there is some evidence that the performance of FSA workers is likely to depend upon job characteristics and workers’ satisfaction with their job. Specifically, Ellingson et al. (1998) estimated a model of agency temporary performance, as rated by a representative of the client firm in which the temporary completed the job assignment. Their findings suggest that an agency temporary worker’s performance is positively and significantly related to assignment length, to the temporary’s overall assignment satisfaction, and to the temporary’s satisfaction with temporary work.21 Drawing on interviews with a small sample of agency temporaries, Rogers (1995) provides anecdotal evidence that agency temporaries experience alienation from their work, from co-workers in both the temporary agency and the client organization, and from themselves. The absence of a control group of nontemporaries, however, suggests caution in attributing the alienation to temporary work. Using a sample of temporary and regular employees working in the same jobs for a single organization, Lautsch (1998) finds no differences between temporaries and regular employees in work-family conflict.

OTHER EFFECTS OF USING FSAs The effects of using FSAs are unlikely to be limited to incumbents. FSA usage may also influence the attitudes, terms, and conditions of employment of regular employees working in the same organization. If FSA workers differ systematically from regular employees in their attitudes and behavior; if organizations systematically treat FSA workers differently than regular employees; or if the cost of utilizing of FSA workers differs from the cost of utilizing regular employees, controlling for productivity; then the use of FSAs may have an impact on organizational performance. Similarly, if FSAs systematically experience

The Effects of Using Alternative Types of Staffing Arrangements

53

different terms and conditions of employment than regular workers or if FSAs influence organizational performance, then the share of the labor force working in FSAs is likely to have an impact on macroeconomic outcomes. While there are a very limited number of studies of the effects of FSA usage on regular employee co-workers, on organizational performance, and on macroeconomic outcomes, including them in our review can assist in pinpointing future research needs. Effects on regular employee co-workers. Surveys of organizations indicate that some firms use FSAs in order to buffer regular employees against demand fluctuations, which suggests that using FSAs may enhance a firm’s ability to make long-term commitments to regular employees (Abraham, 1988; Abraham, 1990; Houseman, 2001b; Kahn et al., 2001). Gramm and Schnell (2001) find that the extent to which an organization relies on temporary lay-offs for regular core employees is inversely related to the use of in-house flexible employment relationships in core jobs. Wiens-Tuers (2001) finds that an establishment is more likely to have a commitment to its core regular employees if it uses short-term hires; in contrast, the use of “indirect and intermediate employment relationships,” which include agency temporaries and contract employees, is not significantly related to the likelihood of a commitment to core regular employees.22 Thus, while both of these studies support the view that FSAs are used to buffer regular core employees against demand fluctuations, the latter study suggests that short-term hires are more likely to be used for this purpose. The use of FSAs may also influence advancement opportunities for regular employees. Barnett and Miner (1992) find that depending upon the characteristics of the internal labor market in which they are used, hiring temporaries can either prolong or reduce the length of time that regular employees wait for promotion. Specifically, they find that hiring temporary workers into jobs that would otherwise provide promotion opportunities for regular employees delays promotions for regular employees, whereas hiring temporary workers into entry-level jobs in internal labor markets with few such entry portals hastens promotions for regular employees. Working along side of FSAs may also influence the attitudes of regular employees. Pearce (1993) finds that regular employees in work units with contractor co-workers reported lower trust in their organization than regular employees performing similar work in employee-only work units. Davis-Blake et al. (2003) find that using temporary workers increased intentions to quit among regular employees in the same job and decreased their perceptions of the quality of their relations with managers, their loyalty, and their desire to unionize. The magnitudes of the temporary worker usage effects on loyalty and relations with manager were lower for high wage regular employees; however, having a high wage did not moderate the effects of temporary usage on regular employees’

54

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

quit intentions or desire for union representation (Davis-Blake et al., 2003). Additionally, the effects of temporary usage on exit intentions and desire to unionize were moderated by the degree of training and supervisory responsibilities of the regular employee (Davis-Blake et al., 2003). The use of contract company employees to perform work associated with the jobs of regular employees also had a negative effect on regular employees’ perceptions of the quality of their relations with managers, although this effect was not moderated by the wage level of regular employees and its magnitude was smaller than the magnitude of the effect of using temporary workers (Davis-Blake et al., 2003). Using contract company employees did not significantly affect regular employees’ exit intentions, loyalty, or desire to unionize (Davis-Blake et al., 2003). Contrary to their hypothesis, Chattopadhyay and George (2001) find that among regular employees, greater work status dissimilarity (i.e. dissimilarity to co-workers in terms of the type of staffing arrangement) is associated with a more positive attraction to co-workers. In work groups comprised of both short-term hires and regular employees but dominated by short-term hires, work status dissimilarity had a negative effect on the organization-based self-esteem, trust in and attraction to peers, and altruism toward peers of regular employees (Chattopadhyay & George, 2001). The use of FSAs appears to influence the job content of their regular employee co-workers. Pearce (1993) finds that when some jobs in a work unit are staffed by contractors, tasks requiring interdependence among co-workers are shifted from the contractors to regular employees. Supervisors reported during interviews that they tried whenever possible to assign simpler and more easily monitored tasks to contractors (Pearce, 1993). In a study applying the Hackman and Oldham (1980) job characteristics model to a sample of regular employees and short-term hires employed in the same types of professional and technical jobs within a single organizations, Allen and Sienko (1998) find differences between the two groups of workers in two of the five core job characteristics identified by the model. Specifically, compared to regular employees, shortterm hires reported higher levels of task identity and higher levels of feedback from job (Allen & Sienko, 1998).23 These results suggest that even when regular employees and FSAs work side-by-side in the same jobs, there may be subtle differences in the tasks they perform and, in particular, that organizations may try to allocate tasks such that the work of FSAs is either more easily monitored or self-monitoring than than the work of regular employees who are ostensibly performing the same jobs. Such subtle differences may account, at least in part, for an organization’s decision to treat FSAs differently than regular employees and for attitudinal and behavioral differences between FSAs and regular employees.

The Effects of Using Alternative Types of Staffing Arrangements

55

The effects of FSAs on the terms and conditions of employment, attitudes, and behavior of regular employee co-workers is unlikely to be uniform across organizations. Evidence from organizational case studies of the use of agency temporaries and contract company employees in low-skilled jobs by auto suppliers, hospitals, and public schools conducted by Erickcek et al. (2002) suggest that the effects of FSA work on regular employees’ wages, benefits, and working conditions vary considerably with the organizational context. They find stronger adverse effects when the FSAs are long-term substitutes for, and receive lower compensation than, regular employees (Erickcek et al., 2002). Organizational performance. Conceptual and theoretical work by organizational scholars suggests that the use of FSAs and the context in which they are used can significantly affect organizational performance (e.g. Matusik & Hill, 1998; Pfeffer & Baron, 1988). A small number of recent studies have begun to explore these effects empirically. In one recent study, Kahn (2000) couples findings based on in-depth interviews with a small sample of HR executives regarding how growth in the use of temporary workers influences profitability and/or productivity, with an industry-level analysis of a broader category of “nonstandard workers.” Both of Kahn’s (2000) analyses indicate that organizations made increasing use of flexible staffing arrangements in the mid- to late-1990s in order to adapt to variability in financial performance. Additionally, Kahn (2000) reports that increases in nonstandard worker usage are positively correlated with productivity but not significantly correlated with profitability. Using data on a small sample of firms (n = 35), Kahn et al. (2001) estimate simple correlations between temporary usage and several measures of organizational performance. The share of temporaries in the firm’s workforces is positively correlated with the change in the share price of the firm’s stock over the five year period and the change in primary earnings per share (Kahn et al., 2001). A measure of strategic changes in the use of temporaries is positively correlated with various profitability measures. Their findings, although exploratory in nature, provide some support for the notion that organizations can use FSAs as part of a staffing strategy designed to enhance organizational financial performance (Kahn et al., 2001). Research by Bendapudi and Ash (1999), moreover, suggests that the effect of using FSAs on organizational performance depends on degree of fit between the choice of staffing arrangements and a firm’s competitive strategy. A recent article by Lepak et al. (2003) explores the effects of four “employment modes” – knowledge-based employment, job-based employment, contract work, and alliances/partnerships – on two measures of firm performance, return on equity (ROE) and market-to-book (RTB). Although they do not define these four employment modes in terms of flexible staffing arrangements, we infer from their

56

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

definitions that knowledge-based employees would consist of regular employees with (implicit or explicit) contracts providing long-term employment with considerable training and development opportunities; that job-based employees would consist of regular employees whose contracts provide less security, training, and development as well as short-term hires and on-call workers; that contract workers would include independent contractors, agency temporaries and consultants; and that alliances/partnerships would include contract firm employees and leased workers. Both of their measures of firm performance increase with the degree of reliance on knowledge-based employment and the degree of reliance on contract work, but are not significantly influenced by the degree of reliance on the other two employment modes (Lepak et al., 2003). Additionally, their findings indicate that how employment modes are combined influences firm performance. Specifically, firms using a mix of knowledge-based employment and contract work perform better than firms relying on only one or neither of these employment modes (Lepak et al., 2003). Moreover, the relationship between reliance on a job-based employment mode and firm performance was positive when the firm also relied heavily on contract work but negative when firms relied less heavily on contract work (Lepak et al., 2003). Keeping in mind the caveat that the links between “employment modes” and categories of staffing arrangements are imprecise, these findings suggest that using a mix of regular employees combination with non-employee forms of FSAs may be desirable. Finally, Lepak et al. (2003) provide some evidence of interaction effects between employment modes and technological intensity. FSA usage can also influence organizational performance if it raises or lowers the cost of doing business relative to the use of regular employees. In contrast to the many individual-level studies comparing the wages and benefits of FSA workers with those of regular employees, there are few organizational-level studies comparing labor costs for FSAs and regular employees. In one such study, Park and Butler (2001) find that one component of labor costs – workers’ compensation costs – are three times greater for leased workers than they are for regular fulltime workers. This result is consistent with evidence of higher injury rates for contract firm employees than for regular workers (Rebitzer, 1998; Rousseau & Libuser, 1997). Park and Butler (2001) attribute their finding mainly to higher claim frequency rates for leased workers. Because the leasing agency is unable to monitor the safety behavior of its employees, and because these workers typically enjoy benefits packages that are inferior to that of regular employees, Park and Butler (2001) argue that leased workers are prone to “claims-reporting moral hazard.” Alternatively, the differential workers’ compensation costs between these two groups may reflect differences in safety-related training provided for or human capital (i.e. trainability) differences between leased workers and regular employees.

The Effects of Using Alternative Types of Staffing Arrangements

57

Macroeconomic outcomes. If employment in some forms of FSAs provides less employment stability than regular employment, then growth in the share of the labor force working in those FSAs is likely to decrease aggregate job stability. Simulations suggest that the rapid growth in U.S. agency temporary employment over the 1986–1987 to 1995–1996 period resulted in a modest decline in aggregate job stability (Houseman & Polivka, 1998). Kahn (1998) finds evidence that the magnitude of decreases in an occupation’s interindustry wage differential increases with the share of employment in an occupation that is self-employed or contracted out. Her analyses suggest, however, that only a small portion of the fall in interindustry wage differentials occurring during the late 1980s and the 1990s are attributable to this relationship. Growth in the share of employment in FSAs also may have implications for the interpretation of macroeconomic indicators. For example, Segal and Sullivan (1995) estimate that half of the measured job loss in manufacturing employment in the early nineties may be attributable to the increased share of leased and temporary workers working in industrial jobs. They also note that the accuracy of estimates of wage inflation and productivity in manufacturing may be affected by the growing share of temporary and leased workers who are working in manufacturing jobs, but are counted as PSS employment in government statistics (Segal & Sullivan, 1995). Similarly, Katz and Krueger (1999) hypothesize that improvements in labor market competition and the efficiency in job matching resulting from the rapid growth of private sector labor market intermediaries, especially temporary help agencies, may help explain the unusual combination of low inflation and low unemployment in the 1990s. Their empirical analyses suggest that states with a greater share of agency temporary employment at the start of the decade experienced lower wage growth in the 1990s (Katz & Krueger, 1999). They further estimate that the temporary help industry may account for a decline of 0.39 percentage points over the past decade in the level of unemployment consistent with stable inflation (the NAIRU) (Katz & Krueger, 1999). Although these findings are, in the authors’ words, “preliminary (and quite speculative),” they suggest that growth in the temporary help sector may explain as much of the decline in unemployment as changes in the demographic composition of the labor market (Katz & Krueger, 1999).

CONCLUSIONS AND RECOMMENDATION FOR FUTURE RESEARCH During the last decade and a half, there has been an explosion of U.S. empirical research investigating the effects of using FSAs. Our review has focused on

58

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

studies investigating the effects of using FSAs on individuals working in such arrangements, on regular employees in the same organization, on organizational performance, and on macroeconomic outcomes. Of these, the body of literature investigating the effects of FSA work on incumbents, especially on their pay, is the most extensive and well-developed, although there is still ample need for further research on such effects. The bodies of literature investigating the effects of using FSA workers on regular employees working in the same organization, on organizational performance, and on macroeconomic outcomes remain quite sparse and offer numerous avenues for future research. Effects of Working in FSAs on Incumbent Workers. Much of the research on this broad topic investigates the compensation effects of working in various types of FSAs, especially agency temporary work. One result is relatively robust across studies using cross-sectional as well as longitudinal data on individuals and after controlling for the effect of work schedule: agency temporaries on average earn lower hourly wage rates than regular employees. The relatively rare exceptions occur in studies of regular employees and agency temporaries working in similar or identical jobs in a single organization, studies of a single profession, or analyses in which the sample is limited to female workers, suggesting a clear need for occupation-specific, organization-specific and gender-specific investigations of this gap. The bodies of research on wage differentials between regular employees and workers in other types of FSAs and on the nonwage effects of FSA work are sparser. Nevertheless, several stylized facts emerge from this literature:
Independent contractors and contract company employees, on average, earn as much or more than regular employees.
Workers in most types of FSAs are less likely to have employer-provided health insurance and pensions than regular employees; however, independent contractors and contract company workers, as well as professionals in working in other types of FSAs, frequently appear to receive hourly earnings premiums to compensate for the lack of benefits.
Although regular employees on average have greater job stability than FSA workers, the majority of FSA workers do not work in contingent jobs.
For some workers, FSA jobs offer training opportunities that they might otherwise not receive.
For many workers, FSA jobs offer bridges to regular employment or to other types of FSA jobs with more favorable terms and conditions of employment. We offer, however, the following caveats to these generalizations. First, most of the existing studies use individual-level, cross-sectional data that control imperfectly for occupation, individual productivity-related characteristics, and work schedules

The Effects of Using Alternative Types of Staffing Arrangements

59

and do not control at all for job-specific and organization-specific factors. There is some evidence that outcomes of FSA work for incumbents may be influenced by these factors, indicating a need to better control for them in future research. Second, most cross-sectional studies assume that workers are randomly assigned to some type of staffing arrangement, implying that estimated differences in outcomes are attributable to the type of staffing arrangement. Available evidence calls into question the validity of this assumption (e.g. Farber, 1999; Polivka et al., 2000), suggesting a need to confirm whether these stylized facts are robust after controlling for sample selectivity. In general, the body of literature investigating the wage gap between agency temporaries and regular employees provides an excellent template for future research on the wage gap between regular employees and other types of FSA workers and on the nonwage effects of FSA-work on incumbents. Further research is needed to establish the robustness of these stylized facts. In particular, additional studies using longitudinal data on individuals, studies isolating the FSA effect from the work schedule effect, organizational level studies examining differences between FSA workers and regular employee co-workers in the same job, and studies examining whether the effects of FSA work differ by gender and other demographic characteristics are needed. Finally, it is useful to consider whether the extant evidence provides an answer to a key empirical question raised by the controversy over growing employment in flexible staffing arrangements: Do FSA jobs provide substandard terms and conditions of employment as adherents of the negative view of FSAs assert? The answer appears to be, not necessarily. Rather, the extant evidence suggests considerable diversity in the quality of FSA jobs. We interpret the available evidence as suggesting that FSA employment can be characterized as having “three faces” reflecting three distinct clusters of job attributes, which we will refer to as: (1) secondary FSA jobs; (2) primary FSA jobs; and (3) transitional FSA jobs.24 Secondary FSA jobs correspond to the negative view’s conceptualization of FSA jobs. Compared to regular employment, these jobs provide low pay, few or no benefits, minimal employment security, little or no opportunities for training or advancement, and other less desirable working conditions. Perhaps one of the most pressing future research needs is to more fully explore what factors influence who ends up in secondary FSA jobs and what steps or interventions might assist such workers in successfully bridging to more favorable types of employment. In terms of their overall quality, primary FSA jobs are comparable to regular employment and are superior to secondary FSA jobs. Some primary FSA jobs provide earnings, benefits, and other working conditions that are similar to those provided by regular employment in the same occupation. For other primary FSA jobs, some elements of the terms and conditions of employment may be inferior

60

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

to those typically provided by regular employment in the same occupation, but such differences are balanced by compensating wage differentials or other job attributes for which the incumbent has a preference. In other words, the value of the total package of employment terms of the primary FSA job is comparable to that of regular employment in the same occupation. Another factor distinguishing primary FSA jobs from secondary FSA jobs is that primary FSA jobs provide either relatively continuous employment (like regular employment) or the opportunity to work with the degree of frequency desired by the incumbent. Thus, some workers may prefer primary FSA jobs over regular employment because they offer relatively high quality employment that accommodates atypical preferences for job attributes, such as intermittent work or a compensation package heavily weighted toward cash. Transitional FSA jobs have some attributes in common with secondary FSA jobs. In particular, they tend to provide lower pay and benefits than regular employment. However, unlike secondary FSA jobs, transitional FSA jobs have attributes that facilitate the incumbent’s transition either into a higher quality type of job (i.e. regular employment or a primary FSA job) or into retirement. Examples include FSA jobs designed to screen and/or train incumbents for regular employment; agency temporary jobs that provide a significant amount of general training and, thus, permit the incumbent to acquire the human capital required to attain a better job; or jobs whose combined attributes address the needs of workers who want to retire from regular employment but not to fully cease working in their occupation or profession. Further research is needed to assess whether actual FSA jobs, based upon their attributes, can be categorized in accordance with the three faces of FSA work that we propose. Conducting such research will require individual-level data with fairly detailed information on job attributes as well as the type of staffing arrangement in which an individual works. If the categorization of FSA jobs as primary, secondary or transitional proves to be empirically meaningful, further research will be needed to estimate the share of the labor force in primary FSA jobs, secondary FSA jobs, and transitional FSA jobs, respectively; and to explore what sorts workers among these three faces of FSA jobs and regular employment. Based on the extant literature, we believe that there may be associations between the type of FSA (i.e, agency temporary, independent contractor, short-term hire, oncall worker, contract company worker) and whether the job can be characterized as a primary, a transitional, or a secondary FSA job. Nevertheless, the universe of jobs in any given type of flexible staffing arrangement is likely to include some primary FSA jobs, some transitional FSA jobs, and some secondary FSA jobs. For example, although contract company workers, agency consultants, and independent contractors are more likely than workers in other types of FSAs to

The Effects of Using Alternative Types of Staffing Arrangements

61

have primary FSA jobs, some contract company workers, agency consultants, and independent contracts will have jobs that are either secondary FSA jobs or transitional FSA jobs. Thus, it will be necessary to investigate what individual traits (e.g. demographic, productivity-related traits, occupation, personal preferences), and organizational traits (e.g. business strategy, reason for using FSAs, job design) influence whether an individual works in regular employment, a primary FSA job, a transitional FSA job, or a secondary FSA job. Other effects of using FSAs. Existing research provides some preliminary evidence that the use of FSA workers may influence the attitudes, performance, job security, advancement opportunities, and work tasks of regular employees working in the same organization. Additional research is needed both to expand our understanding of these relationships and to examine the effects of FSA usage on other terms and conditions of employment for regular employee co-workers. For example, Kandel and Pearson (2001) have hypothesized that several additional terms and conditions of employment of regular employees (e.g. use of incentive pay, screening practices, training investments) may be influenced by the share of the employer’s work force who are FSA workers. Other scholars have suggested that some firms use FSA workers to support efficiency wages for their regular workers, which suggests there will be within-firm compensation differentials favoring regular employees over their FSA co-workers. It is important to note that how FSAs affect regular employees in an organization is likely to depend on the organization’s motive for using FSAs, the types of FSAs used, and the manner in which they are used (e.g. whether the FSAs are used to supplement regular employees in a job or to substitute completely for employing regular employees to perform work). Both organizational scholars and adherents to the positive and negative views of FSAs have articulated propositions regarding consequences for organizational performance of using FSAs. The body of empirical research addressing the link between the firm’s choice of staffing arrangements and firm performance is small and most studies are exploratory in nature. Nevertheless, available evidence is consistent with the view that the mix of staffing arrangements used has implications for firm performance. Because the effects of staffing arrangement choices on firm performance may be contingent on the context in which they are used (e.g. business strategy, job design, technology, motivations for using FSAs), a more systematic exploration of the inter-relationships between FSA usage, other contextual factors, and firm performance is needed. Moreover, most of the existing literature focuses on macro-organizational performance measures (e.g. profitability). Staffing arrangement choices, however, are likely to have their most direct impact on measures of human resource performance in the firm, which in turn, are likely to influence indicators of macro-organizational performance. Thus,

62

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

there is a need to more fully explore the effects of staffing arrangement choices on different indicators of human resource performance. For example, under what circumstances, if any, does the use of FSAs result in lower labor costs than the use of regular employees? Does using FSAs decrease the likelihood of employmentrelated litigation? Does using FSAs to screen workers for regular employment result in better worker-job matches than alternative screening methods? There have been only a handful of studies investigating the macroeconomic effects of the presence of FSA workers in the labor force. Available evidence suggests that changes in the share of the labor force working in FSAs has implications for the accuracy and interpretation of estimates of indicators of macroeconomic performance. These implications warrant further investigation. Additionally, there is a need for research investigating the impacts that the shares of the labor force working in various types of FSAs have on both aggregate earnings distributions and the proportion of the population lacking medical insurance coverage. The lack of time series data on the share of the labor force in most types of FSAs (agency temporaries are the exception) has been the primary barrier to such research. Thus, there is a need for better time series data on the share of the labor force working in FSAs to further explore the relationship between the share of workers in various FSAs and aggregate economic outcomes. A final comment on the positive and negative views of FSAs. In light of the evidence available to date, what can be said about the validity of the positive and negative views of flexible staffing arrangements? The short answer is that neither polar view can be wholly embraced or dismissed. On one hand, consistent with the negative view, there is considerable evidence that some FSA jobs provide less favorable terms and conditions of employment than regular employment. On the other hand, it is clear that not all FSA jobs are substandard. In addition, consistent with assertions by adherents to the positive view, FSA jobs appear to provide bridges to regular employment for some workers who might otherwise not be considered for such employment. Finally, evidence regarding the organizational and macroeconomic effects of FSA usage, in our view, remains too sparse and too exploratory in nature to confirm or refute assertions regarding those effects associated with either view.

NOTES 1. Both regular employment and FSA work may offer either full- or part-time work schedules. Thus, the type of staffing arrangement and the type of work schedule are distinct constructs. See Lee and Johnson (1991).

The Effects of Using Alternative Types of Staffing Arrangements

63

2. These definitions are not without controversy. First, the BLS definitions rely on worker perceptions of the security of their jobs, which may not be consistent with the intentions of the organization at which they work. For example, anecdotal evidence suggests that employers in states recognizing the contract exception to the employment-at-will doctrine sometimes deliberately avoid labeling jobs as temporary or permanent in order to prevent breach-of-contract lawsuits when they terminate employees. Second, Kalleberg et al. (1997, pp. 36–37) argue that even the most inclusive of the BLS definitions is too narrow, asserting that the length of time an individual has worked in an arrangement is minimally relevant to future job security and that the criterion of an expected duration of a year or less is too restrictive, even for independent contractors and the self-employed. Although the BLS criteria may result in under-estimation of the share of employed who hold insecure jobs, the Kalleberg et al. (1997) assumptions – that all workers who are uncertain if their job is temporary, who are uncertain if they can work for their employer as long as they want, and who are on-call/day laborers hold insecure jobs – may result in overstating the share of employed holding insecure jobs. 3. The temporary help industry prefers the labels of temporary help “firm” or “service” to the label of temporary help “agency.” 4. The adoption of the implied contract exception creates an incentive to substitute agency temporaries for regular employees because state courts have not extended that exception’s protections against at-will dismissals to agency temporaries. Two other exceptions to the employment-at-will doctrine – the public policy exception and the implied covenant of good faith and fair dealing exception – create a cause of action for agency temporaries as well as for regular employees (Autor, 2000a). Not surprisingly, a state’s adoption of either of these latter two exceptions to the employment-at-will doctrine does not have a significant impact on THS employment growth after controlling for other covariates (Autor, 2000a). 5. The CAEAS has been conducted every other year beginning in 1995. Most empirical research to date has used the 1995 or 1997 sample. 6. Their base model controls for age, age-squared, gender, race, education, marital status, tenure, tenure-squared, region, city type (center city, suburban, rural), industry, and occupation. 7. Controls are included for four race/ethnicity categories, six education levels, four Census regions, three urban/city categories, age and age squared, two marital status categories, being a leased worker, and whether born in the U.S. 8. Unless otherwise noted, the more complete models estimated in Kalleberg et al. (1997) control for 14 industries, 12 occupations, receipt of either employer-sponsored health insurance or a pension, and union membership or coverage by a union contract, as well as personal characteristics and location. 9. A full review of the theoretical and empirical effects of work schedule on the terms and conditions of employment is beyond the scope of this paper. See Barling and Gallagher (1996) for a review of research on part-time employment. 10. The direction of the bias (i.e. whether the estimated effect overstates or understates the true effect) may vary depending on the dependent variable and the type of FSA employment. For example, assuming that part-time work has a negative impact on hourly earnings, if hourly earnings rates are higher for independent contractors than for regular employees, in hourly earnings models in which the reference category is full-time regular employees and in which the independent contractor variable includes both full- and part-timers, the coefficient on the independent contractor variable will be biased toward zero, thereby

64

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

understating the effect of working in that FSA. Conversely, if hourly earnings rates are lower for agency temporaries than for regular employees, in hourly earnings models in which the reference category is full-time regular employees and in which the agency temporary variable includes both full- and part-timers, the coefficient on the agency temporary variable will be biased away from zero, thereby overstating the effect of working in the FSA. 11. They do not distinguish short-term hires from regular employees. 12. Her models control for age, marital status, education, and number of children. 13. Segal and Sullivan (1995) note that the PSS industry includes employment agencies and other help supply services as well as temporary help supply (THS) firms and that employment in the PSS industry is not all temporary: it includes regular employees who recruit and place the temporary employees as well as the temporary employees. Thus, not all workers classified as employed in the PSS industry fit the standard conceptualization of an agency temporary. They provide evidence, however, that THS employment accounts for the bulk of PSS employment and that only only a small share of PSS firm employees are regular employees. 14. They control for age, age-squared, race, sex, and educational attainment, and allow for year-specific wage inflation rates. 15. Their controls include age, race, sex, education, census division, metropolitan area, union status, part-time status, and one-digit occupation. 16. The latter models control for several observable productivity-related characteristics. 17. See Houseman (2001a) for a detailed discussion of the incentives for employers to use FSAs to save on employee benefit costs. 18. Although not eligible for such benefits from the client organization whose work they perform, leased workers, contract company workers, agency temporaries, and agency consultants may be eligible for such benefits from the intermediary firm that employs them. 19. Kalleberg et al. (2000) control for employment insecurity, unionization, occupational complexity, race, marital status, children, education, age, age-squared, city type (suburban, urban, rural), region, occupation, and industry. 20. The failure to identify short-term hires as a distinct category of alternative staffing arrangement is one of the drawbacks of the CAEAS categorization of staffing arrangements. Estimates from organizational surveys provide evidence that the incidence of using shortterm hires is fairly high – 20–44% (Houseman, 2001b; Lautsch, 1998). 21. Their models control for age, gender, education, work experience, clerical experience, temporary work experience, and measures of the extent to which the individual was voluntarily working as a temp. 22. Wiens-Tuers (2001) defines an organization as having such a commitment if it has an implicit or explicit agreement or in actual practice goes to substantial or great lengths to avoid lay-offs. 23. Task identity is the degree to which a job involves completion of a task from beginning to end with a visible outcome; feedback from job is the degree to which doing the job provides the worker with clear feedback on the effectiveness of his performance. 24. Our labels for the former two clusters borrow terminology from the literature on dual labor markets because their attributes resemble those of jobs in the secondary and primary labor markets, respectively. Categorizing FSA job quality based on clusters of job attributes is consistent with Osterman’s (1987) prediction that work practices and worker outcomes tend to cluster together.

The Effects of Using Alternative Types of Staffing Arrangements

65

REFERENCES Abraham, K. G. (1988). Flexible staffing arrangements and employers’ short-term adjustment strategies. In: R. A. Hart (Ed.), Employment, Unemployment and Hours of Work (pp. 288–311). London: George Allen and Unwin. Abraham, K. G. (1990). Restructuring the employment relationship: The growth of market-mediated work arrangements. In: K. G. Abraham & R. B. McKersie (Eds), New Developments in the Labor Market: Toward a New Institutional Paradigm (pp. 85–119). Cambridge, MA: MIT Press. Abraham, K. G., & Taylor, S. K. (1996). Firms’ use of outside contractors: Theory and evidence. Journal of Labor Economics, 14(July), 394–424. Allen, P., & Sienko, S. (1998). Job motivations of professional and technical contingent workers: Are they different from permanent workers? Journal of Employment Counseling, 35(4), 169–178. Autor, D. H. (2000a). Outsourcing at will: Unjust dismissal doctrine and the growth of temporary help employment. Department of Economics, Massachusetts Institute of Technology (February). Autor, D. H. (2000b). Why do temporary help firms provide free general skills training? Photocopy. Department of Economics, Massachusetts Institute of Technology (April). Autor, D. H., Levy, F., & Murnane, R. J. (1999). Skills training in the temporary help sector: Employer motivations and worker impacts. Final Report to the Department of Labor Education and Training Administration (September). Barling, J., & Gallagher, D. G. (1996). Part-time employment. In: C. L. Cooper & I. T. Robertson (Eds), International Review of Industrial and Organizational Psychology (Vol. 11). Wiley & Sons. Barnett, W. P., & Miner, A. S. (1992). Standing on the shoulders of others: Career interdependence in job mobility. Administrative Science Quarterly, 37(June), 262–281. Belman, D., & Golden, L. (2000). Nonstandard and contingent employment: Contrasts by job type, industry, and occupation. In: F. Carr´e, M. A. Ferber, L. Golden & S. A. Herzenberg (Eds), Nonstandard Work: The Nature and Challenges of Changing Employment Arrangements (pp. 167–212). Champaign, IL: Industrial Relations Research Association. Bellemore, F. A. (1998). Temporary employment decisions of registered nurses. Eastern Economic Journal, 24(Summer), 265–279. Bendapudi, V., & Ash, R. A. (1999). The Shamrock organization: Determinants and consequences of externalization. Academy of Management Proceedings, E1–E6. Carnevale, A. P., Jennings, L. A., & Eisenmann, J. M. (1998). Contingent workers and employment law. In: K. Barker & K. Christensen (Eds), Contingent Work: American Employment Relations in Transition (pp. 281–305). Ithaca, NY: ILR Press. Carrington, W. J., McCue, K., & Pierce, B. (2002). Nondiscrimination rules and the distribution of fringe benefits. Journal of Labor Economics, 20(2), S5–S33. Chattopadhyay, P., & George, E. (2001). Examining the effects of work externalization through the lens of social identity theory. Journal of Applied Psychology, 86(4), 781–788. Christensen, K. (1998). Countervailing human resource trends in family-sensitive firms. In: K. Barker & K. Christensen (Eds), Contingent Work: American Employment Relations in Transition (pp. 103–125). Ithaca, NY: ILR Press. Cohany, S. (1998). Workers in alternative employment arrangements: A second look. Monthly Labor Review (November), 3–21. Davis-Blake, A., Broshak, J. P., & George, E. (2003). Happy together? How using nonstandard workers affects exit, voice, and loyalty among standard employees. Academy of Management Journal, 46(4), 475–485.

66

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

Delsen, L. W. M. (1995). Atypical employment: An international perspective: Causes, consequences and policy. Groningen: Wolters-NoordHoff. DiNatale, M. (2001). Characteristics of and preference for alternative work arrangements, 1999. Monthly Labor Review (March), 28–49. Ellingson, J. E., Gruys, M. L., & Sackett, P. R. (1998). Factors related to the satisfaction and performance of temporary employees. Journal of Applied Psychology, 83(December), 913–921. Erickcek, G., Houseman, S., & Kalleberg, A. (2002). The effects of temporary services and contracting out on low-skilled workers: Evidence from auto suppliers, hospitals, and public schools. Upjohn Institute Staff Working Paper No. 03–90. Kalamazoo, MI: W. E. Upjohn Institute. Farber, H. S. (1999). Alternative and part-time employment arrangements as a response to job loss. Journal of Labor Economics, 17(October), 142–169. Feldman, D. C., Doerpinghaus, H. I., & Turnley, W. H. (1995). Employee reactions to temporary jobs. Journal of Managerial Issues, 7, 127–141. Ferber, M. A., & Waldfogel, J. (1998). The long-term consequences of nontraditional employment. Monthly Labor Review, 121(May), 3–12. Ferber, M. A., & Waldfogel, J. (2000). The effects of part-time and self-employment on wages and benefits: Differences by race/ethnicity and gender. In: F. Carr´e, M. A. Ferber, L. Golden & S. A. Herzenberg (Eds), Nonstandard Work: The Nature and Challenges of Changing Employment Arrangements (pp. 213–234). Champaign, IL: Industrial Relations Research Association. Freedman, A. (1985). The new look in wage policy and employee relations. Conference Board Report No. 865. New York: The Conference Board. Gallagher, D. G., & Parks, J. M. (2001). I pledge thee my troth . . . contingently: Commitment and the contingent work relationship. Human Resource Management Review, 11, 181–208. Golden, L. (1996). The expansion of temporary help employment in the U.S., 1982–1992: A test of alternative economic explanations. Applied Economics, 28, 1127–1141. Golden, L., & Appelbaum, E. (1992). What was driving the 1982–1988 boom in temporary employment? American Journal of Economics and Sociology, 51(October), 473–493. Gottfried, H. (1991). Mechanisms of control in the temporary help service industry. Sociological Forum, 6, 699–713. Gramm, C. L., & Schnell, J. F. (2001). The use of flexible employment arrangements in core production jobs. Industrial and Labor Relations Review, 54(January), 245–258. Hackman, J. R., & Oldham, G. R. (1980). Work design. Reading, MA: AddisonWesley. Hipple, S., & Stewart, J. (1996). Earnings and benefits of workers in alternative work arrangements. Monthly Labor Review (October), 46–54. Houseman, S. N. (2001a). The benefits implications of recent trends in flexible staffing arrangements. Upjohn Institute Staff Working Paper No. 02–87. Kalamazoo, MI: W. E. Upjohn Institute. Houseman, S. N. (2001b). Why employers use flexible staffing arrangements: Evidence from and establishment survey. Industrial and Labor Relations Review, 55(October), 149–170. Houseman, S. N., & Polivka, A. E. (1998). The implications of flexible staffing arrangements for job stability. Upjohn Institute Staff Working Paper No. 99–056. Revised May 1999. Kalamazoo, MI: W. E. Upjohn Institute for Employment Research. James, H. S., Jr. (1998). Are employment and managerial control equivalent? Evidence from an electronics producer. Journal of Economic Behavior and Organization, 36(September), 447–471. Jarmon, R., Paulson, A. S., & Rebne, D. (1998). Contractor performance: How good are contingent workers at the professional level? IEEE Transactions on Engineering Management, 45(February), 11–19.

The Effects of Using Alternative Types of Staffing Arrangements

67

Kahn, S. (1998). Falling interindustry wage differentials: Has contingent work had an impact? In: P. B. Voos (Ed.), Proceedings of the Fiftieth Annual Meeting of the Industrial Relations Research Association (Vol. 1, pp. 259–267). Madison, WI: Industrial Relations Research Association, Kahn, S. (2000). The bottom-line impact of nonstandard jobs on companies’ profitability and productivity. In: F. Carr´e, M. A. Ferber, L. Golden & S. A. Herzenberg (Eds), NonStandard Work: The Nature and Challenges of Changing Employment Arrangements (pp. 235–265). Champaign, IL: Industrial Relations Research Association. Kahn, S., Foulkes, F., & Heisler, J. (2001). Large companies and the changing use of temporary workers: Trends and impacts on financial measures of performance. In: S. Houseman & A. Nakamua (Eds), Life Cycle Working Time and Nonstandard Work. Kalamazoo, MI: W. E. Upjohn Institute. Kalleberg, A. L., Rasell, E., Cassirer, N., Reskin, B. F., Hundson, K., Webster, D., Applebaum, E., & Spalter-Roth, R. M. (1997). Nonstandard work, substandard jobs: Flexible work arrangements in the U.S. Washington, DC: Economic Policy Institute and Women’s Research and Education Institute. Kalleberg, A. L., Reskin, B. F., & Hudson, K. (2000). Bad jobs in America: Standard and nonstandard employment relations and job quality in the United States. American Sociological Review, 65(April), 256–278. Kandel, E., & Pearson, N. D. (2001). Flexibility vs. commitment in personnel management. Journal of the Japanese and International Economies, 15, 515–556. Katz, L. F., & Krueger, A. B. (1999). The high-pressure U.S. labor market of the 1990s. Working Paper No. 416 (May). Princeton, NJ: Industrial Relations Section, Princeton University. Krueger, A. B. (1993). How computers have changed the wage structure: Evidence from Microdata, 1984–1989. Quarterly Journal of Economics, CVIII(February), 33–60. Laird, K., & Williams, N. (1996). Employment growth in the temporary help supply industry. Journal of Labor Research, XVII(4), 663–681. Lapidus, J. (1993). Family structure, flexible employment, and labor market segmentation: Evidence from a study of the temporary help industry. International Contributions to Labour Studies, 3, 91–100. Lautsch, B. A. (1998). Bridging labor markets: A grounded theory of contingent work. Dissertation, Massachusetts Institute of Technology. Lautsch, B. A. (2000). Outcomes for contingent workers in high performance work systems. In: Proceedings of the 52nd Annual Meeting of the Industrial Relations Research Association (pp. 52–60). Urbana-Champaign, IL: Industrial Relations Research Association. Lautsch, B. A. (2002). Uncovering and explaining variance in the features and outcomes of contingent work. Industrial and Labor Relations Review, 56(1), 23–43. Lautsch, B. A. (2003). The influence of regular work systems on compensation for contingent workers. Industrial Relations, 42(4), 565–588. Lee, D. R. (1996). Why is flexible employment increasing? Journal of Labor Research, XVII(4), 543–553. Lee, T. W., & Johnson, D. R. (1991). The effects of work schedule and employment status on the organizational commitment and job satisfaction of full vs. part time employees. Journal of Vocational Behavior, 38(April), 208–224. Lentz, E. A. (1996). Flexible employment: Positive work strategies for the 21st century. Journal of Labor Research, XVII(4), 555–566. Lepak, D. P., Takeuchi, R., & Snell, S. (2003). Employment flexibility and firm performance: Examining the interaction effects of employment mode, environmental dynamism, and technological intensity. Journal of Management, 29(5), 681–703.

68

CYNTHIA L. GRAMM AND JOHN F. SCHNELL

Matusik, S. R., & Hill, C. W. L. (1998). The utilization of contingent work, knowledge creation, and competitive advantage. Academy of Management Review, 23(October), 680–997. McClurg, L. N. (1999). Organizational commitment in the temporary-help service industry. Journal of Applied Management Studies, 8(June), 5–26. Osterman, P. (1987). Choice of employment systems in internal labor markets. Industrial Relations, 26(1), 46–67. Park, Y.-S., & Butler, R. J. (2001). The safety costs of contingent work: Evidence from Minnesota. Journal of Labor Research, 22(Fall), 831–849. Pearce, J. L. (1993). Toward an organizational behavior of contract laborers: Their psychological involvement and effects on employee coworkers. Academy of Management Journal, 36, 1082– 1096. Pfeffer, J., & Baron, J. N. (1988). Taking the workers back out: Recent trends in the structuring of employment. Research in Organizational Behavior, 10, 257–303. Polivka, A. E. (1996). Contingent and alternative work arrangements, defined. Monthly Labor Review (October), 3–9. Polivka, A. E., Cohany, S. R., & Hipple, S. (2000). Definitions, composition, and economic consequences of the nonstandard work force. In: F. Carr´e, M. A. Ferber, L. Golden & S. A. Herzenberg (Eds), Nonstandard Work: The Nature and Challenges of Changing Employment Arrangements. Champaign, IL: Industrial Relations Research Association. Rebitzer, J. B. (1998). Job safety and contract workers in the petrochemical industry. In: K. Barker & K. Christensen (Eds), Contingent Work: American Employment Relations in Transition. Ithaca, NY: ILR Press. Rogers, J. K. (1995). Just a temp: Experience and structure of alienation in temporary clerical employment. Work and Occupations, 22(May), 137–166. Rousseau, D. M., & Libuser, C. (1997). Contingent workers in high risk environments. California Management Review, 39(Winter), 103–123. Schur, L. A. (2003). Barriers or opportunities? The causes of contingent and part-time work among people with disabilities. Industrial Relations, 42(October), 589–622. Segal, L. M., & Sullivan, D. G. (1995). The temporary labor force. Economic Perspectives: A Review from the Federal Reserve Bank of Chicago, 19(2), 2–19. Segal, L. M., & Sullivan, D. G. (1997a). The growth of temporary services work. Journal of Economic Perspectives, 11(Spring), 117–136. Segal, L. M., & Sullivan, D. G. (1997b). Temporary services employment durations: Evidence from state UI Data. Working Paper No. WP-97–23 (December). Chicago: Federal Reserve Bank of Chicago. Segal, L. M., & Sullivan, D. G. (1998). Wage differentials for temporary services work: Evidence from administrative data. Working Paper WP-98–2. Chicago: Federal Reserve Bank of Chicago. Steers, R. M. (1977). Antecedents and outcomes of organizational commitment. Administrative Science Quarterly, 22, 46–56. Stone, T. H., & Gallagher, D. G. (1997). Contingent workers and labour representation: Worker disposition and union prospects. Paper presented at the 3rd International Conference on Emerging Union Structures: Reshaping Labour Market Institutions, Canberra, Australia (December 1). U.S. Bureau of Labor Statistics (1995). Contingent and alternative employment arrangements. Report No. 900. Washington, DC: U.S. Department of Labor. Wiens-Tuers, B. A. (1998). The relationship of race and outcomes of non-standard labor. Journal of Economic Issues, 32(June), 575–585. Wiens-Tuers, B. A. (2001). Employee attachment and temporary workers. Journal of Economic Issues, 35(March), 45–60.

ADOPTION AND USE OF DISPUTE RESOLUTION PROCEDURES IN THE NONUNION WORKPLACE Alexander J. S. Colvin INTRODUCTION This paper investigates the adoption, structure, and function of dispute resolution procedures in the nonunion workplace. Whereas grievance procedures in unionized workplaces have been an important area of study in the field of industrial relations, research on dispute resolution procedures in nonunion workplaces has lagged behind. As a result, our knowledge of the development of nonunion procedures remains relatively limited. Similarly, with a few noteworthy exceptions (e.g. Lewin, 1987, 1990), our knowledge of workplace grievance activity is almost entirely based on research conducted in unionized settings. Given the major differences in the institutional contexts of union and nonunion workplaces in the United States, existing ideas about workplace dispute resolution developed in the unionized setting will likely require significant modification in order to understand dispute resolution procedures and activity in the nonunion workplace. Issues relating to dispute resolution in the nonunion workplace are of increasing importance to public policy given the combination of continued stagnation in levels of union representation and mounting concerns over rising levels of employment litigation in the courts. Knowing what nonunion dispute resolution procedures look like and how they function will help answer the question of what role these procedures may play in the future governance of the workplace.

Advances in Industrial and Labor Relations Advances in Industrial and Labor Relations, Volume 13, 69–95 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0742-6186/doi:10.1016/S0742-6186(04)13003-5

69

70

ALEXANDER J. S. COLVIN

In the contemporary unionized workplace, issues of the adoption and structure of grievance procedures spark relatively little controversy. Development and diffusion of union grievance procedures dates to the period of the 1940s War Labor Board and the post-War consolidation of stable collective bargaining. The structure of union grievance procedures has exhibited remarkable stability since that period, with virtually all current unionized workplaces having multi-stage grievance procedures culminating in arbitration (Eaton & Keefe, 1999). By contrast, nonunion dispute resolution procedures vary widely in both their adoption and structure. As many as half of all nonunion workplaces continue to lack any formal procedures for the resolution of employee complaints (Feuille & Chachere, 1995). Among workplaces that do feature formal dispute resolution procedures, these procedures vary in basic features such as, who is the final decision-maker under the procedure, whether employees are permitted representation, what complaints can be brought under the procedure, and what criteria are used to decide disputes (Feuille & Chachere, 1995; Feuille & Delaney, 1992). An initial issue that needs to be addressed with regard to nonunion procedures is what factors shape the diversity in the structures that are used to resolve disputes in the nonunion workplace. As with other areas of industrial relations and human resource practices, environmental pressures and human resource strategies provide two broad categories of factors that may influence the development of nonunion dispute resolution procedures (Barringer & Milkovich, 1998). Among environmental pressures, concerns about workplace disputes giving rise to litigation and the desire to avoid unionization by providing workers with a mechanism for resolving complaints have both been suggested as possible motivations for nonunion organizations to introduce procedures (Feuille & Delaney, 1992). Some have argued, however, that union substitution no longer plays a major role in the introduction of nonunion procedures due to declining unionization levels and reduced fear of organizing activity (Feuille & Delaney, 1992). Others have suggested that the extent of actual litigation pressures on organizations are also often exaggerated, though fear of litigation may provide an important rhetorical device in convincing managers to adopt procedures as an appropriate organizational “best practice” (Edelman, 1990; Edelman et al., 1992, 1999). An alternative category of explanations suggest that nonunion procedures are adopted as part of human resource strategies that emphasize fostering positive relations between employees and management. Arguments from the perspective of organizational justice theory suggest that by providing greater procedural and distributive justice in decision-making, nonunion procedures may reduce turnover and foster greater employee commitment and effort (Olson-Buchanan, 1996). In addition to the question of what is the respective importance of environmental pressures and human resource strategies in the adoption of procedures, an

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

71

additional question to be addressed is whether there is any interaction between these factors in how procedures are developed. A second set of issues relating to nonunion dispute resolution procedures concern how they are used by employees and what is their function in the workplace. The traditional image of nonunion procedures was that they were little used by employees due to their deficiencies in the area of due process and were largely ineffective as workplace institutions (Slichter et al., 1960). More recent research has found examples of nonunion procedures with more regular employee usage, though at levels well below that of typical union grievance procedures (Lewin, 1987, 1990). In addition to differences in usage levels, the way in which nonunion procedures function may differ significantly from union grievance procedures. In particular, the absence of the institutionalized presence of the union in representing employees means that in a nonunion procedure employees are much more dependent on the goodwill of management in resolving complaints. The question of how nonunion procedures function in practice is closely linked to the issue of what impact these procedures have on the operation of the workplace. For unionized workplaces, persistent high grievance rates have been seen as a major part of embedded high conflict patterns of industrial relations, interfering with the operations of the workplace and reducing productivity (Kochan et al., 1994). By contrast, for nonunion procedures the initial question is not excessive conflict, but rather whether the procedures have enough of an impact on the management of the workplace to make any real difference for employees. In this paper, I investigate these questions of the formation and function of nonunion dispute resolution procedures through a case study of the procedures adopted at the multi-divisional manufacturing firm TRW. TRW provides a useful case for investigating the possible direction of development of nonunion procedures due to its relatively extensive set of procedures for its nonunion employees. Concerning the adoption of procedures, the results of the case study will provide support for the importance of both environmental pressures and human resource strategies. Furthermore, it will suggest that these are not simply alternative explanations for the adoption of procedures, but rather operate as complimentary factors that interact in shaping the structure of nonunion procedures. On the issue of the usage and function of nonunion procedures, the results of the case study suggest that employees do use at least this type of more developed nonunion procedures, but that concerns remain as to due process limitations in the operation of the procedures. At the same time, the procedures do appear to have had some significant impacts on the operation of the workplace. Most notably, procedures in some instances led to increased negotiated resolution of workplace disputes and modification of existing practices or procedures, despite features in the design of the procedures directed at limiting this type of outcome. Lastly, the case study

72

ALEXANDER J. S. COLVIN

results reveal a major division between white-collar and blue-collar employee groups in both the development and operation of procedures. The contrasts found between these two groups indicate the need to be careful about making generalizations about the nonunion workforce, which can be characterized by high degrees of diversity in employment conditions and outcomes.

METHODOLOGY As noted above, this paper investigates the adoption and use of nonunion dispute resolution procedures through a case study of procedures at the multidivisional American manufacturing company, TRW. In the context of dispute resolution, TRW provides a useful case for understanding the development of nonunion procedures due to the relatively extensive and elaborate nature of the procedures that the company has adopted. Rather than being a typical case, it was selected as a “best practice” case in order to investigate the impact of nonunion procedures in an organizational setting where particularly strong efforts have been made to ensure the widespread adoption and use of the procedures. This is not to suggest or advocate that organizations should necessarily adopt the practices that are described here, but rather that we can better evaluate the possibilities and limitations of nonunion procedures from a “best practice” case than from a case with less extensive procedures. By analogy, past research was able to learn much about the impact of employee involvement in the workplace from studying GM’s innovative Saturn division (e.g. Rubinstein & Kochan, 2001) and about employee ownership from the unusually active worker ownership structure used at the Rath packing company (Hammer & Stern, 1986) not because these were typical cases, but rather because they were “best practice” cases that provided strong examples of the phenomena being investigated. The primary data sources for the case study presented in this paper are interviews conducted by the author at the company in 1997–1998 and examination of documents relating to the procedures. Individual interviews were one to one-anda-half hours and followed a semi-structured format. Primary interview subjects included human resource managers at the corporate, divisional, and plant level, and in-house employment and labor law counsel. Secondary information sources included descriptions of the company’s procedures and experience with their usage that have been presented by TRW managers at conferences (e.g. Boxer, 2000) and a profile of dispute resolution procedures at TRW was included in a 1997 General Accounting Office (GAO) report as one of five private sector companies providing leading examples of the adoption of nonunion dispute resolution procedures (GAO, 1997).

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

73

The primary focus in the research was on the management side. The reason for this focus was that management decision-making was central to the research questions being investigated; in particular decisions of management determined the adoption of the procedures, the structure of the procedures, and how they were operated. Although additional perspectives on the issues examined were obtained from union and plaintiff attorney sources, the primary focus on the management side is a limitation of the research emphasis that should be recognized. A more general research caution is inherent in the use of a “best practice” case methodology. As is the case for this study, in following a “best practice” case methodology the researcher selects a research site to represent a strong example of a particular phenomenon. The danger then becomes that the researcher will excessively identify with the interests of the group or organization being studied and this will in turn influence the research findings. This is not to suggest that what the researcher views as “best practice” cases should not be studied, but rather that the researcher needs to be aware of and recognize potential biases. In the present case, as a researcher I approached this case study with the recognition that I was examining what I viewed as likely a best case scenario for nonunion procedures. With that recognition in mind, the case study was written with an effort to identify both the potential weaknesses and the strengths of the company’s procedures and with this initial caution to readers that they are reading about a case that was deliberately selected as a potential “best practice” case for the phenomenon in question.

CASE BACKGROUND The diversified manufacturing company TRW has occupied a prominent role in American industrial relations history and research. During the earlier part of the 20th century, under its former name of Thompson Products, the firm was one of the leading exemplars of the “welfare capitalist” movement among American companies. Despite organizing efforts by the United Auto Workers during the 1930s and 1940s, only a few of the company’s plants became unionized during the period of expansion of unionization following passage of the Wagner Act. In his landmark study of the survival and continuation of welfare capitalism during the post-war period, Modern Manors, Sanford Jacoby (1997) uses Thompson Products, along with Eastman Kodak and Sears, as a critical case study of a company that continued to follow this management philosophy during an era when it was thought to have disappeared from the landscape of American business. Thompson Products’ ability to maintain its predominantly nonunion status during the post-war period was not due solely to the benefits offered to employees

74

ALEXANDER J. S. COLVIN

by welfare capitalism. Jacoby (1997) notes that, unlike Eastman Kodak and Sears, Thompson Products was subject to intensive union organizing, to which the company often responded with more heavy-handed tactics. Although unions succeeded in organizing some TRW plants during the post-war period, by the 1990s, only three or four of the company’s American plants were unionized. The current TRW is a diversified manufacturing company, headquartered in Cleveland, Ohio. Organizationally, it is divided into two business groups. One is the Automotive group, which accounts for 60% of total sales. Auto parts were the traditional center of the company’s business, though it has remained dynamic in this area in recent years through expansion into new product areas such as air bags. The other major group in the firm is TRW’s Space, Defense and Information Technology business group, which contains two units: space and defense; and systems integration. In sharp contrast to the more traditional manufacturing setting of the Automotive group, the Space, Defense and Information Technology (SDIT) group produces highly sophisticated products with a predominantly professional workforce. Among the space and defense unit’s primary products are commercial and military satellites. In addition, it is the primary contractor for such major defense contracts as the airborne laser being produced for the U.S. Air Force. The other major unit within this group is systems integration, which provides management and information system services for complex operations, such as major government research projects. An important factor in understanding developments at TRW is the values and human resource orientation of the company’s management. During the 1930s and 1940s Thompson Products’ retention of welfare capitalist policies was heavily influenced by the employee relations values of the company’s president, Frederick C. Crawford (Jacoby, 1997). This orientation continued during the 1960s, when the company began experimenting with new behavioral science concepts in the management of its workforce, such as innovative compensation and selection systems. In addition, in 1977 the company opened a new nonunion teamconcept plant in which employees were assigned to teams rather than specific jobs and compensation was all-salaried based on a pay-for-knowledge system (Kochan et al., 1994). These human resource policies were developed in the context of the company’s continuing commitment to avoiding unionization. In particular, the innovative practices were concentrated in the company’s newer nonunion facilities and not extended to its few unionized facilities. More recently, the company’s current human resource management orientation is reflected in strong public advocacy by its management of such policies as gain-sharing, pay-for-productivity and employee involvement (Daily Labor Report, 1994 DLR 13 d11). It is also noteworthy that the executive vice-president of human resources at TRW is a member of the firm’s executive committee, indicating the

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

75

relatively strong role of human resource management considerations in the firm’s decision-making. Dispute resolution procedures formed part of the high commitment human resource policies introduced at TRW facilities in the 1960s and 1970s. These procedures initially consisted of provisions for written complaints to be filed with and reviewed by management. These procedures primarily provided a more formal and systematic structure for management review of employee concerns, rather than a more independent process of dispute resolution. In addition, committees composed of employees were established in plants to bring complaints of general concern to management. Although providing a mechanism for raising employee grievances, a weakness of these procedures was that they did not provide a structure for the resolution of disputes beyond the normal management decisionmaking process. However, during the 1990s a series of factors led to the development of more complex dispute resolution procedures that afforded greater independence from the management chain-of-command in the adjudication of employee grievances. Rather than a single process, development of nonunion procedures at TRW in the 1990s actually involved two distinct processes that led to the introduction of two different types of procedures, peer review procedures and nonunion arbitration procedures. Under peer review procedures, employees who are peers of the employee filing the complaint sit on a panel that decides the grievance. By contrast under the nonunion arbitration procedures a neutral arbitrator hears and decides claims by employees against the company concerning violation of employee legal rights. Although there were common elements in the development of these procedures, the development and introduction of each procedure involved a distinct decision-making process in response to separate pressures and motivations. Furthermore, once adopted the peer review and nonunion arbitration procedures retained distinctions in their operation, usage, and functions. For this reason, in the subsequent sections the development and use of the two types of procedure will be examined separately, beginning with the peer review procedures, then turning to the nonunion arbitration procedures. In this respect, the procedures serve as an embedded unit of analysis within the organizational case that permits comparing and contrasting two different processes of procedure development within the same firm.

PEER REVIEW PROCEDURES In the early 1990s, TRW’s Vehicle Safety Systems division introduced peer review for the first time into dispute resolution procedures in the company. Vehicle Safety

76

ALEXANDER J. S. COLVIN

Systems (VSSI) is a division of the automotive business group. Its primary products are air bags and seat belts. This is a relatively new product area for the company, with the initial part of the division entering the company through a 1984 acquisition. With the expansion of the use of air bags in automobiles, the division grew rapidly from a $2 million business in 1984 to an over $2 billion business by the late 1990s. All of VSSI’s American plants are nonunion and are generally recently constructed facilities, with relatively young workforces. The total workforce of the division is 16,000 employees, half of whom are located in the United States, with the remainder spread across plants in other countries, including Mexico, Canada, and Germany. Introduction of a peer review procedure was first proposed at VSSI during the early 1990s based on a series of interrelated reasons why the practice would be advantageous to the division. The first rationale was that peer review would serve as a “counter-defensive” move to avoid VSSI’s plants becoming a target for unionization. Although VSSI had managed to avoid unionization of its American plants, with the continued strength of the UAW in the industry its plants remained a potentially attractive target for organizing. The perception of managers in the division was that their employees were aware of what employment conditions were like in the unionized plants in the auto industry. As a consequence, management actively monitored what was going on in the unionized sector of the industry to ensure that the wages and conditions of employment in VSSI’s nonunion plants were roughly comparable. Peer review was offered as a component of that strategy of union substitution, which would provide a better substitute for union grievance procedures than the division’s existing complaint procedures. A closely related rationale offered for the introduction of peer review was that it would respond to the expectations of the division’s workforce. As noted earlier, employees in the division are relatively young, averaging 25–35 years of age in most plants. Management perceived these employees as expecting a high involvement, team-oriented culture in the workplace, what one manager described as the “video generation who want a say in the business.” Peer review was proposed as a way to respond to these perceived employee expectations that would complement related human resource practices such as performance- and group-based compensation and selfdirected work teams. The complimentarity with VSSI’s overall high commitment human resource strategy was the third rationale offered for the introduction of peer review. Peer review was advocated as being in keeping with the “spirit of a modern company” that TRW was seeking to achieve in its human resource strategy. Peer review was offered as a new “best practice” in the area of dispute resolution procedures, which would be appropriate for TRW to adopt. It is worth noting that although peer review procedures were adopted in other companies, most prominently at General Electric, in the 1970s, they were not part of

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

77

the initial set of human practices adopted when VSSI’s first facilities were opened in the mid-1980s. Instead, peer review first came under consideration as a new type of best practice in the 1990s. This reflects the significant lag times often involved in the diffusion and adoption of new practices, a factor important to consider in examining the causes of change in employment relations. In addition, it reflects the importance of changes in the set of options for behavior under consideration by management. Taking a strategic choice view of employment relations (Kochan et al., 1994) suggests that to understand change it is necessary to consider the specific alternatives under consideration at the relevant point in time. The environmental factors that ultimately led to the introduction of peer review were already present in the 1980s. However it was only in the 1990s that VSSI management brought into consideration the option of introducing peer review as a new best practice that would help avoid the environmental threat of union organizing, respond to the expectations of its workforce, and complement its other human resource practices. Although peer review was presented as a policy that complimented VSSI’s human resource strategy, its initial proposal met resistance from both senior and lower-level line managers. In particular, line managers were concerned about the prospect of having to justify their actions to employees on the peer review panels. A key factor in convincing management to adopt peer review was the assurance offered to line managers that human resource managers would be there to provide support to line managers when they went into peer review. However concerns remained and were exacerbated when the first peer review hearing resulted in the overturning of a line manager’s decision. Although this decision exacerbated line management concerns about peer review, it ultimately proved fortuitous in helping establish early on to employees that management would respect the decisions of the panels, even if unfavorable. The peer review procedure was introduced at VSSI in 1992 as an additional procedure to which employee disputes could be taken to on top of existing open door policies at plants. Usage of peer review was restricted, however, to employee complaints about final warning or termination decisions, excluding terminations due to elimination of the employee’s job or work reduction. Peer review panels consist of five members, two managers and three peer employees. A new panel is formed for each appeal. The definition of peer is grievant specific, so if the employee is a production worker, the peer employee panelists will be selected from all eligible production workers in the plant. In order to serve on a peer review panel, an employee must have been employed for at least one year, not be under corrective counseling, and not have received disciplinary action or corrective counseling in the past year. Additional exclusions from panel membership include: family members of the grievant; anyone in the chain of command of the grievant;

78

ALEXANDER J. S. COLVIN

members of human resources; and “persons who have a vested interest and/or conflict of interest in the outcome of the meeting.” The members of the panel are chosen at random from the pool of eligible peer employees and managers by the grievant. The employee first picks four names from the pool of eligible managers, chooses two of the managers to serve as panel members, then selects one as an alternate, and discards one name. Then the employee picks five names from the pool of eligible peer employees, chooses three employees to serve as panel members, selects one as an alternate, and discards one name. Service on the peer review panels is voluntary and panelists may not serve on two consecutive peer review panels. If a grievant requests, one of the panelists may be chosen from a pool specifically composed of members of a protected minority class of which the grievant is a member. As a general rule, there must be at least 15 members in each pool from which panelists are chosen. If there are not enough members in any group at a facility, employees from other locations may be included in the pool to ensure sufficient numbers. This is a particular concern where there are insufficient numbers of employees in a protected minority class to ensure representation on the panel of peers of the same racial or ethnic group as a minority complainant. For each panel, a human resources representative serves as a “facilitator” (Schwartz, 1994). The role of the facilitator is to handle the administration of the procedure, organize meetings, maintain files, and provide assistance to the panel. In addition to organizational and record keeping duties, the facilitator has a more substantive role in the procedure in meeting with the grievant and supervisor to explain the procedure, determining any conflict of interest questions in the panel selection, and gathering any data or information requested by the panel. The importance of the role of the facilitator is enhanced by the instruction in the procedure that they “may be called upon to provide policy interpretation and general factual information concerning precedents, but he/she is not expected to give detailed information regarding other specific corrective actions, appeals, or precedents.” Given this significant role of the facilitator in the peer review procedure, only senior human resources representatives are allowed to perform this function. In addition, any human resource representative involved in the decision under appeal is not allowed to act as the facilitator. Peer review proceedings are commenced by the employee filing a peer review request form stating whether it is a final warning or a termination case, the nature of the employee’s concern, and the company policy in question. Requests for peer review must be filed within three working days of the final warning or termination decision being appealed. Following the filing of the request form with human resources, a panel will be convened in a time not to exceed 45 days from the request.

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

79

Prior to the hearing, peer review panelists are given a day of training consisting of instruction on the procedure and on the conduct of hearings. A major focus in the training is to emphasize the confidentiality of the procedure. Management had strong initial concerns that employees would be reluctant to use peer review if they thought that confidential information about them would be spread around the plant by the panelists. In practice, this has not been a problem with the panels. During the training period, the panels are also asked to select their own panel leader. The panel leader is responsible for chairing the hearing and making procedural decisions on issues such as the calling of witnesses. The facilitators generally recommend that a peer employee lead the panel and this is the usual practice. At the commencement of the peer review hearing, the human resources facilitator introduces the subject of the complaint. The employee is then given the opportunity to present his/her case, including presenting any documents or witnesses that support his/her case. The employee’s presentation may be oral, written, or both, however no internal or external employee representative is permitted to participate in the hearing. Following the employee’s presentation, the supervisor who made the decision being challenged is given an opportunity to justify his/her actions, including any documents or materials supporting his/her position. The supervisor must present his/her case on his/her own, also without assistance of a representative. The panel can cross-examine witnesses presented by the employee or the supervisor, as well as call any additional witnesses they want to hear from. The panel is entitled to examine company files and records, as well as to “seek advice from experts within the Company regarding policy interpretation, etc.” Following the presentations of the employee and supervisor and when all information the panelists wish to review has been heard, the employee and supervisor are excused from the hearing room and the panel deliberates. During training, the panelists are instructed that they are limited to reviewing the application of company policies and cannot change policy. However, the procedures also note that policies are only guidelines and cannot cover every situation. Thus, panelists are also instructed that “a common sense approach must be used by the panel members in determining the final outcome.” Then, after discussion of the information presented at the hearing, the panel makes its decision by secret ballot. The panel leader conducts the secret ballot, with all panelists voting. Ballots are opened one by one until a majority of votes are cast to accept or reject the employee’s appeal. Remaining ballots are not opened and all ballots are destroyed after the counting of the votes. The unusual procedure of not opening the remaining ballots once a result is reached protects panelist independence by reducing the possibility of determining which way any individual panel member voted – e.g. if four of the five panel members admitted to voting for or against an appeal it would be possible to identify the dissenting panel member on a 4–1

80

ALEXANDER J. S. COLVIN

majority vote. However, in some cases it prevents the employee from knowing what the final vote count was on the panel – e.g. if the first three votes opened were to deny the appeal, the employee will not be able to find out if the final vote count would have been 3–2, 4–1, or 5–0. If an appeal of a final warning is accepted by the panel, the warning is removed. If an appeal of a termination is accepted by the panel, the termination is reversed and the employee reinstated with retroactive pay and benefits. Although the first decision from a VSSI peer review panel accepted the employee’s appeal, the majority of subsequent decisions were in favor of upholding the supervisor’s decision. From 1992 to 1997, only 10 of 160 employees who took cases to peer review were successful. The cases where employees were successful in appealing to peer review panels often involved failures of management to follow the exact procedural details of their own rules, such as the form and timing of warnings, rather than the content of the complainant’s conduct. The division intentionally limited the panels to considering whether or not company policies were correctly applied, though the human resource managers expressed some surprise at the degree to which employees on the panels have held management to the letter of the rules in the conduct of discipline and terminations. This may reflect a reality that if peer employees are placed in a structure where the only basis on which they can overturn management decisions is for lack of compliance with company rules, they then turn to less flexible interpretations of management behavior under the rules in order to obtain the leverage necessary to accept appeals. Although employees have not had a high success rate at hearings, peer review may have a broader impact on the behavior of line and human resources managers. Given the possibility of having to justify their decisions to a peer review panel, managers reported having to be more careful in making termination or final warning decisions. Peer review is thus seen as having an indirect effect on the behavior of managers in making them more accountable and keeping them on their toes.

NONUNION ARBITRATION PROCEDURES Although both represent alternatives to traditional management review in dispute resolution, the development of nonunion arbitration and peer review at TRW involved responses to different environmental pressures involving very different groups of employees. In contrast to the union substitution and human resource strategy motivations that led to the introduction of peer review in VSSI, the development of TRW’s nonunion arbitration procedure was a response to changing patterns of employment litigation against the firm. In addition, whereas peer review was first introduced in the auto parts section of the company, nonunion arbitration

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

81

was developed primarily in response to events in TRW’s Space, Defense, and Information Technology (SDIT) group. In contrast to the traditional manufacturing plants of the automotive business group, the SDIT group operates extremely high technology facilities employing highly skilled workers. Among the best known of the products of the group’s space and defense unit was the NASA probe Pioneer 10, the first probe to reach Jupiter and the first human made spacecraft to escape the gravity of the solar system. On a less ethereal front, the group is building the first tactical lasers for the Army and the Air Force, and is a leader in the area of battlefield information systems. This type of product leads to a strikingly different production facility setting from traditional manufacturing. For instance, at the group’s main “Space Park” location in Redondo Beach, California, one of the production labs is a single production facility in which around 150 engineers and scientists work on producing a single satellite at a time. In some respects, these facilities resemble a high technology version of a shipyard, with each product being worth millions of dollars and production occurring in an almost craft-like fashion. Overall, 60% of the group’s workforce is classified as exempt professional and managerial employees. Until the early 1990s TRW had experienced very little employment litigation. In part, this was likely due to the limited nature of exceptions to employment-atwill. However, a major additional factor was that the company gave employees effective lifetime employment, a practice reflective of the historical welfare capitalist philosophy of the firm. This effective lifetime employment policy was ruptured in the early 1990s when the SDIT group engaged in large-scale layoffs in response to the contraction of the aerospace industry following the end of the Cold War. Although the contraction of the industry may have created the economic necessity for the layoffs, it also made their impact particularly hard for the affected employees. With the whole aerospace industry in contraction, many of the laidoff employees were unable to quickly find alternative employment. Given the inevitable questions of fairness surrounding the dismissal of large numbers of employees, particularly those who believed the company had promised them lifetime employment, a major surge in employment litigation followed the layoffs. The ability of employees to engage in litigation was enhanced by the expansion of substantive employment rights during the 1980s and early 1990s, particularly in California, which increased the willingness of plaintiff attorneys to take on employment cases. As one manager graphically described the situation, they were “laying off employees when employment lawyers were starting to advertise their services on billboards at the side of highways.” At the same time as these developments were increasing the pressures from employment litigation upon the company, changes in the law of arbitration provided new opportunities to use alternative dispute resolution procedures to

82

ALEXANDER J. S. COLVIN

avoid litigation. In the United States Supreme Court’s landmark 1991 decision in Gilmer v. Interstate/Johnson Lane, 500 U.S. 20 (1991), the court had held for the first time that rights contained in employment statutes, in that specific case the Age Discrimination in Employment Act, could be subject to mandatory arbitration under an arbitration agreement. Although the following decade would be filled with decisions attempting to sort out the many implications of Gilmer, what the decision did for the first time was open the door to the possibility that companies could adopt nonunion arbitration procedures that would serve as an alternative to litigation through the courts for claims against them by their employees (Stone, 1996, 1999). One advantage of nonunion arbitration procedures was that the employer could design the type of procedure it desired to fit its organizational objectives. Perhaps most importantly though, arbitration provided a mechanism to avoid being subject to the uncertainties and potentially large damages of the jury system. Although a competent arbitrator might be quite willing to find against the company where the employee had a valid claim, the general perception among management was that arbitrators would not be likely to award the type of extremely high damage amounts designed to punish companies that juries occasionally awarded. Damages would also be further reduced by the potential for arbitration to occur much more quickly than litigation, thereby preventing damages from accruing to the same amounts as could happen during slower moving court proceedings. Recognizing the potential impact of these developments, the group’s labor and employment counsel wrote a memo to TRW’s senior management suggesting that the recent Gilmer decision had opened the door to using a nonunion arbitration procedure to avoid litigation and resolve these disputes at much lower cost. Following the memo, a management team was put together at the corporate level to investigate the development of an alternative dispute resolution procedure to handle employment disputes that might lead to litigation. The team included representatives from legal, human resources, and communications departments and was led by a vice-president of human resources. The team spent six months, occupying its members virtually full-time, investigating the question of what dispute resolution policy to adopt. During its investigations, a major dividing point emerged within the team over the question of how to make sure any agreement establishing a nonunion arbitration procedure was enforceable against employees. Legal department members of the team were concerned that an arbitration agreement had to be clearly enforceable to make sure of protecting the company against potential employment litigation. To do so, they proposed that employees be required to sign arbitration agreements, either at initial hiring or as a condition of receiving salary increases or stock options. Although the courts would likely enforce arbitration agreements without any such additional consideration for the agreement beyond continued employment, legal

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

83

counsel recommended this more conservative approach, which had been followed by another aerospace company that had just introduced a nonunion arbitration procedure. The human resources members of the team were strongly opposed to this proposal. They argued that it would lead to the dispute resolution procedure being viewed as an anti-employee move, which would conflict with TRW’s overall human resource strategy of enhancing employee commitment and involvement. In particular, they objected that whereas other companies might require employees to sign arbitration agreements to get bonuses and stock options, this was inconsistent with the philosophy of employment relations at TRW. The decision on how to handle the problem of ensuring any procedure was binding and enforceable became the crucial stumbling block in TRW’s development of its nonunion arbitration procedure. After nine months of debate, the issue eventually had to be resolved by the company’s General Counsel and its Executive Vice-President of Human Resources, who sided with the view that the nature of the arbitration procedure had to be made consistent with TRW’s overall human resource philosophy. The final design decision reached was that the nonunion arbitration procedure would be mandatory and binding on the company, but that it would be mandatory but non-binding for the employee. The employee would be required to take any legal dispute arising from employment through the arbitration procedure, however if the employee did not accept the arbitrator’s decision they would be free to proceed to litigation. The decision to make the arbitration decisions non-binding for employees makes TRW’s procedure virtually unique among nonunion arbitration procedures. Indeed, it is worth remembering that the key reason Gilmer led to the expansion of nonunion arbitration was that it allowed for the implementation of procedures that would produce binding arbitration awards, eliminating the possibility of taking statutory employment claims to litigation. Despite having been inspired to consider introducing a nonunion arbitration procedure as a result of Gilmer offering the possibility of using it to avoid rising litigation, this decision meant that TRW would be relinquishing one of the primary advantages of arbitration to employers. With the issue of the nonbinding status of the procedure finally resolved, a company policy directive was issued for the implementation of the new dispute resolution procedures (TRW Human Resources Memorandum No. 10, January 1, 1995). Although the policy directive required the establishment of new procedures and set out basic features that had to be included in each procedure, the format of the actual procedures was left up to the individual divisions of the company. The use of a company directive allowing a degree of discretion to individual divisions reflects both the decentralized business group and divisional organization of TRW, as well as the diversity between different parts of the company. Under the directive, each division was required to establish a dispute resolution procedure covering:

84

ALEXANDER J. S. COLVIN

involuntary terminations that could form the basis for a claim in the courts; claims of discrimination or harassment based on protected status; and claims of constructive discharge. Individual divisions could provide that other types of disputes were covered by the procedure, but a number of types of disputes had to be excluded from coverage. The excluded categories included: benefit or pension claims; workers’ compensation; unemployment compensation; claims by the company for injunctive relief; and “the establishment or modification of TRW’s policies or procedures.” Certain due process features had to be included in the procedures. Employees would have the right to representation by counsel. A basic entitlement to discovery of relevant documents and deposition of one individual for each side could be expanded on order from the arbitrator. Remedies equivalent to those available in court could be ordered by the arbitrator. The arbitrator would be required to provide a written decision stating a summary of the claims, the remedies ordered, and the reasons for the award. As had been decided, the decision of the arbitrator would be binding on the company, but not binding on the employee. However, the company reserved the right to introduce the arbitrator’s decision as evidence in any subsequent litigation. Initially, the directive provided that the costs would be split, but that the employee’s contribution would be capped at two days base pay. If the employee were successful, the company would pay the entire costs. This provision was later changed so that employees would only pay a $100 filing fee based on subsequent court decisions rejecting the imposition of the costs of employer mandated nonunion arbitration procedures on employees. As noted above, an interesting feature of the company directive is that it left the format of the procedure up to the division. Prior to the company directive on alternative dispute resolution procedures, the individual divisions of TRW had a variety of internal complaint procedures. These ranged from simple open door policies to the complex peer review procedure in the VSSI division described above. The company directive led to both the introduction of new arbitration procedures and a coincident enhancement of many of the internal complaint procedures. These collateral changes in internal complaint procedures reflect an interesting process in which the focusing of organizational effort on improving existing dispute resolution procedures led to changes beyond those that inspired the initial initiative. This suggests that organizational change is not a continuous process of adaptation to environmental pressures and internal imperatives, but rather that there is a discrete process in which the forces that overcome organizational inertia in a particular area can lead to a series of attendant changes. One of the divisions that introduced more general changes to its dispute resolution procedures in conjunction with the introduction of nonunion arbitration was TRW’s Systems Integration Group (SIG). Prior to the introduction of

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

85

arbitration, SIG had a simple grievance procedure in which employees could lodge written complaints with senior management, who would either accept or reject the complaint. Accompanying the introduction of arbitration, SIG established a new three-stage dispute resolution procedure. The first stage, “management review” consisted of the internal complaint procedure. The employee would begin the first stage by presenting a written statement of the dispute to the immediate supervisor. If unresolved, the next part of stage one would be for the complaint to be reviewed by a higher-level manager who had not been involved in the decision that was the subject of the complaint. During these two parts of stage one the facility’s human resources manager would provide internal mediation to try to resolve the dispute. If unresolved, the employee could take the dispute to stage two, “appeals committee/mediation,” which provided two mutually exclusive options. In the first option, the complaint could be reviewed by an appeals committee, consisting of a five member panel of fellow employees with the senior human resources director serving as a non-voting chair. The appeals committee was a peer review type of panel, though unlike the randomly selected VSSI panels, in SIG employees could nominate their fellow employees to serve on the pool from which the committee is drawn. The employee selected three members from the pool and management selected the remaining two. The other option in stage two was to have the complaint brought to an external mediator who would attempt to facilitate a negotiated resolution of the complaint. As will be discussed later, after a period of experience with the procedure, stage two was modified to eliminate the appeals committee option due to lack of employee usage. If the complaint was unresolved at stage two, the employee could proceed to stage three “alternative dispute resolution.” Where the employee selected the appeals committee option in stage two, they were required to participate in external mediation as an initial part of stage three. If the complaint was unresolved in mediation, the final part of stage three was arbitration. The arbitrators are chosen from American Arbitration Association lists of employment arbitrators. For disputes involving allegations of sexual or racial harassment or termination, the employee had the option to bypass stage one and/or stage two. Disputes involving discipline for violation of sexual or racial harassment policies or legal and ethical conduct policies had to be taken directly to stage three mediation and arbitration. The inclusion of the appeals committee and mediation options in the SIG procedure is particularly noteworthy. The combination of appeals committees and mediation provide an additional dispute resolution structure within the division that goes beyond the basic nonunion arbitration procedure required under the company directive to help shield against litigation. As will be described in more detail below, these additional new institutional structures may have particular importance for the process of resolving disputes within the firm.

86

ALEXANDER J. S. COLVIN

The other divisions of TRW adopted a range of combinations of elements in their procedures. In contrast to the multi-stage procedure at SIG, with its appeals committee and mediation options, the procedure introduced for staff at TRW corporate headquarters simply provided for arbitration of legal claims. Within the Automotive group, the Valve Division introduced both peer review and arbitration procedures in response to the corporate directive. Whereas at VSSI, the division adopted a similar approach to SIG and introduced mediation and arbitration as additional stages to its dispute resolution procedure, supplementing its existing peer review panel procedure. Although the automotive business group has a larger workforce, employees within the SDIT group made greater use of the new mediation and arbitration procedures. During the first three years after the introduction of the new procedures, from 1995 to 1997, out of a total of seventy-two mediation and arbitration cases at TRW, fifty-five were in the Space, Defense, and Information Group. The vast majority of these claims were either resolved or abandoned during or after the mediation stage, with only three of the seventy-two claims during this initial three-year period going to arbitration. A range of different types of cases were brought to mediation and arbitration during this period, including: 27 layoffs involving employees in protected groups; 20 wrongful discharge claims of all types; seven disputed applications of company policies or discipline short of termination; five Americans with Disabilities Act claims; five claims of sexual harassment or discharges for sexual harassment; four promotion denials or demotions; two Equal Pay Act claims; and two failures to rehire following layoffs. With the majority of claims being resolved or abandoned at mediation, this stage assumed particular importance in the impact of the new procedure. The experience with mediation was very positive for management, with most cases being resolved quickly and at low cost in mediation. Time to resolve cases in mediation averaged only 3–4 months from the filing of the initial claim. The average cost to management of resolving cases in mediation has been remarkably low, at around $2000 per case. However, this figure is based on the direct monetary costs of settlements to the company, which does not include indirect costs to the company and often understated the value of the remedy to the employee. A number of the settlements in mediation involved non-monetary remedies. Reinstatement of terminated employees, adjustment of the application of policies, and finding new jobs for employees within the company were common non-monetary elements in settlements that provided substantial benefits to complainants without direct monetary costs to the company. Employees used attorneys in mediation in only around half of the cases. This may reflect an employee perception that they did not need representation at this stage in the procedure or that it was to their advantage to resolve their complaint quickly

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

87

and cheaply without incurring the additional expense of an attorney. However it is also possible that employees may have felt that bringing an attorney into the proceedings would be viewed as overly confrontational and negatively affect their future career in the company. Whatever the motivation of the employees, it is of concern that employees may have been settling potential legal claims in mediation without the benefit of consultation with legal counsel. An interesting additional development in the area of representation is that a number of complainants brought representatives other than counsel with them to mediation. In some cases, spouses accompanied the complainant to mediation. In one instance, the complainant’s priest attended mediation to assist the employee. Although it is difficult to judge the effect of these non-attorneys on the mediation process, they introduce a new and potentially significant element into the dynamics of the dispute resolution procedure. In many cases these nonattorney representatives were very forceful in pressing the complainant’s position at mediation, more so than the complainants themselves. In addition to the direct resolution of cases, the introduction of mediation and arbitration had important effects on the management of employment relations at TRW. For the in-house labor and employment counsels it led to a shift towards sending more of their time on proactive work and less reacting to developments in litigation. One attorney described having shifted from spending 60% of working time before the introduction of the procedures on litigation of cases, to spending 80% of working time after their introduction on counseling managers and helping human resources deal with employee complaints. A major part of shifting to a more proactive focus was the ability to respond to disputes more quickly. The past experience of the company had been that cases would take a couple of years to come to trial, by which point the manager involved might have left the firm. The company might also have initially accepted the manager’s account of events, but then subsequently it was demonstrated at trial that the manager was engaged in illegal conduct and the manager’s behavior had continued on unchecked for the interim period causing further harm and mounting damages. With less than a year being taken to bring cases to mediation and arbitration, the company was able to take remedial action more quickly and managers were more likely to be promptly held responsible for the consequences of their actions.

ANALYSIS The development of peer review and nonunion arbitration procedures at TRW were the product of a combination of external pressures and internal management decisions that led the company to establish these new structures governing

88

ALEXANDER J. S. COLVIN

employment relations within the firm. Institutional environmental pressures and management strategies are often seen as alternative explanations for the development of organizational structures and practices. Yet in this instance, the factors of environmental pressures and management strategies worked together in determining the direction of development of procedures. In a period of declining unionization, deregulation, and globalization, institutional forces are often viewed as being of diminished significance in determining the content of employment relations, particularly in the United States. However it is striking that in the case examined here, which involved a company that is in many respects an exemplar of a “High Performance” or “High Commitment” approach to managing the workforce, institutional pressures played an important role in shaping organizational practices. Furthermore, this was not only the case in the “old” industrial context of automobile parts manufacturing where unions still have a substantial pressure, but also in the “new” context of the high tech aerospace sector where litigation pressures create a similar disciplining pressure on management to that of unions in the older context. Yet it would also be incorrect to solely emphasize the institutional pressures involved and downplay the significance of management strategies. Neither union organizing nor litigation threats necessitate one specific response from management; in either case there was a range of responses the company could take. In each instance, the particular response to the environmental pressure that management chose was a product of the company’s overall human resource strategy. In the context of the potential for union organizing activity, the choice of practices such as the adoption of peer review procedures provides a union substitution approach that fits with a high commitment strategy, as opposed to the alternative of simply relying on suppression of organizing drives. Similarly, in adopting an alternative dispute resolution procedure to respond to litigation pressures, the company could have taken an approach of simply requiring employees to arbitrate claims and designed a procedure that lacked significant due process protections. In this case, the decisions to introduce more elaborate procedures with a major mediation component and, especially, to make the arbitration procedure non-binding for the employee were a product of the desire to balance protection of the company from litigation with a human resource strategy that emphasized fostering high commitment and trust levels with its workforce. Whereas there is a strong common thread in the interaction of environmental pressures and management strategies in the development of the peer review and nonunion arbitration procedures, the most striking contrast in this study involves the dynamics of procedures for white-collar and blue-collar employees. Even though the study examined employees within the same company with a common orientation to human resource strategy, the patterns of development and usage of dispute resolution procedures were distinct and different. For the blue-

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

89

collar employee workforce in the automotive group, the dominant environmental pressure was the threat of union organizing, with litigation by employees playing no significant role. By contrast, for the white-collar employee workforce in the SDIT group, employment litigation was the dominant environmental pressure, with union organizing threats being non-existent and irrelevant. The contrasts in types of environmental pressure produced different directions of development of dispute resolution procedures to respond to the specific nature of the threat involved. To respond to the increase in litigation activity among the white-collar workforce in SDIT, adoption of the nonunion arbitration procedures provided a mechanism to transfer these claims into an alternative dispute resolution structure that could produce faster and cheaper resolution of claims, while substantially reducing the threat of the uncertainties inherent in the jury-driven litigation system. In contrast, in responding to the threat of potential union organizing and related employee expectations for fair treatment among the blue-collar employee workforce of the automotive group, peer review procedures provided a mechanism for employee involvement in the area of dispute resolution that could offer a better alternative to union grievance procedures than simple management decisionmaking in responding to complaints. The contrast between white-collar and blue-collar employee groups in the development and structure of dispute resolution procedures was also reflected in how the procedures were used. Although peer review procedures were available to both blue-collar and white-collar employees in the automotive divisions where it was adopted, in practice it was used almost exclusively by blue-collar employees. Even more strikingly, when peer review was offered as an option that could be used instead of mediation in the procedure adopted for the SIG division’s whitecollar workforce, the peer review option was used so infrequently that the option was later dropped from the procedure. The explanation offered within the division for this lack of usage of the option was that the white-collar employees perceived that mediation would provide them with a more independent dispute resolution processes than peer review. One of the particular concerns with peer review for many white-collar employees is that their effective peer group includes employees who are in management or on career tracks leading into management. If the peers on the panel bring a managerial perspective to the dispute resolution process, much of the distinctiveness of peer review compared to a grievance procedure in which a board composed entirely of managers hears the complaint is lost. This difference has broader implications when we consider the differing nature of employment relations for white-collar and blue-collar employees. Traditionally industrial relations theory and research has emphasized a sharp divide between labor and management as two separate actors in the industrial relations system. In the nonunion context studied here this analysis held up well in looking at the

90

ALEXANDER J. S. COLVIN

blue-collar employee workforce, but broke down in looking at the white-collar employee workforce where the employees did not necessarily identify with their fellow employees as a group with common interests in opposition to managers within the company. For these white-collar employees, external mechanisms of the courts and independent neutrals in dispute resolution provided the supporting institutional structure for them to further their interests, rather than a structure internal to the workforce of the company such as organization either in the stronger institutional form of a union or in the weaker institutional form of worker participation in peer review procedures. Two major criticisms that have been leveled at nonunion dispute resolution procedures are that they are lacking in due process protections and that usage of them by employees tends to be at very low levels. If the TRW procedures represent a “best case” example among current nonunion procedures, what does it tell us about these criticisms? In the area of due process protections, the most negative picture of nonunion procedures is not supported by this study. In both the peer review and nonunion arbitration procedures, substantial protections are included to enhance the neutrality of the decision-makers, ranging from the jurylike selection procedures for peer review panelists to the use of arbitrators and mediators provided by reputable neutral service providers. Balanced against the features aimed at ensuring neutrality of decision-making is the reality that these are procedures structured by management that limit how far decisions can go in affecting organizational rules and procedures. Types of complaints are limited under both types of procedures, with nonunion arbitration and mediation restricted to potential legal claims and peer review to disciplinary and dismissal decisions that are alleged to violate company policies and procedures. An area of concern particular to the peer review procedures is limitations on the ability of parties to bring representatives to the panel hearings. Although this restriction applies equally to the supervisor who made the decision being challenged, the absence of employee representatives is a significant weakness from a due process perspective. In general, the overall picture of the procedures is one of a reasonable degree of due process protections, but at the same time with some substantial limitations on the procedures designed to protect the company. When we turn to the usage of the procedures, the more dire pictures of nonunion procedures again appear unwarranted. Employees used both the peer review and nonunion arbitration procedures to a substantial degree, albeit at a rate well below usage of many union grievance procedures. Although employee win rates were not high under the peer review procedures, employees continued to use them frequently and in a number of cases where management decisions were overturned were able to continue successful careers with the company. Indeed compared to the unionized context, the relatively low win rates in the peer review procedures

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

91

may be somewhat misleading given that there is no effective check on the ability of an employee to file a complaint. Whereas unions typically undertake an evaluation of the strength of a grievance in deciding whether or not it is worth taking the complaint to arbitration, under peer review the employee alone decides whether or not to proceed. Given peer review is a relatively quick and low cost procedure for the employee it is unsurprising that employees would be willing to take a chance on bringing relatively weak complaints to peer review whereas a union might not consider it worth investing the costs of arbitration on a similar weak grievance. It is important to recognize that the most significant effects of these types of dispute resolution procedure may lie in their impact on the general conduct of employee relations by management, rather than on the resolution of specific cases. In this regard, it is noteworthy that some of the strongest concerns about the introduction of peer review came from line management concerned about having their decisions challenged and reviewed by the panels. Indeed, managers noted that one of the main impacts of the adoption of the peer review procedure was to increase the time and care taken in reaching and executing decisions to discipline and dismiss employees. Viewed from this perspective, the procedures served as a tool for human resources within the organization that helped ensure line managers were actually following the policies adopted by the company directed at ensuring fair treatment of employees in the workplace. The nonunion arbitration procedures had a similar impact on helping promote fairer treatment of employees by managers, particularly due to the more rapid resolution of complaints. Given that a manager was likely to have his or her decisions reviewed in a much shorter time frame than had previously been the case, the probability that improper actions would come to light and be remedied was increased. On a broader level, an interesting development was that in some instances employee complaints under the procedures led to modification and improvement of organizational practices and procedures. For example, following a complaint that indicated a weakness in an employee relocation policy, the policy was changed to fix the problem, improving the assistance provided to employees. This occurred even though the procedures themselves specified that decisions could not modify company policies and suggests the utility of the procedures as organizational learning systems for nonunion organizations through which complaints reveal problems and weaknesses in policies and practices that might not otherwise be identified.

CONCLUSION This case study set out to investigate the adoption, structure, and function of dispute resolution procedures in the nonunion workplace. The case examined here does not

92

ALEXANDER J. S. COLVIN

represent the situation in the typical nonunion workplace, rather it was selected as a “best practice” case study to investigate a stronger example of the development of this type of procedure and to see the possible direction of developments in this area. One of the motivations for this investigation was to consider the degree to which nonunion dispute resolution procedures have the potential to provide some element of fairness and voice in the context of a workforce where union representation has declining reach. The answer from this study is that nonunion procedures can serve a partial function in enhancing fair treatment and employee voice in the workplace, but the degree to which they will serve this function is conditional on both the continued existence of external institutional pressures that provide an incentive to companies to adopt procedures and the strategic orientation of nonunion companies in managing their workforces. The absence of either factor will reduce the probability of the development of effective dispute resolution procedures in the nonunion workplace. It is significant that even in this setting of a “best practice” case where the human resource management orientation of the firm was likely to be especially focused on ensuring fair treatment in the workplace, environmental pressures from litigation and unions provided the initial impetus for the introduction of the nonunion procedures. If environmental pressures are a key factor in the adoption of nonunion procedures in the “best practice” case setting where we would expect human resource management strategies to play a stronger role, then we should anticipate that environmental pressures will be even more critical for ensuring fair treatment in the more typical “non-best practice” nonunion workplace. A second lesson from the study is that promoting effective nonunion procedures requires different factors for different parts of the labor force. Whereas an institutional solution of promoting the continued vitality of union representation and collective bargaining is clearly of importance for the blue-collar labor force, other solutions, in particular a robust litigation system combined with alternative dispute resolution mechanisms that enhance rather than undermine rights, are necessary to deal with the problems of white-collar segments of the labor force. For industrial relations research and theory this suggests that while the traditional ideas of ensuring strong institutions in the governance of employment relations remain relevant, they need to be adapted and applied to different types of solutions if they are to be effective across the whole labor force rather than one segment of it. A similar challenge is posed from an industrial relations perspective in the findings from this study that while these nonunion procedures clearly have their weaknesses and limitations, they also are serving an important role for nonunion employees within the organization in giving them a structure that can be used to challenge unfair decisions and helping to shape how managers deal with employees in the workplace. Although not a strongly institutionalized structure for governance of

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

93

employment relations of the form seen in American unionized workplaces with their collective agreements and grievance arbitration procedures, neither does this appear as a situation of unchecked managerial discretion. Rather the combination of external pressures and the company’s development of an internal institutional structure through these dispute resolution procedures creates a situation that can be described as a semi-institutionalized structure for governance of employment relations. If this type of semi-institutionalized structure is viable and effectively excludes external actors such as national unions from the workplace, then it poses a major challenge to the conventional industrial relations system based on strongly organized workplace representation through collective bargaining. Questions that will need to be answered in relation to the development of this type of semi-institutionalized governance structure include: how sustainable are these procedures; how widely will they diffuse across different organizations; and what is the role of employee representation in relation to these procedures? For the union movement, this last question will be particularly critical if they are not to abandon prospects for providing representation across a growing segment of the labor force.

AFTERNOTE Subsequent to the period covered by this study, in 2002 TRW was acquired by the major aerospace company Northrop-Grumman. While the former TRW Space, Defense, and Information Technology (SDIT) group became part of NorthropGrumman, the TRW automotive parts group became an independent firm under the name TRW Automotive. The implications of the acquisition and division of the company for the future of the dispute resolution issues described in this study are yet to be determined, though the shift in corporate ownership may not mean a drastic shift in approach to employee relations. Northrop-Grumman is itself known for its longstanding nonunion grievance procedure that has used arbitration as its final step since the 1940s, so it is by no means clear that the acquisition will result in a major shift in the character of the procedures in the former SDIT group. However, the potential impact of this corporate acquisition is a reminder of the importance of both management strategies and firm history in determining the path of evolution of employment relations within organizations.

ACKNOWLEDGMENTS The author would like to thank Harry Katz, Rose Batt, Katherine Stone, David Lipsky, David Lewin, and anonymous referees for helpful comments. My particular

94

ALEXANDER J. S. COLVIN

thanks are owed to the employees of TRW who gave generously of their valuable time for the interviews for this research project.

REFERENCES Barringer, M. W., & Milkovich, G. T. (1998). A theoretical exploration of the adoption and design of flexible benefit plans: A case of human resource innovation. Academy of Management Review, 23(2), 305–324. Boxer, J. (2000). Arbitration/peer review of statutory employment claims: What are the companies doing? Presentation at Alternative Dispute Resolution in the Employment Arena, 53rd Annual NYU Conference on Labor, New York University School of Law, New York, NY. Edelman, L. B. (1990). Legal environments and organizational governance: The expansion of due process in the American workplace. American Journal of Sociology, 95, 1401–1440. Edelman, L. B., Abraham, S. E., & Erlanger, H. S. (1992). Professional construction of law: The inflated threat of wrongful discharge. Law and Society Review, 26(1), 47–83. Edelman, L. B., Uggen, C., & Erlanger, H. S. (1999). The endogeneity of legal regulation: Grievance procedures as rational myth. American Journal of Sociology, 105(2), 406–454. Feuille, P., & Chachere, D. R. (1995). Looking fair or being fair: Remedial voice procedures in nonunion workplaces. Journal of Management, 21, 27–42. Feuille, P., & Delaney, J. T. (1992). The individual pursuit of organizational justice: Grievance procedures in nonunion workplaces. Research in Personnel and Human Resources Management, 10, 187–232. General Accounting Office (GAO) (1997). Alternative dispute resolution: Employers experiences with ADR in the workplace. Washington, DC: United States General Accounting Office (GAO/GGD97–157 ADR in the Workplace). Hammer, T. H., & Stern, R. N. (1986). A yo-yo model of cooperation: Union participation in management at the Rath packing company. Industrial and Labor Relations Review, 39, 337–349. Jacoby, S. M. (1997). Modern manors: Welfare capitalism since the new deal. Princeton, NJ: Princeton University Press. Kochan, T. A., Katz, H. C., & McKersie, R. B. (1994). The transformation of America industrial relations. Ithaca, NY: ILR Press. Lewin, D. (1987). Dispute resolution in the nonunion firm: A theoretical and empirical analysis. Journal of Conflict Resolution, 31(3), 465–502. Lewin, D. (1990). Grievance procedures in nonunion workplaces: An empirical analysis of usage, dynamics, and outcomes. Chicago-Kent Law Review, 66(3), 823–844. Olson-Buchanan, J. B. (1996). Voicing discontent: What happens to the grievance filer after the grievance? Journal of Applied Psychology, 81(1), 52–63. Rubinstein, S. A., & Kochan, T. A. (2001). Learning from Saturn: Possibilities for corporate governance and employee relations. Ithaca, NY: ILR Press. Schwartz, R. M. (1994). The skilled facilitator. San Francisco, CA: Jossey-Bass. Slichter, S., Healy, J. J., & Livernash, E. R. (1960). The impact of collective bargaining on management. Washington, DC: Brookings Institution.

Adoption and Use of Dispute Resolution Procedures in the Nonunion Workplace

95

Stone, K. V. W. (1996). Mandatory arbitration of individual employment rights: The yellow dog contract of the 1990’s. Denver University Law Review, 73, 1017–1050. Stone, K. V. W. (1999). Employment arbitration under the Federal Arbitration Act. Chapter 2. In: A. E. Eaton & J. H. Keefe (Eds), Employment Dispute Resolution and Worker Rights in the Changing Workplace. Champaign, IL: Industrial Relations Research Association.

THE ORGANISATIONAL OMBUDS: IMPLICATIONS FOR VOICE, CONFLICT RESOLUTION AND FAIRNESS AT WORK Sue Fernie and David Metcalf 1. INTRODUCTION In his review of theoretical and empirical research on grievance procedures, Lewin (1999) states that the “grievance procedure is widely regarded by scholars and practitioners as the centerpiece of union-management relations.” It is somewhat strange, then, that a trawl through British industrial relations publications for the 1980s and 1990s reveals very few dealing with the process for resolving employment disputes in unionised workplaces (usually articles about industrial tribunals, now called employment tribunals). Given this paucity of studies in unionised workplaces, it is less surprising that almost no research has been published recently on how employees and management in non-union firms go about dealing with individual conflict in the workplace today. Whilst industrial relations scholars may have used the excuse of the drastically reduced incidence of collective responses to drop the study of conflict altogether from their recent repertoires, other disciplines still recognise that conflict is an allpervasive element of workplace life. One common conclusion is that conflict can be constructive if managed properly but destructive if not. The positive benefits of conflict include the airing of problems that the employer needs to address and the

Advances in Industrial and Labor Relations Advances in Industrial and Labor Relations, Volume 13, 97–135 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0742-6186/doi:10.1016/S0742-6186(04)13004-7

97

98

SUE FERNIE AND DAVID METCALF

resulting beneficial changes in such areas as policies, procedures and productivity. The negative consequences can include reduced productivity, emotional distress and the escalation of minor differences into major hostilities. This paper deals with an innovative form of alternative conflict resolution – the organisational ombuds, who is “a confidential and informal information resource, communication channel, complaint handler, mediator and dispute resolver who may look into problems informally and work for systems change” (Rowe, 1997). Two threads are woven together: first, is this informal type of problem solving a means to an end – is it just the first step individuals may go through before they embark upon a more formal grievance against their employer – or can the ombuds act as a conduit through which reactive voice can be turned into proactive voice with beneficial opportunities for the organisation? This thread embraces themes one and two in the next section, namely reactive vs. proactive voice and the difference/similarity between the ombud and a grievance procedure. And second, what are the outcomes for workers and organisations? For example, do individuals in (primarily) non-union workplaces who can avail themselves of this alternative form of conflict resolution experience the same sort of outcomes in terms of procedural fairness as those in organisations which use a more formal, mainly reactive form of dealing with grievances? This thread is examined further in theme three below, which discusses different forms of workplace justice. We focus on four case study organisations in the following industries: retail sales (termed “Shop” in what follows); manufacturing and servicing equipment vital to offices and other organisations (“Liftco”); financial services with emphasis on credit card transactions (“Finser”); and an international financial institution with its headquarters in London (“IFI”). These are the only four organisations in the U.K. which have an ombud function, so we essentially had no choice over our case study organisations. Our method of investigation included interviews with the respective ombuds, surveys of users and non-users, and use of documentary evidence. We were not able to conduct “twin” studies of otherwise similar organisations with and without an ombud, nor were we able to undertake a before/after analysis. There is considerable concern in the U.K. that the number of cases going to the labour courts (Employment Tribunals, ET) has doubled over the last decade to some 130,000. These cases cover, inter alia, unfair dismissal and sex and race discrimination. The Employment Act 2002 contained measures to reduce this case load. Employers must now have a written grievance and disciplinary procedure. And an employee must normally exhaust the procedure prior to taking a case to the ET. These provisions have caused many organisations to rethink their voice and dispute resolution procedures, so this study of the workings of the organisational ombud in the four U.K. institutions with such an arrangement is timely because other firms and public sector bodies are contemplating adopting such a system.

The Organisational Ombuds

99

The ombud is able to deal informally with the sort of cases which may otherwise go to a formal ET, for example, dismissal and discipline, but can also handle more interpersonal issues such as the behaviour of line managers and peers, poorly functioning teams and staff development. In the past, some of these matters may have been dealt with by trade unions, but the near halving of union membership in the last two decades has opened the door to alternative voice and dispute resolution mechanisms. In the second section we outline some of the themes concerning the nature of justice and employee voice in organisations. The third section examines in some detail the role of the organisational ombuds and details how this function interrelates with both unions and other aspects of human resource management in organisations. The fourth section describes our case study organisations and in the fifth section we detail the methods of investigation used to study the institution, process and outcomes of the ombuds’ function. The sixth section reports and discusses our evidence on matters like proactive vs. reactive conflict resolution and voice, procedural justice, organisational outcomes and comparisons with formal procedures such as Employment Tribunals. Conclusions, including some implications for the future of unions, are set out in the final section.

2. THEMES Three themes emanating from the literature on employment dispute resolution are used to organise this discussion of voice and conflict resolution. First, the merits of reactive and proactive voice are discussed, and it is suggested that the role of the ombuds is to act as a conduit from the former to the latter. Next we analyse whether the ombuds should be viewed as the first step on the way to a formal grievance and appeal system or as something qualitatively different. Finally, ideas of workplace justice – distributive, procedural and interactional – and the possible mediating effect of the ombuds are presented. 2.1. Reactive vs. Proactive Voice The issue of conflict resolution at work is inextricably bound up with issues of worker voice. Both grievance procedures and many of new forms of work organisation, for example self-managed teams, can be seen as forms of employee voice, and “exit-voice theory” has been used to explain lower turnover in organisations with some form of worker voice. But there is great variety in the types of voice on offer. The grievance procedure is mainly a reactive form of voice, i.e. a complaint is made by an individual that his or her contractual rights have been

100

SUE FERNIE AND DAVID METCALF

violated and the instigation of the grievance is a reaction to a manager’s action. Another type of voice is having the opportunity to influence decisions before action is taken e.g. deciding the method of production with proper employee involvement, and this is termed proactive voice. However, the majority of workplaces do not have this mechanism (Cully et al., 1999, Table 5.5) The Information and Consultation Directive from the European Union, which comes into force in 2007, directs organisations with 50+ employees to establish mechanisms to consult and inform their employees and can be viewed as an opportunity to enhance proactive employee voice (Gospel & Willman, 2003), but for now, most managements unilaterally decide both pay and terms of work in addition to those matters of work organisation not covered by contract. Both types of worker voice are useful. Without the opportunity to react, people do not have recourse to a fair hearing for unjust treatment. Without the opportunity to be proactive, workers have less influence at the workplace nor can management take advantage of their knowledge. Unions are normally a source of reactive voice while high commitment work practices (HCWP) ensure proactive voice. It is worth noting that, in both U.S. and U.K., workplaces with both unions and HCWP have higher labour productivity than counterpart workplaces with just one or with no voice mechanism (see Metcalf, 2003). The ways in which conflict is handled at work will vary according to, among other things, the type of workplace governance (Lewin & Peterson, 1988). The Oxford School pluralists such as Clegg and Kahn Freund, who believed conflict to be an inevitable aspect of working life, saw trade unions as having the collective right to jointly regulate terms and conditions with employers, or even better with employers organisations. This pluralist voice mechanism takes the forms of both collective representation and negotiated grievance procedures. The decline in union power and greater emphasis on individual rights at work during the 1980s, e.g. the right not to belong to trade union, led to a new focus on individual rights. Consequently, the number of applications to Employment Tribunals trebled in the 1990s, just when strike activity hit a post-war low around the millennium. With the advent of high commitment work practices, it is hoped that the management-employee relationship will become less adversarial. The existence of conflict is negated, and “them and us” becomes “us (i.e. workers and management within the firm) vs. them (i.e. the competition).” Commitment wins over control, co-operation prevails over adversarialism (Walton, 1985). Worker participation in decision-making, often in areas traditionally reserved for management such as how to allocate work, team working, flexibility and empowerment are seen as routes to success. To the extent that union-management relations become more co-operative, some trade union reps are backing away from their traditional roles as “prosecutor” or “defence” in resolving grievances. In this context, what happens to workers’

The Organisational Ombuds

101

rights? Do they still receive due process (procedural justice)? As we shall see, one possible solution to ensure such justice is the existence of the organisational ombuds.

2.2. Is the Organisational Ombuds the First Step of a Grievance Procedure, or a Qualitatively Different Thing? “Informal discussion of grievances may be said to have an integrative (problem solving) quality that bespeaks preventive conflict resolution, rather than problemoriented formal grievance filing that bespeak reactive conflict resolution” (Lewin, 1999). Most theories of the grievance procedure at work are predicated on the fact that there has been a deterioration in the aggrieved employee’s relationship with his/her employer, and as a result they concentrate on reactive forms of workplace conflict resolution to the exclusion of any other. We would suggest, on the other hand, that an informal system – yet one which has the power to effect change, like the organisational ombuds – is a qualitatively different thing. It is not just another step in a multi-step grievance procedure which will ultimately end up in a quasi-judicial situation, but a method by which problems can be aired and dealt with in a timely fashion and both behavioural and organisational change can be accomplished. Proponents of a particular form of individual conflict resolution – mediation – use the discourse of interests and needs, and turn attention away from a person as a carrier of rights. But individuals have interpersonal-based problems as well as rights-based problems. For example, an employee who suffers sexual harassment should not be expected to compromise with the harasser to reach an agreement in a mediation. If she chooses not to file a formal complaint, she should know that something will still be done. The ombuds can investigate, can go straight to the CEO, and can publicise the wrongdoing. Therefore the ombuds may not satisfy those who only want fully fledged adversarial litigation, but it addresses more of the critics of ADR’s concerns – especially the lack of involvement of the third party – than does mediation. Where a formal grievance is too much, and mediation is too little, the ombuds may be the optimal solution. S/he is not adversarial like litigation, but s/he can take sides, as opposed to a mediator, who cannot. This does not always mean sides as in favour of one party over the other, but it does always mean “sides” as in favour of honesty, integrity and principles. The idea of third party mediation in general, and the ombuds in particular, in individual grievances is a newish one in U.K. employment relations. Two important factors that will dictate its success are managerial behaviour and trust. In a unionised workplace, a grievance usually means that a union is seeking to

102

SUE FERNIE AND DAVID METCALF

challenge management authority, and management will seek to maintain the status quo. The acceptance of a third party requires the employer and employee to come to some compromise in nearly every case. As such, many managements may fear the existence of a third party as a challenge to their prerogative. Mediation is therefore seen as having a better outcome if it is mutually agreed, but the desire to always press for a win may be indicative of the adversarial nature of employment relations in the U.K. The adoption and success of the ombuds role requires greater co-operation in dealing with conflict. It is therefore likely that the success of an ombuds programme – defined in this case as being the ability to nip problems in the bud informally – will depend on the attitude of senior management to the function and the degree of harmony in employment relations at the workplace. When an organisation establishes an ombud programme, the aim is for disputes/problems to be settled informally through the good offices of the ombud rather than simply being the first step in a formal grievance procedure. Most problems at work are settled in informal discussions between workers and supervisors. In Routes to Resolution the Department of Trade and Industry (DTI) estimates that 900,000 disputes annually could lead to an employment tribunal, but most get solved beforehand. Most tribunal cases are resolved prior to a hearing. Nevertheless, in order to reduce the numbers of cases going forward to tribunal, Routes to Resolution proposed that ALL employers must have written grievance and disciplinary procedures, to a minimum standard. An employee MUST exhaust the procedures before applying to a tribunal (except if s/he is bullied or intimidated at work). The Employment Act 2002 therefore provides a good opportunity for companies without such procedures to think about dispute resolution in the workplace, but it is unlikely that simply producing a list of formal steps is will alter the way in which conflict is dealt with, especially for the 20% of employees who currently work in places with no procedures.

2.3. Workplace Justice – Distributive, Procedural and Interactional Three strands of workplace justice have been identified in the OB literature: distributive, procedural and interactional. Distributive justice is concerned with the fairness of the outcomes of a decision, e.g. selection, pay rises, promotions. Procedural justice is concerned with the extent to which perceptions about the fairness of outcomes are based on processes and procedures used to obtain them (Folger & Greenberg, 1985). Interactional justice is concerned with workers’ sensitivity to the quality of interpersonal treatment they receive during the enactment of organizational procedures (Bies & Moag, 1986), e.g. is the supervisor considerate? Did s/he consider the employee’s rights and deal truthfully with the case? Did s/he explain the decision made? Moorman (1991) found that employees’

The Organisational Ombuds

103

impressions of the fairness of their interactions with supervisors communicated more to them about trust and equity than did fair procedures. Normally an organisational ombud deals with procedural and interactional justice – process and procedure in the workplace. The ombud does not have a direct say in, for example, pay and promotion outcomes – distributive justice – but may help ensure colleagues are content with the process by which such outcomes are reached. We might expect the ombuds to influence both procedural and interactional justice at the workplace, and that outcome variables such as job satisfaction, satisfaction with management and certain citizenship behaviours will be influenced by the mediating impact of the ombuds. Such evidence is presented in the sixth section. Thus the ombud plays a similar role to a union. For example, the union’s role in providing procedural justice via its representation in the grievance system is held to be more significant than other benefits provided by the union (Gordon & Fryxell, 1993). Further, such procedural justice contributes importantly to members’ satisfaction with both their union and their management. But, in addition, the ombud can act as a conduit between reactive and proactive forms of employee voice which, in principle, can enhance outcomes for both employees and management. Further, the majority of workplaces do not recognise a union and therefore the establishment of an ombud programme in such organisations is likely to enhance justice at work.

3. ROLE OF ORGANISATIONAL OMBUDS 3.1. Role and Principles An organisational ombuds is concerned with both conflict resolution and provision of employees’ voice. S/he “fosters values and behaviour such as fairness, equity, justice, equality of opportunity and respect” (Rowe, 1997). The Ombudsman Association (1995) describes the role as follows: “The mission of the organisational ombuds is to provide a confidential, neutral and informal process which facilitates fair and equitable resolutions to concerns that arise in the organisation. In performing this mission, the ombuds serves as an information and communication resource, upward feedback channel, advisor, dispute resolution expert and change agent.” The Ombudsman Association (TOA) is the professional body for corporate and organisational ombuds. It sets out a range of standards of practice and provides training and professional certification. The four key principles are: confidentiality, neutrality, informality and privilege. First, the practice is based on confidentiality. An ombuds does not use the names of individuals without express permission. During the problem-solving process an ombuds may make known information as long as the identity of the individual

104

SUE FERNIE AND DAVID METCALF

contacting the office is not compromised. Second, an organisational ombuds is neutral and is independent of ordinary line and staff structures. He or she serves no additional role within the organisation which would compromise this neutrality. Third, the ombuds is an informal, off-the-record resource. Formal investigations – for the purpose of adjudication – should be done by others. Thus the ombuds does not normally act as an agent for the organisation. Further individual workers are not required to meet with the ombuds – all interactions with the ombuds are voluntary. Fourth, there is a privilege with respect to communications with the ombuds and s/he resists testifying in any formal process inside or outside the organisation. (Privilege is a legal term which describes a relationship which the law protects from forced disclosure. Traditional privileges are client/lawyer, doctor/patient, priest/penitent, husband/wife.)

3.2. Appointment, Cases and Functions In some organisations, the ombuds is appointed from within the institution, and sometimes this is the last job they will hold in that organisation prior to retirement. In other organisations the ombuds is appointed from outside. In yet other organisations, the ombuds remains an outsider, employed on a consultancy basis. Such an arrangement possibly enhances the neutrality and confidentiality principles noted above. An organisational ombuds deals with a great variety of cases. For example, at IFI cases include: job content, working conditions, e.g. downgrading, responsibility; relationship with other staff, including line manager, e.g. bullying, discourteous behaviour; performance evaluation, promotion, salary and bonus, both process and outcome; career development and training; benefits and allowances administration, e.g. rights to specified benefits; harassment and incidents of perceived discrimination. It is also worth pointing out what an organisational ombuds does not deal with (see also Rowe, 1997). S/he is not a safety, ethics, audit or equal opportunities officer. These are all very specific functions whereas ombuds’ cases are often both undefined and complex. Similarly the ombuds is not a counsellor like those running an employee assistance programme nor is s/he a compliance officer, making sure, for example, that the organisation enforces relevant procurement or health and safety rules. Unlike a union steward, the ombuds does not primarily focus on complaints arising under contract. It is apparent that the two key functions of the organisational ombudsman concern dispute resolution and voice (see, for example, American Bar Association, 2000). The manner in which the ombuds differs from and/or fits with more

The Organisational Ombuds

105

Table 1. Disputes and Their Resolution. Degree of Formality

Individual Both Union and Non-Union

Collective Largely Union

Very formal

Collective bargaining procedures (union workplaces only) e.g. re-dismissal

Official industrial action Arbitration • Conventional • FOA

Formal

Law suits e.g. safety Employment tribunal e.g. discrimination Grievance and appeals committee Mediation Conciliation e.g. via ACAS in discrimination cases

Mediation Conciliation

Informal

Ombudsman

Unofficial industrial action

Very informal

Line manager, open door policy

Note: ACAS is state-funded Advisory, Conciliation and Arbitration Service; FOA is final offer arbitration. Source: Adapted from Kaminski (1999).

traditional dispute resolution and voice mechanisms is set out in Tables 1 and 2. Consider first conflict and its resolution (Table 1). An ombuds deals with individuals in an informal manner. There are many other dispute resolution procedures for individual disputes. Formal mechanisms – using established due process procedures – include those established by collective bargaining if there is union recognition, lawsuits, employment tribunals, grievance and appeals committees, mediation and conciliation, perhaps via the state-sponsored Advisory, Conciliation and Arbitration Service (ACAS). The ombuds is an Table 2. Voice Mechanisms.

Source: Adapted from Kaminski (1999), Table 1.

106

SUE FERNIE AND DAVID METCALF

informal mechanism as is, for example, a line manager with an open door policy. The ombuds normally does not get involved with collective disputes. Voice in organisations can be proactive or reactive and can affect both individual and collective relations. The ombuds provides an informal, reactive voice for individuals yet contributes to the proactive resolution of conflict within the organisation. Table 2 puts the ombuds role in context by setting out alternative voice mechanisms which provide both collective and other types of communication. 3.3. Interests, Rights and Power Ury et al. (1988) note that there are three forms of “dispute” which need reconciling. First, interests which involve “needs, desires, concerns, fears.” Dispute resolution here is problem-solving, for example over a poorly carried-out performance appraisal. Second rights, which turn on informal or formal standards – does the employee have the right to a promotion after a specified tenure or a right to a particular benefit? Third, power raises concerns such as the relationship between a line manager and a subordinate. Whilst these notions of interests, rights and power are normally applied to collective disputes, they can, as shown here, also be applied to individual matters. Consider for example a case of alleged gender discrimination in promotion. This would depend, in part or whole, on the perceived power of the line manager and other colleagues who decide on such promotions. Without ombud-ADR it would be resolved via any extant formal grievance and appeal procedures in the organisation, an employment tribunal or litigation – an appeal to rights. But the ombuds focusses on interests. S/he gets the permission of the individual to pursue the case (but only once approached, since an ombuds is not an advocate), talks to the line manager, HR department, peers, etc., perhaps concluding that the colleague was not quite ready for promotion now is likely to find favour shortly. This is precisely the “back and forth communication intended to reach agreement” described by Ury et al. (p. 6). Such back and forth communication is surely cheaper than litigation; both parties are likely to be satisfied with procedures and (probably) outcomes; the parties can continue to work together and the solution is likely to be durable. 3.4. Fit with Unions, Legal Framework, HR Most organisations who have adopted the ombuds function do not also recognise a trade union, but some do (see the fourth section). In such organisations there is normally a division of labour such that the ombuds does not trespass on union territory. At Liftco, for example, the ombuds does not take cases involving pay and conditions covered by collective agreements.

The Organisational Ombuds

107

Many areas of the individual’s employment relationship with the company are defined by law. These include minimum standards for pay, working time, certain benefits such as maternity arrangements, non-discrimination and terminations. An ombuds may on occasion mediate concerning these legal rights but typically the function is much more subtle and complex. Rather than adjudicating on maternity benefits, the ombud can examine the manner in which the organisation’s familyfriendly policy is being implemented by line managers. Similarly, the ombud does not decide minimum safety standards, but it may influence compensation amounts paid to employees for accidents sustained at work. Interactions between the ombuds and HR department are complex. While the ombuds may sometimes be a thorn in Personnel’s side, equally cajoling line managers to take personnel issues seriously is helpful. Perhaps more important is the proactive (internal consultant/change agent) role played by the organisational ombudsman. By commenting on and acting on, for example, the performance appraisal or promotion process, bullying at the workplace or discriminatory procedures, s/he is likely to encourage a more strategic approach to personnel issues within the organisation.

4. THE SAMPLE OF CASE STUDY ORGANISATIONS 4.1. The Organisational Characteristics The characteristics of the organisations and the employees are set out in Table 3. The financial services company (Finser) is a private plc, foreign-owned, which has been in operation since 1850. Shop is worker-owned, British and has operated in the heart of London since the 1880s and subsequently at many other locations. Liftco is a plc with an American parent company that has been operating for 140 years. IFI is London-based, mostly foreign-owned and has only been operating since 1991. Liftco recognises a trade union in the U.K. and both it and Finser have works councils in other European countries. Finser has 8,000 employees in the U.K. and a total of 18,000 in its Europe and Middle East Area. Around two-thirds of employees are female and one-third are managers. Shop has 56,000 employees in the U.K. and none abroad, 60% of whom are female and 16% are managers. Liftco has 60,000 employees worldwide, of whom 2,500 are in Britain. Female employees and managers account for around one-fifth of the total. IFI has some 800 employees at its London base and a further 250 in its countries of operation. Three-fifths of its staff are professional and twofifths are support; just over half are women. Human resource management practices vary enormously among the four organisations. Finser has top-down communication methods, extensive

108

Table 3. Characteristics of the Organisation and Employees. Characteristic Organisation Private/public Sector Foreign/U.K. owned Age of organisation (in 2003) How long at this location Mission

Employees No. of employees % Female % Management

Shop

Liftco

IFI

Private, plc Financial services Foreign 1850 (153 years)

Private Retail U.K. 1929 (Trust), 1880s Oxford Street 122 years “The partnership’s ultimate purpose is the happiness of all its members, through their worthwhile and satisfying employment in a successful business.” No

Private, plc General engineering American parent 140 years

Public IFI Mostly foreign 1991

137 years Sale, installation, service and repair of lifts, escalators and travalators

12 years “Seeks to foster the transition to an open market-oriented economy and to promote private and entrepreneurial initiative in all 26 countries of operation” No

1976 (27 years) “To be world’s most respected brand”

UK: No EU: Yes in Some countries U.K. 8,000 EMEA 18,000 U.K. 60% EMEA 67% 33% Other 67% “associates”

U.K. 56,000 60% 16%

AEEU in UK, Works Councils in Europe U.K. 2,500 Worldwide 60,000 U.K. 20% Worldwide 15% U.K. 20% professional, 20% support and admin Worldwide 15% both

U.K. approx 800 Resident offices approx 250 54% 60% Professional 40% Support

SUE FERNIE AND DAVID METCALF

Union

Finser

Intense, e.g. Committees for comm. Top Mgt. all workforce Store level weekly mts. Partnership/store level newsletters Readers’ letters, trading results in Gazette

Top down culture Cascade communication Team briefing Toolbox talks Employee surveys DIALOG (part of Ombuds program)

Employee surveys eQuality group Staff council minutes HR website & “state of nation” talks

TQM or similar

None

Informal

ACE (Achieving Competitive Excellence) Focus groups

None

Teamworking

No – tried but didn’t work Ad hoc teams for specific projects

Yes, in dept. stores

Yes – Team building “Coverdale” and “Working transitions”

Ad hoc team building and coaching

Performance pay

Yes, e.g. • Sales % of revenue • Team bonuses (no team?) • Performance appraisal (via grid)

Yes • Merit pay, up to 40% of basic • Organisation-wide bonus, % of basic

Annual performance appraisal 360 degree Co-wide PRP axed 2000 Sales/Field incentives Management bonus Executive PRP% of salary × profitability and other key achievements

Professional staff: Annual PA Bonus Merit pay rises Promotion Support staff: Annual PA Merit pay rises Limited 360 degree

Other dispute resolution methods

No, but 2001 “facilitated discussion”

Committees for communication

Registered grievances Appeals against disciplinary ET

Formal grievance and appeal machinery

Other

ee surveys, value surveys, mgt. graded acc. to ee responses

See communication above

ee surveys Gender, bullying, ff etc.

109

Top down, strong ethic Employee surveys Value surveys Mgt. graded according to employee responses

The Organisational Ombuds

HRM practices Communication methods

110

SUE FERNIE AND DAVID METCALF

teamworking which is in the process of being reconsidered, and performancerelated pay (PRP) via team bonuses and the performance appraisal. The ombuds is the main form of conflict resolution and recently extended to “facilitated resolution” – a voluntary process where two parties in conflict discuss informally their concerns, aided by a neutral facilitator. As a result, the parties aim to develop a better understanding of their perspectives and come to a mutually satisfactory resolution. Extensive use is made of employee surveys and individual managers are graded according to employee responses. Shop operates a form of collective industrial relations, similar to works councils, called “committees for communication.” Top management meets all the workforce in a particular shop periodically. In addition, there are weekly store-level meetings, regular newsletters and a lively letters page in the house magazine (which is, however, edited by management). Stores are traditionally organised in a hierarchical fashion, and there is no formal total quality management (TQM) process. Performance-related pay (PRP) is powerful, with merit pay comprising up to 40% of basic pay and an annual organisation-wide bonus determined as a percentage of basic pay. The committees for communication provide a way for staff to express their views, make suggestions for improving the business and influence how it operates. Any member of staff can stand for election, and heads of branches are invited. They meet at least four times per year. Liftco has a variety of communication channels including cascading, teambriefings, toolbox talks, employee surveys and the Dialog programme where the ombuds answers employees’ written queries. TQM is achieved through focus groups “achieving competitive excellence.” There is extensive teamworking. Company-wide performance pay was ended in 2000 so that greater attention is now given to individual performance, including 360 degree appraisals and elements of PRP for sales, field and management employees. The company has a formal grievance and appeals system and a trade union is recognised in the U.K. Until recently HR practices at IFI were mainly concerned with recruitment and separations. There were some ad hoc surveys, e.g. concerning gender or work/life balance. Similarly, team building was via sporadic ad hoc coaching. PRP for professional employees comes through promotion and the bonus (averaging around 20% of basic pay). Support staff receive merit pay awards. IFI has formal grievance and appeal machinery.

4.2. Characteristics and Methods of Operation of the Ombuds The characteristics of the ombuds are set out in Table 4, much of which is self-explanatory. The reasons for adopting the ombuds function is a mixture of

Characteristic Adoption, etc. Why

Finser

Shop

Liftco

IFI

ee surveys incidents (money laundering) US federal sentencing guidelines

Rules 83–95 of Constitution discuss Partners’ counsellor and Registrars (1 per 15 at grocery division) Personnel Director “not too soft” P. C. “not too hard”

People values – voice, confidence, failure of other channels or inappropriate Legal/financial compliance – (Foreign Corrupt Practices Act, Federal Sentencing Guidelines)

No union IFI above the law Staff council only collective issues Grievance and appeal too formal

U.K. 1996

1929

1986

1998

No. of ombuds

2 (1 U.K.)

2 at HQ (P. C.) 40 Registrars

1 per division plus 1 Corporate Ombuds + Otis = 50 DIALOG administrators worldwide

2 (approx 1 FTE 1 day/week)

How long in post

7 years

4 years (HQ)

5 years

5 years

How filled

Internal advert Intensive interviews Representative ee panel Psychometric tests References

By Chairman from within

By invitation

By invitation, agreed by President, Staff Council, VP Personnel

Where can go afterwards

Can go anywhere, it is not pre-retirement job, could even go to HR or work for CEO

Can go anywhere, not necessarily pre-retirement. Previous Omb. → Director of Personnel

Out!

Independent of IFI so no link once appointment finishes

111

When

The Organisational Ombuds

Table 4. Characteristics of Ombudsperson.

112

Table 4. (Continued ) Characteristic

Shop

Liftco

IFI

Linked to specific grade in the organisation

By Chairman

Linked to specific grade

Agreed daily rate

Brighton – physical office plus visit other sites in EMEA area

HQ London Registrars offices

70 countries, 30 languages 2 × offices

Meeting room inside IFI or off-site

How much face-to-face

U.K. cases 50% (phone 50%) Other EMEA (Switzerland) more phone intensive

HQ 90% Registrars almost all

90% Telephone contact 10% F/f country visits

Almost all

Phones, translation, home calls, fax e-mails

Yes, all Completely separate from AMEX systems

Can phone, but normally require face-to-face

Toll free telephone, e-mail, fax, business telephone lines, secretary, intranet Language line, written translation service

Could be all but mostly initiated by e-mail then face-to-face

Documentation

shredded once case completed

HQ – retained in case ET follows

Employee name/address shredded on case closure. Rest of file kept 3–5 months max.

Kept 1 year

Can cases be anonymous

Yes

No (HQ) Occasionally at store level

Yes

Yes, but few are

How is pay set Accessibility, etc. Location, facilities, virtual office, etc.

SUE FERNIE AND DAVID METCALF

Finser

Awareness officer, site visits, newsletters, fictional scenarios, new managers, annual report, quarterly plan

P. C. visits retail locations Via Chronicle

Ombuds and Dialog Intranet, ombud web site, VHQ video conf. participation, country & co. magazine, house journals, newsletters, notice boards, flyers to ees homes, personal visits, mgt. meetings, training meetings, CD training

Intranet, ombud web page, Annual Reports, House magazine, visit to Ros

Who report to

CEO, Annual Report

Chairman, monthly

President at WHQ 2–3 times a year plus quarterly reports to President & senior mgt on activity, volume, trends

President Annual Report: simultaneously to President, Staff Council, VP Personnel

How function evaluated

Annual Report Feedback cards Channel survey to mgt. Goals – landscape

Not formally Informally through Gazette and Chronicle

Measurement of performance against agreed objectives External assessment by corporate omby Feedback from senior mgt. around world

Annual Report discussed by Board, Staff Council, Excom. Nature of satisfaction with cases discussed with Staff Council

The Organisational Ombuds

Publicised

113

114

SUE FERNIE AND DAVID METCALF

instrumentalism and altruism. U.S. law (Federal Sentencing Guidelines) requires firms to guard against corrupt practices, which helps explain the adoption of the ombuds function at Finser and Liftco. But, equally, both organisations emphasise the importance of fair treatment for their staff. The ombuds at Shop, called the Partners’ Counsellor, was introduced as part of its 1929 Constitution. And at IFI – in common with sister organisations like the World Bank and the International Labour Organisation – employees and others cannot sue the organisation, so the ombud provides important protection for employees especially given that no union is recognised. The ombuds has much intimate detail concerning the company and its employees. Yet at both Finser and Shop the ombuds can, once his or her spell ends, find a job in another part of the organisation. Indeed, at Shop the ex-ombud became the Director of Personnel! At Liftco, the ombuds job is specifically the last in the organisation prior to retirement. And at IFI the ombuds is not an employee but an outsider completely independent of the organisation, chosen by both management and staff. Cases are dealt with by a mixture of face-to-face meetings and phone conversations. Information is shredded once the cases are completed except at Shop, where documents are retained in case they are required at an employment tribunal. Cases can be anonymous in all institutions except Shop. In all four organisations the ombuds reports to the CEO/President. This is vital if the function is to be taken seriously by senior management. Evaluation of the ombuds process and outcome varies. Finser uses feedback postcards; Shop has a hard-hitting letters page in its house magazines; Liftco sets performance benchmarks for ombuds and measures performance against these agreed objectives; at IFI there is extensive feedback from employees in general and the President, Staff Council and HR Director in particular. Methods of operation are set out in Table 5, again self-explanatory. Ombuds in each organisation will listen and, if requested, act as a conduit between parties. Fact-finding and conciliation are functions in three organisations. Arbitration is not considered appropriate, except at Shop where the Partners’ Counsellor is the individual of last resort for appeals against demotion and termination. None of the ombuds takes part in exit interviews.

4.3. Interaction with Collective Institutions and Formal Procedures The interaction of the ombuds with collective institutions, formal grievance and appeal procedures and other independent voices is summarised in Table 6. Liftco recognises unions in the U.K. and EU and the ombuds does not trespass on issues

The Organisational Ombuds

115

Table 5. Ombuds’ Methods of Operation. Characteristics

Finser

Shop

Liftco

IFI

Listening Conduit

Yes Yes, shuttle diplomacy

Yes HQ yes, on collective systems and conditions

Yes Yes

Fact finding

No

Conciliation Mediation Arbitration Exit interviews Links with, e.g. equal opportunities

Yes No No No Off the record

Yes (HQ individual) No No Yes No Informal

Yes 25% Yes 65%. Acts as sounding board & refers to appropriate person in org. Yes 10% Yes 10%

Yes Sometimes No No e.g. liaison with opportunity now

No No Informal

Yes

arising from the procedures and substance of collective agreements. Shop, IFI and Finser have no, or almost no, union presence in the U.K. Works Councils are more pervasive than unions. Liftco and Finser have such councils in the EU and at Finser the council has to validate the appointment of the ombuds. Shop has Committees for Communication at store level, chaired by Registrars. These Committees are similar to works councils and the Registrar is equivalent to a lower-level ombuds. At IFI there is a Staff Council but it has little formal status. The ombuds regularly discusses matters of mutual concern with the Council. The ombuds function interacts with formal grievance and appeals machinery in each organisation. At IFI the formal system has three stages: line manager, Vice President Personnel, internal tribunal chaired by an independent expert. An employee can run the informal route – the ombuds – in parallel with stages one and two. But if the case goes to stage three, the Tribunal, the ombuds washes his/her hands of the case. By contrast, at Shop the ombuds initially arbitrates on demotions and terminations, but then acts as “prosecutor” if the employee subsequently takes the case to an Employment Tribunal. At Finser and Liftco the ombuds provides informal help to any employee taking a case to a tribunal. Each of the four organisations have a compliance officer and an auditor, but they have quite distinct functions from the ombuds. In most of what follows we analyse the ombuds function across the four organisations. But we pay special attention to IFI, where we have a close relationship with the ombuds. This permits us to go into detail concerning terms of reference, incidence of cases and issues, fit with grievance and appeals

116

Table 6. Processes and Interactions. Characteristic

Finser

Interaction with collective institutions Unions No union in UK. Where union in EU, avoid substance of collective agreements but are involved in procedures

Works council

Liftco

IFI

No union in shops Residual membership in stores taken over

No cases or enquiries handled involving c.b. agreement issues 100% membership in U.K. and most western Europe. No appreciable change last 5 years.

No union

Registrars chair C for C PC distils minutes and reports to Chairman

Yes, in many Euro companies EWC H & S in most Euro cos.

Staff council deals with collective matters

Officially no active role in any formal process Informally might help coach for formal disciplinary & grievance procedure

Once case goes to formal appeal Ombuds withdraws

No. Only Ombud and Dialog Admin truly independent

Compliance Officer Audit

Interaction with formal grievance and appeal procedures Process On request, help ee prepare PA is formal appeal (not coaching) no ajudicator. Becomes involvement if ET “prosecutor” if case goes to ET Other independent voice

Compliance Officer, audit but ombuds keeps distance

As Finser

SUE FERNIE AND DAVID METCALF

Validates ombuds appointment

Shop

The Organisational Ombuds

117

procedure, and documentation. The latter is important because the ombuds has been proactive, helping to encourage new policies on areas like job families, performance appraisals, review of benefits, bullying, gender, career progression of support staff, and the medium term HR strategy.

5. METHODS OF INVESTIGATION A variety of methods of investigation were used to analyse the ombuds process and related outcomes. They involve comparing across the four organisations and providing considerable detail concerning the ombuds function in one of the organisations. Interviews with the ombuds working in Finser, Liftco, Shop and IFI were a central component of our work. An initial day-long scene-setting was followed up by a number of subsequent interviews. We also participated in the bi-annual formal meetings of the European Ombudsman Association and attended training courses in conflict resolution specifically designed for ombuds by the USA Ombudsman Association. Next, we were able to draw on surveys conducted by external consultants in three of the organisations which document satisfaction with employment relations in general and the grievance/ombuds procedures in particular. An obvious obstacle to empirical research in this area is the tradition of confidentiality, and the consequent lack of access to data. The sensitivity of interpersonal conflict at the workplace makes it almost impossible to observe, in the majority of cases, actual meetings between the ombuds and enquirer. But we did gain access to much documentary evidence, mostly of a confidential nature, and were able to construct scenarios for hypothetical resolution by the four ombuds. We formed a close relationship with the ombuds at IFI. This permitted a detailed analysis of cases, a clear understanding of the fit between the ombuds function and other strands of HR, and a knowledge of the way in which the ombuds reports and discussion of the themes behind the individual cases contribute to changes in personnel policy on matters like bullying, conduct of performance appraisals, staff development and work-life balance. This research design is far from perfect. It permits a description of the ombud function in the only four organisations in Britain with an organisational ombud. The description also yields insights into such matters as procedural justice, the costs and benefits of different types of dispute resolution, and how voice mechanisms operate and develop. We recognise that, ideally, it would have been good to compare “twin” organisations – one with and without an ombuds – or to study one organisation before and after establishing the ombuds function. It has not been possible to do such research – such “twins” do not exist and the research reported

118

SUE FERNIE AND DAVID METCALF

here has taken three years of detailed investigation inside our four co-operating organisations. Similarly, we had no contact with these organisations prior to them establishing the function. We return to this limitation in our research design when providing cautious suggestions for future research at the end of the paper. Further, the organisations do vary by union status, human resource management systems, industry and public/private ownership, so some tentative associations between such variables and the ombud process/outcomes do emerge.

6. EVIDENCE AND DISCUSSION In this section we discuss evidence from our four organisations concerning: the incidence and composition of cases; whether the ombuds is reactive or proactive; whether the ombuds should be viewed as an end in itself or as part of a wider conflict resolution mechanism; the extent to which the ombuds helps achieve justice in the workplace for employees, and various outcomes for the organization. We also make comparisons with costs associated with other dispute resolution mechanisms.

6.1. Incidence and Composition of Ombuds’ Cases The incidence and composition of cases flow in large part from the nature of employee-management relations in our four organisations, set out in Fig. 1. Such industrial relations can be summarised along two continua: first, how important is the collective element? Second, how well-developed is human resource management? Liftco recognises a trade union and negotiates collective agreements on pay and conditions. Human resource management is also progressive with an emphasis on staff development, suggestion schemes and continuous improvement in work practices. Shop has a very limited union presence, primarily in its warehousing and distribution functions, but its employment relations are best described as collective because it operates de facto works councils in each establishment. Personnel management does not emphasise formal teamworking or integrating HR strategy into overall business strategy. Finser does not recognise a trade union, nor does it have a works council (in the U.K.). Rather, it emphasises team working – although this initiative is currently being reviewed – TQM, and specific training. Finally, IFI has no trade union and a staff council with only a modest presence. Until very recently personnel management was primarily focussed on recruitment and separations and made little contribution to the development of employees and their

The Organisational Ombuds

119

Fig. 1. Extent of High Commitment Work Practices and Collective Industrial Relations in Our Case Study Organisations.

everyday concerns. This might be because IFI is “above the law” and employees cannot sue the organisation or take their case to an employment tribunal. Table 7 shows that the incidence of cases at IFI is much higher (11%) than at the other organisations. This is exactly in line with expectations. There is no trade union and the ombuds also deals with many matters which, in other firms, would be handled by personnel. At Liftco, which has both collective industrial relations and best practice HR, the incidence of full cases reaching the ombuds is only 1.6% in the U.K. (and 1% worldwide). At Liftco the lower-level ombuds also responds to Dialog cases which are mainly queries about company policy, but even including these cases does not raise the incidence to that in IFI. Finser, with its struggling team structure, has a U.K. incidence of 3.1%. There is no consistent pattern in the incidence of cases by gender. Women are disproportionately represented at Liftco and IFI but underrepresented at Finser. The composition of cases is also set out in Table 7. Recall that the ombud does not decide directly on pay and promotions – distributive justice – but influences employees’ perceptions about the fairness of the process behind such decisions. At IFI the composition of cases is heterogeneous and includes both those flowing from grievances and from employee voice. This mixture reflects the lack of welldeveloped collective industrial relations or human resource management. Onethird of cases concern line or other managers’ relationships with their subordinates

120

Table 7. Incidence and Composition of Cases. Characteristics

Shop

Liftco

IFI

U.K. 246/8,000 = 3.1% EMEA 323/18,000 = 1.8% Worldwide 3344/70,000 = 4.8%

(Details for HQ only) 140 in last 18 months, of which 68 appeals against dismissal, demotion Registrars: not known

Dialog and Ombuds U.K. 200 = 8% Rest of world 1200 = 5% Ombuds alone U.K. 40, 1.6% WW: 240, 1%

U.K. 88 = 11% (plus 2 visits to ROs)

Days worked

EMEA 2 f/t, say 440

HQ, 2 × f/t, say 440

FT say 220 50 Dialog administrators all FT ees who do this

Approx 35

Cases per day

0.73

HQ 0.32

1.3

2.5

% Female

Cases 58% Population 67%

Cases 60% Population 60%

Cases 35% Population 15%

Cases 62% Population 54%

Compensation and perf related 40% Mgt. practices/discrim 15% General, sexual harassment, bullying, 15% Training, quality, safety 20% Working conditions, environment, other 10%

Grading + pay + benefits 24% or 12% Perf. Appraisals 11% or 9% Job content/staff devt 9% or 15% Line/other manager, teams 26% or 33% Work-life balance 12% Recruitment and separations 11% Miscellaneous 8%

Number of cases Say average of 3 years 2000–2002

Issues Issues %

(HQ only) 43% Leadership effectiveness, e.g. staff devt, integrity, respectfulness (50%) 28% Job, e.g working hours, practices (19%) 18% Meritocracy, e.g. pay benefits (or 15%) 11% Co. assets, e.g. safety, criminal activity (16%)

Pay, promotion, perf. appraisals 20% Terminations and demotions 61% Line manager, e.g. family-friendly, bullying 12% Sexual harassment 2% Staff devt 5%

SUE FERNIE AND DAVID METCALF

Finser

The Organisational Ombuds

121

plus malfunctioning teams. A further fifth are about grade, pay, benefits and performance appraisals. The remainder cover job content, work/life balance and recruitment and separations. The ombuds at Liftco does not deal with issues of rights such as employees’ grievances concerning pay and benefits, which belong in the union’s remit. But, taking both Dialog and full ombuds cases, Table 7 shows that the ombuds does deal with many queries concerning the compensation, training and safety policy of the company. Via the Dialog programme, the ombuds office provides backup information about company policy which the union may not easily be able to communicate to its members because of their diverse and constantly changing work locations. The other cases concern matters of interest and power including bullying and sexual harassment, management practices including discrimination, and working conditions/environment. It is clear that the vast bulk of the case-load at Liftco concerns interests rather than rights. At Finser under a fifth of cases are about the job – staffing levels, selection, conditions, discipline and terminations. This low fraction concerning work allocation reflects the team ethos, 360 degree appraisal and intense recruitment methods. Rather a high proportion of cases – as compared with the other organisations – are about the assets of the company including threatening or criminal activity and breaches of the code of conduct. This whilstleblowing element accounts for one-case-in-six. What the teams cannot fully deal with is poor leadership, which accounts for around half the cases, and this inability to countenance inter-personal or intra-team issues is one reason why the team experiment at Finser has been modified. Bullying, discrimination, thieving, demotions and terminations comprise the bulk of cases at Shop. There is no union in most Shop workplaces and personnel management is somewhat paternalistic. The composition of cases at Shop reflects the reactive nature of the ombuds function there. 140 cases came to the attention of the ombuds during 2002, of which 68 led to employment tribunals, during which time the ombuds acts for management.

6.2. Is the Ombuds Reactive or Proactive? Employee voice can be either or both proactive and/or reactive and various dimensions of proactive and reactive voice and representation were set out in the second section. The implementation of the EU Directive on information and consultation can, for example, be interpreted as a push to a more proactive voice for employees (Gospel & Willman, 2003). Similarly, if a union shifts from an adversarial relationship with management to a partnership approach, union voice

122

SUE FERNIE AND DAVID METCALF

adds a proactive element to its traditional reactive function. What kind of voice does an institutional ombuds provide? We investigate this issue by, first, comparing the role of the ombud at Shop and IFI, and second, by contrasting the ombud with trade union representation. At Shop the ombuds is mainly reactive. At local level the registrar (say in a large store) might be asked to deal with, typically, instances of bullying, sexual harassment or a demotion. S/he resolves the case as best as s/he can and normally that is the end of the matter. In a minority of instances – mainly concerning proposed terminations – the case is appealed to the Partners’ Counsel at HQ. (If the worker is still not content and s/he takes the case to an Employment Tribunal, the Partners’ Counsel switches from being judge at the internal appeal to prosecutor at the ET. This is an unusual juxtaposition of roles not replicated by other ombuds). Although the Partners’ Counsel visits the registrars periodically and then reports to the CEO on trends in the cases there is no specific mechanism for translating this reactive detail on individual cases into a company-wide proactive personnel policy. By contrast, consider the operation of the ombud function at IFI. The terms of reference require a biannual report to the President, Vice President for Human Resources and Staff Council. The written reports are discussed in lengthy meetings, including Board level. The full Annual Report deals not only with the cases but with wider themes, for example any lack of alignment between HR processes and the strategy of the IFI or malfunctioning teams. These ombuds reports have had a profound influence on various IFI policies. For example, the IFI has initiated or altered policies in the following areas in the last three years: use of job families in pay setting, paying more attention via “job families” to the external labour market; management of the annual performance appraisal; promotion process in Banking; review(s) of benefits on housing, education, pensions and relocation; availability of benefits for same-sex partners; bullying; sexual harassment; worklife balance including flexible hours and homeworking; career progression of secretaries; use of the internet; (re-)organisation of HR department; establishment of staff development committee emphasising inter alia the importance of HR skills for line managers. Of course, the ombuds reports are not the only reason for such initiatives. But in each case they have either been the catalyst or a central element in the HR policy process. Thus at IFI the ombuds both reacts to individual cases and plays a proactive role in the development of institution-wide personnel policy. To the extent that the ombud is a conduit by which individual cases flow through into a more institution-wide personnel policy, a comparison with unions may be illuminating. At Liftco, for example, the union is encompassing but primarily reactive. It helps employees who allege bullying by the line manager or who face

The Organisational Ombuds

123

dismissal. This reactive voice is essential to the vitality of the union. We were told that if the union went too far along the partnership route its members would see it as a management poodle and it would lose their trust. But the ombuds at Liftco, who also handles cases involving discipline, bullying and dismissal, is able to integrate with management more easily and, again, be the conduit to translate case information into company policy. In summary, the ombud voice cannot be uniquely categorised as reactive or proactive – it depends on the type of case and the niche the ombud has inside the organisation. But it is plausible that the ombud provides added value when s/he contributes to the process of developing human resource policy via detailed knowledge of staff concerns.

6.3. Is the Ombuds the First Step in a Grievance Procedure, or an End in Itself? In the second section, we proposed that the success of the ombuds as an end in itself would be dependent on two things: first, the management’s attitude to and support for the office of the ombuds, and, second, the climate of employee/management relations at the workplace. Both Finser and Liftco adopted the ombuds function for the same reason, i.e. as a response to the U.S. Federal sentencing guidelines, and so it may be constructive to compare the status of the ombuds role in those two organisations. Finser does not recognise a trade union, whereas Liftco does. Responses from employee surveys show that 65% of Liftco employees as compared to only 37% of those from Finser believe that staff morale is good or better. The ombuds office at Liftco has recently benefited from an injection of 50 new Dialog officers and the promotion of the ex-U.K. based ombuds to worldwide. At Finser, on the other hand, the ombuds function has recently been downsized to one full-timer, who has had all office facilities withdrawn and now works from home. By the ombuds’ own admission, employees fear retaliation if they are found to be using the ombuds’ services, as the degree of trust between management, particularly line management, and workers is poor. Just as with any other HR inititive, it is pointless to put in place the ombuds function with neither trust nor support from top management. As a result, Finser has experienced several high profile ET cases in the last three years which were impossible to deal with within the company’s existing conflict resolution structure. On the other hand, Liftco’s ombuds has as one of his objectives a duty to reduce the use of ETs and, indeed, over a three year period, the number of ET cases has fallen from 15 to 4. As a response to Article 8 of the European human rights legislation, IFI set up a new grievance and appeals procedure but, in the 18 months of its operation,

124

SUE FERNIE AND DAVID METCALF

only four cases have been brought and no cases have found their way past the first stage, while the ombuds’ cases remain constant at about 10% of employees per year. There is therefore no evidence that the ombuds’ cases are being referred instead to the formal procedure. Therefore the use of the ombud can normally be thought of as a discrete process rather than the first step en route to more formal grievances. A well-functioning ombud will nip problems in the bud. But if the ombud is embedded in a wider low-trust situation many cases will bypass the ombud and go directly outside the organisation to an ET.

6.4. Outcomes for Individuals – Procedural and Interactional Justice Does the existence of an ombuds impact upon the organisation’s procedures and practices such that employees feel a greater sense of justice when their concerns are aired? To answer this question we analysed responses from employee surveys of two of our case study organisations – Liftco and IFI. At Liftco the 2002 employee survey indicated that four-fifths of employees knew that the company had a Code of Ethics (dealing with personal and business issues) and two-thirds were aware that the ombud/Dialog program is available to raise business/personal-related issues in confidence. The existence of the code and the ombud are sufficient to assuage any apprehension about raising issues for the majority of workers: three-fifths of respondents stated that they could report a violation of the code without fear of retribution; and three-quarters said that they would feel comfortable using the ombud/Dialog program to raise a business-related matter, as opposed to the one-quarter who said that they would feel confident raising such issues with their union representative. These findings hint that the institutions of the Code and the ombud contribute to good perceptions of informal due process – and possibly by implication procedural justice – but a full before/after study would be required to confirm this. IFI conducted detailed staff surveys in March 2000 and March 2003. For the 2000 survey, 1122 questionnaires were distributed and the response rate was 70%. In 2003 the response rate was slightly lower at 65%. We do not have information on which survey respondents visited, or did not visit, the ombuds prior to the first survey. But for the 2003 survey respondents can be split between those availing themselves of the ombuds in the prior three years and those making no such use of the ombuds. In the three years March 2000 to March 2003, the ombuds had 255 clients. Of these 255 employees, 61 no longer worked at IFI by March 2003, so the potential sample of ombuds-users was 194. Their response rate to the survey was 72%, giving an actual sample of ombuds users of 140.

The Organisational Ombuds

125

Table 8. Outcomes for Individuals – IFI (%). Strongly Agree

Agree to Some Extent

Disagree to Some Extent

Strongly Disagree

Promotion decisions always follow formal procedures and the reasons for exceptions are clearly made 2000 8 33 34 15 2003 10 40 30 20 2000–2003 ombuds client Yes/No 15/8 45/38 23/33 17/21 With regard to your annual appraisal: your manager conducts the process well 2000 24 38 20 2003 26 41 17 2000–2003 ombuds client Yes/No 28/25 43/40 16/18

18 16 13/17

The IFI is good at communicating with its staff 2000 10 2003 12 2000–2003 ombuds client Yes/No 12/12

19 17 19/16

40 41 40/42

31 30 29/30

Notes: (i) Sample sizes: 2000 survey n = 789. 2003 survey n = 813; of these 813 employees who had answered the 2003 survey, 140 were ombud users in the three years March 2000 – March 2003 and 673 did not use the ombud. (ii) Row 2000 refers to survey results from 2000 (n = 789); row 2003 refers to survey results for 2003 (n = 813); row 2000–2003 reers to 2003 staff survey results. Left-hand figure is ombuds clients (n = 140), right-hand figure is non-ombuds clients (n = 673), e.g. in 2003, 15% of ombuds clients strongly agreed that promotion decisions always follow formal procedures, compared to only 8% of non-clients. (iii) Significance tests. For each of the three questions in the table we aggregated “strongly agree” plus “agree to some extent”. We then asked (a) does the percentage agreeing in 2003 differ significantly from that in 2000? The difference is significant at 1% for questions 1 and 2 but not significant for question 3; and (b) does the percentage agreeing differ as between those who were ombuds clients in the three-year period 2000–2003 and those who were not ombud clients? Again, the difference is significant at 1% for questions 1 and 2 but non-significant for question 3.

Evidence on two major areas of procedural justice – promotion procedures and the appraisal process – is set out in Table 8, which also provides information on communications. The three rows on each item refer to: (i) the cross-section sample 2000; (ii) the cross-section sample 2003; and (iii) responses for 2003 from the panel of ombuds-users (2000–2003) compared with the panel of non-users (2000–2003). The first point to note is the considerable overall improvement in each distribution of responses over three years. In 2000 only 41% of respondents strongly agreed or agreed to some extent that promotion decisions always followed formal procedures and the reasons for exceptions were clearly made. By 2003

126

SUE FERNIE AND DAVID METCALF

the corresponding percent agreeing had risen to 50%. Similarly, the proportion agreeing that the manager conducts the appraisal process well rose from 62% in 2000 to 67% by 2003. Both these differences are significant at 1%. And the proportion agreeing that IFI is good at communicating with its staff rose from 50% in 2000 to 53% in 2003 (difference not statistically significant). It is plausible that these improvements in procedural justice and communications stem partly from the proactive role of the ombuds: the promotion and appraisal process, and the importance of improving face-to-face communication, were major themes in the various reports the ombuds made to the Board and the President during this period, which contributed much more management time and resources being devoted to promotions, appraisals and communications. Second, on promotions and appraisals there is a clear additional “ombud effect” because the panel who visited the ombud at any time during the 3 years to March 2003 score more favourably than the panel who were not clients of the ombud. Thus, of the sample of ombud clients, 60% agreed that promotion procedures were followed and 71% agreed that performance appraisals were done properly. These percentages are substantially higher (and statistically significant differences at 1%) than non-ombud clients (42% re promotions, 65% re appraisals), suggesting that the ombud plays a pivotal role in enhancing perceptions of procedural justice within IFI. There is much less variation in responses of the two groups concerning communications (the difference is responses is not statistically significant) – the wording of this question, however, referring to the IFI not the line manager, such that the interactional element of justice is missing. Further evidence that the ombud at IFI strengthens perceptions of procedural justice comes from an internal evaluation of the ombud function for 2001. Some 10% of the workforce (88 employees) were clients. Around a third of these employees remained unhappy with the outcome of the case after the involvement of the ombud. This is hardly surprising as the issues include, for example, nonpromotions and terminations. More importantly, 80 out of the 88 employees were satisfied with the process (the remaining 8 felt the ombud should have pursued their case more vigorously). Thus, among IFI users, procedural justice was significantly deepened by the existence and use of the ombud institution both in terms of wider IFI issues such as promotions and appraisals and the more specific matter of their own case.

6.5. Outcomes for the Organisations What do organisations and their employees get from an investment in the ombuds function? We investigated this via discussions with the relevant ombuds and staff

The Organisational Ombuds

127

surveys. At IFI we also discussed this matter with HR staff and senior management including the President. Our findings are set out in Table 9. In each of the organisations the ombuds acts as a quasi consultant on personnel matters, but to top management rather than the HR department. This was stated explicitly: “providing evidence on trends [in personnel issues] and anticipating problems” at Finser, and “educating management” in the form of a monthly meeting with the CEO at Liftco. At IFI this role had stronger impact because the verbal and written reports of the ombuds were the catalyst for change in items like bullying, same-sex benefits equal opportunities and family-friendly policy. At Shop the HQ ombuds attends meetings of the consultative committees at workplace level which provides an early warning of local and organisational concern. At both Liftco and IFI the ombuds does not trespass on collective issues dealt with by the union or staff council respectively. But the interaction between the ombuds with his or her individual worker clients and the collective institution is believed to strengthen employee relations in each institution. At Liftco union membership is close to 100%. It is plausible that such an encompassing union dealing with collective issues coupled with an ombuds for individual cases is a fruitful structure for voice and conflict resolution. Similarly, at IFI the ombuds reported the concerns of some resident office colleagues on pension, education and housing benefits and these were subsequently pursued on a bank-wide basis by the staff council. Just as there is a debate about the compatability of welldeveloped HRM systems and trade union recognition, it is also possible that the management of a firm might decide to opt for an ombud – providing voice and alternative conflict resolution – to try to undermine or forestall union recognition. This has not happened in our four organisations where the ombuds function coexists happily beside the staff council/works council/trade union, in large part because unions and ombuds fulfil different functions. But the nature of the tasks undertaken by workplace representatives (shop stewards) is changing: a recent survey of 877 workplace representatives found that individual conflict resolution was the most common activity they carried out in the previous year, comprising 84% of their allotted time (TUC, 2002). It is plausible that some managements will, in future, opt for an ombuds to “crowd out” such activities of the workplace representative. All four organisations state that the existence of the ombuds lessens the use of both formal internal grievance and appeal procedures and external employment tribunals (discussed immediately below). In addition, both Finser and Liftco state that the ombuds is crucial to them in connection with the 1991 U.S. Federal Sentencing Guidelines (see Howard & Furtado, 1999). They get credit for having a programme that allows employees to report suspected criminal activity “without fear of retribution.” So, if their firm appeared before the U.S. courts and was

128

Table 9. Outcomes for Organisations. Characteristic Softer outcomes Organisational

Individual

Daily absenteeism rate %

Shop

Liftco

IFI

“Safety net” awareness/HR problem solving move issues up agenda, e.g. taping phone calls, use of IT screens Quasi consultant to CEO Trends, anticipates problems Stop “another Barings Bank”

Registrars: See trends by store Formal link to C for C PC: Distils trends Anticipates problems Essentially monitors Personnel

To create a more caring co. with higher standards and values To reduce turnover and retain high performers Alternative to legal action, reduce Ets To educate management

Conduit to CEO, Board, Staff Council about HR issues Catalyst for change e.g. equal opps, bullying, benefits, family-friendly

ees definitely aware of ombuds role, e.g. HR, fraud but fear of retribution/retaliation by line mgr inhibits full trust relationship

Definitely aware of PC and Registrar No desire for union

To provide good, fair working environment in which ees can speak up, contribute, develop potential to full benefit of both ind. and co.

Provide neutral, independent contact Listening and “insurance” Talk through issues

15 (9) 5.1 (4.5)

26 (19) 4.2 (3.4)

12 (13) 5 (3)

12 (7) 5 (4.7)

Note: Turnover rate refers to number of voluntary resignations per 100 employees. For both the turnover rate and daily absence rate the first number is for the institution and the number in brackets is the industry average, from Cully et al. (1999), Table 6.6.

SUE FERNIE AND DAVID METCALF

Hard outcomes Turnover rate %

Finser

The Organisational Ombuds

129

found guilty of some corporate criminal activity, their punishment would be lower because the existence of the ombuds protects such whistleblowers. Further, at Finser the ombuds is used by any worker who suspects a co-worker of defrauding the company. The matter can then be investigated discreetly, with the identity of the “accuser” remaining completely confidential. It may well be that U.K. companies adopting an ombuds will be in a similar position regarding the Public Interest Disclosure Act 1998. It is plausible – but we do not have the information to confirm our speculation – that the organisation also gains because the workforce is more satisfied as a result of the benefits to individual workers from the ombuds institution. For example, first, it sends a signal that the organisation “provides a good, fair working environment in which employees can speak up, contribute, and develop their potential to the full to the benefit of both the individual and the company” (Liftco). Next, not all individuals desire union representation but most do wish to have a voice (Charlwood, 2003) and the ombuds is a fruitful institution for individual voice, independent of any collective voice. Third, many individuals view the ombuds as an insurance mechanism. If a worker is having difficulties with a line manager – say over job content and work allocation – they may wish to put their concerns on the record, even if the employee does not initially wish the matter to be pursued. Often the ombuds is used for such a task because the employee has concerns about the (lack of) confidentiality in the HR department. Fourth, the ombuds can get things done on behalf of the employee. At IFI in 2002, for example, the ombuds had successes in: stopping bullying; improving the functioning of three separate teams; ensuring performance targets for colleagues who had received a written warning were stretching but achievable; making sure that the annual performance appraisal was conducted according to procedures laid down; confirming and helping implement eligibility to various benefits and family-friendly initiatives including job shares and homeworking. It is normally expected that a voice mechanism lowers labour turnover and absenteeism, to the extent that the latter reflects dissatisfaction at work (Bemmels, 1997). Contrary to expectations, this appears not to be so for our four organisations. In each case the absence rate is above the industry average and in three of the four organisations the voluntary turnover rate is also above the corresponding industry average. Further, at IFI it is generally held that the pay of bankers is below that in the local labour market while the relative pay of other professionals (e.g. in law, HR, IT) and support staff is above their counterparts locally. On this basis we might expect bankers, when they have a problem, to exit to a higher paying outside opportunity whereas other staff, with less good outside options, would stay and use the voice mechanism provided by the ombuds. We investigated this for the year 2000 and found the reverse. The incidence of ombuds cases among bankers was

130

SUE FERNIE AND DAVID METCALF

high, and the leaving rate low relative to the IFI average. By contrast, the incidence of ombuds cases was relatively low for non-bankers and their exit rate relatively high. This little cameo refutes the normal expectations concerning exit and voice.

6.6. Comparison of Cost of Ombuds with Cost of Employment Tribunal Once disputes or grievances between an individual employee and the employer go beyond the organisation, Employment Tribunals (ETs) comprise the major formal system of dispute resolution. ETs deal with numerous strands of the employment relationship, but the bulk of their cases arise from allegations of unfair dismissal or discrimination by race, sex and (dis)ability (see Shackleton (2002) for a thorough discussion of ETs). A successful ombuds would reduce the number of cases going to formal grievance and appeal systems including ETs. It is therefore useful and informative to compare the cost of the ombuds with the cost of operating the ET system. Shackleton notes that there are a number of different costs in the operation of ETs: direct resource costs, compensation, compliance costs and indirect costs. Direct resource costs for 2000/2001 are set out in Table 10. It will be seen that these alone total around £1 billion. Given approximately 100,000 cases per year (of which around a quarter get as far as the tribunal itself) the resource costs per case equals £10,000. This figure is higher than the amount quoted by DTI (2001b) which only covers management time and legal costs (i.e. only item 3 in Table 10). Such management time and legal costs are detailed in Table 11. Taking Table 10. Estimates of Resource Cost of Employment Tribunal System in 2000/2001. Item (1) Employment Tribunal Service (2) ACAS conciliation of individual disputes (3) Management time and legal costs (4) Extra recruitment costs (5) Cost to applicants of bringing cases (6) Other institutions’ costs (7) Employers compliance costs Total

Cost (£m) 52 30 292 341 10 30 250 1005

Notes: (1) Employment Tribunals Service Annual Reports and Accounts; (2) ACAS Annual Report, Shackleton’s estimate of ET proportion; (3 and 4) CBI estimates; (5) Shackleton’s estimate based on material in Tremlett and Banerji (1994); and (6 and 7) Shackleton’s estimate. Source: Shackleton (2002).

The Organisational Ombuds

131

Table 11. Average Legal Costs and Management Time Spent by Respondents on Tribunal Cases. All Cases

Median senior management time dealing with case (hours)

Withdrawn

Settled Privately or via ACAS

Tribunal Hearing

16

10

16

27

Mean legal costs, including where no legal costs incurred (£)

1,000

1,000

700

1,450

Mean legal costs, excluding cases where no legal costs incurred (£)

2,250

3,000

1,700

3,000

Notes: (i) 42% of employers incur legal costs when hearing a tribunal case (ii) legal costs are in 2000/1 prices rounded to nearest £50. Source: DTI (2001b) Table 6.

all cases, the legal cost averages £1,000 but if the sample is limited to those cases where legal costs are actually incurred the corresponding figure is £2,250. These legal costs, and management time, rise substantially for the quarter of cases which reach the tribunal hearing stage. Indeed, both Liftco and Finser state that difficult cases coming before a tribunal can cost upwards of £100,000 in legal fees and management time. Compensation payments are not included in the above amount because compensation is not a direct resource cost; rather they are similar to a transfer payment like social security or a gift. Shackleton calculates that in 2000/2001 compensation employers paid as a result of tribunal systems was £25m plus a further £20m for pre-tribunal settlements. Compliance costs occur because of the measures employers take to make sure they do not breach employment law. The DTI (2001a) are keen on formal dispute resolution procedures: “Employers with written procedures are more likely to reach a settlement or have an application against them withdrawn than employers without written procedures. At hearing, employers with procedures are far more likely to defend a case successfully than those who do not” (quoted in Shackleton, 2002). Costs include management time and some proportion of the costs of the personnel department. Under the Employment Act 2002 all firms are now encouraged to establish such formal procedures: tribunals will award higher compensation against firms without procedures. Further, ET cases should normally only be brought by the employee once such internal procedures have been exhausted. Although the existence of the ombuds does not constitute formal procedures – s/he

132

SUE FERNIE AND DAVID METCALF

is, rather, an informal complement – it is plausible that ETs will look more favourably on organisations with such ADR in place when making judgements on cases. Finally, indirect costs are incurred as a result of tribunal cases, including staff stress, damage to the reputation of the employer and damaged workplace relations. The incidence of such concerns is especially noticeable in sex/race discrimination cases (DTI, 2001a), precisely the area where the existence of an ombuds in a firm would be helpful in nipping the problem in the bud prior to the case reaching the formal stage. Individuals also face short and long run costs from making a tribunal application. The DTI (2001b) states that “many find their employment prospects worsen, around half report lower status or lower paid employment subsequent to an application, and almost a quarter were unemployed subsequent to the case” (p. 10). Now consider the costs of having an ombuds. In the four organisations mean annual ombuds pay is below £50,000. Assume office space, travel, etc. totalling £100,000 (per ombuds, per year). This gives a total annual cost of £150,000 or £600 per working day. The number of cases dealt with per day (see Table 5) ranges from 0.32 at Shop to 2.5 at IFI implying a cost per case between £240 and £1,800. Even comparing these figures with the amounts in Table 10 – solely legal fees and management time associated with ET cases, rather than the higher figure from Table 9 – suggests that an organisational ombuds is likely to be very cost effective. The ombuds may also have other indirect benefits such as enhancing trust (e.g. where employees do not believe that the Personnel Department will be scrupulous over confidentiality), improving the functioning of teams and acting as a de facto sounding board/consultant on various HR initiatives.

7. SUMMARY AND CONCLUSIONS An organisational ombud is concerned with both conflict resolution and provision of employee voice, operating in a confidential, neutral and informal process. The evidence in this paper suggests that the ombud provides both reactive and proactive voice and is an end in itself – enhancing procedural and interactional justice, rather than being simply the first step in a formal grievance procedure. Crucially, the ombud is also able to be the conduit to turn reactive voice into proactive voice, to the benefit of both employees and the organisation. The evidence also suggests that an ombud can favourably influence outcomes for workers and the firm. For example, at IFI those visiting the ombud in the threeyear period 2000–2003 had a more favourable view of both the promotion and performance appraisal process than those that did not. And, for the organisation, it

The Organisational Ombuds

133

was shown that investing in an ombud is normally much more cost-effective than letting cases reach a formal employment tribunal. Probably there is no “best system” of voice and conflict resolution which would fit all organisations. But on the limited evidence of this paper we conjecture that the elements of a good system might comprise some combination of high commitment work practices, an encompassing union and an organisational ombuds. Some such permutation provides for both reactive and proactive voice. For example, it is important for worker representatives to always provide some reactive voice – to retain the trust of employees – but an ombuds acts as a conduit providing proactive voice and conflict resolution for the good of the whole organisation. Further, while it is always possible for employees to use formal grievance procedures an informal voice/conflict resolution system – such as an ombuds – comprising an end-in-itself (rather than a first step towards the formal route) may well have advantages for both the worker and the organisation. In particular, third, it is possible that such an informal system is associated with employees’ perceptions of greater procedural, distributive and interactional justice in the workplace and potentially superior organisational outcomes. Future research (see below) will determine whether such speculation is confirmed. It is also likely that in future such informal procedures will be viewed favourably should an organisation be taken to an Employment Tribunal. Given these potential advantages associated with the existence of an informal voice and conflict resolution mechanism it is plausible that the institution of the organisational ombuds will be adopted more widely in future. What implications does this have for trade unions? Modern systems of conflict resolution and voice imply two types of interaction with employers – a reactive function dealing with grievances and more proactive function emphasising co-operation and partnership. But if a trade union emphasises its partnership role excessively, employees may start to see the union as a sweetheart. An organisational ombuds helps redress the balance. S/he permits the union to retain its more traditional functions and/or can take cases which may fall to the bottom of the hard-pressed shop steward’s in tray. The composition of tasks undertaken by workplace representatives has altered over the last two decades such that they now spend more time on individual grievances. So an organisational ombuds lightens the load of such representatives but might be seen as usurping their functions. It might also be that the ombuds is better able than a union to get an individual issue (say the availability of benefits for same-sex partners) dealt with at the highest level of the organisation. Finally, it is worth re-emphasising the limitations of this study, some of which may be resolved by further research. The discussion in this paper simply describes and compares the ombud function in the four British organisations where it exists. Ideally, future research will be able to compare matched pairs of organisations with

134

SUE FERNIE AND DAVID METCALF

and without an ombud or one organisation before and after the ombud is established. This will permit a thorough analysis of the contribution of the ombud to outcomes for individual employees – both objective like absenteeism and subjective like job satisfaction – and for the organisation, including savings compared with cases ending up at an Employment Tribunal and a possible impact on the performance of the organisation.

ACKNOWLEDGMENTS We are indebted to the Ombuds at the four organisations studied for their kind cooperation, openness, attention to detail and intellectual encouragement, and to the editors/referee of this journal for extensive and thorough comments on an earlier version.

REFERENCES American Bar Association (2000, July). Standards for the establishment and operation of Omubdsman offices. Section of Administrative Law and Regulatory Practice and Section of Dispute Resolution. Report to the House of Delegates. Bemmels, B. (1997). Exit, voice and loyalty in employment relationships. In: D. Lewin, J. B. Mitchell & M. A. Zaidi (Eds), The Human Resource Management Handbook (Part II). Greenwich, CT: JAI Press. Bies, R., & Moag, S. (1986). Interactional justice: Communication criteria of fairness. In: R. Lewicki, B. Sheppard & B. Bazerman (Eds), Research on Negotiation in Organisations (Vol. 1, pp. 43–55). Greenwich, CT: JAI Press. Charlwood, A. (2003). Light at the end of the tunnel? Unionisation preferences of non-union workers: A review of theory and evidence. In: H. Gospel & S. Wood (Eds), Representing Workers: Trade Union Representation and Membership in Britain. London: Routledge. Cully, M., Woodland, S., O’Reilly, A., & Dix, G. (1999). Britain at work. London: Routledge. Department of Trade and Industry (2001a, June). Routes to resolution: Improving dispute resolution in Britain. Department of Trade and Industry (2001b, June). Dispute resolution in Britain: A background paper. Folger, R., & Greenberg, J. (1985). Procedural justice: An interpretive analysis of personnel systems. In: K. Rowland & G. Ferris (Eds), Research in Personnel and Human Resources Management (Vol. 3, pp. 141–183). Greenwich, CT: JAI Press. Gordon, M. E., & Fryxell, G. E. (1993). Voluntariness of association as a moderator of the importance of procedural and distributive justice. Journal of Applied Social Psychology, 19, 993–1009. Gospel, H., & Willman, P. (2003). Dilemmas in worker representation: Information, consultation and negotiation. In: H. Gospel & S. Wood (Eds), Representing Workers (Chap. 8). London: Routledge. Howard, C., & Furtado, T. (1999). The United States sentencing guidelines. Dallas, TX: Ombudsman Association.

The Organisational Ombuds

135

Kaminski, M. (1999). New forms of work organization and their impact on the grievance procedure. In: A. Eaton & J. Keefe (Eds), Employment Dispute Resolution and Worker Rights in the Changing Workplace (Chap. 7). Champaign, IL: IRRA. Lewin, D. (1999). Theoretical and empirical research on the grievance procedure and arbitration: A critical review. In: A. Eaton & J. Keefe (Eds), Employment Dispute Resolution and Workers Rights in the Changing Workplace. Champaign, IL: IRRA. Lewin, D., & Peterson, R. (1988). The modern grievance procedure in the United States. New York: Quorum Books. Metcalf, D. (2003). Unions and productivity, financial performance and investment: International evidence. In: J. Addison & C. Schnabel (Eds), International Handbook on Economics of Trade Unions (Chap. 5). London: Edward Elgar. Moorman, R. (1991). Relationship between organizational justice and organizational citizenship behaviors: Do fairness perceptions influence employee citizenship? Journal of Applied Psychology, 76(6), 845–855. Ombudsman Association (1995). Standards of practice. Dallas, TX. Rowe, M. (1997). What is it like to be an organizational ombudsman? Perspectives on Work, 1(2), 60–63. Shackleton, J. R. (2002). Employment tribunals. London: IEA Tremlett, N., & Banerji, N. (1994). The 1992 survey of industrial tribunal applications. Employment Department Research Series, 22. TUC (2002, November). Union reps – winning respect at work. London: TUC. Ury, W., Brett, J., & Goldberg, S. (1988). Getting disputes resolved: Designing disputes to cut the costs of conflict. San Francisco: Jossey-Bass. Walton, R. (1985). From control to commitment in the workplace. Harvard Business Review, 63(2), 57–74.

BUILDING AND SUSTAINING LABOR-MANAGEMENT PARTNERSHIPS: RECENT EXPERIENCES IN THE U.S. Susan C. Eaton, Saul A. Rubinstein and Robert B. McKersie ABSTRACT Since the 1980s the U.S. has experienced a variety of partnership arrangements between labor and management focused on improving industrial relations and organizational performance (Ichniowski et al., 1996; Kochan et al., 1986; McKersie, 2002; Rubinstein & Kochan, 2001). Yet there is an absence of research comparing these partnerships across industries and evaluating the factors that: (a) contribute to their success; (b) seem to be barriers to achieving their stated goals; or (c) predict which ones will stand the test of time. (For exceptions see Preuss & Frost, forthcoming 2003; Rubinstein, 2001b). This paper summarizes recent U.S. experience with partnerships; identifies factors that seem to influence the formation and sustainability of partnerships, including the development of network ties across traditional boundaries; and suggests theoretical and empirical implications of this experience in building and sustaining partnerships at work. We draw on a variety of types of evidence from the authors’ cumulated experience and research with more than 50 such partnerships in the U.S., spanning multiple industries and multiple decades. Advances in Industrial and Labor Relations Advances in Industrial and Labor Relations, Volume 13, 137–156 © 2004 Published by Elsevier Ltd. ISSN: 0742-6186/doi:10.1016/S0742-6186(04)13005-9

137

138

SUSAN C. EATON ET AL.

This study seeks to evaluate critically the development and sustainability of labormanagement partnerships in the United States during the last two to three decades, including some ongoing active partnerships. The authors use a combination of original case study research, literature review, meta-analytic and network analysis techniques, and participant-observation research to identify factors common to the apparent success and sustainability, as well as factors that serve as barriers to success in such partnership arrangements. The study differs from previous studies because it is grounded in in-depth experience and research with multiple original cases as well as relying on existing studies, covers more than one or even a few partnerships, and seeks to find patterns across time, unions, industry sectors, and types of partnership arrangements. While it cannot offer definitive conclusions for each case mentioned, it offers both practitioners and researchers a framework and tools for further analysis. In this paper we offer a brief historical perspective to summarize recent U.S. experience with partnerships; identify factors that seem to influence the formation and sustainability of partnerships, including the development of network ties across traditional boundaries; and suggest theoretical and empirical implications of this experience in building and sustaining partnerships at work. We draw on various types of evidence from the authors’ cumulated experience and research with a total of 50 partnerships in the U.S., spanning multiple industries and multiple decades.

INTRODUCTION AND HISTORICAL PERSPECTIVE ON U.S. LABOR-MANAGEMENT RELATIONS The United States has historically exhibited adversarial labor-management relationships. In the U.S, unlike in some parts of Europe, a corporatist structure never emerged that allied the governing classes with an organized labor movement peacefully participating in social governance and national level bargaining. This dominant conflictual style of relationships was somewhat muted during World War II, when the War Labor Board issued rulings that both promoted industrial peace and established some unions as providers of significant wage and benefit increases in their core industries. In the late 1940s, labor turmoil erupted, and more economic gains were made for workers, but soon after that, in the early 1950s, union membership reached its peak percentage ratio in the U.S. (about 35% membership) and the parties settled into a relatively stable period of institutionalized adversarial relations and economic growth. In the 1950s and 1960s, broad economic growth, working-class “family” wages in organized industries, and productivity grew in tandem. While most unions continued to engage in oppositional posturing and negotiations, significant gains

Building and Sustaining Labor-Management Partnerships

139

were made for members, including in the newly organized public sector workplaces during the 1960s and early 1970s (Freeman & Ichniowski, 1988). Most American unions engaged in what has been termed “business unionism,” leaving most of the initiative to the employer to “manage the workplace.” Most unions appeared to see their role as organizing and bargaining to take wages out of competition, grieving management’s alleged unfair practices or violations of the contract, and making regular gains for organized members in both economic areas and non-economic working conditions such as safety and health. But management tended to insist that managements’ rights clauses be all-encompassing, leading to the need to negotiate specifically over many relatively rigid job classifications and rules. Most unions did not bargain over the work process or product, nor at the strategic level of the business decisions, but confined their activity to collective bargaining arenas around a relatively narrow set of “mandated” subjects for bargaining (Kochan et al., 1994; McKersie, 2002). This began to change in the 1980s after a period of high inflation when management succeeded in eliminating cost of living increases from many contracts during negotiations or after strikes. Also, unions as well as companies began to experience the onset of globalization, low wages overseas, high technology penetration and aging plant structures in the industrial sector, and a large increase in low-wage service jobs as well as high-wage independent professional jobs in typically non-unionized sectors. During this decade, management in core industries began to seek partnerships with labor to relieve itself of some of what it then experienced as overly rigid work rules and job descriptions. Some employers adopted techniques for management that they viewed as leading to the success of the Japanese industries, particularly in auto and steel manufacturing. These included total quality management initiatives, off-line quality and process teams, on-line self-directed work teams, and other efforts to engage frontline workers, sometimes without the explicit involvement of the unions who represented them (see e.g. Adler, 1992; Appelbaum et al., 2000; Cooke, 1990). Further, the distinction between blue-collar and white-collar work began to blur, as more frontline workers effectively became computer operators, and as skill levels increased even while productivity growth slowed. Also in 1980 an era of concession bargaining began, captured for many by the image of President Ronald Reagan successfully firing U.S. federal air traffic controllers (and breaking their union) when they went on strike for more staffing and higher pay. Even though the non-union replacement controllers eventually re-organized their own new union, the broad postwar period of labor-management stability was over, and employers began coming to the bargaining table with changes that they wanted from their unions, including give-backs and take-aways of previously negotiated wages or benefits, higher health insurance premiums and

140

SUSAN C. EATON ET AL.

co-payments and fewer, broader, job classifications. Few unions were prepared to accept concessions at first, and while some did, U.S. labor unions as a whole continued to decline in membership. Some unions lost members during this period who felt forced out on strike and were then replaced with permanent new workers, other unions lost members when old manufacturing plants closed and workers were laid off, and some unionized jobs were lost to overseas job transfers. Gradually a new, contested, industrial relations approach began to emerge. Some companies wanted to try a new relationship with their unions, one in which frontline workers would be more empowered (Kochan & Osterman, 1994). Some union leaders began to envision new relationships with employers that would be less adversarial but would involve workers in higher-level decisions. Many union leaders distrusted management’s efforts to engage frontline workers and saw this as union-busting at worst and undermining the union at best. Also, many union leaders were not trained or prepared to engage with the new management techniques, particularly those which had been developed and promoted mostly in non-union environments. However, after a decade of downsizing and increasing competitive pressures, many union leaders reasoned that their earlier posture of “management acts, we react” was not suitable for the foreseeable future. To minimize the possibility of more cutbacks, and particularly to gain some say in how changes at work were made to deal with competition and changing environments, unions needed to be involved at the strategic level. They also needed to be participating in efforts to make the business more viable, which in at least some situations meant cooperating with programs to improve productivity. This laid the groundwork for union readiness to consider partnerships. On the management side of the equation, a readiness had also developed, at least where unions were strong, to involve them in dealing with the tough problems of the business. In industries such as airlines, autos, and steel, unions were not going to go away, and to continue the adversarial system with the downside of strikes and hassles over change programs was recognized by some as not wise on management’s part. In a way, both sides had something very important to gain, and they had already experienced what could be lost by continuing to negotiate and operate in a conflictual mode. This set the stage for at least some of the partnerships we discuss here.

THE U.S. EXPERIENCE WITH PARTNERSHIPS Although no hard data are available at this point, a number of partnerships have been operating for significant periods of time in key sectors, including

Building and Sustaining Labor-Management Partnerships

141

airlines, steel, telecommunications, electronics, clothing, and healthcare. There is little consensus in the industrial relations literature about the definition of a partnership arrangement. Clearly, they often involve sharing of information and open communication about business developments, in the interests of allowing the parties to develop a high level of trust (McKersie, 2002). Usually they involve finance and operations people from the management side, in addition to labor relations or human resource and legal experts. Many partnerships involve the union leadership at an early stage of decisions, as this quote from the Communications Workers of America illustrates: Union management participation (partnership) is an expansion of collective bargaining by union representatives at all company levels. It can include strategic decision making, process planning, reengineering, and how work is done. Participation implies increased union access to strategic information and involvement in process design before key corporate decisions are made (quoted in McKersie, 2002).

Most documented partnership studies have been focused at the facility level (Batt, 1990; Cooke, 1990; Frost, 2000; Hammer & Stern, 1986; Rubinstein, 2000) but a few have been corporate or industry-level efforts (Cutcher-Gershenfeld, 1988; Korshak, 1995; Kochan et al., 2002; Preuss, 1999; Rubinstein, 2003). Most partnership arrangements have included greater decision making on the part of labor in determining and performing work tasks and more involvement of labor institutions in managerial and strategic business decisions (Ichniowski et al., 1996). Partnerships typically result in closer working relations between labor and management institutions during the entire term of a contract, which may also be longer than the traditional three years. Partnership arrangements tend to make widespread use of both on-line and off-line teams to manage and improve operations (Rubinstein et al., 1993). For this reason they are sometimes confused with TQM (Total Quality Management) efforts. But, while TQM is focused on improving performance in part by widespread involvement of the workforce, it is essentially a management driven/dominated process. Also, under the partnership banner, many labor-management relationships have established structures for union involvement and representation at the highest levels of business decision making, whether corporate boards or governing committees. Again, a possible confusion here mixes partnerships with ESOPs, such as that at United Airlines, where union representatives hold a majority of seats on the board of directors. Yet an ESOP does not necessarily change the structure of work relationships below the strategic level. While partnerships share several characteristics with TQM and ESOP’s – to mention two examples of the genre that has been a theme of labor-management relations over a long time span – the distinguishing feature of the current

142

SUSAN C. EATON ET AL.

development is the emphasis on participation by union leaders and their members at all levels, especially at the strategic level. Historically, many examples exist of labor and management working together and sharing authority outside of the normal activities of collective bargaining – i.e. contact negotiations and grievance handling – however, these earlier examples tended to be limited to specific subjects such as worker displacement, training and health and safety. If the focus of the cooperation were productivity improvement, the union did not engage management on important operational decisions, nor on strategic ones such as budgeting, capital allocations and corporate alliances. Just how the participation of union representatives is structured at the strategic level varies considerably. At Kaiser Permanente representatives from the Coalition of Unions meet regularly with key management officials in a top-level council. At Saturn partnership was realized with the UAW having membership in the strategic action council. In the steel industry the union has negotiated the right to nominate members to the board of directors, thus enabling union concerns and viewpoints to be shared at the highest level of the enterprise. Below, we assess the staying power and significance of these arrangements as a whole. Among the factors we have found in our research that influence the sustainability and effectiveness of partnerships are labor and management leadership and turnover, resources available to the partnership, training, organizational structure, organizational learning, and diffusion. We have also noted a consistent tension that can arise between traditional union representation functions (like grievance handling and contesting management decisions) and partnership activities. Other factors we consider, though they cannot be analyzed in detail due to space constraints, include the role of shared and useful performance metrics, local union institutional history, organizational politics, the challenges that lead to crucial turning points or “pivotal events” in the relationships, corporate governance, and board representation.

RESEARCH METHODS The data for this paper come from research conducted by the authors on more than 50 partnerships in a variety of industries over two decades. In addition, one of the authors is a union-nominated director of a steel company that has entered into a labor-management partnership. The partnerships provide a variety of experiences in various types of innovative work systems across a number of industries including health care, automotive, steel, defense, telecommunications, airlines, papermaking, pharmaceuticals, higher education and the public sector. This research uses a mix of case studies, interviews, social network analysis and

Building and Sustaining Labor-Management Partnerships

143

surveys in an attempt to analyze partnership dynamics in depth as well as draw comparisons across a highly diverse group of workplaces.

Key Examples of Partnerships Studied We describe below five examples of partnerships that the authors have studied, partly to give the reader a sense of the extensive variation in patterns of initiating and sustaining partnerships. These partnerships are not presented as “typical” in any way, but as demonstrating the wide scope of both similarities and differences in some of the 50 or more partnerships whose experiences have informed our analysis. Saturn Corporation: Between the UAW and General Motors, perhaps the most extensive example of “co-management” where the UAW is an institutional partner and participates in consensus decision making from the shop floor to the strategic level of the enterprise. The partnership has produced higher-quality vehicles, better job security, and rapid implementation of decisions (see Rubinstein, 2000; Rubinstein & Kochan, 2001). CWA and ATT: These partners developed a plan and contract for “The Workplace of the Future,” and see labor-management participation as an expansion of collective bargaining by union representatives at all company levels, according to a CWA officer. Many accomplishments can be attributed to this partnership in multiple locations, but the parties have struggled to maintain it with limited success given the structurally fragile state of the telecommunications industry and a large number of downsizing and layoff decisions made by management. Kaiser Permanente and the Coalition of Kaiser Permanente Unions: In 1997 this large health maintenance organization with 90,000 employees signed a partnership agreement with a coalition of 26 local unions and 8 international unions that called for employment and income security, and joint work redesign efforts, but not for changes to collective bargaining. In late 2000, the parties signed a five-year “National Agreement” that provides the framework for all participating 75,000 unionized employees and their local union contracts. Unions are assured of consultation and input on key strategic matters, particularly on joint staffing and patient care. Anticipated benefits include the lowering of costs (such as reduced workers’ compensation claims), more flexibility, higher wages and employment security, and increased union membership (Kochan et al., 2002). The partnership is currently experiencing a transition to new leadership on the management side. Steel Industry – This partnership agreement emerged between the United Steelworkers of American (USWA) and the major steel companies in 1993. It was renewed in 1999 for another five years. Important elements of this partnership

144

SUSAN C. EATON ET AL.

included: employment security, training, information sharing, joint structures for problem solving productivity and quality problems, consultation on strategic decisions, and union-nominated directors (see Frost, 2000; McKersie, 2001, 2002; Rubinstein, 2003). U.S. Federal Government – many employees were involved here including members of the American Federation of Government Employees (AFGE) and the National Treasury Employees Union (NTEU). President Clinton issued an executive order in 1993 (rescinded by President G. W. Bush in 2001) that directed agency heads to create labor-management partnerships that would involve employees and their union representatives in identifying and solving problems. At the top level, a National Partnership Council was to coordinate and stimulate the formation of local partnership councils. Despite President Bush’s rescinding of the Executive Order, a majority of federal employees work in units where the councils have been formed. Their attention has been focused to date mainly on quality improvement, reorganization, and training programs (McKersie, 2002).

FACTORS THAT INFLUENCE THE FORMATION, SUSTAINABILITY, AND FAILURES OF PARTNERSHIPS From our 50-plus cases, we have identified a number of factors that contribute to the formation of partnerships. (1) First, many partnerships are initiated in the context of a crisis, either an impending confrontation between labor and management, a financial crisis of existence for the company, or other dramatic events in the relationship. Most had an economic root. Bethlehem Steel’s partnership – and in fact most of those in the steel industry (partly because of the influence of a national leader, another key factor we discuss later) – arose from the deteriorating competitive situation of steel production in the United States, which had led not only to concessions but to layoffs and plant closings. Similarly, in the auto industry, the Saturn partnership arose because no U.S. domestic automaker could produce a competitive, quality small car – and it became an effort to see if this was possible under a completely new set of work systems and industrial relations. The Fremont, California partnership between General Motors, the UAW, and Toyota (New United Motors Manufacturing, Inc. or NUMMI) came from the same basic cause, but represented a different strategy – using Japanese team-based management systems with American UAW-represented workers in a non-greenfield plant (Adler, 1992).

Building and Sustaining Labor-Management Partnerships

145

In health care, partnerships also arose because of competition and costcutting, though not because of the export of service sector jobs (which of course are much less able to be moved overseas to reduce labor costs, being locally based in many instances). In Minneapolis, the hospital association and the nursing and other healthcare unions came together to deal with the ravages of managed care that were affecting staffing, reimbursement, patient care patterns, and their capacity to stay in existence. Eventually the partnership deteriorated for some of these same reasons, however this is another story we will tell later in the paper. At Kaiser Permanente, the same crisis in health care cost containment, and an increasing level of competition from other HMOs, along with the non-profit organization’s first-ever annual loss (in the 100s of millions of dollars) led to concession bargaining that was variably successful across regions – but it also led the unions at Kaiser to form a coalition which was prepared to try to devastate the organization’s reputation and credibility. The potential success of this coalition’s destructive plans brought Kaiser leaders to the table (or, in this case, literally to the Dallas Fort Worth airport) to discuss doing things differently through a joint effort. (2) The second factor common to the formation of successful partnerships was extraordinarily dedicated and talented leadership on both sides – usually “champions” of the cause whose presence and working relationship with each other became a critical factor for the launching of the partnership – and whose turnover sometimes led to partnership declines. Such champions played a key role in the formation of partnerships at Saturn, Kaiser Permanente, Xerox, NUMMI, ATT-CWA, and others. In some industries, such as steel, the prominent status of the union champion (Lynn Williams, international president of the United Steel Workers of America at the time) ensured that partnerships would cover more than a single plant and would involve multiple local unions. (3) A third factor related to the formation of successful partnerships is union capacity to engage in managerial decision-making while maintaining the ability to balance individual representation with representing the interests of the collective in the partnership (see Frost, 2000; Rubinstein, 2001a for detailed examples at the local union level). One pattern we have seen repeatedly is an increased tension within unions, principally over how the union can continue to perform its representational function and at the same time engage in a new way in business or strategic decisions. One potential solution has been a new role for collaborative union leaders, sometimes called a “labor-management liaison” (as in the Kaiser Permanente case), or an “activist” (as in the Harvard Union of Clerical and Technical Workers case described in Eaton, 1996), or “coordinators” (in the

146

SUSAN C. EATON ET AL.

steel industry cases), or “partners” (in the Saturn case). These often newly active leaders require training in business processes, including interest-based bargaining, problem-solving, leading meetings, and sometimes finance and planning. Most unionists cannot manage a high level of “partnership” activities and maintain their full-time jobs and family lives, unless some part of their job becomes their new role in the partnership. In the case of Saturn, these unionrepresented “partners” served as co-managers of departments comprised of self-directed work teams. Additionally union leaders served on joint labormanagement governance bodies at the plant, site, and strategic level of the company. Some of these leaders were elected and some were appointed, which became an issue in contention internally in the union (Rubinstein & Kochan, 2001). (4) Partnership formation has also benefited from the presence of knowledgeable, skillful, and well-accepted facilitators, either developed from internal union or management leaders, or from outside the management or union through third party consultants or academics. Sometimes these facilitators are hired by the parties as consultants, as in the case of Restructuring Associates, Inc. at Kaiser Permanente. A key finding of the research work that has been conducted at Kaiser Permanente by a team from MIT and Harvard is that facilitation is especially crucial in helping the parties move from the traditional configuration wherein management acts and the union reacts to a decision making process characterized by interest based problem solving and consensus decision making. (5) A fifth factor that can aid partnership formation is strong enabling language that motivates, guides and directs organizational leaders (both union and management) to enter into such arrangements. One example is the Executive Order issued by President Clinton for federal agencies. Another example is the negotiation of Cooperative Partnership Agreements (CPA) across the U.S. basic steel industry, although the results have been mixed (Rubinstein, 2003). The CPA agreements appear to have promoted improvements in quality, safety, cost reduction, union participation in decision-making, and increased sharing of business information. However, for plants that had well-established local partnerships prior to the negotiation of the national partnership agreement, the prescriptive language in the CPA may have undermined the locally designed efforts. On the other hand, in plants unable to begin a partnership on their own prior to the negotiation of the agreement, the power and legitimacy of the national agreement may have helped them over start-up barriers. The steel industry experience shows that negotiated language is most effective if it focuses on key principles for

Building and Sustaining Labor-Management Partnerships

147

partnership arrangements, rather than on specifying detailed standards for organizational structures and membership at each level of the organization. Also, national contract language is most effective if it accommodates and builds upon workplace innovations that may already be in place at local facility levels.

Sustainability We identified a number of factors that are key to the sustainability of partnerships. (1) First and perhaps most interesting, our research on labor-management partnerships demonstrates that sustaining effective partnerships requires building dense communications networks. These are particularly important at the level of the frontline worker, but also require vertical communication between workers and management as well as union-represented coordinators or facilitators, and between the institutions of labor and management. Workers must also be able to communicate across traditional department lines. We have learned from in-depth analysis of several cases how much difference these networks can make to a partnership’s success. This is particularly clear in the strongest partnership examples that come closest to “co-management.” In the case of Saturn, network analysis was used to understand and model the labor-management partnership structure (Rubinstein, 2000). Among the variables found to have had the most significant effect on quality performance were interdepartmental communications and coordination between union leaders in managerial positions, and the alignment between these union leaders and their non-represented partners. The local union was found to function within the partnership as a dense communications network among its represented partners who filled managerial roles, solving quality and other operational problems. This network was supported by a variety of institutional forums that made up the governance process of this local union and functioned as the backbone of the partnership adding significant economic value to Saturn (Rubinstein, 2000). A network analysis of the impact of the union-management partnership at Bristol-Myers Squibb showed positive effects on the pattern and structure of employee communications (Rubinstein & Eaton, forthcoming). Horizontal communication among peers increased as did communication between workers and managers. Centralization measures the level of participation or involvement in a network and gives us an indication of the level of hierarchy and control by measuring the ability of a few actors in a network to control access to

148

SUSAN C. EATON ET AL.

information. Networks which are less participative and more hierarchical tend to be more centralized. Centralization indices at Bristol-Myers Squibb declined after the formation of the partnership indicating the communication networks became less hierarchical and more egalitarian as participation increased. At the same time the level of customer satisfaction improved across all categories – reliability, responsiveness, timeliness, professionalism and overall customer satisfaction from the time the partnership training was introduced in 1999 to the present. Further, the communication between union and management representatives also improved as employees reported more influence over daily activities and departmental decisions. More workers began communicating directly with managers and also with research scientists, which seems to have improved the quality and quantity as well as the value of the communications. This in turn led to labor’s increased ability to influence management decision making as shown by surveys and interview data (Rubinstein & Eaton, forthcoming). In the Kaiser Permanente partnership, registered nurses and support staff like nurses’ aides and licensed vocational nurses (LVNs) often find themselves with different kinds of information and perspectives on problems with patients than registered nurses (RNs) have. They may observe depression symptoms, for example, or particular kinds of skin conditions that are not apparent to the RNs, who typically spend less time directly communicating with patients in many settings. These workers in different occupations are usually represented by different unions. To deal with repeated kinds of miscommunication or perceived lack of respect, sometimes the unions gather their members across classification boundaries to take on projects like opening a new hospital (as in the case of Baldwin Park) with co-management and non-traditional cross-functional teams (see Kochan et al., 2002). Alternatively, RNs and their union leaders often are aware of problems with physicians that operations leaders are not – in one example, two similar medical clinics had quite different performance and outcomes metrics for no apparent reason, until the nurses’ union representative pointed out that one chief medical officer used and worked well with nurse practitioners, whereas the other one did not. This provided new insight to operations managers and they were able to influence the behavior of doctors where the nurses themselves could not (Kochan et al., 2002). Similarly, in a university setting, faculty assistants organized through their union sat down with faculty at a joint council meeting and planned a “new faculty and staff” orientation that would ease typical tensions in working together. The administrative staff that the faculty assistants reported through could not have called such a meeting with faculty, given university politics. The orientation functioned in part through enacting reverse role plays, for example, with the assistants (now acting as faculty) giving their faculty members (now acting as assistants) express mail

Building and Sustaining Labor-Management Partnerships

149

packages to be posted within five minutes, while asking for multiple copies of the syllabus and new reference materials from the library at the same time (Eaton, 1996). New networks and relationships were created during the orientation sessions and the joint council meetings that facilitated getting non-partnership educational work done at later times. (2) A second key factor related to sustainability seems to be confronting and overcoming key “pivotal events” or “turning points” where the relationship hinges on a successful resolution of some pending potential disruption.1 These may be related to leadership turnover in either labor or management, to labor disputes or economic crises and pressures that fall outside the partnership, to key decision points where someone or the parties take a big risk, to a membership ratification or governing body meeting, or to any of a number of other areas of partnership. A strike in Portland Oregon nearly derailed the budding Kaiser Permanent labor-management partnership in 1999, for example. Also, when a new union president was elected at ALPA on a platform of getting full parity in pilot salaries, this caused problems for the partnership with United Airlines. One critical event that always seems to present extreme challenges to partnerships occurs when the union leadership challenges management’s willingness to support the union’s institutional goals. This has derailed partnerships in cases where managers would not oppose “right to work” laws or grant recognition by card check, for example (see Preuss & Frost, forthcoming). (3) Third, it is critical that the alliances and coalitions on both sides of multiparty relationships continue to grow and thrive. Where these are threatened, sometimes a key supporting column of the partnership is removed. For instance, in Minnesota, the original employer coalition split as a result of industry changes into competing managed care networks and the partnership could not be held together under those circumstances. Similarly, the Kaiser Permanente partnership has had a harder time succeeding in negotiating a new local union agreement in Atlanta, where the larger community is predominantly non-union, in contrast to California, where unions represent a major portion of the workforce and constitute a large percentage of HMO members. (4) Fourth, industry sustainability and improved productivity or quality are essential elements to creating joint gains to be shared by the parties. In bad times, partnership can be weakened if old tensions re-emerge, or if each party blames the other for problems, or if internal tensions on one or both sides are magnified by the partnership. (See Preuss & Frost, forthcoming 2003) For example, the implosion of the telecommunications industry has made it

150

SUSAN C. EATON ET AL.

very difficult for the CWA partnerships with AT&T to sustain, given massive layoffs and pressures to reorganize. (5) In addition, maintaining the tension between collaborative work and conflictrelated or distributional issues in a productive way is more easily said than done. Partnerships link organizations with shared power, mixed motive relations and histories of low trust. (Walton & McKersie, 1965; Walton et al., 2000) To be effective they require enhanced communications, information sharing, increased trust, joint problem solving, and attention to the interests of multiple stakeholders. This includes the development of processes that facilitate alignment on goals and responsibilities (Rubinstein, 2000). Working out whether the partnership is really a joint collaboration with consensus decision making, co-management, cooperation, an off line consultative arrangement, or some other complex relationship, is never easy, and may in fact be different at different levels of the relationship. In the Harvard University partnership with HUCTW, its union for clerical and technical workers, the negotiators created “joint councils” for building relationships in major workplaces and discussing significant issues like work reorganization, while they identified “problem-solving” committees to handle what would typically have been considered grievances (Eaton, 1996). The Kaiser Permanente partnership sponsors joint committees at multiple levels, from nationally to regionally, from the level of the medical center or clinic to some departments. The higher up in the organization the committee is located, the more likely its work will be “off line” and require much more work at local levels to implement “on line” (Kochan et al., 2002). (6) Successful partnerships have also learned to build capabilities to manage their relationships across organizational boundaries with parent companies and national unions. For example, the failure to do this over time was a key factor in the lack of expansion of the Saturn partnership beyond Spring Hill, Tennessee. The success of some steel industry partnerships is likely due to the support and even demand by the international union for locals’ participation (Frost, 2000) as well as some local unions’ positive capacities and resources to participate. Managers at Xerox in Rochester, NY who were involved in partnership with their ACTWU (later UNITE) colleagues needed to maintain support for the partnership in the rest of Xerox, which was and remains predominantly nonunion (Cutcher-Gershenfeld, 1988). (7) Related to the preceding point about building capacity is the reality that a partnership will not last unless this capacity is present at all levels of interaction between the union and management. And more is involved than training and education – at the heart of a successful partnership is a relationship built on mutual understanding and respect.

Building and Sustaining Labor-Management Partnerships

151

If management at the operating level commits to working collaboratively with the union on programs for maintaining a competitive edge, while at the same time fighting the same union in its attempts to organize new ventures of the company, then the partnership concept will not survive. One of the reasons that several unions, especially ALPA and the United Steelworkers have sought to place their nominees on corporate boards is to ensure that strategic decisions are aligned with the premises of the partnership and that interests of shareholders do not dominate the interests of other stakeholders. Failures We have seen many partnerships thrive for some period of time, then plateau and either fail or become much less active (or less joint). Examples are familiar to industrial relations scholars and practitioners – for example, the Caterpillar/UAW partnership, the Twin Cities Hospitals and Minnesota Nurses’ Association as well as other unions involved; Corning, Inc. and the American Flint Glass Workers Union; ATT and CWA; and Inland Steel and the United Steelworkers Local 1010, are among them. While some of these are related directly to industry problems that were mentioned above as a possible derailing factor, others were more directly tied to the partnership itself. (1) A profound issue that many partnerships face is engaging middle managers and union stewards in the partnership. We have seen, as in Kaiser Permanente, extraordinary levels of communication and leadership at the “top” of both labor and management, but we have also seen some middle managers who “put their heads down and hope this too will pass,” as a high-level manager put it. Also, some union stewards are so trained in the “confrontational” or grievance mode that partnership is problematic for them, or at least they do not feel confident or possess the skills required to engage in it productively. At Kaiser Permanente some physicians have supported and participated in partnership projects with enthusiasm, while others seem to hope that the partnership will result in “happier” workers who smile more at members and that patient evaluations will improve as a result. One doctor complained that the partnership training was supposed to occur “on our own time;” while that was not the parties’ intent, the fact is that many doctors feel they have virtually no free or discretionary time while at work (Kochan et al., 2002). (2) This can lead to a second kind of pivotal event, challenges by members who feel that the union leaders have abandoned their traditional “individual representational” role for something akin to attending management meetings

152

SUSAN C. EATON ET AL.

and talking about “running the business” issues. These tensions arose in Saturn and led to changes in the local union partnership structure (Rubinstein & Kochan, 2001), and they have also occurred in the steel and health care partnerships. Similar issues are being raised in some, but not all, of the Kaiser Permanente union coalition members’ organizations. (3) A third area where failure is more likely seems to occur in the lack of sufficient union or management capacity to carry on the sophisticated partnering at three levels (workplace, collective bargaining, and strategic) that is required for success. For example, a union or union coalition may be partnered at the bargaining level, as at Harvard, but still be struggling to reach a majority of rank-and-file workers to encourage them to use the partnership on a day-to-day basis (Eaton, 1996). A similar problem can occur when management has not fully spread the partnership through its middle management structure to ensure activity at an operational level. One way to do this is by making middle managers’ collaborative evaluations and/or compensation at least partially contingent on successful partnership activity, a step Kaiser Permanente has recently taken. (4) Further, as might be expected from the conditions for success noted above, an unsuccessful transition to new leadership, a political or economic decline of the industry, and other failures to successfully navigate “pivotal events” can also lead to declines. Examples include the change in Presidential leadership from Clinton to Bush, when the new president rescinded the Executive Order supporting partnerships in the Federal sector. Another example occurred when Levi Strauss decided after extensive joint work with ACTWU (now UNITE) to relocate significant production offshore. Similarly, ATT decided to open non-union wireless facilities, threatening its partnership with CWA. In short, when precipitous restructuring, leadership changes, or other political pressures take a partnership by surprise, the parties can find themselves returning to an adversarial relationship. (5) A final reason why partnerships may not last involves the tension between confidentiality and representative democracy. Most partnerships involve the union at the strategic level. With this access comes confidential information about business plans. Some union leaders are nervous about having information that they cannot share, and if the plans call for changes that are averse to the short run interests of the members, they fear they may get blamed for the decisions. In addition, they feel that they have some obligation to report back to the membership who elected them. This is one reason why some unions ask outside directors to serve on boards in lieu of union officers. There is a problem on the flip side of this involvement, as well. Union leaders expect as a result of the partnership to be apprised of all-important strategic decisions.

Building and Sustaining Labor-Management Partnerships

153

When management holds back information, for example because of Securities and Exchange Commission (SEC) rules, the union leadership feels marginalized and basic partnership relationships are threatened. This is precisely what happened on several occasions at Inland Steel; top management was exploring acquisitions and felt it could not share this with the union leadership who in turn viewed the holding back of information as a violation of the spirit of the partnership. A recent in-depth study looking at the experience of the Minnesota Nurses’ Association and the Minneapolis hospitals identified four paradoxes that the authors proposed are likely to lead to the demise of a partnership (Preuss and Frost, forthcoming). They noted that constraints imposed by cooperation across employer sites (in a multi-employer partnership) seemed to limit the topics and potential solutions the partnership could address. For instance, local worksite issues differed from hospital to hospital depending on distinct approaches to patient care, and most partnership meetings did not occur at individual hospital levels but rather with a multi-employer committee. Second, they suggested that the institutions of cooperation in that case were static and unable to evolve to meet shifting issues, contexts, or problems: the parties cannot grow out of the “initial event” that caused them to come together. Third, they believed the employees eventually “turn” on cooperation, and find in it a ready target for anything that is problematic for them at the workplace. This is a version of expectations for partnership exceeding reality, something that is clearly a problem for many participants. Finally, they thought cooperation was vulnerable to any conflict in the relationship of the parties (even having little to do with cooperation). We agree that these are areas of challenge for many partnerships.

CONCLUSION A multi-year research program tracking a network of local unions involved in long-term partnership arrangements found that union roles, structures, and leadership had changed significantly over time (Rubinstein, 2001b). While a wide variety of experiences, tensions, and controversies had occurred, a set of common characteristics emerged from the 20 cases examined in this research. These local unions have continued to fulfill their responsibilities for contract administration, negotiations, and grievance handling, yet they were also increasingly taking on roles in firm governance, challenging traditional managerial prerogatives. Union leaders saw that the collective economic and security interests of their members could be met by organizations that were both flexible and responsive to the demands of a changing global economy. Not wanting to leave the responsibility for developing these capacities to management alone, they saw both the need

154

SUSAN C. EATON ET AL.

and the potential power that can be achieved by representing members through processes of joint governance and co-management. The power and influence of these local unions has increased at both the strategic and operational levels through knowledge acquisition and increased control and responsibility for managing the business. When management is seen less as a class of employees and more as a legitimate avenue for unions to represent the interests of their membership, local leaders have found themselves needing to expand their leadership abilities and repertoires and the organizational capacities of their unions. This has involved rethinking local structures, reorganizing resources around these multiple roles, and acquiring new skills. These skills go beyond technical ones to strategic and process skills, providing leaders with the capability to participate in firm governance, to add value through managerial decision making, and to manage change in complex organizations, including both the company and the local union. Local union leaders are increasingly trying to create a vision for the union that balances the responsibility for running the business and collective bargaining, with representing individuals. They are also finding that the capability to managing across organizational boundaries is increasingly important in order to maintain healthy relationships with other stakeholders such as the corporation, national union, other locals, customers, and suppliers. Most important, and most challenging, to survive in this new role, union leaders have found they needed to develop the political skills needed to balance these different roles (Rubinstein, 2001b). Labor-management relationships have always involved a mixture of hard bargaining (forcing) and joint problem solving (fostering) (Walton et al., 1994), in fact giving all negotiations scholars an interesting lens through which to view interests in general. Labor-management partnerships are an effort to institutionalize the “fostering” side of the relationship without completely abandoning the “forcing” side – where interests are not coincident. The most fertile soil for labormanagement partnerships seems to be in places where there is a strong union presence (so avoidance is not an option), a competitive or otherwise changing environment (so management has an interest in jointness), a willingness to break with the past, strong leadership on both sides, and perhaps the possibility of damaging each other beyond repair, in a “lose-lose” crisis. Each partnership’s success ultimately rests on the kinds of successful communication and coordination that can be attained where the results of collaboration meet the basic interests of each party, whether employment security, voice, organizational performance, or something else. While we can quantify some of these relationships with network analysis, there are deeply human connections involved with each party sometimes coming to see the perspective of the other in new ways. One doctor at Kaiser Permanente remarked that at times he could

Building and Sustaining Labor-Management Partnerships

155

not tell which person was from labor and which was from management at joint meetings, since each was articulating a joint concern. Each partnership seems to confront (and may arise from) a series of pivotal events, where actors make decisions that lead either toward further, stronger partnership, or away from it and back toward adversarialism. We have seen examples of both – and we have used network analysis to show that possibilities for work systems redesign, participation in decision making, and reductions in hierarchy, as well as other positive outcomes, can depend very much on the strength and nature of connections within and across the partnerships. Diffusion is always a concern for innovators and leaders – and unions provide a tremendous potential tool for diffusion, as long as their members’ needs and interests are also served by the partnership outcomes and processes. Yet partnerships face challenges, whether leadership turnover (and the inevitable breaking of strongly forged network links), external environment changes and pressures, and political pressures from within and without. Our research also suggests that network methods can be helpful in analyzing the internal processes of labor-management partnerships in a new and deeper way. Network analysis, combined with more richly textured case studies and classic industrial relations research linking workplace relations to organizational outcomes, can help us understand the nature of these challenges and perhaps even help the parties to overcome them.

NOTE 1. We are grateful to Tom Kochan for this concept and for emphasizing this point in our discussions.

REFERENCES Adler, P. (1992). The learning bureaucracy. In: L. L. Cummings & B. M. Staw (Eds), Research in Organizational Behavior (Vol. 15, pp. 180–205). Greenwich, CT: JAI Press. Appelbaum, E., Bailey, T., Berg, P., & Kalleberg, A. (2000). Manufacturing advantage: Why highperformance work systems pay off. Ithaca, NY: Cornell ILR Press. Cooke, W. N. (1990). Labor-management cooperation. Kalamazoo, MI: Upjohn Institute. Cutcher-Gershenfeld, J. (1988). Tracing a transformation in industrial relations: The case of Xerox Corporation and the Amalgamated Clothing and Textile Workers Union. Washington, DC: U.S. Department of Labor. Eaton, S. C. (1996). The customer is always interesting: Unionized Harvard clericals renegotiate working relationships. In: C. Macdonald & C. Sirianni (Eds), Working in the Service Society (pp. 291–332). Philadelphia: Temple University Press.

156

SUSAN C. EATON ET AL.

Freeman, R., & Ichniowski, C. (1988). When public sector workers unionize. Chicago: The University of Chicago Press. Frost, A. C. (2000). Explaining variation in workplace restructuring: The role of local union capacities. Industrial and Labor Relations Review, 53(4), 559–578. Hammer, T. H., & & Stern, R. N. (1986). A yo-yo model of cooperation: Union participation in management at the Rath Packing Company. Industrial and Labor Relations Review, 39(3), 337–349. Ichniowski, C., Kochan, T., Levine, D., Olsen, C., & Strauss, G. (1996). What works at work: Overview and assessment. Industrial Relations, 35(3), 299–333. Kochan, T. A., Eaton, S. C., & McKersie, R. B. (2002). The Kaiser labor-management partnership: The first five years. Unpublished MS. Kochan, T. A., Katz, H. C., & McKersie, R. B. (1986). The transformation of American industrial relations. NY: Basic Books. Kochan, T. A., & Osterman, P. (1994). The mutual gains enterprise. Boston: Harvard Business School Press. Korshak, S. R. (1995). Negotiating trust in the San Francisco hotel industry. California Management Review, 38(1), 117–137. McKersie, R. B. (2001). Labour’s voice at the strategic level of the firm. Transfer (Autumn). McKersie, R. B. (2002). The power of labor-management partnerships. Address presented at the Lima Peru Conference on Growth and Equity. Preuss, G. (1999). Committing to care: Labor-management cooperation and hospital restructuring. Economic Policy Institute Working Paper. Washington DC. Preuss, G. A., & Frost, A. C. (Forthcoming Winter 2003). The rise and decline of labor-management cooperation: Lessons from health care in the twin cities. California Management Review. Rubinstein, S. A. (2000). The impact of co-management on quality performance: The case of the Saturn Corporation. Industrial and Labor Relations Review, 53(2), 197–218. Rubinstein, S. A. (2001a). A different kind of union: Balancing co-management and representation. Industrial Relations, 40(2), 163–203. Rubinstein, S. A. (2001b). The local union revisited: New voices from the front lines. Industrial Relations, 40(3), 405–435. Rubinstein, S. A. (2003). Partnerships of steel? – Forging high involvement work systems in the U.S. steel industry: A view from the local unions. Advances in Industrial and Labor Relations, 12. Rubinstein, S., Bennett, M., & Kochan, T. (1993). The Saturn partnership: Co-management and the reinvention of the local union. In: B. Kaufman & M. Kleiner (Eds), Employee Representation: Alternatives and Future Directions. Madison, WI: IRRA Press. Rubinstein, S. A., & Eaton, A. E. (Forthcoming). The effects of high involvement work systems on employee and union-management communication. Unpublished MS. Rubinstein, S. A., & Kochan, T. A. (2001). Learning from Saturn. Ithaca, NY: ILR Press, Cornell University. Walton, R. E., Cutcher-Gershenfeld, J., & McKersie, R. B. (1994). Strategic negotiations. Ithaca, NY: Cornell ILR Press. Walton, R. E., & McKersie, R. B. (1965). A behavioral theory of labor negotiations. New York: McGraw-Hill.

UNION AND EMPLOYER TACTICS IN ONTARIO ORGANISING CAMPAIGNS Felice F. Martinello and Charlotte Yates ABSTRACT Cluster analysis is applied to the union and employer tactics used in a sample of Ontario organising campaigns to identify the combinations of tactics or strategies that are used most often. Seven union organising strategies and five employer resistance strategies are revealed. Contingency table analysis shows that the union and employer strategies are not independent of one another. More active campaigns by one side (in terms of more tactics used) are met by more active campaigns by the other side. Regression analysis is used to estimate the effects of the strategies on the outcome of the organising campaign. The most active strategies, including intensive communication with workers and worker committees, work best for the employers. For unions, strategies emphasising personal communication through house calls are the most effective.

INTRODUCTION The decline of union density in the United States, and more recently in Canada, has led labour unions in these countries to invest increased time and resources into organising the unorganised. This renewed emphasis on organising raises a number of questions for union practitioners and industrial relations analysts alike about the linkage between union organising and rates of union membership. Can Advances in Industrial and Labor Relations Advances in Industrial and Labor Relations, Volume 13, 157–190 Copyright © 2004 by Elsevier Ltd. All rights of reproduction in any form reserved ISSN: 0742-6186/doi:10.1016/S0742-6186(04)13006-0

157

158

FELICE F. MARTINELLO AND CHARLOTTE YATES

union organising reverse the decline in union density, and therefore play a central role in union renewal? What role do union, as well as management, strategies play in determining the success of union organising efforts? Are certain union and management strategies more effective than others in influencing the outcome of organising campaigns? What impact, if any, do structural factors such as the type of workforce, union or workplace play in determining union or management tactics and the outcome of the campaign? Several scholars have offered analyses in answer to these questions. Out of these analyses have emerged several debates over the relative importance of union and management strategies vs. structural factors for union success in organising campaigns and the transferability of U.S. organising practices to other national labour movements and experiences. Without discounting the importance of union organising, recent stock-flow analyses by Farber and Western (2001) for the United States and Johnson (2002) for Canada show that increases in organising alone are unlikely to reverse the decline in union density. This and other related research (Card & Freeman, 1993; Martinello, 2000; Riddell, 2001; Rose & Chaison, 1990) point to the importance of legislative and economic factors in determining levels of union membership. At the same time, recent research on the incidence and effects of both union and employer tactics during an organising drive (Bronfenbrenner, 1997 and the references cited therein; Peterson et al., 1992; Reed, 1989) underlines the importance of union and management behaviours in influencing the outcome of organising campaigns. Yates (2000, 2001) uses the data set that we use to estimate the effects of gender and selected union tactics on campaign success. Bronfenbrenner and Juravich (1998, p. 21) go so far as to conclude about Bronfenbrenner’s earlier work that “. . . union tactics as a group play a greater role in explaining election outcome than any other group of variables . . .”. Still other research points to bargaining unit characteristics such as size and gender and racial/ethnic composition of the workforce as crucial in determining the outcome of organising campaigns in the United States (Crain, 1994; Farber, 2001; Milkman, 2000). Despite these ongoing debates over the determinants of union organising success and membership renewal, unions and industrial relations scholars in other countries have embraced the work done in the United States and pushed union reform in their own countries based on an American organising model (Carter, 2000; Fairbrother & Yates, forthcoming 2003; Heery, 1996; Heery et al., 2000; Oxenbridge, 1997). Although national variations in industrial relations are accounted for in these transfers of practices and ideas, there has been virtually no empirical validation of whether the union strategies used in the United States work elsewhere. This paper makes three contributions to these debates. First, Lawler (1990) and Bentham (2002) argue that individual tactics may not be the most relevant factor for organising success and that analysing combinations of tactics, which in turn reflect

Union and Employer Tactics in Ontario Organising Campaigns

159

overall strategies, may be more appropriate. By applying cluster analysis to union and employer behaviours, this paper uses empirical techniques that shed new light on union and management strategies. The cluster analysis moves us away from studies of individual tactics to the identification of distinct combinations of tactics, which serve as a better proxy for understanding union and management strategy. Through this analysis, we identify and describe seven union strategies and five employer strategies as well as the distinct characteristics of the bargaining units where such strategies are employed. Second, using contingency table analysis, we test the proposition made by a number of authors (Bronfenbrenner, 1997; Peterson et al., 1992, p. 379) that unions and employers adjust their strategies in response to those deployed by their opponent. The contingency tables overwhelmingly reject the independence of union and employer behaviour and show which union and employer strategies tend to be used against one another. Thirdly, the data set used to analyse these issues is a relatively new one based on survey data collected on organising drives in the province of Ontario in Canada. Using this Canadian data set, we replicate some of the work done by Bronfenbrenner and others on the effect of union and employer strategies on the outcome of organising campaigns. We find that union and management tactics that rely on direct communication with workers have the greatest influence on the outcome of an organising campaign. We do, however, reach different conclusions from Bronfenbrenner about the efficacy of mass communication methods, such as mailing letters to workers, for union organising. This analysis thus contributes to the debate about the transferability of the American organising model.

DESCRIPTION OF THE INSTITUTIONAL SETTING, THE SURVEY AND THE SAMPLE Most collective bargaining in Ontario occurs under the authority of a certificate issued by the Ontario Labour Relations Board (OLRB). Voluntary recognition of unions by employers is allowed, but it is rare, especially in the sectors considered below. Union organisers start an organising drive or campaign and file an application for certification with the OLRB when they believe that they have sufficient support from the workers. Upon receiving the application, the OLRB schedules a representation vote and a hearing. At the hearing, the OLRB “disposes of” the application by granting or dismissing it, based on the results of the representation vote. If the OLRB grants the certification application, the union wins the right to represent the workers, and we refer to this as a successful organising campaign. The employer, of course, may also campaign throughout the organising drive to convince workers to vote against the union.

160

FELICE F. MARTINELLO AND CHARLOTTE YATES

Data on the tactics used by unions and employers were collected through a 1999 survey of union organisers. Surveys covering 1,281 organising campaigns were sent to the union organiser primarily responsible for the campaign. The campaigns correspond to the certification applications disposed of by the OLRB between August 1996 and July 1998. 677 or 52.8% of the surveys were returned. Due to limitations in resources and an original primary focus on union strategies, our information on employer behaviours comes from the union organisers. To address possible questions about the reliability of our employer data, it is useful to compare incidences of employer behaviours found in our sample, with that of authors who have surveyed employers directly. Frequencies from our data on hiring a consultant or lawyer, or filing a charge of unfair labour practice, are similar to those reported by Bentham (2002) and Thomason (1994). Our data register higher incidences of captive audience meetings (61% vs. 49%) and anti-union literature (46% vs. 36%) than those found in Thomason and Pozzebon (1998) but the differences are barely significant at the 10% level. Further, they can likely be explained by changes in the legislative and overall political environment in Ontario between the later 1980s and early 1990s when Thomason and Pozzebon collected their data and the late 1990s when our data were collected. See Martinello (2000) for a description and analysis of these changes. The biggest differences in reported employer behaviours are found in comparing our data with Bentham’s (2002) on propensity to challenge the bargaining unit (29% of campaigns vs. 46% here) and employer promises to improve working conditions or threats to relocate or contract out (both less than 5% in Bentham, 2002 vs. 43% and 24% respectively in our data). Given the highly charged nature of the latter employer behaviours, and the fact that they arguably constitute unfair labour practices on the employers part, one cannot draw definite conclusions about which source of information is more reliable. Overall, then, the comparison of our data collected from union organisers with studies based on data collected from employers demonstrates a fair degree of consistency and points to the reliability and validity of our data. The sample used in the analysis is a subset of the returned surveys. Specifically, organising campaigns directed at workers already represented by a union (i.e. raid campaigns) are excluded from the sample, as are surveys with any incomplete or “Don’t know” information on the union or employer campaign tactics. Organising campaigns covering workers in the construction industry are also excluded because the rules governing certification in that industry are quite different from those in the rest of the economy. Finally, campaigns covering public sector workers are excluded, yielding a final sample of 237 organising drives.1 Basic descriptive statistics of the organising drives in the sample are shown in Table 1. In Table 1, International unions refers to unions that operate in Canada but are part of a larger American union (e.g. the United Steel Workers of America).

Union and Employer Tactics in Ontario Organising Campaigns

161

Table 1. Descriptive Statistics of the Organising Campaigns. Sample (%)

Survey (%)

OLRB Certification Applications (%)

Successful organising campaigns

67.5

68.6

61.4

Campaigns with bargaining unit size 1 to 9 employees 10 to 19 employees 20 to 39 employees 40 to 99 employees 100 to 199 employees 200 to 499 employees over 500 employees

11.8 19.4 22.8 29.5 10.6 4.6 1.3

24.8 19.9 20.5 20.8 8.6 4.0 1.5

32.8 21.3 18.5 18.7 5.4 1.5 1.9

26.3 73.7

N.A. N.A.

Public sector Private sector

0 100

Campaigns by major industry Construction Health and social services Manufacturing Other services Retail trade Food, beverage and accommodation Transportation and storage

0 16.9 29.9 16.0 12.2 8.9 6.3

21.0 19.7 14.8 10.5 7.0 4.3 4.4

25.6 16.5 16.6 16.1 4.5 5.1 1.9

Campaigns by major union United Food and Commercial Workers United Steel Workers Canadian Auto Workers Ontario Nurses Association CUPE Teamsters

24.5 21.5 9.7 6.3 5.5 5.1

11.8 11.1 5.9 5.3 10.6 2.8

8.7 7.9 4.4 2.7 6.8 4.5

International unions Average % of females in the bargaining units

69.2 45.0

64.4 38.5

N.A. N.A.

Number of organising campaigns

237

677

1,320

Note: N.A. indicates not available.

Table 1 also shows descriptive statistics for all of the survey responses and the campaigns whose applications were disposed of by the OLRB over the same period.2 The percentage of successful organising drives in the sample is slightly higher than for the other Ontario campaigns at that time and the sample organising drives cover larger bargaining units. The distribution of organising campaigns by industry in the sample is strongly affected by the removal of the construction and

162

FELICE F. MARTINELLO AND CHARLOTTE YATES

public sector campaigns, and the manufacturing and retail trade industries are overrepresented. The distributions of campaigns by union shows that the larger unions are over represented in the sample, but their proportions relative to one another are similar to the population.

DESCRIBING UNION AND EMPLOYER BEHAVIOUR DURING ORGANISING CAMPAIGNS This section describes the behaviour of unions and employers during organising drives. Cluster analysis is employed to identify the combinations of tactics that union organisers use most often in organising campaigns. Once identified, popular combinations of tactics are considered union organising strategies or categories of behaviour. The analysis is repeated using the employers’ tactics to identify common employer strategies or categories of behaviour. The unit of observation, however, is the organising campaign and not the tactic. Thus the cluster analysis actually identifies organising campaigns where similar combinations of tactics are used, and those similar campaigns are put into clusters.3 Consider a cluster made up of campaigns with similar union organiser tactics. All of the campaigns in the cluster are deemed to use the same organising strategy or belong to the same category of union organiser behaviour, since they all use a similar combination of tactics. An examination of the tactics used in the clustered campaigns reveals the composition of that strategy, and the number of campaigns in the cluster indicates the popularity of that strategy. The combinations of tactics used in the largest clusters show the composition of the most common union organising strategies. This interpretation holds in the same way for campaigns clustered according to the employer tactics. Cluster analysis is a form of “data reduction.” It is necessary because of the large number of union and employer tactics. No feasible data set is large enough to even span all of the possible combinations of union or employer tactics. This situation is even more difficult because the tactics tend to be used in combinations. There is simply not enough independent variation in the use of the tactics to estimate their individual effects reliably. Cluster analysis is required to reduce the large number of tactics to a few strategies or categories of behaviour for unions and employers that characterise most of the campaigns in the sample. As in all data reduction techniques (e.g. factor analysis) some judgement is required. With cluster analysis one must decide how many clusters of campaigns to allow. This, in turn, determines how many campaigns each cluster will contain and how similar the tactics in those campaigns will be. For example, if there are more clusters, the clusters will contain fewer campaigns and the tactics used in

Union and Employer Tactics in Ontario Organising Campaigns

163

those campaigns will be more similar. The opposite holds if there are fewer clusters. These decisions are analogous to factor analysis decisions about how many factors to include and what factor loadings are significant and to be included. In addition, some grouping or splitting of clusters should be done manually since the cluster analysis treats all tactics equally, with no understanding of their content. For example, tactics like distributing leaflets and sending letters to workers are similar. Thus campaigns that have a high incidence of one but not the other could reasonably be put together into the same cluster. Other tactics, like distributing leaflets and appearing to fire someone for union activity are very different. Campaigns with different incidences of these two tactics should definitely be split into separate clusters.

Union Strategies Table 2 shows the union tactics included in the data, listed in descending order of frequency of use. The first numerical column in Table 2 shows the percentage of campaigns in the sample where that tactic is used. Before proceeding to the results of the cluster analysis, note that the union tactics form a number of distinct groups based on their overall frequency. This grouping by overall frequency is completely different from the results of the cluster analysis reported below, but it helps to explain the results of the cluster analysis. The first three tactics; using an inside organising committee, telephoning workers, and small group meetings; are used in over 80% of the campaigns. A second group of tactics; letters, leaflets, and house calls; are each used in 50–60% of the campaigns. Large group meetings are held in 43% of the cases and rank and file organisers from other bargaining units are used 29% of the time. The frequency declines steadily for the remaining tactics and nine of the tactics are used in less than 10% of the campaigns. A measure of the similarity of union campaign tactics must be defined before the cluster analysis can be applied. In this case the similarity between the union tactics in any two campaigns is the proportion of tactics where the organisers’ behaviour matches (i.e. the tactic is used in both campaigns or the tactic is not used in both campaigns). Given this measure of similarity, the cluster analysis is restricted to tactics that are used more than 5% of the time. This exclusion is required because, if the rarely used tactics are included, the huge number of negative matches (the tactic is not used in either campaign) overwhelms the positive matches and there is little discrimination across the different sorts of organiser behaviour in the campaigns. The seven rightmost columns of Table 2 show the seven strategies or categories of organiser behaviour identified by the cluster analysis and the combination of tactics that are used in each. Each category is given a name reflecting its dominant

164

Table 2. Categories of Union Organiser Behaviour. Tactic

# Of campaigns

All

#1 Small Grp Meetings

237

17

85 84 81 59 54 54 43 29 20 17 9 8 4 4 3 3 3 1 1

12 76 94 18 6 0 18 24 0 0 0 0 0 0 0 0 0 0 0

#2 Large Grp Meetings

#3 Letters

#4 House Calls

#5 Rank & File Organisers

#6 Offers Education

#7 The Maximum

18

40

37

13

13

52

100 100 92 69 54 100 15 15 100 69 0 0 0 15 15 0 8 0 8

96 96 94 92 85 90 90 40 13 13 33 12 6 8 4 8 6 2 2

Percentage of Campaigns Using that Tactic 78 100 97 100 50 100 89 92 33 95 92 100 11 80 49 46 11 50 46 38 0 3 100 92 100 10 8 23 11 13 0 100 22 35 0 0 22 0 19 54 0 0 0 0 0 0 14 0 0 0 8 15 0 0 0 8 6 0 3 8 0 3 3 8 0 3 0 0 0 0 0 8 0 0 0 0

FELICE F. MARTINELLO AND CHARLOTTE YATES

Use inside organising committee Telephone workers Have small group meetings Mail letters to workers’ homes Distribute leaflets at the workplace Make house calls to workers homes Have large group general meetings Use rank & file organisers from other BUs Offer education (e.g. OH&S) to workers’ Newsletters Have solidarity (button or t-shirt) day Use email, web site, electronic communication Social events Leaflets or letters to public and/or clients Videos Use print, radio, or television media Used other unions or labour solidarity groups Rallies, public forums and/or debates Used community groups

Categories of Union Organiser Behaviour

Union and Employer Tactics in Ontario Organising Campaigns

165

characteristic or tactics. The seven categories account for 190 or 80% of the campaigns. For example, the rightmost column of Table 2 shows one cluster of campaigns where the union organisers use similar tactics. The cluster contains 52 organising campaigns. The union organisers in these campaigns use the largest number of tactics, compared to the campaigns in the other clusters, so this category of union organiser behaviour is called The Maximum. The numbers in the column show the percentage of campaigns in the cluster where that tactic is used. An inside organising committee, for example, is used in 96% of the campaigns in that cluster. Unfortunately, no test has been developed to determine whether the cluster analysis did a good job or a poor job of grouping campaigns into clusters. One cannot test whether differences in tactic use across the categories are statistically significant because the organising drives are sorted by their use of tactics and not randomly sampled from (possibly) different populations. Thus the usual rules of statistical inference cannot be applied. That said, a minimum standard for judging the performance of the cluster analysis is whether the differences between the clusters would be statistically significant if there had been random sampling. Tests of this minimum standard were done using the binomial distribution for the proportions of organising drives where the tactic is used. The details of the test results are not reported, but they show that even with the small numbers of campaigns in some categories, the differences in tactics use described below are all large enough that they would be statistically significant at conventional levels. One can test whether other characteristics of the campaigns are significantly different across the categories. Table 3 reports the values of other characteristics of the campaigns for each of the categories and the two tailed p-value for the hypothesis test that they are not different from the rest of the sample. Univariate, anova type tests are used with no adjustments for the levels of the other variables. Table 3 also shows the industries that are over and under represented in each category at 5% or 10% significance levels with one tailed hypothesis tests.4 The test for industry concentrations is limited to industry/category combinations where the expected frequency is at least five occurrences to maintain the validity of the test. This requirement accounts for the lack of over or under represented industries in the small categories. In general the union strategies identified by the cluster analysis distinguish themselves along the dimensions of: (a) the overall level of activity (i.e. how many tactics are employed in the campaigns in the cluster and how often) and

166

Table 3. Other Characteristics of the Campaigns in Each Organiser Category. Categories of Union Organiser Behaviour All

∗ Indicates

#3 Letters

#4 House Calls

#5 Rank & File Organisers

#6 Offers Education

#7 The Maximum

67.5

64.7 (0.8)

77.8 (0.34)

67.5 (0.99)

81.1* (0.06)

76.9 (0.46)

46.2* (0.09)

63.5 (0.48)

31.5

8* (0.06)

17.5 (0.22)

31 (0.0)

29 (0.78)

26 (0.72)

57* (0.07)

78** (0.0)

69.2

76.5 (0.5)

50* (0.07)

67.5 (0.8)

75.7 (0.36)

53.8 (0.22)

92.3* (0.06)

76.9 (0.17)

45.0

27.2** (0.03)

61.7* (0.06)

62.5** (0.00)

34.4* (0.05)

51.9 (0.50)

61.4* (0.09)

36.9* (0.07)

Health*

Manufact*

significant at the 10% level. significant at the 5% level.

∗∗ Indicates

#2 Large Grp Meetings

Manufact**

Other services**

FELICE F. MARTINELLO AND CHARLOTTE YATES

% Of organising campaigns successful(p-value of test whether different from others) Median size of the bargaining unit(p-value of test whether different from others) % With international unions(p-value of test whether different from others) Average % of females in bargaining units(p-value of test whether different from others) Industries significantly over-represented in the category: Industries significantly under-represented in the category:

#1 Small Grp Meetings

Union and Employer Tactics in Ontario Organising Campaigns

167

(b) the use of more personal communication (e.g. small group meetings, house calls) as compared to more impersonal communication (e.g. distributing leaflets, sending letters and holding large group meetings) although the content of the communication (offering education) discriminates one small category and the use of rank and file organisers discriminates another, also small, category. Table 2 shows that all of the strategies, except the first and second, use an inside organising committee, telephone workers, and hold small group meeting in most campaigns; as is to be expected given the high overall frequencies for these tactics. Arranged from least to most active, the categories of union organiser behaviour are: (1) Just Small Group Meetings. Organisers in this category are the least active. They hold small group meetings, telephone workers, and use rank and file organisers at roughly the sample averages, but use very few other tactics. Table 3 shows that this strategy is used in bargaining units that are much smaller than the rest of the sample, so the small number of tactics may be sufficient for the situation. Certainly the success rate of these campaigns is not different from the overall average. This strategy is also used in bargaining units with much higher proportions of male workers. (2) Just Large Group Meetings. This category is very different from all of the others because it uses large group meetings in every campaign, and telephoning workers and small group meetings much less often than the rest of the sample. Organisers use inside organising committees, education and newsletters similarly to other campaigns in the sample but far fewer of the other tactics. In particular, letters, leaflets and house calls are hardly used. Table 3 shows that this strategy is used less often by International unions and more often in bargaining units with higher proportions of females. The success rate of these campaigns is higher and the bargaining units are smaller than the rest of the sample, but the differences are not statistically significant. This category also tends to occur more often in the social and other services industries (except business services), but there are too few expected observations in those cells to make the differences statistically valid. (3) Sends Letters. Organisers are much more active in this category. In addition to the common tactics of inside organisers, telephoning, and small group meetings, organisers send letters to workers’ homes more often than the overall sample, and more often than all but the most active category. Organisers distribute leaflets and offer education at roughly the sample averages, but use few other tactics.

168

FELICE F. MARTINELLO AND CHARLOTTE YATES

The success rate of the campaigns, the bargaining unit size, and the proportion of international unions in this category are all very similar to the overall sample. This set of tactics is used more often in the health industry and therefore, not surprisingly, in bargaining units with higher proportions of females since bargaining units in the health industry tend to be female dominated. (4) Makes House Calls. The defining feature of this category is that house calls are made in every campaign. Large group meetings are rarely held and letters and leaflets are used slightly less than in the overall sample. The common tactics of inside organising committees, telephoning, and small group meetings are used more often than average. All of this suggests a greater emphasis on personal, one-on-one communication compared to the previous categories of organiser behaviour. Table 3 shows that union campaigns using this set of tactics have a much higher rate of success. The campaigns occur more often in the manufacturing industry and in bargaining units with lower proportions of females. (5) Uses Rank & File Organisers. This small category is quite similar to the previous one in that it uses house calls in almost every case; fewer leaflets, letters and large group meetings than the average; and more small group meetings, telephoning and inside organising committees. Thus organisers show the same emphasis on personal, one-on-one communication with the workers. The distinguishing feature of this category is the use of rank and file organisers. No other category comes close to using rank and file organisers as often. This strategy tends to be used less by International unions although the difference is not significant due to the small number of observations in the category. None of the other characteristics of this category are significantly different from the sample averages. (6) Offers Education. This is another small but distinct category. It is similar to the previous two categories in that it emphasises personal, one-on-one communication with house calls and telephoning in every case, many small group meetings, and few large group meetings. However it is also a much more active category; with leaflets, letters, and newsletters in 50–70% of the campaigns. It is, in fact, almost as active as the most active category. The unique feature of this category is that organisers offer education (e.g. on occupational health and safety, sexual harassment, rights on the job) to workers in the bargaining unit in every campaign.

Union and Employer Tactics in Ontario Organising Campaigns

169

This is the only category where there is little diversity in the unions conducting the organising campaigns. Only four unions are represented and 77% of the campaigns were run by one international union, accounting for the extremely high proportion of international unions shown in Table 3. Despite its small size, all of the characteristics of this category are significantly different from the sample. Table 3 shows that this strategy is used in larger bargaining units with higher proportions of females and that the success rate is substantially lower than average. This combination of tactics is more likely in the retail trade, and food beverage and accommodation industries but the small number of expected occurrences makes the differences statistically invalid. (7) The Maximum. As shown in Table 2, organisers in this category use the most tactics the most often. House calls and small group meetings are used in most cases along with all of the mass communication methods (letters, leaflets, large group meetings). Also notable is the use of solidarity days (where union supporters display union caps, t-shirts, or buttons in the workplace to generate support for the union) in one third of the campaigns. Solidarity days are not used at all in any of the other organiser strategies. This category is actually comprised of two clusters. The first one contains 45 campaigns whose characteristics are very similar to those shown in Table 2. The second cluster contains 7 campaigns that are grouped into a different cluster because they are even more active than first. Thus it is reasonable to combine the two clusters into one, most active, category. It is not surprising that these “all out” campaigns are used in the largest bargaining units with a median size more than twice the overall average. These campaigns are more common in the very competitive manufacturing industry and in bargaining units with low proportions of females. They are much less common in the other services industry. The campaigns in this category are very average with respect to their win rate, and involve only an insignificantly higher proportion of international unions.

Employer Strategies The same method of analysis is applied to the tactics used by the employer during the union’s organising campaign. Table 4 lists the employer tactics in approximately descending order of frequency of use. Employers are reported to use fewer tactics than unions during the organising drives and there is not the same

Tactic

Categories of Employer Behaviour All

# Of campaigns

237 57 46 61 50 46 43 41 40 24 20 19 16 11 10 9 8 4.6 4 3 3 3 2 2

#1 Do Nothing 15 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

#2 Advice, CBU Some Comm

#3 Promises

#4 Just Communication

#5 Do Everything

45

14

47

44

Percentage of Campaigns Using That Tactic 91 79 19 93 71 9 42 71 98 29 14 96 42 21 77 4 100 60 11 7 77 13 0 60 7 21 17 0 36 15 2 14 21 16 0 6 2 0 6 4 0 2 0 0 0 4 0 4 0 7 9 0 0 2 0 0 4 7 0 0 2 0 0 2 0 2 0 0 2

93 64 98 98 73 68 82 80 48 61 27 41 34 27 30 25 7 16 12 5 7 5 9

FELICE F. MARTINELLO AND CHARLOTTE YATES

Hires consultant and/or outside law firm Challenges the union’s proposed bargaining unit Holds captive audience meetings in working hours Distributes anti-union leaflets at work Mails anti-union letters to workers’ homes Promises improvements in wages, hours, . . . Spreads rumours about union and key supporters Sets up a workers’ anti-union committee Threatens to close, relocate, or contract out work Appears to use bribes or favours to get votes Makes “positive” changes in mgmt personnel Unfair labour practice charge filed by union Makes improvements in wages, benefits . . . Appears to fire someone for union activity Initiates an employee involvement plan Lays off workers during the campaign Promotes key union leaders during the campaign Organises social event with an anti-union message Shows anti-union movies or videos Important community members speak vs. Union Files unfair labour practice against the union Makes press statements or took out media ads Contracts out work during the campaign

170

Table 4. Categories of Employer Behaviour.

Union and Employer Tactics in Ontario Organising Campaigns

171

clear groupings of tactics according to how often they are used in the full sample of campaigns as there is for union organisers. The cluster analysis is again restricted to tactics used in more than 5% of the campaigns to improve the discrimination across campaigns. Even with this restriction, however, employer tactics are used much less often than the union tactics. This causes the large number of negative matches (i.e. the tactic is not used in either campaign) to overwhelm the number of positive matches (i.e. the tactic is used in both campaigns) in the similarity measure and poor discrimination across campaigns results. To solve this problem a similarity measure that counts the number of positive matches and ignores the negative matches is used.5 Note that an unfair labour practice charge is included as an employer tactic. It is clearly a union decision, but it is usually in response to a management behaviour and the literature almost always treats it as an indicator of employer behaviour. Table 4 shows the five employer strategies or categories of behaviour identified by the cluster analysis (with some manual grouping noted below) and the combination of tactics that are used in each. Table 5 shows the other characteristics of the campaigns in each category and p-values analogous to Table 3. The categories account for 165 or 70% of the campaigns. In general, the categories of employer behaviour identified by the cluster analysis distinguish themselves from one another along the following dimensions: (a) the overall level of activity, (b) whether the employer sought legal or management advice and challenged the proposed bargaining unit, and (c) the content of the communications with the workers. Arranged from least to most active, the categories of employer behaviour are: (1) Do Nothing. In 15 of the 237 organising campaigns employers use absolutely none of the tactics. Table 5 shows that this occurs more often in smaller bargaining units, with higher numbers of females, and that the union campaigns are more likely to be successful. But none of these differences are statistically significant. (2) Advice, Challenge the Bargaining Unit (CBU), and Some Communication. In this category the employer hires a lawyer or consultant and challenges the bargaining unit in almost every campaign. In addition the employer communicates directly with the employees. Letters are sent at roughly the sample average. Meetings with workers and leaflets are also used, but less frequently than in the sample as a whole. Very few of the other tactics are used. As for the content of the communications, they appear to be squeaky clean.

172

Table 5. Other Characteristics of the Campaigns in Each Employer Category. Categories of Employer Behaviour

∗ Indicates

significant at the 10% level. significant at the 5% level.

∗∗ Indicates

#1 Do Nothing

#2 Advice, CBU Some Comm

#3 Promises

#4 Just Communication

#5 Do Everything

67.5

80 (0.29)

84.4** (0.01)

57.1 (0.4)

59.6 (0.2)

47.7** (0.0)

31.5

21 (0.49)

25 (0.31)

25.5 (0.61)

38 (0.42)

73** (0.0)

69.2

66.7 (0.83)

51.1** (0.0)

64.3 (0.68)

89.4** (0.0)

68.2 (0.87)

45.0

59.3 (0.13)

51.3 (0.19)

50.1 (0.59)

40.6 (0.35)

43.0 (0.69)

Health**

Manufact*

Retail*

Retail*

Other services*

FELICE F. MARTINELLO AND CHARLOTTE YATES

% Of organising campaigns successful ( p-value of test whether different from others) Median size of the bargaining unit ( p-value of test whether different from others) % With international unions ( p-value of test whether different from others) Average% of females in bargaining units ( p-value of test whether different from others) Industries significantly over-represented in the category: Industries significantly under-represented in the category.

All

Union and Employer Tactics in Ontario Organising Campaigns

173

There are very few promises of improvements, threats of closure or relocation, appearances of bribes, or rumours about the union or key supporters. Given this it is surprising that unfair labour practices are filed against the employer at the overall sample average. This category is created by combining two clusters consisting of 34 and 11 campaigns, respectively. The tactics in the larger cluster are very similar to those shown in Table 4. The campaigns in the smaller cluster are also very similar with the exception that they use a different combination of leaflets, letters, and meetings. The organising success rate in this category is much higher than it is in the other categories. This strategy is used more often in the health industry and less often in the retail industry. It is also used in campaigns with a significantly lower proportion of International unions. The bargaining units have higher proportions of females but the difference is not significant. (3) Promises. This small but distinct category is similar to the previous one in that the employer hires advice, mounts a bargaining unit challenge, and communicates directly with the employees in most cases. The advice and bargaining unit challenges, however, are used less frequently than in the second category, and there are differences in the methods of communication. In this category employers use more direct and personal methods of communication, with more meetings and fewer letters and leaflets, but the differences are not large enough to be significant under the minimum standard discussed above. The big difference from the previous category is the content of the communication. Employers make promises of improvements in wages and working conditions in every case and appear much more likely to offer bribes to gain votes against the union. The employer also makes more threats of closure or relocation, but the difference is not nearly as large. This strategy may be effective as the organising success rate is below average, but the difference is not statistically significant. Table 5 shows that none of the other characteristics in this category are substantially different from the sample averages. (4) Just Communication. In this category the employer is much more active than in the previous categories, but far less likely to hire a consultant or challenge the bargaining unit. Instead, the employer communicates directly with the workers much more intensively than in the previous categories or in the sample as a whole. The employers hold captive audience meetings and distribute leaflets during working time in almost all cases and mail letters to workers’ homes in over three quarters of the campaigns. An anti-union committee of employees is established in 60% of the cases. As to the content of the communications,

174

FELICE F. MARTINELLO AND CHARLOTTE YATES

employers make promises of improvements and spread rumours about the union or rank and file supporters in over three quarters of the organising drives. While not squeaky clean, the content of the communications are not blatantly illegal. There are few reported threats of closure or relocation, bribes, or charges of unfair labour practices. This category consists of two clusters; one with 39 campaigns and another with 8 campaigns. The smaller cluster is different from the larger, in that it uses only leaflets and no letters, instead of the roughly equal proportions of leaflets and letter in the larger group. Table 5 shows that this combination of tactics is used much more often against international unions and in the manufacturing industry, and less often in the retail industry. The success rate is less than in other campaigns, but not significantly. (5) Do Everything. Employers are much, much more active in these campaigns. Tactics are used, on average, almost twice as often as in the previous category. In addition to using all of the tactics that characterise the previous categories, employers in this group: – make threats of closure, contracting out, or relocation, – appear to offer bribes or favours to obtain anti-union votes, – make improvements to wages or working conditions during the campaign, – initiate an employee involvement program, – lay off workers during the campaign, and – appear to fire workers for union activity. These tactics are used much more often than any of the other categories or in the overall sample. Not surprisingly, all of these activities lead to a much higher frequency of charges of unfair labour practices. The success rate for campaigns in this category is much lower than for all of the other categories and significantly lower than the overall sample. This strategy is used in bargaining units that are much larger than the sample median and less likely to be in the “other services” industry. This category also tends to occur in the manufacturing and the food beverage and accommodation industries, but the differences are not statistically significant.

Discussion of Union and Employer Strategies Tables 3 and 5 show that union and employer strategies vary systematically with the characteristics of the bargaining unit and the union. Both employers and unions choose more active strategies the larger the bargaining unit. This is consistent with optimising choices by the employer and the union. The larger benefits or costs

Union and Employer Tactics in Ontario Organising Campaigns

175

associated with the certification of a larger bargaining unit make it rational to expend more resources for or against the organising drive. Recall that the level of activity is measured in terms of the number of tactics used or the proportion of campaigns using the tactics, so activity is not mechanically related to the size of the bargaining unit. The most striking feature of Table 3 is that the proportion of females in the bargaining units varies substantially across the categories of union organising behaviour. It is surprising that the less personal Large Group Meetings and Letters categories are much more common in bargaining units with higher proportions of females while the more personal, one-on-one communication categories of Small Group Meetings and House Calls occur in bargaining units with much lower proportions of females. The proportions of females in the bargaining units are not significantly different in any of the employer strategies individually. The fourth numerical row of Table 5, however, shows a clear trend. The proportion of females tends to fall as the employer’s strategy becomes more active. Regression analysis on the proportion of females by employer category cannot confirm this downward trend at a 5% significance level, but it is close, and the trend is statistically significant at the 10% level. One should not conclude that this is a causal relationship since there are other inter-correlated variables, such as industry, but it is worth noting and a subject for future investigation. There are few statistically significant results with respect to strategy by industry because of the small sample sizes. It seems clear, however, that campaigns in the manufacturing industry are hotly contested with high levels of resources expended by both sides. On the other hand, organising drives in the other services industries are much less likely to be hotly contested, with both sides unlikely to use their most active campaigns. Finally, certification drives in the health industry appear to be fairly polite affairs with meetings, letters, and leaflets; and much lower incidence of the other tactics.

RELATIONSHIPS BETWEEN UNION AND EMPLOYER STRATEGIES A contingency table can be used to show whether the union and employer strategies are related to one another.6 Unfortunately with five employer categories, seven organiser categories, and only 237 observations, a contingency table provides credible inferences for only a few combinations of union and employer strategies. To solve this problem the first two (least active) union categories are collapsed into a new category called Do Little. The campaigns in the Rank & File Organisers category are merged into House Calls, the Offers Education campaigns are added

176

Table 6. Observed Minus Expected Frequencies by Behaviour Category (Expressed as Percentages). Categories of Union Organiser Behaviour

#2 Advice, CBU Some Comm.

#3 Promises

#4 Just Communication

#5 Do Everything

0.84 (0.29) 1.05 (0.26) 0.45 (0.44) −2.33* (0.08)

2.62* (0.099) 0.32 (0.49) 0.8 (0.41) −3.74* (0.08)

0.67 (0.3) 0.36 (0.45) 0.01 (0.59) −1.05 (0.31)

−0.2 (0.56) −0.49 (0.5) −1.0 (0.41) 1.69 (0.3)

−3.93** (0.01) −1.24 (0.36) −0.26 (0.54) 5.42** (0.03)

Note: The probability of obtaining an observed frequency that far or further from the expected, using a binomial distribution, is shown in parentheses. It is also the p-value for a one tailed hypothesis test that the observed frequency is not different from the expected frequency. ∗ Significantly different from zero at the 10% level in a one tailed hypothesis test. ∗∗ Significantly different from zero at the 5% level in a one tailed hypothesis test.

FELICE F. MARTINELLO AND CHARLOTTE YATES

# 1, 2 – Do Little # 3 – Letters # 4, 5 – House Calls (+ R & F Organisers) # 6, 7 – The Maximum (+ Offers Education)

Categories of Employer Behaviour #1 Do Nothing

Union and Employer Tactics in Ontario Organising Campaigns

177

to The Maximum, and the Letters category is left unchanged. This yields a new total of four union behaviour categories. Table 6 shows, for each of the employer and new union categories, the difference between the observed frequency of that combination of categories and the expected frequency under the null hypothesis that union and employer categories are independent.7 The frequencies are expressed as percentages and each cell is the observed percentage minus the expected. Thus positive numbers show that that combination occurs more often than one would expect if the union and employer categories were independent. The number in parentheses shows the probability of obtaining an observed frequency that far or further from the expected frequency using a binomial distribution. Thus it is also the p-value for a one tailed hypothesis test that the observed frequency is not different from the expected frequency. Table 6 shows that more active behaviour by one side is matched by more active behaviour from the other side. The corresponding relationship also holds for the less active categories. When one side chooses a less active strategy the other side is also likely to use a less active strategy. These correlations are shown by the large positive values in the upper left and lower right corners of Table 6. Table 6 also confirms that the converse is true. When one side chooses a more active strategy, the other side is unlikely to use a less active strategy. This is shown by the large negative values in the lower left and upper right corners of Table 6. Finally, it is notable that when employers do absolutely nothing during the campaign, union organisers are less likely to choose the augmented The Maximum category, but there are no other significant differences in their campaign strategies. A 7 × 5 contingency table is used to test the independence of the original behaviour categories. A Pearson chi squared test overwhelmingly rejects the hypothesis that the union and employer choices of strategy are independent of one another. Since the behaviour categories are ordered by increasing levels of activity, Kendall’s tau-b and Goodman and Kruskal’s gamma provide measures of association between the activity levels of the union and employer strategies. Both measures show a positive association that is overwhelmingly statistically significant. Thus the original union and employer behaviour categories confirm that union organisers are more active (i.e. use more tactics) when the employers are more active more tactics and vice versa.8

UNION AND EMPLOYER STRATEGIES AND ORGANISING CAMPAIGN SUCCESS Theory investigating how and why tactics and other factors affect union organising success is well covered in sources such as Lawler (1990), Bronfenbrenner (1997)

178

FELICE F. MARTINELLO AND CHARLOTTE YATES

and Freeman and Kleiner (1990). Thus, for the sake of brevity, we do not review that material here. Bentham (2002) provides a review of recent empirical work on the effects of employer tactics and other factors on union organising success. Lawler (1990) and Heneman and Sandver (1983) provide extensive reviews of the earlier literature. The left panel of Table 7 reports tobit coefficient estimates where the proportion of votes cast in favour of the union is the dependent variable. Tobit is a maximum likelihood estimation procedure that makes explicit adjustments for campaigns where the proportion of votes equals one or zero, corresponding to representation elections that are unanimously in favour or opposed to the union. It also treats the proportion of votes cast as a single piece of information, regardless of the size of the bargaining unit or the total number of votes cast in the election.9 The right panel of Table 7 reports maximum likelihood estimates from a logit regression, where the dependent variable is 1 or 0, corresponding to whether the organising campaign was successful or not. Control variables are included in both regressions and are listed in Table 7.10 The behaviour category variables are dummy variables that equal one for campaigns where the unions or employers use that strategy. Unclassified campaigns are the omitted category for both union and employer strategies so the regression coefficients show the estimated effect of that strategy compared to the unclassified campaigns. The relative sizes of the coefficients within each regression are also informative since they show how that strategy affects union campaign success, compared to the other strategies. Differences in the relative sizes of the effects are described below only if the differences are statistically significant. Comparisons of the sizes of the estimated coefficients across regressions are not valid because the dependent variables are different. The union and employer strategies are arranged in descending order of their coefficients in both regressions. This corresponds to going from the strategies most favourable to a successful organising campaign to the strategies that are least favourable.

Employer Strategies The two regressions yield very similar estimates of the effects of the employer behaviour categories. Both regressions show that more active management resistance is associated with lower campaign success even after adjusting for union organiser behaviour, industry, bargaining unit size, and other variables. In the tobit regression, the two most active behaviour categories, Just Communication and Do Everything, significantly reduce the proportion of votes in favour of the union

Tobit

Logit Dependent variable: Campaign successful => 1; Campaign not successful => 0

Dependent variable: Proportion of votes in favour of the union 0.043 (0.068) 0.026 (0.046) −0.088 (0.071) −0.124** (0.051) −0.127** (0.049)

# 2 – Advice, CBU, Some Comm. # 1 – Do Nothing # 4 – Just Communication # 3 – Promises # 5 – Do Everything

0.412 (0.597) −0.143 (0.804) −1.130** (0.566) −1.221* (0.712) −1.306** (0.509)

Union behaviour categories # 1 – Small Group Meetings # 4 – House Calls # 3 – Letters # 2 – Large Group Meetings # 7 – The Maximum # 5 – Rank & File Organisers # 6 – Offers Education

0.189** (0.071) 0.106** (0.054) 0.095* (0.052) 0.073 (0.071) 0.047 (0.050) 0.007 (0.080) −0.117 (0.081)

# 4 – House Calls # 7 – The Maximum # 1 – Small Group Meetings #3 – Letters # 5 – Rank & File Organisers # 2 – Large Group Meetings # 6 – Offers Education

1.402** (0.656) 0.452 (0.556) 0.389 (0.731) 0.264 (0.581) 0.222 (0.858) −0.531 (0.817) −0.542 (0.835)

Control variables Manufacturing industry Health industry Other services industry Worker called union Organiser initiated contact Organiser experience in years Female organiser

−0.094** (0.042) −0.002 (0.056) 0.067 (0.053) −0.067 (0.044) −0.016 (0.060) 0.001 (0.003) −0.028 (0.046)

Manufacturing industry Health industry Other services industry Worker called union Organiser initiated contact Organiser experience in years Female organiser

−0.648 (0.445) 0.249 (0.691) 0.184 (0.567) −0.230 (0.480) 0.259 (0.673) 0.027 (0.031) −0.540 (0.502)

179

Employer behaviour categories # 1 – Do Nothing # 2 – Advice, CBU, Some Comm. # 3 – Promises # 4 – Just Communication # 5 – Do Everything

Union and Employer Tactics in Ontario Organising Campaigns

Table 7. Regression Estimates: Union and Employer Strategies (Listed from Most to Least Favourable for Organising Success).

180

Table 7. (Continued ) Tobit

Number of campaigns Note: Standard errors are in parentheses. ∗ Indicates significant at the 10% level. ∗∗ Indicates significant at the 5% level.

0.162** (0.042) 0.032 (0.037)

−0.043 (0.042) −0.029 (0.040) −0.000 (0.000) 0.001 (0.001) −0.000 (0.001) −0.009 (0.031) 0.594 (0.078) 206

Union committed to drive because of strong worker interest Union committed to drive because it represents workers in competing workplaces

1.209** (0.466)

Previous organising drive in unit International union Bargaining unit size Percentage females in unit Percentage part time More than 20%, English is not their primary language Constant

−0.398 (0.438) −0.195 (0.437) 0.001 (0.002) 0.014** (0.006) 0.002 (0.007) 0.190 (0.368)

Number of campaigns

0.982** (0.462)

−0.549 (0.850) 221

FELICE F. MARTINELLO AND CHARLOTTE YATES

Union committed to drive because of strong worker interest Union committed to drive because it represents workers in competing workplaces Previous organising drive in unit International union Bargaining unit size Percentage females in unit Percentage part time More than 20%, English is not their primary language Constant

Logit

Union and Employer Tactics in Ontario Organising Campaigns

181

compared to the Do Nothing and Advice CBU Some Communication strategies, and the tactics in the unclassified campaigns. The logit equation shows the same effects with one exception. The Do Nothing category has a huge standard error and so its effect is statistically indistinguishable from the any of the other categories or the unclassified campaigns. The Promises category also reduces the probability that the campaign is successful in the logit regression compared to the second category and the unclassified campaigns. It is not estimated to have a statistically significant effect compared to the first two strategies in the tobit equation, although it is close. The effects of the three most active strategies are not statistically different from one another in either regression. The Do Nothing and Advice CBU Some Communication categories do not have statistically significant effects compared to the unclassified campaigns. Their positive effects on organising success, relative to the more active categories, are stated above. The results suggest that, after controlling for union behaviour and other variables, an effective anti-union strategy involves intensive communication with workers. Captive audience meetings appear to be effective when there is some real content to the communications. The marked success of the last two categories also suggests that setting up worker anti-union committees is effective. The use of management consultants is a widely cited and controversial employer tool for resisting union organising drives. Lawler (1990) cites earlier work showing that consultants reduce support for the union but concludes (p. 117) that they have an ambiguous effect overall. Bronfenbrenner (1997) and Peterson, Lee and Finnegan (1992) also report statistically insignificant effects for management consultants. Our results support the conclusion that management consultants have an insignificant effect on organising success. Consultants are used intensively in both effective and ineffective strategies, and hardly used at all in one of the most effective employer categories. Unfair labour practices are another controversial employer tactic. Table 7 shows that the Just Communication and Do Everything categories have very similar effects even though they have very different rates of appearing to fire workers for union activity, laying off workers and charges of unfair labour practice. This suggests that unfair labour practices are not very effective. The cluster analysis also has implications for some estimates of the effects of unfair labour practices found in the literature. Recall that charges of unfair labour practice occur mostly in the Do Everything category, where many other tactics are also used. Thus an estimate of the effects of unfair labour practices should include controls for the other employer tactics used. If unfair labour practices is included without controls for other employer tactics as in, for example Thomason (1994),

182

FELICE F. MARTINELLO AND CHARLOTTE YATES

Riddell (2001), or Freeman (1986), the estimate includes the cumulative effect of all of the tactics used in the campaigns and not just the unfair labour practice. The large negative impact of the Just Communication category shows that the other tactics, such as direct communication with workers, also affect campaign success. Thus the reported impact of unfair labour practices may be overestimated.

Union Strategies The results from the two regressions differ more for the union behaviour categories, but both show a significant positive effect for the House Calls strategy compared to the unclassified campaigns. In the logit equation the estimated effect of the House Calls category is huge and significantly greater than the effects of most of the other strategies. In the tobit equation the estimated effect of House Calls is smaller and significantly larger than only the worst, Offers Education, category. The union strategies other than House Calls have large standard errors in the logit equation so the estimates convey little additional information. The effects of these other strategies are not statistically distinguishable from one another, or from the tactics in the unclassified campaigns. The tobit equation shows a few more statistically significant results. Small Group Meetings has a large positive effect on the proportion of votes in favour of the union compared to the unclassified campaigns and the three worst union strategies (numbers 7, 5 and 6). The Letters category also has a positive significant effect on votes that is quite close to the House Calls category. This is surprising since it contrasts with the findings of Bronfenbrenner (1997) and Peterson et al. (1992), who report that traditional mass communication organising strategies such as letters and leaflets are associated with lower support for the union. This is certainly not a refutation of the “rank and file intensive strategy” recommended by Bronfenbrenner (1997), however, since the more personal, one-on-one intensive strategies of House Calls and Small Group Meetings are by far the most successful strategies in both sets of estimates. It simply means that, in Ontario, the traditional mass communication method of sending letters also increased support for the union. In contrast to employer strategies, the three most active union categories do not distinguish themselves as effective for winning campaigns or increasing support in representation elections. The relatively poor performance of Rank & File Organisers is noteworthy since it is another tactic often suggested as important for organising in the current more hostile environment. Bronfenbrenner (1997), however, also reports that it has an insignificant effect; although she uses a different, narrower, definition. Finally, the Offers Education category is the worst behaviour category for the union in both regressions and it is significantly worse than most

Union and Employer Tactics in Ontario Organising Campaigns

183

of the other union categories in the proportion of votes regression. Again, this is surprising since it is a very active category with lots of house calls and other personal communication methods. However the result may not be robust as it is a small category dominated by one union.

Outcomes for Two Combinations of Organiser and Employer Strategies Table 7 shows the estimated effects of the different strategies, adjusting for the strategy chosen by the other side and the factors measured by the control variables. Given that the most active campaigns by one side tend to be matched against the most active campaigns by the other side, it is interesting to estimate the net effect of both sides using their most active strategies. To accomplish this, campaigns where the employer uses Do Everything and the union organiser uses The Maximum or Offers Education are identified by a dummy variable and the tobit and logit regressions shown in Table 7 are re-estimated with the new dummy variable included. The two union categories are combined as in the fourth section to provide more observations for the category. The numerical estimates are not reported, but they show a large and statistically significant decrease in the probability that a campaign will be successful when both sides use their most active strategies. The estimates also show a decrease in the proportion of votes in favour of the union, but the effect is estimated to be smaller and not statistically significant. These results are consistent with the coefficient estimates reported in Table 7. The negative impacts of the employer’s Do Everything strategy are estimated to be much larger than the effects of the union’s The Maximum or Offers Education strategies.11 The net effect of both sides using their most effective strategies; namely, direct personal communication with the workers, is also estimated. A dummy variable identifying campaigns where the employer uses Just Communication and the union organiser uses House Calls or Rank & File Organisers is defined and included in the regressions shown in Table 7. In this case the effects of the union and employer strategies largely offset one another and there is no net effect, as would be expected given the Table 7 estimates.12

Control Variables Control variables are included in both regressions to capture the effects of other factors on support for the union in the representation election and the outcome of the organising drive. The only variable that is statistically significant in both

184

FELICE F. MARTINELLO AND CHARLOTTE YATES

regressions is a dummy variable indicating whether the union committed to the campaign because of strong worker interest. This variable provides a measure of the initial level of worker support for the union, before any tactics were employed by the union or employer. Not surprisingly, strong initial worker interest is associated with a higher proportion of votes in favour of the union and a higher probability that the campaign is successful. Another control dummy variable shows that when the union commits to the organising drive because they also represent workers in competing workplaces, the organizing drive is more likely to be successful. Support for the union in representation elections also increases in this case but the effect is not statistically significant. The variables provide no information as to why this occurs, but the competing unionised workplace may provide a comparison group that the organiser can point to and help resolve workers’ uncertainty about what the effects of union representation would be. There is also evidence that support for the union is lower in the manufacturing industry and that campaigns are more successful in bargaining units with higher percentages of females, but few other significant results. The gender and experience of the organiser, who initiates first contact (the organiser or the workers), whether an international union is involved, and other characteristics of the bargaining unit (e.g. size, percentage part-time, whether there was a previous organising drive, and whether there are more than 20% non-English speakers) all have statistically insignificant effects in both regressions when the union and employers strategies and the other control variables are included.

CONCLUSIONS Cluster analysis shows that there are distinct models or categories of union and employer behaviour. The categories are differentiated by their level of activity and the nature and content of their communication with workers. The analysis also shows, however, that even with these clusters of campaigns with similar tactics there is considerable diversity in union and management behaviours. Each category of behaviour includes a relatively small number of campaigns and the different categories display a very wide range of behaviours. In addition there are 20% or 30% of organising campaigns which do not fit into any of the union or employer categories, respectively. A second conclusion is that union and employer strategies vary with the characteristics of the bargaining unit, the industry, the type of union and the strategy category chosen by the other side. Union and employer strategies are not independent of one another, and notwithstanding the message of diversity

Union and Employer Tactics in Ontario Organising Campaigns

185

above, a more active campaign by union organisers tends to be met with more active resistance from employers and vice versa. Less aggressive, more “polite” campaigns by both unions and employers are more likely to be found in the health sector and in those workplaces with higher proportions of female employees. More active strategies tend to be used in larger bargaining units, and are more likely to be employed in the manufacturing sector. The finding about larger bargaining units confirms a proposition offered by Farber (2001, p. 345) that “it is possible that large employers fight more strongly against union organisation” in part because of the greater stakes for both parties in winning or losing such a contest. While our data analysis does not offer reasons for why more active strategies are used more often in manufacturing, one possible explanation may lie in this industry’s highly competitive nature. Marginal changes in wage rates, which are the likely result of unionisation, can have a significant impact on profitability and competitiveness in the manufacturing industry. Thus, there is good reason for employers to fight hard against an organising campaign. Further, manufacturing establishments are more likely to be large workplaces, which also tend to be the subject of more intense campaigns by both employers and unions. Union and employer strategies also tend to vary with the proportion of females in the bargaining unit, although this is more significant for union strategies. There are a number of possible reasons for this differential which have been explored by Yates (2001). Attitudes held by unions towards women workers regarding their militancy, desire for unionisation and so on may affect the strategies that unions use when organising workplaces with significant numbers of women. Thirdly, given the multiple possible endpoints for organising drives and the evidence that firm and union tactics are endogenous, simple regression estimates of the effects of tactics on certification success should be regarded cautiously.13 More work is required to model the interactions between union and employer tactics and the certification process.14 Keeping this caveat in mind, however, our regression results relating union and employer strategies to campaign success in Ontario are generally consistent with findings based on U.S. data. On the employer’s side they show that more active anti-union campaigns are effective, as is direct communication with the workers. On the union side they show that one-on-one, more personal methods of communication are the most effective. At the same time, our finding on the impact of the Letters category tends to differ from the results of Bronfenbrenner (1997) and Peterson, Lee and Finnegan (1992). When both sides use their most active strategies, the employer’s strategy is estimated to be the more effective. When both sides use direct personal communication with the workers, the effects of the union and employer strategies basically cancel one another out. In general, the findings confirm the validity of conclusions based on research done in the U.S. that certain union strategies work better than others, and offer some

186

FELICE F. MARTINELLO AND CHARLOTTE YATES

evidence in support of the proponents of transferring the American organising model to other countries.

NOTES 1. We removed public sector campaigns from the sample due to their distorting effect on the data analysis. Employers in the public sector do very little to oppose unionisation. The effect of this “do nothing” employer behaviour tended to overwhelm the statistical differences in union and employer strategies across the campaigns in the private sector. 2. The descriptive statistics for the campaigns whose certification applications were disposed of by the OLRB were obtained from the Board’s annual reports for the two fiscal years covering May 1996 to April 1998. The fiscal year dates do not match the months of the survey exactly, but it is reasonable to expect that the campaigns in the few months before and after the overlapping period would not be systematically different. 3. To be more precise, hierarchical, agglomerative, average linkage, cluster analysis is used. In the first round the analysis calculates, for every pair of campaigns, a measure of the similarity of the tactics used in the two campaigns. The pair of campaigns whose tactics are most similar are then put together into a cluster. This procedure is then repeated many times, with clusters of campaigns now treated as single cases. When clusters of campaigns are involved, the average similarity over all of the campaigns in the cluster is used to calculate similarity. As the procedure is repeated single campaigns are merged into new clusters, single campaigns are put into existing clusters, or existing clusters are merged; all according to highest value of average similarity in that round. Consequently, after each round, there are fewer clusters, the clusters contain more campaigns, and the tactics used in those campaigns are more heterogeneous; since less and less similar campaigns or clusters of campaigns have been merged. If allowed to continue, the process eventually produces just one cluster consisting of all of the campaigns. Thus the researcher must stop the agglomerating process at an appropriate point. This is equivalent to deciding (simultaneously) how many clusters there will be, how large the clusters will be, and how similar the campaign tactics will be in each cluster. Johnson and Wichern (1998) or most texts on multivariate analysis provide a detailed description of cluster analysis. 4. The test for industry concentrations uses the proportion of campaigns in that industry, and that category (separately), to calculate the number of campaigns expected in that industry/category combination if industry and category are independent. The binomial distribution is then used to calculate a p-value for the test, which is the probability of the observing the observed frequency (or a more extreme frequency) given the expected frequency. 5. To be more specific, the similarity of tactics is measured by the number positive matches as a proportion of the number of positive matches plus the number of mismatches (the tactic is used in one campaign but not another). This measure is not used for union tactics because it provides worse discrimination. In that case the large numbers of positive matches for the frequently used union tactics dominate the variation over the other tactics and one large group with fringe groups result. 6. See Everitt (1992) for an exposition of the analysis of contingency tables.

Union and Employer Tactics in Ontario Organising Campaigns

187

7. The expected frequency under the null hypothesis is calculated from the overall sizes of the categories and it is just the product of their respective proportions. 8. The Pearson chi-squared test statistic is 47.3 with 24 degrees of freedom so the independence hypothesis is overwhelmingly rejected. Tau-b is 0.258 with an estimated asymptotic standard error of 0.056 while gamma is 0.328 with an asymptotic standard error of 0.07. Both are overwhelmingly statistically different from zero and there is a positive association between the behaviour categories as they are ordered. Virtually the same results obtain if the unclassified campaigns are included as separate categories in the contingency tables, and the conclusions are unchanged. 9. A more common method of estimation for the proportion of votes is weighted least squares, with or without a logit transformation of the proportion (e.g. Bronfenbrenner, 1997; Dickens et al., 1987; Maranto & Fiorito, 1987; Reed, 1989). This method is problematic when some elections are unanimous in favour or against the union because the logit transformation is not defined and the weights on some organising drives may be zero, or even negative, depending on how they are calculated. These problems can be eliminated by using the maximum likelihood version of the weighted least squares procedure. See Greene (2000, pp. 834-837) for a description of the econometrics involved. A more important issue is that weighted least squares (and its maximum likelihood equivalent) treat each worker’s vote in the representation election as a separate and independent observation on the effects of the right hand side variables. A binomial distribution for votes is assumed so the results from larger elections are assigned lower variances and higher weights in the regression than the results from smaller elections. The differences in variances and weights are linear in the number of votes cast so, given the large range of bargaining units sizes, there are huge differences in the variances and weights assigned to specific organising campaigns. In addition, the logit transformation is highly non-linear. It produces disproportionately large (or small) values for the dependent variable when the proportion of votes is close to one (or zero). This means that the coefficient estimates are influenced much more by campaigns with proportions of votes close to zero or one. In practice, this all leads to coefficient estimates that are implausibly precise (i.e. their standard errors are implausibly low) and too sensitive to outcomes in large bargaining units and outcomes with very high or low proportions of votes. It seems questionable that individual votes in a representation election are independent trials that provide, one-for-one, separate and equal pieces of information on the effects of the tactics and the other control variables. The truth likely lies somewhere between the two extremes of tobit and weighted least squares. The outcome from larger representation elections must convey more information than a smaller election and the tobit estimation, by treating all elections equally, ignores this. Weighted least squares, on the other hand, likely overestimates the extra information conveyed by larger or near unanimous elections and, therefore, places too much weigh on them. We choose to err on the more conservative side and report results from a tobit estimation. 10. There are only 221 observations (rather than 237) in the logit won/lost regression because of missing observations in the control variables. The 206 campaigns in the proportion of votes regression is 221 minus the 15 campaigns where there was no election or the election data was unavailable. The 15 campaigns that are excluded from the tobit regression, but included in the logit regression, could explain some of the differences between the two sets of results reported in the text. It turns out, however, that the differences

188

FELICE F. MARTINELLO AND CHARLOTTE YATES

between the regression samples explain only a little of the differences between the logit and tobit estimates. This occurs because there are only a small number of campaigns withdrawn or dismissed without a vote, and they are spread fairly evenly across the union and employer behaviour categories. A “hot deck” procedure was used to impute the missing observations in the control variables in both regressions. When the regressions are run with the imputed data some of the coefficient estimates are different and many of the estimated t-statistics are larger but the same basic results obtain. Given that the results are so similar, another layer of statistical complexity seems unwarranted and the results with no imputation are reported. 11. The numerical estimates from the re-estimated tobit and logit equations are available from the authors. The net effect of the employer using the Do Everything strategy and the union using only The Maximum is also estimated. The results are quite similar, with the exception that the negative impact on the probability of campaign success is only significant at the 16% level. 12. Again, the numerical estimates from the re-estimated tobit and logit equations are available from the authors. Identical results are obtained if the net effect of the employer using Just Communication and the union using House Calls is estimated. This occurs because there are no campaigns where the combination of Just Communication and Rank & File Organisers is employed. 13. An organising drive could end: (a) with no certification application filed; (b) with the application withdrawn by the union or dismissed by the OLRB before the representation vote is held; or (c) with the application granted or dismissed after the representation vote. 14. Two important econometric challenges should be addressed in future work. First, the analysis shows that union and management choices of tactics are not exogenous. Thus they should be modelled as endogenous choice variables that depend on what the other side is doing, the characteristics of the bargaining unit, and other variables. If not, the estimates, like most found in the literature, suffer from simultaneous equation bias that affects both the value of the estimates and the estimates of their statistical significance. Freeman and Kleiner (1990) provide an exposition of the sources and size of some simultaneous equation biases as well as an instrumental variable correction. In general, instrumental variable estimation can adjust for the endogeniety problem, but it is difficult to implement because there are too few instruments for the union and employer tactics, or in this case, their strategies. Sample selection is the second challenge. It is not as well documented in the literature, but likely more serious. Tactic choices by unions and employers cause campaigns to end at different stages in the certification process and the different endpoints affect their inclusion in or exclusion from the sample. For example, an employer tactic may be very effective and cause organising drives to end with the application dismissed by the Board or withdrawn by the union before the representation election is held. These campaigns would be excluded from any study using election results and so the effect of that very effective tactic would be seriously underestimated.

ACKNOWLEDGMENTS The authors thank Greg Auton, Bilal Abu Zeid, and Scott Devereaux for research assistance and acknowledge financial support from SSHRC grants 410–2001–1232

Union and Employer Tactics in Ontario Organising Campaigns

189

(Martinello) and 410–98–1302 (Yates). Thanks also to the union organisers who made collection of the survey data possible.

REFERENCES Bentham, K. (2002). Employer resistance to union certification: A study of eight Canadian jurisdictions. Relations Industrielles/Industrial Relations, 57(1), 159–185. Bronfenbrenner, K. (1997). The role of union strategies in NLRB certification elections. Industrial and Labor Relations Review, 50(2), 195–212. Bronfenbrenner, K., & Juravich, T. (1998). It takes more than house calls: Organizing to win with a comprehensive union-building strategy. In: K. Bronfenbrenner, S. Friedman, R. Hurd, R. Seeber & R. Oswald (Eds), Organizing to Win: New Research on Union Strategies (pp. 19–36). Ithaca, NY: ILR Press. Card, D., & Freeman, R. (Eds) (1993). Small differences that matter: Labor Markets and income maintenance in Canada and the United States. Chicago: University of Chicago Press. Carter, B. (2000). Adoption of the organising model in British trade unions: Some evidence from manufacturing, science and finance (MSF). Work, Employment and Society, 14(1), 117–136. Crain, M. (1994). Gender and union organizing. Industrial and Labor Relations Review, 47(2), 227–248. Dickens, W., Wholey, D., & Robinson, J. (1987). Correlates of union support in NLRB elections. Industrial Relations, 26(3), 240–252. Everitt, B. S. (1992). The Analysis of Contingency Tables (2nd ed.). New York: Chapman and Hall. Fairbrother, P., & Yates, C. (2003). Trade union in renewal: A comparative study. London, United Kingdom: Continuum Publishers. Farber, H. (2001). Union success in representation elections: Why does unit size matter? Industrial and Labor Relations Review, 54(2), 329–348. Farber, H., & Western, B. (2001). Accounting for the decline of unions in the private sector, 1973-1998. Journal of Labor Research, 22(3), 459–485. Freeman, R. (1986). The effect of the union wage differential on management opposition and union organizing success. American Economic Review, 76(2), 92–96. Freeman, R., & Kleiner, M. (1990). Employer behavior in the face of union organizing drives. Industrial and Labor Relations Review, 43(4), 351–365. Greene, W. (2000). Econometric analysis (4th ed.). Upper Saddle River, NJ: Prentice Hall. Heery, E. (1996). The new new unionism. In: I. Bearwell (Ed.), Contemporary Industrial Relations: A Critical Analysis (pp. 175–202). Oxford: Oxford University Press. Heery, E., Simms, M., Delbridge, R., Salmon, J., & Simpson, D. (2000). Organizing comes to the UK. Employee Relations, 22(1&2), 38–58. Heneman, H., & Sandver, M. (1983). Predicting the outcome of union certification elections. Industrial and Labor Relations Review, 36(4), 537–559. Johnson, R., & Wichern, D. (1998). Applied multivariate statistical analysis (4th ed.). Upper Saddle River, New Jersey: Prentice Hall. Johnson, S. (2002). Canadian union density 1980–1998 and prospects for the future: An empirical investigation. Canadian Public Policy, 28(3), 333–351. Lawler, J. (1990). Unionization and deunionization. Columbia, South Carolina, USA: University of South Carolina.

190

FELICE F. MARTINELLO AND CHARLOTTE YATES

Maranto, C., & Fiorito, J. (1987). The effect of union characteristics on the outcome of NLRB certification elections. Industrial and Labor Relations Review, 40(2), 225–240. Martinello, F. (2000). Mr. Harris, Mr. Rae and union activity in Ontario. Canadian Public Policy, 26(1), 17–34. Milkman, R. (2000). Organizing immigrants: The challenge for unions in contemporary California. Ithaca: ILR Press. Oxenbridge, S. (1997). Organizing strategies and organizing reform in New Zealand service sector unions. Labor Studies Journal, 22(3), 3–27. Peterson, R., Lee, T., & Finnegan, B. (1992). Strategies and tactics in union organizing campaigns. Industrial Relations, 31(2), 370–381. Reed, T. (1989). Do union organizers matter? Individual characteristics and representation election outcomes. Industrial and Labor Relations Review, 43(1), 103–119. Riddell, C. (2001). Union suppression and certification success. Canadian Journal of Economics, 34(2), 396–410. Rose, J., & Chaison, G. (1990). New measures of union organizing effectiveness. Industrial Relations, 29(3), 457–468. Thomason, T. (1994). The effect of accelerated certification procedures on union organising success in Ontario. Industrial and Labor Relations Review, 47(2), 207–226. Thomason, T., & Pozzebon, S. (1998). Managerial opposition to union certification in Quebec and Ontario. Relations Industrielles/Industrial Relations, 53(4), 750–771. Yates, C. (2000). Staying the decline in union membership: Union organizing in Ontario, 1985–1999. Relations Industrielles/Industrial Relations, 55(4), 640–674. Yates, C. (2001). Debunking the myths about organizing women into unions. Paper prepared for presentation at Annual Meeting of Canadian Industrial Relations Association, Quebec City, Quebec (May).

LABOR’S “LAST STAND” IN NATIONAL POLITICS? THE CAMPAIGN FOR STRIKER REPLACEMENT LEGISLATION, 1990–1994 John Logan The threat of permanent replacement hangs like a ‘Sword of Damocles’ partially paralyzing the union movement. Union Attorney, 1991.1

INTRODUCTION The campaign for striker replacement legislation, which began in the late 1980s and had effectively ended by the mid-1990s, was the most important political battle over labor legislation since the defeat of the Labor Law Reform Bill in 1978. Striker replacement was the AFL-CIO’s top legislative priority in the early 1990s and, coming quickly after the passage of NAFTA, which labor had opposed, the defeat of its campaign solidified organized labor’s reputation for failure in legislative battles. As yet, however, the political campaign for striker replacement legislation has attracted surprisingly little attention from industrial relations scholars. This paper provides the first detailed scholarly analysis of the AFL-CIO’s campaign to gain legislation proscribing the use of permanent replacements. It is based on research in the records of the AFL-CIO (especially those of the Advances in Industrial and Labor Relations Advances in Industrial and Labor Relations, Volume 13, 191–243 © 2004 Published by Elsevier Ltd. ISSN: 0742-6186/doi:10.1016/S0742-6186(04)13007-2

191

192

JOHN LOGAN

Industrial Union Department, which spearheaded labor’s campaign), management associations (especially the umbrella organization, the Alliance to Keep Americans Working, a coalition group that led the employer campaign against the legislation), government agencies (the NLRB, Department of Labor, and the Federal Mediation and Conciliation Service), and other contemporary sources (newspapers, the Daily Labor Report, union and trade association journals, and interviews with union, management and government participants). The paper is divided into two main sections. Section One provides a detailed narrative of the AFL-CIO’s two campaigns for striker replacement legislation – the first in 1991–1992, the second in 1993–1994. It examines not only organized labor’s campaign, but also the business lobby’s campaign against the proposed legislation, the role played by key politicians, and the main factors behind the AFL-CIO’s failure to win the necessary 60 Senate votes in both 1992 and 1994. The second section analyzes the key issues that dominated the debate on striker replacement legislation and examines organized labor’s failure to refute decisively accusations from the business lobby that it was promoting special interest legislation that would promote strikes, undermine investment, and destroy American jobs. The campaign’s significance, however, extends beyond the critical issue of striker replacement policy. It illuminates organized labor’s on-going problems when seeking labor law reform: the resolute and cohesive opposition of the Republican Party; often lukewarm and ineffectual support from a fractured Democratic party; the AFL-CIO’s failure to persuade non-unionists to consider labor rights as “basic human rights” rather than as special interest legislation; a union movement divided on its labor law reform priorities and distracted by other important policy issues; and, most importantly, the power of the business lobby to block pro-union reforms when it is determined and united in its opposition. The AFL-CIO had hoped that the striker replacement campaign would serve as a stage-setter for comprehensive reform of the NLRA, but ultimately served only to illustrate the AFL-CIO’s inability to win favorable legislation over vigorous opposition from the business community.

THE POLITICAL CAMPAIGNS The First Campaign for Striker Replacement Legislation, 1991–1992 The striker replacement issue did not first emerge in the 1990s. Employers had enjoyed the legal right to hire permanent replacements during economic strikes since the Supreme Court’s 1938 Mackay Radio decision, but until the 1970s, the use of permanent replacements was relatively rare. In the intervening

Labor’s “Last Stand” in National Politics?

193

decades, a number of developments enabled employers to operate more easily during strikes – technological innovations, the increasing number of multi-plant enterprises, and the relative ease with which firms could recruit and train employees willing to work as replacements. The external environment of labor-management relations also changed dramatically in the 1970s and 1980s. As a result of the growing union/non-union wage differential, heightened international and domestic non-union competition, and the increasing importance of workplace flexibility, employers started to resist unions more aggressively at the bargaining table and during organizing drives. In the 1980s and 1990s, one particularly overt manifestation of this intensification of employer opposition to unionization was the use, and threatened use, of permanent replacements. Nor were the striker replacement campaigns of the 1990s organized labor’s first attempts to gain legislation outlawing permanent replacements.2 The AFL-CIO had sought to change the law on striker replacements during its campaign for comprehensive labor law reform in 1977–1978. The initial version of the Labor Law Reform Bill had included a provision stating that permanent replacements could not be used in strikes involving disputes over the negotiation of first or second contracts, and that replacements could not vote in decertification elections. But the Carter administration rejected the provision as politically unfeasible and the AFL-CIO effectively dropped the striker replacement issue for another decade.3 By the late 1980s, as a result of their belief that a growing number of employers were retaining strikers at the end of strikes, unions started to pressure the AFLCIO to lobby for legislation prohibiting the use of permanent replacements.4 Several large unions, including the Teamsters, Paperworkers, Electrical Workers, Steelworkers and Machinists, believed that the campaign to enact striker replacement legislation should be the AFL-CIO’s top legislative priority. The International Union of Electronic Workers stated that striker replacement was an issue of “life or death to the labor movement,” while the Chemical Workers Union warned that, with the increasing use of permanent replacements, “the entire labor movement is now facing an insidious threat from employers, the likes of which we’ve never seen before.”5 United Mine Workers President Richard Trumka warned that without a change in the law on permanent replacements, unions would “almost certainly lose their only true weapon – the right to strike. Without that weapon, organized labor in American will soon cease to exist . . . We must draw the line now and make a stand if we are to survive in the twenty-first century.”6 The issue also appeared to enjoy widespread support among ordinary union members. The Washington Post reported that striker replacement “appears to be the one exclusively labor issue to galvanize union rank-and-file support in more than a decade.”7

194

JOHN LOGAN

In response to this increasing pressure for action, the AFL-CIO once again took up the issue of permanent replacement in the late 1980s. Agreeing with affiliates that the economic strike had now been “turned on its head” and converted into “a tool to destroy unionism,” the AFL-CIO made plans to introduce a bill outlawing permanent replacements. At the urging of the president of the paper workers’ union, representative Joseph Brennan of Maine (site of a high-profile strike involving permanent replacements at the International Paper Company) introduced the first bill seeking to restrict the use of replacements, which would have banned permanent replacements during the first 10 weeks of a strike, on May 10, 1988. But that bill, which was not supported by the AFL-CIO, failed to make it out of committee.8 When organized labor started its striker replacement campaign, the chances of victory appeared slim, but the AFL-CIO drew inspiration from the campaign for Plant Closing legislation in the 1980s. In the early 1980s, it recounted, winning passage of a plant-closing bill “appeared to be a dream,” but by 1988, support for the bill “was so strong that it passed both houses overwhelmingly, immune to a veto.” Thus, the AFL-CIO concluded that, while winning striker replacement legislation would “not be easy,” it should by no means be considered “unrealistic.”9 Coordinated by the Industrial Union Department (IUD), the striker replacement campaign was intended to serve as a “stage-setter” for the long-range goal of comprehensive labor law reform, which would include substantial revision of the NLRA’s certification procedures. The International Brotherhood of Teamsters (IBT) believed that the enactment of a striker replacement law would send a “first positive signal” to the American public after years of sustained government and employer attacks on collective bargaining and “help to create an atmosphere” conducive to broader labor law reform.10 Opponents of the legislation also warned that the striker replacement bill would merely serve as a precursor for the AFL-CIO’s larger goal of “radical” overhaul of the NLRA. One prominent opponent of the bill, Senator Orren Hatch of Utah, warned that if Congress were to pass the strike bill, the AFL-CIO would not stop “until they get all of those heinous aspects of labor law reform that will . . . unionize America against its will.”11 The campaign for striker replacement legislation started to gain public notice when several union newspapers dramatically stepped up their coverage of the issue shortly before the House and Senate Subcommittees on Labor held oversight hearings in early 1990. In August 1990, the AFL-CIO formed a special committee to raise public awareness of permanent replacement, the Committee on Workplace Fairness, whose members included former Labor Secretaries, Willard Wirtz and Ray Marshall.12 In late 1990, Representative William Clay (D-Missouri) and Senator Howard Metzenbaum (D-Ohio) prepared to introduce into the House and

Labor’s “Last Stand” in National Politics?

195

Senate bills proscribing the use of permanent replacements. The AFL-CIO believed that the timing of its campaign had initially caught the business lobby off guard. It reported that employers were “not yet geared up to oppose the legislation. We are there first.”13 Support for striker replacement legislation quickly became a “litmus test” for union assistance in political campaigns. In 1990, the IBT announced that it would not contribute to the campaigns of candidates who refused to support reform on permanent replacements, and the following year, the Machinists followed suit.14 The Communication Workers of America (CWA) also ended support for candidates who refused to endorse the strike bill and withdrew an invitation to Labor Secretary Lynn Martin to address the union’s convention after she had announced her opposition. And at the AFL-CIO convention in 1991, President Lane Kirkland told delegates that commitment to a ban on permanent replacements would be one of organized labor’s “acid tests” when it came to endorsing a candidate for the next presidential election. Indeed, Bill Clinton’s support for the striker replacement bill and the opposition of his opponent for the Democratic nomination, Massachusetts Senator Paul Tsongas, was one factor contributing to the AFL-CIO’s decision to endorse the Arkansas governor in May 1992.15 However, unions were not alone in their belief in the critical importance of the striker replacement question. For business organizations, too, striker replacement was a “no compromise” issue of paramount importance. By the early 1990s, defeat of the striker replacement bill had become one of the top legislative priorities for several national employer and anti-union organizations, including the Chamber of Commerce (CoC), National Association of Manufacturers (NAM), and National Right to Work Committee (NRTWC). The CoC attacked the striker replacement bill as one of the most “frightening legislative initiatives to emerge from Congress in many years” and claimed that its 215,000 members were “unanimous in their unbending opposition” to the legislation.16 The NAM predicted that the campaign over striker replacement legislation “would not be a quiet battle,” while the NRTWC lambasted the “push-button strike bill” as “pure economic poison” that was designed to make “calling and winning a strike as easy as pushing a button.”17 The employer campaign against the striker replacement bill was spearheaded by the “Alliance to Keep Americans Working,” an umbrella organization representing more than 300 member companies and trade associations. Members of the Alliance included the CoC, NAM, National Federation of Independent Business (NFIB), Society for Human Resource Management, and Labor Policy Association (LPA), American Mining Congress, American Trucking Association, Printing Industries of America, and Air Transport Association, Caterpillar, American Airlines, Dow Chemical Company, Phelps Dodge Corporation, and USX Corporation.18 The CoC also formed a separate organization to conduct grassroots campaigns in the

196

JOHN LOGAN

districts and states of key House and Senate members, titled GAIN (Grassroots Action Information Network). It quickly became apparent that the business community intended to oppose the striker replacement bill with the same vehemence that it had fought the Labor Law Reform Bill in the late 1970s. Business representatives were quietly confident of their ability to mount an effective counter-campaign, with one management publication reporting on the existence of “much behind the scenes activity” – the NAM was putting together an “industry lobbying coalition” that would be able to mobilize against the bill at “a moment’s notice.”19 The AFL-CIO also reported that the business lobby was “gearing up to derail” labor’s campaign and it adopted several strategies designed to “soften the impact” of the anticipated attack, including seeking out friendly employers who supported the striker replacement bill. As the campaign progressed and employer opposition intensified, the IUD warned of the pressing need to identify supportive employers because the “tide of business pressure in the other direction is rising quickly.”20 The AFL-CIO succeeded in finding a handful of employers who endorsed the striker replacement bill, such as Bernard Rapoport, President of the American Income Life Insurance Company, who believed that the practice of permanent replacement had “severely undermined” the right to strike.21 But the overwhelming majority of American employers did not share Rapoport’s position and, for the most part, business opposition was both vigorous and resolute. The initial striker replacement bill proposed to revise only the NLRA, but after complaints from the Machinists and Airline Pilots Association (ALPA), the AFL-CIO arranged for the bill to be considered in Congress as part of a package that would also ban permanent replacements under the Railway Labor Act (RLA). In 1988, a Supreme Court decision had extended the Mackay doctrine to cover employees under the Railway Labor Act22 (primarily railroad and airline workers) and ALPA stressed that three of the most egregious incidents of permanent replacement in recent years had occurred at Continental, TWA, and Eastern Airlines. Since supporters of the strike bill repeatedly cited these disputes as examples of why reform was needed, ALPA could not understand why its industry had been “left out” of the proposed legislation.23 Likewise, the Machinists stated that revision of the RLA was essential because these strikes had demonstrated that what once were “economic disputes of short duration” had now become “desperate life and death struggles” because employers had been “permitted to utilize the poison gas of permanent replacement.” Not surprisingly, airline industry representatives rejected the notion that legislation was needed to deal with the issue of permanent replacement. The Air Transport Association stressed that the worst examples of permanent replacement in the airline industry had involved carriers who were no longer in business or now under different ownership and argued that

Labor’s “Last Stand” in National Politics?

197

the prohibition of permanent replacements would provide “temptation” for unions to make “excessive bargaining demands,” ultimately forcing more airline carriers into bankruptcy.24 Other unions wanted the AFL-CIO to follow the lead of the Canadian province of Quebec, by lobbying for a ban on both temporary and permanent replacements, especially after an NLRB decision that had allowed employers to hire temporary replacements during offensive lockouts.25 The American Federation of Grain Millers, which was involved in a six-year lock out with BASF, complained that the proposed legislation did not address a “greater threat to employees . . . when the board gave employers the equivalent of an atomic bomb” by establishing their right to hire temporary replacements during a lockout. But the AFL-CIO rejected its proposal to outlaw temporary replacements. While the IUD agreed that the recruitment of temporary replacements could weaken a union’s effort to secure a “fair settlement,” it explained that if winning legislation to ban permanent replacements would be “extremely difficult,” banning temporary replacements would be “impossible.”26 In June 1991, the AFL-CIO launched a $500,000 advertising campaign in support of the striker replacement bill.27 From the start, labor’s campaign concentrated on certain “swing” Congressional members, thereby hoping for “a clear payoff in an otherwise uncertain vote.” While the AFL-CIO distributed news releases and educational materials on a nation-wide basis, its “intensive activities” and the bulk of its funds were “carefully targeted” at key members of Congress.28 Initially the campaign had a strong southern focus (with the only new AFL-CIO staff member hired on the campaign operating out of Nashville) because of the vital importance of getting southern Democrats in Congress to co-sponsor the controversial legislation. The AFL-CIO immediately recognized the importance of enlisting the support of “key non-labor allies in the community.”29 The IUD worried that non-unionists would view its campaign as “purely a union issue” and believed that support from non-labor organizations would be “critically important,” while that of traditional labor allies would have only “limited effect.” IUD President Howard Samuel concluded that success would depend on active support that went “far beyond” the ranks of organized labor: “If the campaign develops as a broad community issue, rather than as a ‘narrow’ union issue, our chances of success will be considerably increased.”30 Thus, in an effort to create a broad-based grassroots campaign, the AFL-CIO sought alliances with religious, civil rights, women’s, academic, and community organizations.31 Support from civil rights organizations constituted a key component of the AFL-CIO’s grassroots strategy. Since minority workers “benefit enormously” from union membership, the IUD believed that civil rights organizations “should

198

JOHN LOGAN

recognize the importance of passing this bill in the interests of their constituents.”32 Companies with a high proportion of black or Latino employees that had recruited permanent replacements included Greyhound, Delta Pride Catfish, Diamond Walnut Growers and the Frontier Hotel. Several civil rights organizations immediately signed up for the striker replacement campaign. The National Association for the Advancement of Colored People (NAACP) announced that it “would be pleased to be associated” with labor’s campaign because permanent replacement represented a “severe threat” to the livelihood of black workers, while the Black Leadership Committee for Workplace Fairness (whose members included the Southern Christian Leadership Conference and the A. Philip Randolph Institute) believed that the practice of permanent replacement had “disproportionately and adversely affected” black workers.33 In all of its targeted states, the AFL-CIO established broad grassroots campaigns, for which one union had overall responsibility, but well-placed local organizations played leading roles in each state. In Georgia, for example, the strong Atlanta Labor Council and Jobs with Justice led the campaign, while in Minnesota, the Evangelical Lutheran Church, Catholic Church and Jobs with Justice played central roles. The ultimate objective of these grassroots campaigns, however, was not to “impress the community or build general support,” but to influence the votes of key lawmakers.34 The best practice model for all of the state campaigns was that run by mineworkers’ president Richard Trumka in Alabama. The group of over 50 religious, civil rights, women’s, community and social justice organizations supporting the striker replacement campaign in Alabama included the following: Southern Christian Leadership Conference, NAACP, National Organization for Women, Mobile Catholic Social Services, Southern Rainbow Education Project, Alabama Coalition Against Hunger, Alabama Civil Liberties Union, People for Peace, League of Women Voters of Alabama, and Jobs with Justice.35 The first objective for the grassroots campaigns was to win a substantial majority in the House in order to gain the momentum necessary to get the bill through the Senate.36 On July 15, 1991, the House voted 247–182 to ban permanent replacements, but fell 43 votes short of the 290 votes required to override a threatened presidential veto. Sixteen Republicans broke ranks and voted for the bill, while 33 Democrats voted against it. The AFL-CIO welcomed the vote as a “pretty good” result, but opponents claimed that it was a “hollow victory” for organized labor that indicated that the strike bill was “going nowhere.” Unions had viewed the House vote as a test of how much they would have to amend the bill to get it through the Senate and, although gaining a few more votes than expected, they privately admitted that the final tally meant that significant compromise was inevitable if the bill were to have any chance of passing the upper chamber.37

Labor’s “Last Stand” in National Politics?

199

Opponents of striker replacement had no doubt that the business lobby was responsible for the substantial House vote against the striker replacement bill. Business organizations had initially set themselves a target of winning 146 “no” votes – the number required to demonstrate veto strength – but some in their ranks even considered that a long shot; thus, they heralded the 182 votes against the bill as an important victory. The International Paper Co. (IP) boasted that the House vote had demonstrated the ability of the business community “to affect legislative decisions when we take the time and effort to get actively involved.” IP believed that were it not for the business campaign, the striker replacement bill would have swept through the House with enough support to override a Presidential veto. Instead, the day before the vote, over 50 House Democrats had urged the party leadership to pull the bill from the calendar because, IP claimed, they now understood that the bill would destroy competitiveness and undermine investment. Congressional opponents of the bill also praised employers’ efforts. The leader of the Republican opposition in the House, Congressman Mickey Edwards of Oklahoma, applauded the campaign against the strike bill as “the best effort I have seen the business community undertake” since its defeat of the Labor Law Reform Bill in 1978.38 But the business lobby did not act alone. Employers also received external help in coordinating their campaign from the White House. That the Bush Administration opposed the striker replacement bill and made known its intention to veto it was no surprise. But supporters of the bill were surprised by how early in the campaign the White House had announced its opposition and even more taken aback by the unequivocal character of that opposition, which appeared to leave no possibility of compromise. In March 1991, Labor Secretary Lynn Martin told the House Subcommittee on Labor-Management Relations that she would recommend a veto because the bill provided “little incentive” for unions to moderate their bargaining demands and would eliminate a “major check on precipitous striking.”39 Union leaders found it “astounding” that Martin would recommend a veto so early in the legislative process, while the House sponsor of the bill, William Clay of Missouri, believed that the White House had made its position “crystal clear” – it had left “no room to negotiate.”40 Even so, in an election year, the AFL-CIO believed that President Bush could be pressured into signing the striker replacement bill. The IUD was unsure whether or not Bush would “in fact veto this bill when it reaches his desk; it will be a presidential election year and a veto of an issue of critical importance to working Americans is by no means certain, in spite of the advice of his senior advisors.” If the bill passed the Senate, the IUD intended to enlist the support of several religious leaders who had personal access to the President, including Edmond Browning of the Episcopal Church, John O’Connor, Cardinal Archbishop of New York, and the Rev. Billy Graham. It also hoped that pro-union clerics in the Catholic

200

JOHN LOGAN

Church might influence Labor Secretary Martin, who was a Roman Catholic. And if Bush did veto the bill, its supporters believed that the controversial action might undermine his electoral popularity. Iowa Senator Tom Harkin wanted the Senate to pass the bill then let the President “answer to the American people.”41 White House opposition to striker replacement did not end with its promise to veto the bill if it reached the President’s desk. Labor Secretary Martin also played a critical role in ensuring that business opposition was both vigorous and cohesive. When the bill was first introduced into the House, the NFIB had considered dropping its opposition if organized labor agreed to incorporate certain amendments that would make the legislation more palatable to small businesses. But Secretary Martin immediately contacted the NFIB and persuaded it of the necessity of defeating the strike bill in all its guises. NFIB chief lobbyist John Motley recalled that Martin had “got right on the phone” to him: She was saying that you can’t fix this bill – the analysis that these amendments will help small businesses is illusory because once it is passed, labor will have a tremendous organizing tool. If we stand any chance of defeating this bill, she said, we have to have a united business community.42

Motley praised Martin’s decisive role in “orchestrating” business opposition to the bill and stressed that he had “never gotten that [kind of personal involvement] from a Secretary before.” A pivotal member of the Alliance to Keep Americans Working, the NFIB consistently argued that a prohibition of permanent replacements would have a deleterious impact on “labor intensive” small firms; if key employees struck and could not be replaced, these firm’s ability to survive would be “seriously jeopardized.” Shortly after labor’s Senate defeat, the NFIB announced that a survey of its membership had revealed that 81% opposed striker replacement legislation. By the time that the NFIB fell into line, defeating striker replacement legislation had become “an article of faith” for most employers, opposition to the bill was virtually “unshakable,” and the business community was more unified on striker replacement than it was on any other issue of labor policy.43 The AFL-CIO and its allies also faced robust opposition from the mainstream media, particularly in the form of newspaper editorials opposing the strike bill. Between 1990 and 1994, over 170 newspapers throughout the country published editorials opposing legislation banning permanent replacements. Newspapers that ran editorials attacking the striker replacement bill included traditionally liberal papers such as the New York Times and Washington Post, as well as the Detroit News, Chicago Tribune (which was generally considered a good barometer of the tone of editorials across the country) and Wall Street Journal.44 The Washington Post, which had itself hired permanent replacements during an acrimonious strike

Labor’s “Last Stand” in National Politics?

201

in the 1970s, argued that the current ambiguity of the law on economic strikers – that they could be permanently replaced but not sacked – was “less of a defect than a virtue.” Irate AFL-CIO officials suspected that the business lobby had cajoled the newspaper’s owners “not to come out for any compromise position” on the strike bill. Whatever the reason for its opposition, the fact that a bastion of the liberal establishment such as the Post had attacked labor’s favorite bill was used to great effect by opponents of the legislation.45 The Senate vote was always going to be a much tougher test for the bill’s supporters than was the House vote. After only 33 Senators signed up to cosponsor the legislation by summer 1991, Business Week reported that unions blamed this relative poor showing on a “tepid push” from the Democratic leadership in the Senate.46 It was increasingly apparent that several senior Democrats viewed organized labor’s policy issues as “losers” and as an embarrassment to the party; and they considered striker replacement a particularly divisive issue that they would rather not have to go on record as having endorsed and run the risk of antagonizing the business community, especially during an election year. As the Senate vote approached, the supporters of the bill recognized that they were several votes short of the super-majority of 60 required to invoke “cloture” (i.e. to end debate and bring the bill to a vote on the floor). In an eleventh hour attempt to win additional votes, the AFL-CIO agreed to a narrower, more qualified prohibition on permanent replacements proposed by Senator Bob Packwood (R-Oregon). The Packwood amendment provided for voluntary arbitration under certain circumstances in return for a ban on permanent replacements; it would have mandated government-supervised mediation in most major labor disputes through three-member fact-finding panels agreed to by both unions and employers. Most labor leaders welcomed the Packwood bill, believing that it now represented the best hope for getting the bill through the Senate.47 In addition to receiving the support of the AFL-CIO and many affiliated unions, the Packwood amendment was endorsed by five former Labor Secretaries (Willard Wirtz, John Dunlop, Ray Marshall, William Usery, and George Schultz) who believed that it provided the country with a “unique opportunity to greatly improve the prospects for rational and peaceful resolution of collective bargaining disputes.” But business associations and their political allies adamantly opposed any compromise on the striker replacement bill. The LPA warned that, like the original measure, the Packwood amendment would “tilt the law in organized labor’s favor, producing more strikes and inflationary labor agreements,” while the CoC argued that labor’s last gasp effort to glean a few additional votes in the Senate was “unwise, unworkable, and totally unacceptable.”48 Senators Orrin Hatch (R-Utah), Don Nickles (R-Oklahoma) and Bob Dole (RKansas) led the Republican filibuster against the bill, attacking the Packwood

202

JOHN LOGAN

amendment as an “untested” proposal that would “insert the federal government into virtually every labor dispute.” According to Hatch, the compromise bill would produce the same dire consequences as the original for American employers, “only this time instead of being shot the punishment will be hanging,” and he defended the filibuster strategy as being in the “best interests” of the country. Republicans also pointed out that it was not only labor’s opponents who had, on occasion, resorted to filibuster; the AFL-CIO had regularly encouraged friendly Senators to vote against cloture on issues such as product liability reform, capital gains tax reductions, and the balanced budget amendment.49 On June 16, 1992 the Senate voted 57–42 for cloture but fell short of the required 60 votes. Five southern Democrats sided with the business lobby and voted against cloture: Hollings (South Carolina), Boren (Oklahoma), Sanford (North Carolina), and Bumpers and Pryor (both Arkansas).50 Five Republican Senators, all from northern or Pacific states, broke ranks and voted for cloture: Specter (Pennsylvania), Stevens (Alaska), D’Amato (New York) and Hatfield and Packwood (both Oregon). Three of the dissenting Republicans – Specter, D’Amato, and Packwood – were facing tough re-election battles in the fall and did not want to antagonize the influential labor lobby in their states. Business organizations and their political allies welcomed the Senate’s rejection of an “economically destructive bill,” but also recognized that the vote had come close to going against them. According to the NAM, the business community had gone “toe-to-toe” with the unions – “they blinked and we won.” Labor Secretary Martin also applauded the outcome of the vote, calling the result a victory for not just President Bush, but for “all working men and women.”51 Despite its failure in the Senate, the AFL-CIO’s “base-broadening” grassroots strategy appeared to have operated reasonably well. After the 1991 House vote, the AFL-CIO reported that in several “critical” cases, support from religious, civil rights, and community organizations had been “essential to our ability to get the votes of tough House members,” particularly those from Alabama, Arkansas, Kentucky, and Tennessee.52 In the 1992 Senate vote, however, the results of labor’s state campaigns were mixed. Among the targeted Republican Senators, the grassroots campaigns had succeeded in delivering the votes of Stevens, Specter, and D’Amato, but had failed to secure support from Durenberger (Minnesota), Cohen (Maine) or Chaffee (Rhode Island). More seriously, in the state with the strongest union and community campaign, Arkansas, labor lost the votes of the state’s two Democratic Senators, Dale Bumpers (who had also cast a decisive vote against the Labor Law Reform Bill in 1978) and David Pryor.53 Nonetheless, after the 1991–1992 campaigns, the AFL-CIO believed that its grass-roots strategy had established “successful patterns of activity” that the labor movement could build upon during its next big push for striker replacement legislation.

Labor’s “Last Stand” in National Politics?

203

The Second Striker Replacement Campaign, 1993–1994 After the 1992 presidential election, certain labor leaders pressed the AFL-CIO to abandon the striker replacement issue and instead mount a major campaign for comprehensive labor law reform. With the election of Bill Clinton, the AFL-CIO believed that there now existed, for the first time in fifteen years, the “very real possibility” that it would be able to enact “sweeping changes” to the NLRA.54 Thus, immediately following the election, it asked affiliated unions to outline their priorities for labor law reform during the next Congress. In response, several large industrial unions argued that it was essential that the AFL-CIO continue to pursue striker replacement as stand-alone legislation without delay. Mineworkers president Richard Trumka argued that striker replacement legislation should be the “first priority of the entire labor movement”; the IBT asked that it be “moved ahead separately as completion of ‘old business’ before other aspects of labor law reform are pursued”; the Paperworkers believed that the bill was of “paramount importance” as it represented the most effective vehicle for reversing the “harm done” to collective bargaining during the Reagan and Bush years; the Rubberworkers insisted that striker replacement legislation should be “our number one priority” and “must not be integrated into any other legislation”; and the UAW argued that organized labor had “been on the defensive for far too long . . . The issue is right – the time to act is now!”55 But not all unions or their political allies shared this unbridled enthusiasm for a second major striker replacement campaign. Several senior unionists questioned the priority that the AFL-CIO had given to striker replacement and argued that other aspects of labor law reform were of more critical importance to the future of the labor movement. The Amalgamated Clothing and Textile Workers Union (ACTWU) and Service Employees International Union (SEIU), for example, believed that a ban on permanent replacements would not assist with what should be the labor movement’s top priority – organizing the unorganized. Since the labor movement would have to mount another all-out campaign if it were to win striker replacement legislation, unions such as ACTWU and the SEIU urged the AFL-CIO to “go for more” – i.e. lobby for substantial reform of the NLRA’s certification provisions. Likewise, Illinois Senator Paul Simon, a co-sponsor of the bill in the upper house, believed that the AFL-CIO had made a tactical error by investing so much effort in a bill that dealt with only one aspect of labor law reform. Simon believed that, in the face of intransigent employer opposition, a more effective strategy would be to campaign for reforms to the NLRA that would facilitate organizing.56 Within the AFL-CIO, too, senior officials differed on the desirability of a second striker replacement campaign. The IUD, which would once again spearhead labor’s

204

JOHN LOGAN

effort, warned of the need to guard against those within the AFL-CIO who would postpone the striker replacement campaign until a broad consensus on labor law reform had been achieved.57 After the 1992 Presidential election, several top AFLCIO officials believed that labor should shift its focus to achieving comprehensive reform of the NLRA and considered striker replacement a political liability that they could ill-afford. As a result, the IUD worried that AFL-CIO leaders intended to delay its next campaign and thus it would not have sufficient time to build the necessary support for the Congressional vote, whereas the business lobby would already have mobilized against the bill. One IUD official warned that a successful campaign would take “months to plan and execute,” but feared that by the time the AFL-CIO leadership and its Senate allies had given the green light for its campaign to commence, “we may not have the time we need to do the job.”58 These concerns appear to have been well founded. Although the IUD and its allies in affiliated unions prevailed on the debate over a second striker replacement campaign, the AFL-CIO did not begin its Congressional campaign until March 1993, just two months before the House vote.59 The business lobby also questioned the AFL-CIO’s commitment to a second striker replacement campaign. Employers claimed that organized labor’s major effort against NAFTA and support for healthcare reform meant that it was unwilling and unable to devote the same resources to passing the strike bill that it had during 1991–1992. The Chamber of Commerce reported seeing “less pressure on striker replacement because of the strong focus” on NAFTA and healthcare reform. But the AFL-CIO denied that its involvement with other important policy issues had necessitated a scaled-down striker replacement campaign; instead, it claimed that an all-out effort was not needed this time around because the groundwork had already been prepared during the previous campaign and labor was now targeting only a handful of key lawmakers.60 In its early stages, the campaign appeared to be proceeding well. On June 15, 1993, the bill passed the House by a margin of 239–190, slightly lower than the total in the 1991 House vote and well short of veto strength; but this time the White House supported the legislation. After the House vote, the striker replacement campaign targeted just seven “swing” Senators in six states – Bumpers and Pryor (both Arkansas-D), Nunn (Georgia-D), Cohen (Maine-R), Durenberger (Minnesota-R), Chaffee (Rhode Island-R), and Jeffords (Vermont-R) – but also sought to reach a broader national audience through its “DC-oriented message.”61 The AFL-CIO re-established its grassroots campaign in those six states and hoped that, together with aggressive support from the White House and the threat of an “old-fashioned filibuster” (seven days a week, 24 hours a day for as long as it takes to have an up or down vote based on merit), it would gain enough votes to achieve cloture. The IUD believed that its best chance of victory in the Senate lay with

Labor’s “Last Stand” in National Politics?

205

winning the support of moderate Republicans who had opposed cloture in 1992, but had voted with organized labor on other important policy issues.62 Determined to defeat labor’s campaign for a second time, however, the business lobby “pulled out all the stops” to prevent the bill reaching the President’s desk.63 Business associations targeted the same seven Senators as did organized labor, stating that they “needed reinforcement,” and intensified their national campaign against the bill. Two months before the Senate vote, the NRTWC launched a $100,000 advertising campaign – fronted by former Screen Actors Guild President, Charlton Heston – aimed at public opinion in the home states of swing Senators.64 The CoC, NFIB, and NAM arranged for hundreds of business owners to call lawmakers to express their opposition and organized what their opponents disparagingly called a “CEO parade” of business leaders from around the country to converge on the Capitol. The Alliance to Keep Americans Working ran full-page “Don’t Let Strikes Hit Home” newspapers ads in the home states of key Senators and in the National Journal, paid for commercials on Rush Limbaugh’s radio show, mailed out over 200,000 postcards to businesses asking them to sign the petition and return them to Congress, arranged for coverage of its position on C-Span and the McLaughlan Report, and got USA Today to present its case in the paper’s “Point Counterpoint” column. Towards the end of the campaign, one senior member of the CoC stated that his organization had opposed the striker replacement bill since 1991 and he didn’t “know a whole lot more we could do.”65 The business lobby’s campaign to win the support of swing Senators paid great dividends. One of the AFL-CIO’s top targets was Senator James Jeffords of Vermont (one of the most liberal Republican in Congress) and its grassroots campaign operated at full tilt in Vermont. The IBT and UE held sit-down demonstrations at Jefford’s office, the Vermont Labor Federation organized numerous press conferences and public meetings on striker replacement, and other unions and community organizations mounted “massive” letter writing and phone calling campaigns in the state. But organized labor’s efforts proved fruitless. Jeffords indicated that he would consider supporting a more limited version of the strike bill that prohibited permanent replacements when employers had demanded “contract give backs” (i.e. concession bargaining) and during negotiations over first contracts. Like most other “swing” Senators, however, Jeffords firmly opposed legislation outlawing all permanent replacements and he voted against the bill for a second time in two years. But Jeffords also sought to downplay his position, hoping that it would not become a major issue at the time of his next bid for re-election.66 Nor were labor’s problems restricted to these swing Senators. Business organizations argued that several senior Democrats were far from enthusiastic in their support for the strike bill. The Printing Industries of America – an organization

206

JOHN LOGAN

that had fought the striker replacement bill “tooth and nail for the past three years” – believed that many top Democrats did not want to face a second vote on striker replacement (even though they had supported cloture in 1992) and reported that they had tried unsuccessfully to convince the party leadership not to bring the bill to the Senate floor. Under tremendous pressure from the agricultural industry in her home state, for example, California Senator Diane Feinstein attempted to persuade the Senate leadership to withdraw the legislation.67 In addition to opposition from the business lobby, senior Democrats faced the wrath of the mainstream media for their support of the strike bill, with one editorial lambasting their “seemingly terminal reflex” of pandering to special interests at the expense of the public interest.68 The IUD acknowledged that some of its “reluctant allies” in the Democratic Party would “like nothing better than to . . . avoid dealing with the issue of permanent replacement.”69 But organized labor persisted with the campaign because this time it had a new, potentially decisive weapon on its side – the President. Union leaders attempted to put pressure on Bill Clinton to campaign actively on behalf of the striker replacement bill. AFSME President Gerald McEntee hoped that Clinton would “expend some political capital to keep a promise” made during the primary campaign, while IBT president Ron Carey argued that the President “ought to get out there where he could be quite persuasive on this.”70 Opposing any amendment to the bill, Carey stressed to Lane Kirkland the importance of getting Clinton to lobby vigorously for labor’s top priority: It is our hope that before we consider any effort to compromise on the substance of our highest legislative priority, President Clinton can be persuaded to use the considerable power of his Presidency . . .. The President’s lobbying efforts on the recent NAFTA vote reflect both the power and the persuasiveness of a committed President. We hope that we . . . can count on the same level of commitment.71

Jay Mazur, president of the International Ladies’ Garment Workers’ Union, welcomed the administration’s verbal support for striker replacement, but insisted that Clinton must approach this legislation with the “same energy” and the “same resourcefulness” that he had applied to ensure the passage of NAFTA over labor’s opposition. And at a public rally in support of striker replacement legislation in March 1994, one senior AFL-CIO official told Clinton that, after the split with labor over NAFTA, “You owe us one, and this is the one we want.” In response to unionists’ demands for assistance, Labor Secretary Robert Reich (the bill’s strongest supporter within the Clinton Administration) claimed that the President had been actively lobbying the AFL-CIO’s targeted Democratic Senators.72 Two months before the Senate vote, the AFL-CIO still hoped that vigorous support from the White House would help secure the required support. The IUD reported that Clinton had “given his staff their marching orders on the issue” and

Labor’s “Last Stand” in National Politics?

207

that both the White House and the Labor Department were “gearing up for a big push” in the final weeks of the campaign. But the IUD also conceded that it was unsure about exactly what kind of effort to expect from the President: “The intangible is what degree of pressure might be asserted by Clinton. If it is at all comparable to that exerted for NAFTA we . . . [see] a very good chance for victory.”73 Shortly before the Senate vote, Clinton wrote to targeted Democrats stating that he was “deeply concerned” by reports that they were considering upholding the Republican filibuster and urged them to support his efforts “to secure the timely enactment” of the striker replacement bill.74 The AFL-CIO repeatedly asked Clinton to lobby the two Democratic Senators from Arkansas, Dale Bumpers and David Pryor, believing that strong White House pressure could at least deliver Pryor’s vote.75 Within the President’s home state, the coalition supporting the bill attempted to establish “a local predisposition that a vote against the bill was a vote against the (very popular) President,” in the hope that this strategy would “help position Clinton to do something” with his fellow Arkansans.76 One unionist reported that at a recent meeting of the Arkansas laborcommunity alliance: A lot of weight was put on the importance of the degree of the Clinton administration’s support for the bill. People felt that if Clinton were out front, actively supporting the bill, our chances of getting Senator Pryor and other swing Democratic votes would greatly improve.77

Members of the religious community also tried to get the President to do more. The Arkansas Interfaith Conference on Workplace Fairness, a religious coalition formed to campaign for the bill, attempted to influence the President to lobby aggressively in his home state, telling Clinton that Bumpers and Pryor would feel “compelled to follow your clear lead on this issue.”78 Business organizations also lobbied intensely Arkansas’s two Senators. While the CoC was confident that both Bumpers and Pryor were leaning towards opposition to cloture, it feared that a last-minute call from Clinton might still change their votes. The NRTWC warned that Bumpers and Pryor (whom it described as being “at the epicenter of a political earthquake in Washington”) were “showing signs of cracking” and considering supporting a labor-backed compromise bill. Both Senators enjoyed strong labor support when first elected to the Senate and shared a similar political outlook to and were strong allies of the President. Until the last moment, Bumpers and Pryor refused to disclose how they would vote, but stated that they intended to “act together” on the issue. In the end, neither would confirm whether or not Clinton had contacted them directly but, for the second time in two years, both voted with the Republicans against cloture.79 Business associations recognized that support from the White House would be critical to the success of labor’s campaign, but suspected that the President was

208

JOHN LOGAN

lukewarm on the bill and would be happy to see it fade away without having to take a public stand. One business publication reported that there was “no way” that the strike bill would pass without “all-out administration backing. Support exists, but the White House has reservations. The legislation probably won’t happen.”80 While Clinton was committed to signing the bill if it reached his desk, the LPA noted that it had not “seen him up on the Hill pushing it.” Congressional Republicans also questioned Clinton’s level of commitment, with the ranking minority member on the House Labor Committee, William Goodling of Pennsylvania, reporting that he had not heard the President “running around encouraging people to pass this legislation. He has . . . not been pressuring anyone.”81 By March 1994, the bill’s Senate sponsor, Howard Metzenbaum, believed that he had 57 guaranteed votes and thus needed only three more votes to achieve cloture. With the Republicans “holding totally solid,” however, the additional votes were not materializing. By April, Metzenbaum had initiated intensive behind-thescenes discussions to negotiate a compromise on the legislation.82 Many unionists supported a compromise, if that were required to get legislation enacted, and urged the AFL-CIO to build support for an amended bill as quickly as possible. If Senator Metzenbaum believed that it would be impossible to break the Republican filibuster, the Paperworkers urged the AFL-CIO to support a compromise based on a six-month moratorium on the hiring of permanent replacements, similar to the law that had operated in Ontario since 1970. The union would rather have “some protection rather than none at all” and it pushed the AFL-CIO to “gain as much lead time as possible to shop the compromise around the Senate” to make sure it could overcome the Republican filibuster.83 While they were confident that organized labor lacked support to pass the bill in its original form, business groups worried that the AFL-CIO would negotiate a last-minute compromise (similar to the 1992 Packwood amendment) allowing them to gain additional votes and break the Republican filibuster.84 The NAM warned its members that the bill’s Senate sponsor, Howard Metzenbaum, was a “master at ‘let’s make a deal,” and argued that a vote for compromise was a de facto vote for the original bill because “any changes would be lost in a House-Senate conference.”85 Proposed amendments to the 1994 striker replacement bill included the following:
A moratorium of several weeks before allowing businesses to hire permanent replacements.
Forcing businesses to appeal to the NLRB before hiring permanent replacements.
Setting up an arbitration system for major strikes. The business lobby rejected all of the proposed changes. The CoC described the striker replacement bill as a “no-compromise issue” and stated its

Labor’s “Last Stand” in National Politics?

209

categorical opposition to the amendments: “It is an issue that is not open for negotiation . . . Under no circumstances are any of them acceptable.”86 In particular, management associations attacked amendments that proposed a moratorium of some length (usually 1–8 weeks) on the recruitment of permanent replacements after the start of a strike. The CoC insisted that a moratorium “of any length” would serve only to encourage strikes; it claimed that no business could shut down for even a few weeks without “potentially fatal” consequences and that small businesses would be “especially vulnerable” to such interruptions.87 The CoC publicized incidents in which firms had been forced out of business as a result of brief economic strikes, such as the Churchill Truck Lines company of Missouri, which folded after a week-long strike by the IBT in April 1994, leaving 2,000 workers, most of them Teamsters, unemployed.88 The LPA attacked amendments that would allow the NLRB to decide whether or not the recruitment of permanent replacements was a matter of “business necessity” because this did not allow for the short time frame within which employers had to make decisions concerning replacements and would substitute the judgment of the NLRB for that of the employer. The only reasonable trade-off for banning permanent replacements, the LPA argued, would be to outlaw all economic strikes.89 In the final event, however, Senator Metzenbaum’s efforts were to no avail and no compromise bill was offered before the final vote. By June 1994, both he and the AFL-CIO realized that the bill was heading for defeat, but labor hoped that it could at least force a change in the Senate filibuster rule.90 Business groups and their political allies, in contrast, were supremely confident of their ability to defeat the striker replacement bill. One prominent opponent in the Senate, Malcolm Wallop of Wyoming, boasted that Republicans “can’t wait for a vote” and were so certain of victory that they would consider bringing the issue to the floor themselves, if the Democratic leadership developed cold feet. The CoC believed that even if unions intensified their “lackluster” campaign, it “wouldn’t make any difference. . . They’ll still be running in place on a treadmill.”91 In all likelihood, however, a compromise bill would not have altered the outcome of the Senate vote. While some of its opponents had been “hinting at compromise,” the IUD argued this was “no more than window dressing” intended to placate organized labor.92 Although it had received signals from lawmakers that certain amendments would “make the bill more attractive,” the IUD dismissed their words as merely “lip service from Senators trying to relieve pressure” rather than a serious attempt to resolve the issue. And it warned state affiliates that were attempting to win over the swing Senators: Be under no illusions! These Senators would like nothing better than for us to bargain against ourselves, to weaken our position and then still vote against us because we have not weakened our position enough . . .. We must let them know that these words are cheap as long as they

210

JOHN LOGAN

take sides with the most backward elements of the business community who have made “No compromise, no negotiation” the keystone of their campaign.93

The IUD’s words of warning proved prophetic: none of the AFL-CIO’s targeted Senators changed their votes as a result of labor’s intensive lobbying efforts. Shortly before the Senate vote, the AFL-CIO read out on the steps of the Capitol the names of 19,722 workers who had lost their jobs to replacements since 1980, which symbolized the “unconscionable impact” of permanent replacement on “tens of thousands of employees, their families and communities.” But time had run out for labor’s campaign and one White House official predicted that the striker replacement bill was certain to “go down in flames.”94 As expected, on July 12, 1994, the Senate voted 53–46 for cloture, again falling short of the required 60 votes, and the striker replacement bill died. In the end, the outcome represented a virtual replay of the “death-by-filibuster” vote two years previously. All but three Republicans (Hatfield of Oregon, Specter of Pennsylvania, and D’Amato of New York) opposed cloture, while six Democrats voted with the business lobby: Pryor and Bumpers of Arkansas, Matthews of Tennessee, Nunn of Georgia, Boren of Oklahoma, and Hollings of South Carolina. As in 1992, the Democrats opposing cloture came from Southern states, where the business lobby was strongest, while the Republicans who voted for cloture represented industrialized northern states or Pacific states, where permanent replacement was a critical issue for local unions and where labor’s political influence was strongest.95 Supporters of the bill remained upbeat. Vice President Al Gore pledged the Clinton Administration’s continuing commitment to the bill, telling a union audience that, while the outcome of the Senate vote was a “major disappointment,” it was “just a temporary setback.”96 Senate sponsors of the legislation, Howard Metzenbaum and Tom Harkin, announced that they would try to thwart the Republican filibuster by attaching striker replacement to an unrelated bill supported by Republicans – such as the product liability bill – a procedure that would require that they win only a straightforward majority in the Senate. That way, both labor and management would “walk away with something” if the bill were to pass.97 However, Democratic loses in the November 1994 elections made impossible any attempt either to reintroduce striker replacement or attach it to a measure supported by Republicans. The AFL-CIO attributed the fact that it won fewer Senate votes in 1994 than it had done in 1992 to a number of factors: tight party discipline among the Republican “guardians of gridlock,” a breakdown of Democratic Party discipline, new faces in the Senate after the 1992 elections, and the fact that in 1992 the Senate had voted on the Packwood-amended version of the strike bill. However, the AFL-CIO also

Labor’s “Last Stand” in National Politics?

211

underestimated the enormity of the task it faced in the Senate. In March 1994, the IUD had reported that its “best evaluation” was that it had between 56 and 58 votes for cloture, “with several Senators still undecided, as well as others we will try to influence,” and concluded that the bill had a “very good chance” of passing the upper house this time around.98 But it was probably never that close to winning the super-majority needed to end debate and bring the bill to a vote. In 1992, conservative Democrats, such as Sam Nunn of Georgia, had considered their vote for cloture a “free vote,” safe in the knowledge that supporters of the bill were well short of the total required to override the promised presidential veto. Thus, Nunn was able to appease organized labor without incurring the wrath of the business community by putting on the statute books a bill that it had categorically opposed. Nunn subsequently claimed that he had voted for cloture in 1992 only because it was clear that “substantial revision” of the bill was required to win White House approval or gain the votes needed to override the veto. Getting the bill past the Republican filibuster, he argued, would have facilitated this process of revision. But Nunn stressed that the situation was quite different in 1994: The atmosphere surrounding [striker replacement] has changed since 1992 with President Clinton’s indication that he would sign the bill as is into law. If cloture is invoked, there will be little chance that the bill will be changed on the floor since only 50 votes will be required for passage.99

Thus, Nunn stated that now he would support the bill only if it were amended to set a time limit on the prohibition against permanent replacements or to include binding arbitration. Recognizing that Nunn was a key player among the Southern Democrats in the Senate, the business lobby also campaigned vigorously for his support. Several days before the 1994 vote, the Printing Industry Association of Georgia accused Nunn of “playing both sides” on the strike replacement issue, but it was nonetheless confident that he was leaning “slightly towards our position.” Nunn did indeed vote with the business lobby in 1994; the AFL-CIO had considered his support “essential” and its failure to win the Georgia senator’s endorsement proved a major blow to its campaign.100 In addition, opposition to the striker replacement bill in Congress had probably hardened after the 1992 defeat, making it less likely that the original bill would ever become law. Several “swing” Senators argued that it had been a tactical mistake to insist on a vote in 1992, believing that the AFL-CIO had done so only to embarrass President Bush by sending him a controversial bill that he had promised to veto during an election year. It had always been unlikely that the bill’s supports would garner sufficient votes to overcome Bush’s promised veto; indeed, prior to the 1991 House vote, AFL-CIO president Lane Kirkland had conceded that overriding a veto would probably be impossible.101

212

JOHN LOGAN

The AFL-CIO leadership concluded that the striker replacement bill had failed for a second time because the filibuster rule had allowed a “willful” but cohesive minority in the Senate to frustrate popularly supported legislation. The “unfortunate bottom line,” according to Lane Kirkland, was that labor’s targeted Senators had “remained with the business lobby and with the forces of gridlock.”102 But opponents of the bill defiantly stood by their controversial filibuster strategy. One of the leaders of the filibuster, Kansas senator Bob Dole, insisted that, on this occasion, gridlock was “good politics,” while the Alliance to Keep Americans Working claimed that a little gridlock in Congress today was infinitely preferable to “years of job-killing economic gridlock throughout the country” in the future.103 For business associations the outcome of the 1994 Senate vote was entirely predictable, a foregone conclusion that merely reflected the decline in the AFLCIO’s political influence both in Washington and in the nation as a whole. The CoC believed that the result was as “expected” and a “codification of what the public has been witnessing for several decades. Organized labor has lost its clout. It can’t sign up people in the marketplace, and it can’t convince politicians to front for them.”104 After the Senate defeat, the AFL-CIO leadership applauded the efforts of the Clinton Administration. AFL-CIO President Lane Kirkland praised Clinton’s “outspoken support,” stated that the White House had done “everything possible” to win over hostile Democrats, and thanked the President for the letters of support he had sent to targeted Senators.105 The Clinton Administration also robustly defended its own efforts, with Labor Secretary Reich insisting that the Administration had “pulled out all of the stops. Nobody can say that we didn’t try.”106 Despite their protestations, several union leaders and their political allies blamed the defeat partly on Clinton’s lackluster lobbying. Because it had managed to deliver only one vote from the four Democratic Senators from the home states of President Clinton and Vice President Gore, several senior unionists concluded that White House efforts on behalf of the strike bill had been limited and ineffectual. IBT President Ron Carey complained that Clinton had not invested the same effort on behalf of striker replacement that he had done to ensure the passage of NAFTA over labor’s opposition, while Iowa Senator Tom Harkin believed that the bill would have passed if the White House had “put in one-tenth of the effort it did to pass NAFTA.” The New York Times agreed with the President’s critics, reporting that labor’s top priority had “never inspired the midnight phone calls and political arm twisting the White House had lavished on other difficult political issues.”107 But Clinton’s ineffectual campaigning on behalf of the strike bill was not the only factor contributing to labor’s second defeat in two years. Once again, the business lobby had proved to be united, robust, and fiercely opposed to any compromise on the legislation. In addition to tepid support from the White House and resolute

Labor’s “Last Stand” in National Politics?

213

opposition from the business community, labor’s failure had deeper, underlying causes. Despite creating alliances with numerous civil rights, women’s, religious, and community organizations, the AFL-CIO failed to persuade skeptics that they should consider the protection of strikers’ jobs as a “fundamental human right,” rather than simply as special interest legislation. Shortly after the 1992 Senate vote, an outraged Ron Carey stated that the right to strike without fear of losing one’s job was a “fundamental right . . . It’s like the freedom of speech . . . there should be no debate about it.”108 Unlike the campaign for plant-closing legislation, however, the striker replacement campaign failed to reach out to a broad non-labor base, and non-unionists, for the most part, did not consider the right to strike without fear of replacement as analogous to fundamental rights like freedom of speech. Ultimately, the AFL-CIO failed to convince non-unionists of the importance (if not the justice) of its legislation and even some unionists questioned the priority given to the striker replacement campaign within labor’s policy agenda. While conducting grassroots campaigns at the state level, moreover, the AFL-CIO failed to fully exploit its potential support among non-labor organizations. After the second Senate defeat, the ACLU wrote that the level of cooperation and communication between organized labor and its progressive allies in the community was “far less” than it might have been: “Many groups, ours included, were willing to do far more on striker replacement legislation than they were called upon to do.”109 And, finally, organized labor failed to win decisively the debate on several critical issues that dominated the striker replacement campaign. Instead, the AFL-CIO was consistently forced to respond to business accusations concerning its self-interested motives and predictions of the disastrous impact of the strike bill on economic performance and business activity. The next section of the paper provides a detailed analysis and evaluation of the significance of the debate on these major issues.

THE MAJOR ISSUES IN THE CAMPAIGN The Impact of Striker Replacement Legislation on Economic Competitiveness and Business Performance Perhaps the most contentious issue during the campaign was the question of the legislation’s likely impact on job creation, investment, and economic competitiveness. Management associations recognized that striker replacement involved highly emotive issues – the freedom to strike and the loss of employment during a period of economic uncertainty – and thus they had to tread carefully when arguing in favor of the right to permanently replace strikers. As a result, their campaign against the bill focused primarily on the “hard” issues of economic

214

JOHN LOGAN

competitiveness and business necessity, and largely sought to avoid the emotive or moral aspects of the striker replacement debate. Employer associations claimed that by sanctioning “risk free” strikes in support of unreasonable bargaining demands, striker replacement legislation would impose “unilateral economic disarmament” on management and undermine the ability of American firms to compete in the international economy. The law would reverse the tentative economic recovery of the early 1990s and thus the stark choice facing American lawmakers was one of “more jobs or more strikes.” One employer representative argued that the reassurance that one can strike without fear of permanent replacement would be little solace “if one couldn’t find a job in the first place,” while the CoC believed that the best context for understanding the deleterious impact of “Job Destruction Bill Number One” was the precarious economic climate of the early 1990s and warned: “Recession looms.”110 New technology and heightened international competition had increased the necessity for maintaining production during strikes, according to employers, and any legislation that interfered with the ability to continue operations would accelerate the relocation of American businesses overseas.111 One management representative claimed that the strike bill would encourage firms to move to countries with less intrusive labor laws: “This bill provides a unique incentive to American companies to locate within the borders of countries that are known for being favorable to the interests of business.”112 Other opponents of the bill believed that restrictive labor laws in European countries, such as those prohibiting permanent replacements, had contributed to the determination of European firms to operate union free in their U.S. transplants. The Detroit News warned those tempted to support the bill that “assisted economic suicide” would be the result: If you doubt the effects of this labor protectionism, just observe Germany, a nation that bans permanent replacement workers . . . Mercedes Benz and BMW are seeking to escape the high cost of these German labor protection laws by starting plants in the United States, with BMW’s factory to be in decidedly non-union South Carolina.113

Employers’ contention that the striker replacement bill would destroy jobs and encourage firms to flee the country seriously damaged labor’s campaign, especially after the November 1993 enactment of NAFTA, which had already heightened fears about plant relocations. The decision to hire permanent replacements was frequently a question of business necessity, employers argued, because many firms would be unable to recruit temporary replacements during strikes. Skilled workers would refuse to switch jobs without the assurance of permanent employment and many potential replacements would be unwilling to cross picket lines without the prospect of longterm jobs. Somewhat ironically, supporters of striker replacement legislation used

Labor’s “Last Stand” in National Politics?

215

Charles Perry’s book, Operating During Strikes (which had previously been used by Phelps Dodge as a strike-breaking operating manual), to sustain their claim that firms could survive strikes without permanently replacing economic strikers. According to the bill’s advocates, Operating During Strikes had categorically demonstrated that “companies that want to do so are able to continue operations” without resorting to permanent replacements.114 Unions and their allies recognized that it would be critically important to make the economic case for striker replacement legislation in order to “blunt the attack” that was sure to come from “business-backed economists and think tanks.”115 They pointed out that most large industrial economies – including Japan, Germany and France – had laws protecting the jobs of strikers, thereby refuting the notion that such legislation was “anti-competitive.” Given the negative impact of permanent replacements on workplace relations, argued two of the bill’s Congressional sponsors, it was “no wonder” that these countries recognized that permanent replacements were “unwise as a matter of both public policy and good business.”116 Moreover, firms would not gain competitive advantage by replacing an experienced unionized workforce with an inexperienced non-union workforce, the AFL-CIO argued, and it pointed to the fates of companies like Eastern and Continental Airlines, the Daily News, and Greyhound, which had hired permanent replacements en masse but subsequently performed poorly in the marketplace.117

The Impact of Striker Replacement Legislation on the Character of Labor-Management Relations In the 1980s and 1990s, a growing number of academics, lawmakers, union leaders, and management representatives stressed the need for government policy to encourage labor-management cooperation and promote workplace innovation designed to boost productivity and help businesses to compete in the global economy. Unions argued that outlawing permanent replacements would facilitate co-operation and innovation, while the business lobby responded that it would promote strikes and undermine employers’ efforts to transform the workplace. Business groups claimed that the “strike-breeder bill” would promote adversarial labor relations and discourage cooperation because it provided no incentive for unions to moderate their bargaining demands. If union members could walk out without fear of being replaced, one employer representative argued, the strike weapon would be used “as often as a Saturday Night Special”; current policy on permanent replacements, in contrast, had contributed to historic low levels of strike activity.118 The Bush Administration agreed with the business lobby that the strike

216

JOHN LOGAN

bill would almost certainly result in “increased labor strife, weakened businesses, diminished competitiveness, and a less dynamic work force.”119 Unions and their allies claimed that permanent replacements were incompatible with the type of high performance workplace and labor-management cooperation needed to compete effectively in the global economy. Congressional sponsors of the strike bill, Howard Metzenbaum and William Ford, argued that the U.S. could not hope to compete in global markets while allowing the permanent replacement of economic strikers: “Hiring permanent replacements sends an unmistakable message that workers are disposable, reducing employee morale and lowering productivity.” After the 1992 election, the new Democratic Administration concurred; Labor Secretary Robert Reich told the Senate that permanent replacement exemplified the “practices and attitudes that make real cooperation between labor and management impossible.” Thus, as representatives of both labor and management were calling for the creation of a less adversarial system of industrial relations, the time had come to “do away with” the practice of permanent replacement.120 The bill would discourage bad management practices, its supporters argued, by rectifying the current imbalance in power in labor-management relations that had encouraged “low-road” strategies of profit maximization. If employers could instigate strikes and then replace their existing workforce, they would likely respond to heightened international competition by cutting wages and benefits rather than by improving productivity through investment in skills, technology and workplace innovation. AFL-CIO President Lane Kirkland told President Clinton that opponents of striker replacement legislation welcomed the deterioration of real wages as “necessary to enhance our country’s ‘competitiveness’ . . . They prefer a low-wage America.” Providing employees the right to strike without fear of replacement, in contrast, would encourage efficient and competitive firms, but not at the “disastrous cost” of turning the U.S. into a low road economy.121 The prohibition of permanent replacements would therefore discourage employers from competing on the basis of cheap labor costs and instead encourage greater investment in skills, technology and workplace innovation. According to organized labor, strikes in which employers had hired permanent replacements tended to be longer in duration, more intense, and more contentious than were those in which they had hired temporary replacements or had ceased operating during strikes.122 Moreover, the recruitment of permanent replacements frequently transformed limited economic disputes into broader struggles and even in those cases where the threat to recruit permanent replacements had hastened the end of strikes, it often “deeply damaged” relations between management and labor for years after the conclusion of strikes. During the 1992 Caterpillar strike, for example, the company’s threat of permanent replacement had forced the striking

Labor’s “Last Stand” in National Politics?

217

UAW members back to work without a contract, thereby ending the strike, but had “poisoned” the company’s relationship with its employees and union for years afterwards.123 A few days before the 1994 Senate vote, President Clinton wrote to the CEO of Caterpillar, which was still involved in a dispute with the UAW, stating that the threat or implementation of permanent replacements had a “poisonous affect on relationships between workers and employers and does great damage to the collective bargaining process.” The practice of permanent replacement, Clinton concluded, “must stop, because it deters the type of . . . cooperative work forces that we need to prosper in the new world economy.”124 In an attempt to demonstrate the system of collective bargaining working at its best, the AFL-CIO sought to highlight economic strikes in which employers had recruited only temporary replacements or no replacements at all and the strikes had been settled quickly with the resumption of normal labor-management relations.125 Labor’s arguments concerning the destructive consequences of permanent replacement received an unexpected boost when Bernard DeLury, head of the Federal Mediation and Conciliation Service (FMCS) and a Bush appointee, expressed his belief that permanent replacement had a deleterious impact on labor-management relations. To the chagrin of management associations, DeLury claimed that the permanent replacement of economics strikers undermines the collective bargaining process by making it difficult for unions and management to settle their differences in a peaceful and orderly manner. DeLury’s replacement at the FMCS, John Calhoun Wells, also favored a change in the law because permanent replacements “chilled meaningful bargaining.”126 The AFL-CIO could credibly point to instances where permanent replacements had irreparably damaged labor-management relations, but the relationship between the threat of permanent replacement and the decline in the frequently of strikes proved a more problematic issue. The number of strikes involving 1,000 or more workers had fallen dramatically from 424 in 1974 to only 44 in 1990. Strike levels in the 1980s were about one-half of the number for the 1970s and that figure continued to fall in the 1990s; by the mid-1990s, the level of strikes was at its lowest for the entire postwar period.127 While it believed that the threat of permanent replacement was largely responsible for declining strike levels, as workers were increasingly reluctant to put their jobs on the line, the AFL-CIO avoided making this argument for fear of assisting its opponents, for whom the current “deterrent effect” of the law was a strength, not a weakness.128 Indeed, when the Labor Research Association approached the AFL-CIO for assistance with a study exploring the relationship between public policy and the decline in strike activity, the IUD expressed concern that this was also the argument advanced by the NRTWC.129 Although it saw a calm born of fear, rather than one born of harmony (as its opponents would suggest), the AFL-CIO tried to avoid a debate over the relationship between striker replacement

218

JOHN LOGAN

policy and the diminishing utility of the strike weapon. Nevertheless, employers’ prediction that the bill would lead to a sharp increase in strike activity and business failures severely damaged labor’s campaign.

The Rights of Strikers vs. the Rights of Replacements, Non-Strikers, Non-Union Members, Consumers, and Employers Part of the debate over striker replacement legislation concerned the question of whose rights most needed legal protection. The business lobby claimed that the legislation represented a blatant power grab by “Big Labor” that would promote the rights of striking union members at the expense of those of replacements, nonstrikers, non-union members, consumers, and employers. The CoC argued that unions were desperately attempting to win through legislation what they had failed utterly to achieve at the workplace – i.e. recruit more dues-paying members and regain muscle at the bargaining table. Organized labor was seeking to “accomplish through politics what it couldn’t accomplish in the marketplace” and reverting to the “destructive strategy of seeking a legislative bailout as a quick fix to its problems.”130 Opponents of the bill insisted that the current interpretation of the law on striker replacements had struck the appropriate balance between the rights of strikers and the rights of employers: it provided workers with the absolute right to strike, prohibited management interference with that right, gave strikers a life-long right to return to their jobs after the strike ends and positions become available, and, crucially, permitted firms to operate during strikes, hiring permanent replacements if necessary. Like unions, the NRTWC believed that the striker replacement issue was primarily a question of rights – but the rights of replacements and of the overwhelming majority of employees in the nation, who were not union members, rather than the rights of striking union members. The bill would adversely affect non-union employees, the CoC argued, because employers would be forced to favor striking union members over workers who had chosen not to strike and because they would suffer from the downturn in business activity that would inevitably result from the bill.131 According to the business lobby, employers’ right to hire permanent replacements was entirely unrelated to the recent decline in union density, which was the result of long-term structural and demographic changes. The alleged greater resort to permanent replacements in recent years, argued one management lawyer, was a “symptom, not a cause” of the decline in union power and thus the onus was on supporters of the legislation to demonstrate why a successful 50 year-old labor policy should be overturned.132 Business organizations and their

Labor’s “Last Stand” in National Politics?

219

political allies also sought to characterize the striker replacement bill as a “radical” and “unwarranted” overhaul of current labor policy. When the House and Senate labor committees approved bills outlawing permanent replacements on May 5, 1993, the House Committee’s ranking Republican, Bill Goodling of Pennsylvania, remarked: “This is Karl Marx’s birthday. He couldn’t ask for a better birthday present.”133 Unions believed that the use of permanent replacements as an offensive weapon by management had resulted in thousands of workers “suffering capital punishment in the workplace” for exercising what was supposed to be a federally protected right. One Machinists official said of the impact of permanent replacements on labor’s strike weapon: “We have nothing to bargain with now. Labor has an empty gun.”134 The AFL-CIO saw no practical difference between the discharge and the permanent replacement of strikers – it was a “distinction without a difference.” It also pointed out that the legal environment was now quite different (and more hostile to unions) than it was when the Court issued its Mackay decision. During the intervening half century, the legal status of strikers and their replacements had changed significantly, and unions saw no reason why the law on replacements should not continue to evolve. While the practice of permanent replacement had created a severe power imbalance at the workplace, striker replacement legislation would “level the playing field,” thus benefiting all employees, union and non-union.135 But organized labor did not make much headway on this issue. It was always going to be more difficult for unions to argue for the overturn of a policy that had operated for 50 years than it was for employers to defend the status quo.

The Level of Public Support for Striker Replacement Legislation and the Use of High-Strikes Involving Permanent Replacements In order to gain the upper hand in the propaganda war, supporters and opponents of the legislation both commissioned opinion polls intended to demonstrate majority support for their position on permanent replacements (see Table 1 for summary of major polls) and the poll results themselves quickly became a contentious issue during the campaign. Business associations dismissed the AFL-CIO’s claim that the polls clearly demonstrated public support for legislation protecting the jobs of strikers. Typical of the conflict over poll results was the business lobby’s response to the AFLCIO-sponsored poll in June 1994, just one month before the second Senate vote. The “National Survey of Attitudes on Striker Replacement” asked 1,000 registered voters the question, “Once a majority of employees vote to strike, should companies

220

JOHN LOGAN

Table 1. Major Polls on Permanent Replacements, 1990–1994 (Compiled by the Author). Poll

Year

Major Finding

Committee for Workplace Fairness/Penn & Schoen Associates

1990

65% of respondents were against firing strikers; 26% supported employers’ right to fire.

Committee for Workplace Fairness/The Roper Organization

1990

42% of respondents believed that the hiring of permanent replacements for striking workers was a “bad thing”; 23% thought it a “good thing.”

Time-CNN/Yankelovich Clancy Shulman

1990

57% of respondents believed that Greyhound, which had permanently replaced strikers, “should be required to rehire all workers”; 22% sympathized with the company.

Committee for Workplace Fairness/The Roper Organization

1991

67% of respondents had an unfavorable view of “the hiring of permanent replacements for striking workers.”

Economic Policy Foundation/Penn and Schoen

1991

54% of respondents opposed firing replacements at the end of strikes; 34% favored such an action.

Time-CNN/Yankelovich Clancy Shulman

1992

29% of respondents favor a law that would prohibit employers from hiring permanent replacements for striking workers; 60% opposed such a law.

National Right to Work Committee/Marketing Research Institute

1993

57% of respondents opposed the striker replacement bill; 30% supported it.

Committee for Workplace Fairness/FingerhutGranados Opinion Research

1994

65% of respondents believed that companies should not be allowed to fire and permanently replace striking workers; 31% believed companies should enjoy that right.

be allowed to fire employees for striking and hire permanent replacements to take their jobs or should companies not be allowed to hire permanent replacements?” The survey found that by a margin of 65–31% respondents believed that employers should not be permitted to fire and permanently replace striking workers.136 Accusing it of attempting to cloud the real issue with “spurious claims,” the CoC dismissed the AFL-CIO’s “ironic attempt” to show public support for its strike bill: What the survey in fact shows is that Americans support the current law . . .. Organized labor’s claim that the survey proves public support for [striker replacement legislation] is a deliberate effort to distort the facts in a last-minute attempt to muster support for the bill. Organized labor

Labor’s “Last Stand” in National Politics?

221

should be shocked . . . that nearly one third of Americans, by its own estimation, think that employers should be allowed to fire strikers!137

Business associations also repeatedly cited the results of a 1992 Time/CNN poll (conducted during the widely reported Caterpillar strike) that had found that 60% of respondents opposed striker replacement legislation, with only 29% supporting a ban on permanent replacements. This result contrasted with those of AFL-CIOcommissioned polls conducted by the Roper Organization in 1990 and 1991 that had found that that a majority of the public supported workers’ right to strike without fear of losing their jobs to replacements. The Roper Organization believed that the explanation for this discrepancy lay with the fact that the Time/CNN poll had provided only one possible response to a “very complex” subject. If the respondents were given the choice of supporting a ban on permanent replacements under certain circumstances (such as during strikes over safety conditions, pensions or health benefits), Roper suspected that most respondents would favor a “more moderate” position on striker replacement that “tilts the advantage more toward labor than toward management.”138 However, both sides had to tread carefully on the issue of public opinion and neither labor nor management could credibly claim unqualified support for their positions. Labor recognized that the public generally disliked strikes, especially disputes over higher wages. Union leaders believed that public support would be “critical” to the success of the campaign but readily conceded that they could not rely upon the level of public support that they had once enjoyed in political confrontations with management because many Americans now viewed organized labor as a “greedy special interest.”139 One union leader explained labor’s fundamental problem: The public is relatively ignorant about the entire issue. They see striker replacement as a “union” or “labor” issue, with all the preconceptions and biases these terms bring. They do not see it as a “public” or “moral” or “fairness” issue that everybody has to take a stand on. This is clearly a weakness . . . This must be a coalition effort or it will be seen as just “special interest” legislation.140

In addition to public distaste for strikes and ignorance about the issue, labor had to contend with the problem of widespread indifference towards the plight of replaced strikers. With the President of the United States as their inspiration, many firms had resorted to the permanent replacement of economic strikers in the 1980s, “with minimal negative reaction from the public.”141 Indeed, many non-union employees now viewed the recruitment of permanent replacements as a legitimate and standard management strategy, not as deviant and unacceptable behavior that required a legal remedy. Organized labor also faced the challenge of how best to explain the rather technical issue of permanent replacement to a non-union audience. In its public

222

JOHN LOGAN

campaign, the AFL-CIO talked of employers who fired (rather than permanently replaced) workers for exercising their legal right to strike, because permanent replacement was “just plain gibberish” to non-unionists, an “insider’s language” that hinged on the legal distinction between temporary and permanent replacement workers under the NLRA and various Supreme Court decisions. This allowed unions to discuss striker replacement in a way that would hopefully elicit public understanding and support. While the language of permanent replacement was of “limited value in communicating with non-adepts,” the public did understand that employers’ ability to fire strikers “completely undermines” the right to strike. The AFL-CIO also attempted to present the issue as one of the wrongs done to replaced strikers, not one of the legal status of permanent replacements, because the latter issue would allow opponents to claim that the bill denied all rights to an important but often neglected group – replacements hired during the strike.142 Unions recognized that if the striker replacement debate were framed simply as one of labor vs. management, their opponents would have a “much easier time” defeating the legislation. Instead, the bill’s supporters needed to stress its “applicability to all Americans,” which was “key to getting the degree of public support” necessary to ensure its passage. Particularly important was whether or not people believed that permanent replacement “could happen to me or somebody like me.” In addition, polling among non-unionists had disclosed strong support for the rights of strikers when the main issues were those of workplace safety, pensions or health benefits because “virtually all Americans” worried about “their own health care or pensions being cut, or their workplace being unsafe.”143 While these issues were “big winners” in the striker replacement debate, several other arguments in favor of reform did not enjoy widespread support. Focus group polling had revealed a “strong lack of support” for the rights of strikers when the main issue was wages because many non-union employees believed that union members made “too much money,” to the extent of harming their employers and causing consumers to pay inflated prices for their products. Less than one third of respondents to one AFL-CIO survey agreed that new legislation was required because the ability to hire permanent replacements provided employers with an “unfair advantage in collective bargaining.”144 And while a majority of the public agreed with the statement that companies generally “don’t care about” their employees, they stated this out of a sense of realism rather than out of anger. As a result, the “disposable worker” thesis – that permanent replacements illustrate that firms don’t respect their employees – could not be discarded, but neither was it “a clear winner.”145 Thus, the AFL-CIO needed to frame its campaign in a way that both stressed the benefits of striker replacement bill to all employees and rebutted business efforts to label it “special interest” legislation.

Labor’s “Last Stand” in National Politics?

223

In an attempt to refute the business lobby’s accusation that it was behaving like a greedy special interest, organized labor’s campaign repeatedly affirmed “traditional” American values, namely, “fairness, family, and the fact that workers’ rights are based on a long-standing legal foundation.”146 The bill was not a power grab by “Big Labor,” unions insisted, but an attempt to prevent a new breed of unscrupulous corporate profiteers like Frank Lorenzo (CEO of Eastern Airlines) from abusing the permanent replacement loophole. By protecting hard-working families against acts of corporate villainy, it would assist their efforts to fight for vital issues such as health care, workplace safety, pensions, and a decent standard of living. Since labor’s goal was, in part, to motivate non-union households to support the bill, its message needed to have broad appeal and avoid parochialism. Thus, campaign materials promoted a “positive message” and avoided not only “negative” language such as “scabs” but also all “labor movement terms.”147 The object was to promote a straightforward message (permanent replacement undermines the legal right to strike and harms the entire system of collective bargaining), force opponents of the bill onto the defensive, and avoid raising issues that would assist the business lobby. The principal objective of the Congressional campaign was to frame the issue as being well within the political mainstream, to give targeted legislators a sense of the broad potential base of support for the legislation (by conveying a message that would appeal even to those who were unlikely ever to strike) and to tarnish lawmakers as anti-working family if they dared vote the wrong way. Indeed, in an attempt to stress the legislation’s broad appeal and contest “special interest” accusations, the AFL-CIO titled its Congressional bill, the “Caesar Chavez Workplace Fairness Act.”148 Business groups also needed to exercise caution when claiming public support for the right to replace strikers. Mainstream employer associations acknowledged that the strident behavior of firms such as Eastern Airlines, International Paper, and Phelps Dodge had caused widespread public disconcertion, and recognized the importance of lobbying vigorously against the striker replacement bill without creating the impression that they opposed the right to strike or supported union busting. For the CoC, NAM, and NFIB, this meant distancing their campaign from the more bullish and confrontational approach of the NRTWC. Indeed, at a founding meeting of what would become the Alliance to Keep Americans Working, one House Republican urged the business community to avoid the “appearance of being an arm of Right to Work.”149 Management associations particularly wanted to avoid the campaign turning into a debate on the policy implications of a few high-profile strikes involving permanent replacements, fearing that these often bitter confrontations would persuade the public of the need for legislation protecting strikers (see Table 2 for list of high-profile strikes in 1980s and 1990s). Unions fully understood the propaganda

224

JOHN LOGAN

Table 2. High-Profile Strikes Involving Permanent Replacements in the 1980s and 1990s (Compiled by the Author). Company

Year

Number of Strikers

Federal Air Traffic Control Brown and Sharpe Manufacturing PRC Recording Continental Airlines Phelps Dodge Magic Chef Greyhound Chicago Tribune Hormel Meat Trans World Airlines Colt Firearms Boise Cascade mill General Dynamic Corp. International Paper Company Port of New York Eastern Airlines Kraft/SS Pierce Central Industries Greyhound Ravenswood Aluminum Delta Pride Catfish Inc. Textron New York Daily News Caterpillar Diamond Walnut Growers Frontier Hotel Bridgestone-Firestone Pirelli Armstrong

1981 1981 1982 1983 1983 1983 1983 1985 1985 1986 1986 1986 1987 1988 1988 1989 1989 1989 1990 1990 1990 1990 1990 1991 1991 1992 1994 1994

12,000 1,600 250 13,000 2,400 1,300 12,000 1,000 1,500 6,000 1,100 1,200 3,500 2,500 2,100 30,000 188 230 9,300 1700 2,000 400 2,200 12,600 520 550 4,200 1,000

value of highlighting acrimonious strikes that the business lobby viewed as a “worst case scenario,” such as the strike at the New York Daily News: Business Associations blame Charles Brumback [combative CEO of the Chicago Tribune, parent company of the Daily News] for raising the public consciousness about the replacement worker issue – the last thing that unfair employers want. Recent polls show that the more the public learns about permanent replacement, the more they favor legislation to protect workers.150

As a result, unions attempted to make controversial strikes the focal point of the debate, using them to illustrate how permanent replacements turned limited labor disputes into all-out wars (precisely what the collective bargaining process was

Labor’s “Last Stand” in National Politics?

225

intended to avoid) that threatened the stability, and even the continued existence, of the enterprise. One AFL-CIO official wanted to make the Daily News the “centerpiece” of its campaign activities because that dispute provided the ideal opening to discuss why permanent replacements “have to go . . . They turn every fight into a fight to the death.” AFL-CIO President Lane Kirkland raised the issue of high profile strikes at every opportunity, arguing that while only a minority of employers practiced permanent replacement, its use by nationally recognized firms such as Greyhound presented “a direct threat to the vitality” of the entire collective bargaining system.151 The AFL-CIO repeatedly cited high-profile disputes as examples of egregious incidents of permanent replacement that demonstrated the pressing need for legislation and used them to demonstrate the wide-ranging nature of the issue: employers who had initiated strikes to unload unwanted unions, avoid signing first contracts or demand sweeping concessions to existing contracts; strikes over health benefits, pensions or safety conditions, often provoked by employers seeking to shift costs onto employees; employers who advertised for, recruited, and even trained permanent replacements during contract negotiations;152 and strikes where the recruitment of permanent replacements had needlessly “poisoned” labormanagement relations for years after the termination of the dispute. In short, the AFL-CIO claimed that striker replacement legislation was necessary because an increasing number of labor disputes now involved corporate strategies to use the threat or implementation of permanent replacements in a way that fundamentally undermined the process of collective bargaining: For decades both workers and employers viewed the strike as an interruption, not a termination, of the labor-management relationship. And when the strike was settled, workers returned to their jobs, and the dialogue between labor and management continued. Now, too many employers see a strike as the management weapon of choice . . .. Companies that hire permanent replacement workers have one goal – to break a union.153

Shortly before the 1992 Senate vote, the Frontier Hotel in Las Vegas and Diamond Walnut Growers in Stockton, California replaced the Daily News as the posterchildren for the striker replacement campaign. Prior to the 1990s, many companies that had hired permanent replacements were experiencing tough economic times and thus were seeking to reduce their labor costs. But Frontier and Diamond Walnut (the country’s largest walnut processor that employed mostly Latino women, earning $5–6 an hour) were financially successful firms that had provoked strikes by demanding sweeping reductions to wages and benefits established through collective bargaining.154 Two years later, a strike at Bridgestone-Firestone became the new focus of labor’s campaign. Beginning in July 1994, just days before the Senate vote, the strike escalated dramatically after the company announced its

226

JOHN LOGAN

intention to recruit permanent replacements in January 1995. President Clinton subsequently announced that Bridgestone’s actions had demonstrated exactly why striker legislation was needed: “When companies replace their workers . . . they sow the seeds of distrust and resentment which can extend far beyond their company, undermining labor-management relations across the land.”155 Organized labor’s allies in Congress also sought to spotlight the practices of influential companies that had recruited permanent replacements, but met with limited success. The House Subcommittee on Labor-Management Relations was forced to cancel hearings on firms that had hired permanent replacements when representatives of Eastern Airlines, Greyhound, Phelps Dodge, International Paper, Chicago Tribune, and Kraft/S. S. Pierce refused to testify.156 In an attempt to deflect attention away from the potentially damaging issue of these high-profile strikes, opponents of the bill acknowledged that certain employers had on occasion abused the law on permanent replacements, but denied that these exceptional cases justified a major overhaul of the law. Labor Secretary in the Bush Administration, Lynn Martin conceded that particular employers had behaved in a way that “brings support from no one,” but that “did not and should not” mean that Congress should change the rules on permanent replacements.157 The Republican whip in the House, Newt Gingrich, also acknowledged the existence of disputes where it was “clear that management has deliberately maneuvered to destroy the union” with permanent replacements, but explained that Congressional Republicans were working with the Bush Administration to “develop a true balance between the two opposing concerns and market realities.”158 Despite the efforts of the business lobby and its allies to play down the significance of high profile strikes, the issue generally benefited the bill’s supporters by allowing them to document the individual human tragedy that was often a consequence of permanent replacement. The debate on opinion polls was more of a mixed blessing for labor: polls demonstrated that, while the public often sympathized with the plight of replaced strikers and was uneasy with, if not disapproving of, the actions of firms such as Greyhound and Caterpillar, those favorable attitudes did not necessarily translate into support for a blanket ban on permanent replacements.

Lessons from the Experience of Striker Replacement Legislation in Canada Labor and management both argued that the experience with striker replacement laws in Canada offered backing for their predictions on the likely impact of similar legislation in the U.S. Drawing on the conclusions of two academic studies,

Labor’s “Last Stand” in National Politics?

227

employer organizations repeatedly claimed that the Quebec experience with striker replacement legislation (the province had proscribed striker replacements since the mid-1970s) had demonstrated that it had increased both strike incidence and strike duration in that province. The CoC concluded that these studies indicated that outlawing permanent replacements would significantly heighten industrial strife and asked, “Shouldn’t the United States learn from others’ mistakes?”159 In response, the AFL-CIO pointed out that Quebec’s legislation outlawed both temporary and permanent replacements, while the proposed U.S. legislation would ban only permanent replacements, and stressed that several U.S. studies had concluded that strikes involving permanent replacements lasted longer than those that did not.160 All eleven jurisdictions in Canada had legislation or jurisprudence protecting the rights of economic strikers, moreover, indicating that lawmakers believed that permanent replacements were incompatible with free collective bargaining; and most of that legislation dated back to the 1970s and 1980s – a period during which Canada’s record of job creation had outstripped that of the U.S.161 Thus, supporters of the bill argued that the Canadian experience suggested that legislation banning permanent replacements was compatible with both positive economic performance and the Wagner model of free collective bargaining. Nevertheless, the comparison with Canada, whose strike and unemployment rates were higher than those of the U.S., allowed the business lobby to raise questions about the desirability of legislation outlawing permanent replacements.

The Role of the State in the Regulation of Collective Bargaining The striker replacement campaign was, at a fundamental level, a conflict over two competing visions of the role of the state in the regulation of collective bargaining. Not surprisingly, both sides argued that their position on permanent replacements was consistent with the intent of the NLRA. Management representatives claimed the bill would unjustifiably overturn a half-century of labor policy that had consistently protected both unions’ right to strike and employers’ right to recruit permanent replacements and destroy the “delicate balance” of power in labormanagement relations that had been established by the NLRA. A ban on permanent replacements amounted to a dramatic change in the relationship between labor and management that could be enforced only through an excessive degree of state intrusion in the process of collective bargaining. Government policy should not attempt to influence the outcome of collective bargaining by sanctioning “risk free” strikes, employers claimed, but establish the rules and procedures of bargaining, then allow the “free play of economic forces” to determine the outcome of the process.162 The Washington Post believed that it was “one thing to try to keep

228

JOHN LOGAN

the collective bargaining system functioning fairly, quite another to get into the business of trying to ordain results.”163 And former NLRB General Counsel turned management lawyer Peter Nash claimed that the striker replacement bill ran counter to the “very purpose” of the NLRA, which had never intended to dictate a particular outcome but instead sought to promote negotiations “within the confines of the marketplace.” Employers’ right to hire permanent replacements, Nash insisted, was simply “part and parcel of the free market interplay” anticipated by the authors of the NLRA.164 Supporters of reform argued that the Mackay doctrine had created a “legal loophole” that had enabled employers to render powerless labor’s strike weapon and which now threatened to undermine entire postwar system of collective bargaining. One union leader believed if a worker did not enjoy right to strike without fear of replacement, “then the whole collective bargaining system collapses” and warned that if the campaign to outlaw permanent replacements were to fail, “those who wish to see America as a union-free society will surely win. This may prove to be organized labor’s last stand in national politics!”165 In making these arguments, however, both sides effectively distorted the intent of the NLRA. Employers were correct in claiming that Senator Wagner and his allies had deliberately restricted legal regulation of the economic weapons of both unions and employers during the process of bargaining; thus, the Mackay decision was not strictly a “loophole” and permanent replacements were technically compatible with the law’s model of limited state intervention. But the authors of the NLRA had limited state intrusion because they viewed the purpose of the law as promoting, not interfering with, free collective bargaining; the manner in which certain employers had used permanent replacements in the 1980s and 1990s, in contrast, had clearly impeded the process of collective bargaining.166

The Extent of the Actual Use of Permanent Replacements Labor and management clashed repeatedly over the extent to which employers actually hired permanent replacements, drawing opposite conclusions from major studies on striker replacements conducted by the General Accounting Office (GAO) and the Bureau of National Affairs (BNA). Supporters of the bill insisted that more firms were recruiting permanent replacements, especially after the 1981 PATCO strike, which had emboldened employers to resist strikes more aggressively. The AFL-CIO cited the BNA’s special reports on “Employer Bargaining Objectives” as evidence of how widespread had become the practice of using or threatening to use economic strikers. The BNA’s 1991 report found that 82% of employers would use or consider using replacements in the event of a strike,

Labor’s “Last Stand” in National Politics?

229

while only 16% of employers stated that they would not use replacements; the following year, 80% of employers surveyed stated that they would use replacements or consider doing so, with 18% saying they would not hire replacements during a strike. In 1993, the BNA report showed that 79% of employers would replace strikers or would consider doing so (19% would hire permanent replacements), while the following year, 82% of respondents stated they would attempt to replace strikers, with 21% saying they would hire permanent replacements.167 For the IUD, these results were “very significant” because they documented “just how widespread has become the employer tactic of firing strikers fostered by the Reagan and Bush Administrations.” And it believed that, by demonstrating the “almost universal” threat of permanent replacement, the BNA reports would provide “invaluable” evidence in their efforts to convince Congress and public of the urgent need for legislation protecting strikers.168 The GAO study, however, was more problematic for organized labor. It had found that employers announced that they would hire permanent replacements in about one-quarter of strikes in 1985 and 1989 (23% in 1985, 30% in 1989). Two-thirds of employers and almost 90% of unionists surveyed, moreover, believed that firms had recruited permanent replacements more frequently in the late 1980s than they had done a decade earlier.169 The AFL-CIO argued that, like the BNA reports, the GAO study had demonstrated that the use and threatened use of permanent replacements was widespread and increasing; but it also recognized that the business lobby would seize upon the GAO figures to back up its claim that the actual incidence of permanent replacement was surprisingly low and that drastic overhaul of labor policy was therefore unwarranted. The IUD warned that given the probability that the GAO numbers would be “re-cooked in a way that seriously undermines our case,” labor must emphasize the frequency with which the threat of permanent replacement arises at the bargaining table and during organizing campaigns – “i.e. in the vast majority of cases.”170 For the AFL-CIO and most affiliates, the permanent replacement issue was always as much about the threat of replacement as it was about the actual use of replacements, though this proved a difficult concept to convey to lawmakers and the public. By the early 1990s, the threat of replacement had become an almost standard feature of both bargaining campaigns and organizing drives. In bargaining campaigns, employers threatened replacement to demand concessions from existing wages and benefits, to prevent unions securing first contracts, or to unload unwanted unions.171 Employers threatened replacement during organizing drives to dissuade employees from voting for unionization. The AFLCIO frequently cited Caterpillar’s threat to replace 12,600 striking UAW members as the deciding factor in the collapse of that union’s bargaining position, while many unions reported having similar experiences with the threat of replacement.

230

JOHN LOGAN

The Utility Workers union explained that their members had failed to support strike actions on numerous occasions due to the threat of replacement and in every case the employer “unilaterally implemented its last, best and final offer.” Employers had frequently instructed bargaining committees of the Paperworkers union to “take it or leave it . . . and raised the right to permanently replace employees if they struck in every campaign.”172 And one management representative cogently expressed the problem of the threat of replacement during organizing campaigns: Aside from the diminution of a worker’s right to strike, a more chilling impact of this phenomenon exists in the organizing context. In considering whether to vote for or against unionization, a worker often votes “no” because of the fear associated with being replaced in the event of a work stoppage. This message is invariably articulated by management and can often mean the difference... between winning and losing NLRB elections.173

The business lobby also stressed that the prohibition of permanent replacements would provide unions with a powerful weapon during organizing campaigns. The American Mining Congress insisted that striker replacement legislation would unfairly assist unions attempting to organize non-union mines: Unions would be able to use the law as an important part of their organizing activities. Promising higher wages and shorter hours (which they could deliver because of their unlimited right to strike), union leaders would be able to persuade many miners to vote for union representation where none existed before.174

The Alliance to Keep Americans Working also worried that the strike bill would result in “coerced organizing” because unions could gain a majority of signatures on authorization cards, then stage recognition strikes that inevitably would succeed. The effect would be to “greatly lower the threshold for organizers,” thus enabling them to unionize workers who would “never choose to do so themselves” in secret ballot elections. The CoC castigated the bill as a cynical attempt by desperate unions to “energize their moribund organizing activities,” and predicted that labor leaders would exploit their expanded rights to “terrorize countless union-free employers.” And the NFIB opposed the strike bill, in part, because Congress “should not be in the business of enhancing union organizing tools.”175 According to the business lobby, employers’ use of permanent replacements had not increased significantly since the 1970s and the actual incidence of permanent replacement was relatively low. It accused unions of exaggerating the extent of the permanent replacement “problem” and exploiting public concerns over a few high profile strikes to promote special interest legislation. As the IUD had predicted, employers used the GAO study to sustain their contention that striker replacement legislation was a “solution in search of a problem.” The GAO report had indicated that only 3% of strikers were replaced during strikes and that a majority of them returned to work within a few weeks, the CoC argued.

Labor’s “Last Stand” in National Politics?

231

In 1991–1992, only 40,000 economic strikers were replaced (0.03% of the total workforce), and almost two thirds of them were later reinstated; thus, the 14,000 workers who were permanently replaced that year represented just 0.01% of the total U.S. workforce. The “hue and cry” over permanent replacement, one employer representative concluded, was a “classic” incident of lawmakers overreacting to a handful of high profile strikes.176 Business organizations also cited a study by management attorney Daniel Yager that found no major increase in the frequency with which employers had recruited permanent replacements during the halfcentury since the enactment of the NLRA. Supporters of the bill dismissed Yager’s conclusions, but the “numbers debate” damaged labor’s contention that legislation was required urgently and unions frequently encountered a “why now” response when lobbying “swing senators.”177 In response to the AFL-CIO’s campaign, one of its top Senate targets, James Jeffords of Vermont, stated that the “undisputed data shows that permanent replacement of strikers occurs only very rarely. . . . I simply cannot conclude that the legislative reversal of more than fifty years of judicial precedent is necessary.”178 An almost complete lack of hard data on the use of permanent replacements made it doubly difficult for the AFL-CIO to refute the contention that the actual recruitment of permanent replacements was rare and not a growing trend. Few state labor departments compiled accurate figures on permanent replacements and while individual unions were keenly aware of the issue, they seldom kept records on its scale.179 Despite this paucity of information, the AFL-CIO was convinced that, by focusing exclusively on large strikes, studies such as those by the BNA and GAO had consistently understated the scale of the permanent replacement problem. It contended that smaller companies frequently hired permanent replacements, but their actions went largely unnoticed, except in the local community. Union officials also believed that the number of employers using permanent replacements would be much higher were it not for the precipitous decline in economic strikes, which itself could be attributed to the threat of permanent replacement.

CONCLUSION In 1992 and 1994, a majority of both Houses of Congress supported striker replacement legislation, and in 1994 the White House also favored the bill. So, why did organized labor lose the struggle for legislation outlawing permanent replacements? The immediate reason that the AFL-CIO’s campaign ended in defeat on both occasions was its inability to win the 60 Senate votes required to overcome the Republican-led filibuster, including those of several key southern Democrats and moderate Republicans, several of whom resented being forced to take sides on such

232

JOHN LOGAN

a controversial issue. Opposition to the bill in the Senate hardened after 1992 and the one thing that might have delivered the necessary votes – enthusiastic support from the White House – was not forthcoming. Although officially in favor of the legislation, the Clinton Administration never expended the same political capital to ensure its passage that it had on behalf of NAFTA. Indeed, Bill Clinton and Al Gore were able to deliver only one vote in favor of the 1994 bill from the four Democratic Senators representing their home states, Arkansas and Tennessee. In addition, the AFL-CIO had trouble convincing non-unionists that the striker replacement bill was more than simply special interest legislation and failed to win decisively the argument on the impact of the bill on economic competitiveness and business performance. Rather than serve as a stage-setter for more extensive legal reforms, as the AFL-CIO had originally hoped, labor’s second failure within two years damaged the campaign for comprehensive revision of the NLRA and reinforced its reputation for failure in legislative battles. These defeats ultimately offered further evidence of organized labor’s impotence in the face of sustained employer opposition and provided a trial run for the battle over the Dunlop Commission. It remains to be seen whether the defeat of strike replacement legislation effectively proves to be organized labor’s last major campaign for labor law reform.

NOTES 1. Daily Labor Report, “Management Attorney Says Striker Replacement Measure will Tilt Labor Laws Against Employers,” March 6, 1991, A-2/3. 2. From the 1950s–1990s, several states and cities passed laws limiting employers’ use of permanent replacement. Laws enacted in the 1980s and 1990s included the following: St. Louis and several other cities enacted laws requiring that employer inform replacements that they were taking the place of strikers; Rhode Island, Maine, Minnesota and Columbus, Ohio banned permanent replacements but the courts subsequently ruled that the laws were pre-empted by the NLRA; New York considered legislation that would have required the termination of replacements no later than 14 days after the end of the strike. Most of these laws attempted to restrict the way in which employers could hire or locate replacements; once the employer had hired replacements, federal regulations pre-empted any local laws. However, most of these laws were overturned in the courts (because they were pre-empted by the NLRA) or subsequently fell into disuse. On earlier state efforts to limit replacements, see Joseph Sales, “Replacing Mackay: Strikebreaking Acts and Other Assaults on the Permanent Replacement Doctrine,” Rutgers Law Review, Vol. 36 (1984), pp. 861–886. 3. In the 1990s, opponents of striker replacement legislation claimed the Carter Commerce Department had concluded that a ban on permanent replacements would promote strikes, contribute to inflation and upset the balance of power between unions and management in collective bargaining. Barbara Townley, Labor Law Reform in US Industrial Relations (Brookfield, VT, Gower, 1986); Daily Labor Report, “Labor-Management Group Addresses Replacement Hiring Issue in Report,” March 19, 1991, A-7/9.

Labor’s “Last Stand” in National Politics?

233

4. Daily Labor Report, “AFL-CIO Stokes Publicity Campaign for Bill to Ban Strike Replacements,” August 14, 1990, A-8. 5. Frank D. Martino, President, International Chemical Workers Union, letter to Howard Samuel, IUD, AFL-CIO, September 25, 1990; Industrial Union Department records (unprocessed collection), George Meany Center archives, Silver Spring, MD. Unless otherwise stated, manuscript sources are from this collection. 6. Richard Trumka, President, United Mine Workers of America, “A Grass Roots Campaign Strategy in the State of Alabama for Passage of HR 3936 and S 2112,” (1991). 7. Frank Swoboda, “The AFL-CIO Begins Campaign on Replacement Workers,” Washington Post, June 21, 1991, F1. 8. Roger Keeran and Greg Tarpinian, “Public Policy and the Recent Decline of Strikes,” Policy Studies Journal, 18:2 (Winter 1989–1990), pp. 461–470. 9. “Passing Striker Replacement Legislation: Strategy Outline for Field Activities,” Industrial Union Department, AFL-CIO, May 31, 1990; Daily Labor Report, “Unions Must Make Labor Law Reform Issue In Congressional Elections, Donahue Says,” May 17, 1994, A-14/15. 10. International Brotherhood of Teamsters, “Labor Law Reform: IBT Working Group – Initial Discussion,” (December 1992). 11. Daily Labor Report, “Employers Converge on Capitol to Oppose Senate Passage of Striker Replacement Bill,” July 15, 1993, A-11/12. 12. Daily Labor Report, “AFL-CIO Stokes Publicity Campaign for Bill to Ban Strike Replacements,” August 14, 1990, A-8. 13. IUD Report on Striker Replacement Campaign, July 20, 1990. 14. Robert Thompson, “An Anti-Worker Labor Bill,” Wall Street Journal, August 31, 1990; Daily Labor Report, March 5, 1991, A-11. 15. During the primary campaign, Clinton had walked the picket line with strikers at Caterpillar; the company subsequently threatened to permanently replace the striking UAW members. Daily Labor Report, June 26, 1991, A-20; Daily Labor Report, “AFL-CIO’s Kirkland Seek Commitments From Democratic Presidential Hopefuls,” November 13, 1991, A-2/4. 16. Robert Reno, “Rights of Strikers: Some People Just Don’t Understand,” Star Tribune, May 12, 1993, 17A; Chamber of Commerce, “Business Says ‘No’ to Strike Bill, ‘No’ to Any Compromise,” (June 14, 1994). 17. Scot Lautenschiager, “Unions Back Bill to Ban ‘Scabs,’“ State Journal (Madison, WI), September 23, 1990, p. 29; Daily Labor Report, “Right to Work Targets Senators in Campaign to Defeat Striker Replacement Legislation,” May 23, 1994, A-6/7. 18. Alliance to Keep Americans Working, Membership list. 19. William N. Miller, “New Year’s Big Labor Issue,” Industry Week, July 2, 1990, pp. 56–57. 20. Howard D. Samuel, IUD Executive Council, January 7, 1991; Brian Turner, IUD, Memo to Howard Samuel, IUD, April 26, 1991. 21. Bernard Rapoport, Chief Executive Officer, American Income Life Insurance Company, letter to Jack Rosenthal, New York Times, February 13, 1991. 22. Trans World Airlines v. Independent Federation of Flight Attendants, 489 U.S. 426 (1989). 23. Henry A. Duffy, President, Air Line Pilots Association, letter to Howard Samuel, IUD, March 29, 1990.

234

JOHN LOGAN

24. Daily Labor Report, “Striker Replacement Bill Clears House, Senate Panels,” May 6, 1993, AA-1/4. 25. 280 NLRB 597. 26. Robert Willis, President, American Federation of Grain Millers, letter to Howard Samuel, March 27, 1990; Howard Samuel, IUD, letter to Robert Willis, April 4, 1990. 27. Daily Labor Report, “AFL-CIO Launches Campaign for Striker Bill,” June 21, 1991, A-19. 28. IUD, “Passing Striker Replacement Legislation: Strategy Outline for Field Activities,” May 31, 1990. 29. IUD Report on Striker Replacement Campaign, July 23, 1990. 30. Howard Samuel, President, IUD, letter to William Lucy, Secretary-Treasurer, American Federation of State, County, and Municipal Employees, August 9, 1990. 31. Labor’s campaign recruited the support of so many religious leaders that its opponents accused unions of implying that “God was on their side.” 32. Howard Samuel, President, IUD, letter to William Lucy, Secretary-Treasurer, American Federation of State, County, and Municipal Employees, August 9, 1990. 33. Benjamin L. Hooks, Executive Director/CEO, NAACP, letter to Howard Samuel, IUD, June 3, 1991; Black Leadership Committee for Workplace Fairness, May 30, 1991. 34. IUD, “Targeted States for Expanded S. 55 Grassroots Activities,” (April 1994); Howard Samuel, memo to Joe Uehlein, AFL-CIO, March 7, 1990. 35. Richard Trumka, President, United Mine Workers of America “A Grass Roots Campaign Strategy in the State of Alabama for Passage of HR 3936 and S 2112,” (1991). 36. UAW Legislative Department, Congressional Legislative Advisory, June 6, 1991. 37. Michael Basharah, Memo to Bill Klinefelter, RE: H.R. 5 Vote Analysis, July 18, 1991; International Paper Company, Washington Newsletter, July 19, 1991. 38. International Paper Company, Washington Newsletter, July 19, 1991. 39. Daily Labor Report, “Employer-Sponsored Poll Finds Majority Think Striker Replacements Should Keep Jobs,” March 12, 1991, A-3. 40. Peter Kilborn, “Ban on Replacing Strikers Faces Veto Threat,” New York Times, March 7, 1991. 41. Michael Szpak, memo to Mike Gildea, AFL-CIO, RE: Towards an Executive Branch Strategy: Life After Congressional Passage of the Workplace Fairness Bill (no date); Brian Turner, Executive Assistant to the President, IUD, Campaign Coordinator, Americans for Workplace Fairness, letter to Gail Dratch, S. 55 State Coordinator for New York, Political Director, Garment Workers Union, October 9, 1991; Tom Harkin, U.S. Senate, letter to Arkansans for Workplace Fairness, September 11, 1991. 42. Kirk Victor, “Disorganized Labor,” National Journal, September 21, 1991, pp. 2256–2261. 43. National Federation of Independent Business, “NFIB Opposes H.R. 5, The Strike Bill” (1992); Statement of Jackson Faris, President of the NFIB Before the Senate Subcommittee on Labor on S.55, Striker Replacement Legislation, March 30, 1993; Frank Swoboda, “Labor Faces Senate Defeat on Striker Replacements,” Washington Post, July 12, 1994, C1. 44. Alliance to keep Americans Working, Editors Against Striker Replacement Legislation (no date). 45. “The Striker Replacement Bill,” editorial, Washington Post, April 27, 1993, A16. 46. “Labor and the Democrats,” Business Week, July 15, 1991, p. 41.

Labor’s “Last Stand” in National Politics?

235

47. Given the AFL-CIO’s historical opposition to legal restraints on the right to strike, Packwood expressed surprise that it was prepared to sacrifice certain aspects of the freedom to strike. Not all unions supported the amendment, however; IBT President Ron Carey later wrote that the Packwood amendment “was and continues to be unacceptable” to his union. Ron Carey, President, International Brotherhood of Teamsters, letter to Lane Kirkland, President, AFL-CIO, March 7, 1994. 48. Daily Labor Report, “Senate Fails to Invoke Cloture on Striker Replacement Bill,” June 12, 1992, A-10/12; AFL-CIO, Department of Information, News Release, June 12, 1992; Labor Policy Association, “The Striker Replacement Bill: A Compromise in the Works,” April 27, 1994. 49. Daily Labor Report, “Republican Senators, Business Leaders Oppose to any Compromise on Striker Bill,” May 5, 1994, A-1/2; Daily Labor Report, “Senate Vote Kills Bill to Restrict Use of Permanent Striker Replacements,” June 17, 1992, A-9/11. 50. Prior to the vote, Presidential candidate Clinton had “assured” Paperworkers president Wayne Glenn that Arkansas’s two Senators would vote with labor on cloture and final passage. Even after the first vote, in which both Senators sided with the business lobby, the AFL-CIO believed that Bumpers and Pryor were “committed to voting right on a second cloture vote.” But both again voted against cloture on the second vote. Brian Turner, IUD, Memo to S. 55 Workplace Fairness Campaign State Coordinators, RE: Urgent: Final Weekend Push for Victory in Senate Cloture Vote, June 11, 1992. 51. Daily Labor Report, “Senate Vote Kills Bill to Restrict Use of Permanent Striker Replacements,” June 17, 1992, A-9/11; Peter Kilborn, “Ban on Replacing Strikers Faces Veto Threat,” New York Times, March 7, 1991; Frank Swoboda, “House Bans Use of Permanent Substitute Workers in Strikes,” Washington Post, July 18, 1991, p. B9; “Death Knell May Be Ringing For Union Strikes,” HR Focus, October 1992, p. 1. 52. Brian Turner, Memo to Howard Samuel, IUD, RE: Evaluation of House “Don’t Fire Strikers” Grassroots Campaign: New Line Implications for the Senate Campaign, July 22, 1991. 53. The AFL-CIO expressed “surprise” that it had failed to gain the votes of Bumpers and Pryor, and believed that, had they supported the bill, several other swing Senators “would have come under pressure to give us the 60th vote.” Brian Turner, Memo to Howard Samuel, RE: Evaluation of the HR 5/S 55 Workplace Fairness Grass-Roots and Coalition Campaign (no date). 54. Harold McIver, Organizing Director, AFL-CIO, Memo to IUD Organizing Directors, RE: Meeting to Discuss Changes in Labor Law and Regulations, November 17, 1992. 55. Richard L. Trumka, President, United Mine Workers of America, letter to Lawrence Bankowski, President, American Flint Glass Workers Union, November 25, 1992; International Brotherhood of Teamsters, “Labor Law Reform: IBT Working Group – Initial Discussion,” (December 1992); Wayne E. Glenn, President, United Paperworkers International Union, letter to Elmer Chatak, President, IUD, December 2, 1992; Kenneth L. Coss, International President, United Rubber Workers, letter to Elmer Chatak, President, IUD, December 10, 1992: Bill Casstevens, Secretary-Treasurer, United Auto Workers, letter to Lane Kirkland, July 27, 1993. 56. IUD, Committee on Organizing and Labor Law, Staff Meeting – December 9, 1992, Summary Notes; Daily Labor Report, “Senator Simon Says that Striker Bill Unlikely to Pass Senate,” May 6, 1994, A-18.

236

JOHN LOGAN

57. Howard Samuel, President, Industrial Union Department, AFL-CIO, Memo to Jay Mazur, President, International Ladies’ Garment Workers Union, January 10, 1990. 58. Joe Uehlein, AFL-CIO, memo to Elmer Chatak, Pete DiCicco, RE: Striker Replacement, December 23, 1992; William Klinefelter, Director of Legislation, AFL-CIO, Memo to Elmer Chatak, President, IUD, January 5, 1993. 59. Later that month, IUD President Elmer Chatak (who had replaced Brian Turner) requested that unions send as a matter of urgency information on recent or current striker replacement situations to Senator Metzenbaum’s office. Elmer Chatak, President IUD, AFLCIO, Memo to Members, IUD Executive Council, RE: Striker Replacement Legislation, March 31, 1993; Daily Labor Report, “Organized Labor Launches New Effort for Passage of Striker Replacement Bill,” March 17, 1994, A-11/13. 60. “Diverting Labor’s Lobbyists,” Legal Times, October 4, 1993. 61. Steven Hahn, IUD, AFL-CIO, Memo to Elmer Chatak and Joe Uehlein, RE: Public Relations Plan for the S.55 Campaign, March 21, 1994. 62. Daily Labor Report, “Organized Labor Launches New Effort for Passage of Striker Replacement Bill,” March 17, 1994, A-11/13. 63. Steven Hahn, IUD, AFL-CIO, Memo to Elmer Chatak and Joe Uehlein, RE: Public Relations Plan for the S.55 Campaign, March 21, 1994. 64. Kevin Murphy, IUD, memo to IUD Coordinating Committee, RE: Campaign Update, April 18, 1994; Daily Labor Report, “NAM Targets Senators on Striker Replacement,” April 21, 1994, A-15/16; Daily Labor Report, “Right to Work Targets Senators in Campaign to Defeat Striker Replacement Legislation,” May 23, 1994, A-6/7. 65. Stephen Franklin and Michael Arndt, “Time Grows Short in Battle Over Striker Replacement Bill,” Chicago Tribune, June 14, 1994, A1; Robert M. McGlotten, Director, Department of Legislation, Memo to State Federations, Regional Directors, COPE Regional Directors, Central Labor Councils, May 24, 1994; Daily Labor Report, “Employers Converge on Capitol to Oppose Senate Passage of Striker Replacement Bill,” July 15, 1993, A-11/12. 66. James Jeffords, U.S. Senate, letter to Paul Rohmann, June 14, 1994. 67. Daily Labor Report, “Defeat of Striker Replacement Bill A Victory for Business Coalition,” July 14, 1994, AA-1/3; Paul Massey, Printing Industry Association of Georgia, Inc., Memo to All PIAG CEO’s, July 7, 1994. 68. “Pandering to Big Labor,” Editorial, Albuquerque Journal, July 23, 1991. 69. Kevin Murphy, IUD, Memo to IUD Coordinating Committee for S.55, RE: Campaign Update, June 22, 1994, original emphasis. 70. Kevin Salwen, “Foley Says House Will Vote to Shield Jobs of Strikers,” Wall Street Journal, February 16, 1993. 71. Ron Carey, President, International Brotherhood of Teamsters, letter to Lane Kirkland, President, AFL-CIO, March 7, 1994. 72. Daily Labor Report, “Organized Labor Launches New Effort for Passage of Striker Replacement Bill,” March 17, 1994, A-11/13; Daily Labor Report, “Senate Striker Replacement Vote Expected by May,” March 15, 1994, A-15. 73. Elmer Chatak and Peter DiCicco, Memo to Members, IUD Executive Council, RE: S.55 – Striker Replacement Campaign, March 18, 1994. 74. President Bill Clinton, letter to Senator Dale Bumpers, June 21, 1994. 75. Daily Labor Report, “Backers of Striker Replacement Bill Look to Clinton for Help in Senate,” June 17, 1993, A-11/12.

Labor’s “Last Stand” in National Politics?

237

76. Pete DiCicco, AFL-CIO Memo to Elmer Chatak, RE: S.55 Senate Cloture Vote Evaluation, May 6, 1994. 77. Liann Ainsworth, Arkansas AFL-CIO Memo to Elmer Chatak, Joe Uehlein, RE: Arkansas Report Summary, July 27, 1993. 78. Mimi Dortch, Executive Director, Arkansas Interfaith Conference, letter to President Bill Clinton, May 11, 1994. 79. Reed Larson, “Arkansas Senators at Center of Major Labor Law Fight,” The Times (North Little Rock, Arkansas), March 12, 1994; Daily Labor Report, “Right to Work Targets Senators in Campaign to Defeat Striker Replacement Legislation,” May 23, 1994. 80. National Institute of Business Management, Business Briefing, 45:10 (March 7, 1994). 81. Daily Labor Report, “Backers of Striker Replacement Bill Look to Clinton for Help in Senate,” June 17, 1991, A-11/12; Mark Levinson, “One for the Rank and File,” Newsweek, July 19, 1993. 82. BNA Conference Report, Labor Relations Week, March 23, 1994; Daily Labor Report, “S.55 Compromise Being Discussed, Metzenbaum Says,” April 20, 1994, A-24. 83. Wayne E. Glenn, President, United Paperworkers International Union, letter to AFLCIO Executive Council, June 25, 1993. 84. In particular, the CoC viewed Senators Exon (Nebraska), Nunn (Georgia) and Graham (Florida) as possible brokers of a compromise bill. “Diverting Labor’s Lobbyists,” Legal Times, October 4, 1993. 85. NAM, Legislative Alert, “Time to Strike Down the Striker Bill,” (April 1994). 86. Richard Lesher, Chamber of Commerce, “The Voice of Business: Knowing When to Compromise,” May 2, 1994; Chamber of Commerce, “Moratorium = Strikes R U.S.,” (April 1994); Daily Labor Report, “Polls Show Two-Thirds of Americans Oppose Replacing Workers Who Strike,” June 15, 1994, AA-1/3; Daily Labor Report, “Senate Vote to End Filibuster on Striker Replacement Fails 53-47,” July 13, 1994, AA-1/2. 87. Richard Lesher, President, U.S. Chamber of Commerce, “Knowing When to Compromise,” The Voice of Business, May 2, 1994, original emphasis. 88. Peter Kilborn, “Struck Trucking Company Goes out of Business,” New York Times, April 12, 1994. 89. U.S. Chamber of Commerce, “Moratorium = Strikes R U.S.,” (April 1994); Labor Policy Association, “The Striker Replacement Bill: A Compromise in the Works,” April 27, 1994; Daily Labor Report, “Labor Targets Senators in Quest for Bill to Ban Striker Replacement,” April 11, 1994, C-1/3. 90. Wayne E. Glenn, President, United Paperworkers International Union, letter to AFLCIO Executive Council, June 25, 1993; Daily Labor Report, “Labor Targets Senators in Quest for Bill to Ban Striker Replacement,” April 11, 1994, C-1/3 91. Daily Labor Report, “No Act Scheduled in Senate on Striker Replacement Bill,” June 21, 1994, A-13/14. 92. Kevin Murphy, IUD, AFL-CIO, Memo to Gene Bruskin, RE: Background on S. 55 Vote for Jesse Jackson, May 13, 1994. 93. Kevin Murphy, IUD, Memo to IUD Coordinating Committee for S.55, RE: Campaign Update, May 9, June 22, July 13, 1994. 94. Elmer Chatak, IUD, Memo to Members of the U.S. Senate, RE: Workplace Fairness Act, July 10, 1994; Frank Swoboda, “Labor Faces Senate Defeat on Striker Replacements,” Washington Post, July 12, 1994, p. C1.

238

JOHN LOGAN

95. Taylor E. Dark, The Unions and the Democrats: An Enduring Alliance (Cornell UP, 1999), pp. 176–178. 96. Daily Labor Report, “Gore Pledges Another Try at Striker Replacement Ban,” July 19, 1994, A-1. 97. Helen Dewar and Frank Swoboda, “Republican-Led Filibuster Kills Striker Replacement Bill in Senate,” Washington Post, July 14, 1994, A7. 98. Elmer Chatak, Memo to All IUD Staff, RE: Senate Vote on Workplace Fairness . . . A Go! March 3, 1994. 99. Sam Nunn, U.S. Senate, letter to Betty Smith, August 12, 1994. 100. Howard Samuel, IUD, letter to Jack Sheinkman, President, ACTWU, September 3, 1991; Paul Massey, President, Printing Industries Association of Georgia, Inc., Memo to All PIAG CEO’s, July 7, 1994. 101. Daily Labor Report, “Poll Show Two-Thirds of Americans Oppose Replacing Workers who Strike,” June 15, 1994, AA-1/3; Bruce Vail, “Two-Thirds Back Bill That Protects Strikers,” Journal of Commerce, February 20, 1991. 102. Lane Kirkland, President, AFL-CIO, letter to Elmer Chatak, President, IUD, July 19, 1994. 103. Daily Labor Report, “Republican Senators, Business Leaders Opposed to Any Compromise on Striker Bill,” May 5, 1994, A-1/2. 104. Daily Labor Report, “Defeat of Striker Replacement Bill A Victory for Business Coalition,” July 14, 1994, AA-1. 105. Lane Kirkland, president, AFL-CIO, letter to President Bill Clinton, July 18, 1994. 106. Frank Swoboda, “Labor Faces Senate Defeat on Striker Replacements,” Washington Post, July 12, 1994, C1. 107. Daily Labor Report, “Defeat of Striker Replacement Bill A Victory for Business Coalition,” July 14, 1994, AA-1/3; Catherine Manegold, “State Republicans Deal a Major Defeat to Labor,” New York Times, July 13, 1994. 108. Daily Labor Report, “Senate Vote Kills Bill to Restrict Use of Permanent Striker Replacements,” June 17, 1992, A-9/11. 109. Lewis Maltby, Director, National Taskforce on Civil Liberties in the Workplace, letter to Joe Uehlein, IUD, November 17, 1995. 110. Doug Bandow, Cato Institute, “Pushing all buttons for strikers’ rights,” Washington Times, June 25, 1994, D3; Robert T. Thompson, “An Anti-Worker Labor Bill,” Wall Street Journal, August 31, 1990. 111. Capital Economics, “The Case for Maintaining Current Balance Between Management and Labor,” report prepared by David Kapan and Mark Gluek. 112. Stephen J. Cabot, “Labor Unions and Their Supporters Are Trying to Turn Back the Clock,” American Banker, September 17, 1993. 113. “Assisted Economic Suicide” (editorial), Detroit News, May 17, 1993. 114. Charles Perry, Andrew Kramer and Thomas Schneider, Operating During Strikes, Labor Relations and Public Policy Series No. 23 (Philadelphia: University of Pennsylvania, The Wharton School, Industrial Research Unit, 1982); Teresa Yates, Field Coordinator, ACLU National Taskforce on Civil Liberties in the Workplace, letter to Senator John Chafee, September 29, 1993. 115. Jeff Faux, President, Economic Policy Institute, letter to Frank Hanley, President, International Union of Operating Engineers, September 21, 1990.

Labor’s “Last Stand” in National Politics?

239

116. Howard Metzenbaum and William Ford, “Support for the Workplace Fairness Act,” Letter to Washington Post, May 7, 1993. 117. Stephen Sleigh, Michael Kapsa and Chris Hall, The Cost of Aggression (IUD, AFLCIO, Washington, DC, August 1992); Aaron Bernstein, “Busting Unions Can Backfire on the Bottom Line,” Business Week, March 18, 1991, p. 108. 118. Daily Labor Report, “Union and Management Advocates Spar over Need for Striker Replacement Law,” October 21, 1991, A-13; Robert T. Thompson, “An Anti-Worker Labor Bill,” Wall Street Journal, August 31, 1990; Chamber of Commerce, “Position Statement on S.55/H.R.5” (1994). 119. Daily Labor Report, “Striker Bill Unwise Policy For U.S., Bush Advisor Says,” May 16, 1991, A-6. 120. Statement of Robert B. Reich, Secretary of Labor, Before the Subcommittee on Labor Committee on Labor and Human Resources, U.S. Senate, March 30, 1993; Howard Metzenbaum and William Ford, “Support for the Workplace Fairness Act,” letter to Washington Post, May 7, 1993; Julius Getman, Professor of Law, University of Texas, Statement in Support of S. 2112, Before the Subcommittee on Labor Committee on Labor and Human Resources, U.S. Senate, March 30, 1993. 121. Lane Kirkland, President, AFL-CIO, letter to President Bill Clinton, July 18, 1994; Wayne E. Glenn, President, United Paperworkers International Union, letter to Elmer Chatak, President, IUD, December 2, 1992. 122. Cynthia Gramm, “Employers’ Decisions to Operate During Strikes: Consequences and Policy Implications,” in William Spriggs, ed., Employee Rights and the Changing Economy (New York: M.E. Sharpe, 1991). 123. Arthur Coia, president, Laborers International Union of America, letter to John H. Chafee, U.S. Senate, May 27, 1994. 124. President Bill Clinton, letter to Donald Fites, Chairman and Chief Executive Officer, Caterpillar Incorporated, July 12, 1994. 125. Joseph Uehlein, Memo to Howard Samuel, Elmer Chatak, RE: Don’t Fire Strikers Update, December 14, 1990. 126. Incensed business groups responded that his letter in support of the bill “may as well have been written by the AFL-CIO” and would undermine business confidence in the neutrality of the FMCS. They also later castigated NLRB Chairman Bill Gould for expressing support for a change in the rules on permanent replacement. Daily Labor Report, “Chief U.S. Mediator Says Use of Permanent Strike Replacement Make Bargaining Difficult,” October 21, 1991, A-5/7; John Calhoun Wells, Director, Federal Mediation and Conciliation Service, letter to Howard Metzenbaum, Chairman, Subcommittee on Labor, U.S. Senate, May 27, 1994; Daily Labor Report, “No Action Scheduled in Senate on Striker Replacement Bill,” June 21, 1994, A-13/14. 127. Bureau of Labor Studies, “Strike Statistics”; Peter Kilborn, “The Daily News Strike Tests The Will of Weakened Labor,” New York Times, January 27, 1991; “Trends in the Number of Strikes and Use of Permanent Strike Replacements in the 1980s,” Statement of Franklin Frazier, Director of Education and Employment Issues Human Resource Division, General Accounting Office, Before the Subcommittee on Labor, Senate Committee on Labor and Human Resources, June 6, 1990. 128. Daily Labor Report, “House Approves Bill That Would Ban Permanent Replacement of Economic Strikers,” June 16, 1993, AA-1. The BNA reached a more cautious conclusion, stating that it was “uncertain” whether employers’ more frequent use of

240

JOHN LOGAN

permanent replacements had contributed to declining strike activity, but union leaders were “clearly concerned that strikers will lose their jobs.” BNA Plus, “Replacement Workers: Evidence from the Popular and Labor Press, 1989 and 1990 (BNA, Inc.: Washington, DC, 1991). 129. Greg Tarpinian, letter to Brian Turner, IUD, AFL-CIO, August 28, 1991; Roger Keeran and Greg Tarpinian, “Public Policy and the Recent Decline of Strikes,” Policy Studies Journal, 18:2 (Winter 1989–1990), pp. 461–470). 130. Daily Labor Report, “Senate Vote to End Filibuster on Striker Replacement Fails, 53–47,” July 13, 1994, AA-1; Robert T. Thompson, “An Anti-Worker Labor Bill,” Wall Street Journal, August 31, 1990; Peter Eide, Manager of Labor Law, U.S. Chamber of Commerce, quoted in Roland Klose, “Replacing Strikers Increases as Option,” Commercial Appeal (Memphis, TN), December 30, 1990, p. 38. 131. NRTWC, “Issue Briefing Paper: Pushbutton Strike Bill” (February 1991); Bruce Josten, Senior Vice President, Membership Policy Group, Chamber of Commerce, letter to Edward M. Kennedy, U.S. Senate, June 29, 1994. 132. Daily Labor Report, “Debate on Hiring of Strike Replacements Featured at New York Labor Conference,” June 12, 1990, A-1/3. 133. William N. Miller, “New Year’s Big Labor Issue,” Industry Week, July 2, 1990, 56–57; Senator Dan Coats (R-Indiana) quoted in Daily Labor Report, “Senate Labor Committee Clears Striker Replacement Legislation,” June 20, 1991, A-15; Daily Labor Report, “Striker Replacement Bill Clears House, Senate Panels,” May 6, 1993, AA-1/2. 134. Janice Castro, “Labor Draws an Empty Gun,” Time, March 26, 1991, pp. 56–59. 135. Arthur Coia, president, Laborers International Union of America, letter to John H. Chafee, U.S. Senate, May 27, 1994. 136. Vic Fingerhut, President, Fingerhut/Granados Opinion Research Co., Memo to Elmer Chatak, IUD, AFL-CIO, June 2, 1994. 137. Bruce Josten, Senior Vice President, Membership Policy Group, Chamber of Commerce, letter to Edward M. Kennedy, U.S. Senate, June 29, 1994. 138. Ed Keller, Executive Vice President, The Roper Organization, letter to Dennis Beal, Assistant to the President for Communications, AFL-CIO, April 12, 1992. 139. Joseph Uehlein, letter to Howard Samuel and Elmer Chatak, RE: Abuses of Labor Law, February 26, 1987. 140. New York City Central Labor Council for H.R. 5/S. 55 (no date). 141. The Kamber Group, memo to Howard Samuel, RE: Replacement Workers, August 30, 1989. 142. Brian Turner, IUD, Memo to Howard Samuel and Elmer Chatak, RE: Birmingham and Dallas Focus Groups: First Lessons, May 18, 1990. 143. The AFL-CIO believed that an increasing number of strikes involving permanent replacements were primarily strikes by employees attempting to resist efforts by their employers to shift the cost of health care benefits, with one study concluding that health care was a central issue in 55 percent of strikes in 1990. Report Prepared by the Department of Employee Benefits, AFL-CIO, June 1991, “The Permanent Replacement of Workers Striking Over Health Care Benefits in 1990”; Daily Labor Report, “AFL-CIO Study Indicates Increase in Permanent Replacement of Strikers,” June 13, 1991, A-3. 144. Brian Turner, Memo. to the DFS Campaign Files, RE: Striker Replacement Messages, February 7, 1991. 145. Brian Turner, IUD, Memo to Howard Samuel and Elmer Chatak, RE: Birmingham and Dallas Focus Groups: First Lessons, May 18, 1990.

Labor’s “Last Stand” in National Politics?

241

146. The Kamber Group, Memo to Committee for Workplace Fairness, RE: Media Campaign, December 14, 1990. 147. Richard Trumka, “A Grass Roots Campaign Strategy in the State of Alabama for Passage of HR 3936 and S 2112,” (1991); Greer, Margolis, Mitchell, Grunwald, & Associates, Memo to AFL-CIO Strikebreaker Legislation Working Group, RE: Proposal for Media Campaign to Support the Passage of Workplace Fairness Legislation, June 3, 1991. 148. Brian Turner, IUD, Memo to Ken Young, IUD, July 26, 1990. The public campaign for striker replacement legislation employed populist slogans such as “Don’t Fire Strikers” and made no direct reference to the issue of temporary versus permanent replacement. 149. Daily Labor Reporter, “Management Groups Form Coalition to Oppose Striker Replacement Bill,” January 28, 1991, A-1. 150. The Kamber Group, Memo to Committee for Workplace Fairness, RE: Media Campaign, December 14, 1990. 151. “New York City Central Labor Council Campaign for H.R. 5/S. 55”; Frank Swoboda, “Labor Vows Fight Over Strike Workers,” Washington Post, February 19, 1991. 152. In the 1990 Greyhound strike, for example, the company started advertising for permanent replacements in several major newspapers throughout the country three weeks prior to the expiration of the collective agreement. 153. Howard Samuel, President, and Barbara Warden, Director of Legislation, IUD, AFL-CIO, letter to William Ford, Chairman, House Education and Labor Committee, April 17, 1991. 154. Daily Labor Report, “International Union Report Places U.S. Among Violators of Basic Labor Rights,” April 30, 1993, C-1/2. 155. The White House, Office of the Press Secretary, “Statement by the President,” January 13, 1995. 156. Daily Labor Report, “Rep. Clay Chides Employers for Refusing Invitation to Testify on Strike Bill,” September 24, 1990, A-9. 157. Peter Kilborn, “Ban On Replacing Strikers Faces Veto Threat,” New York Times, March 7, 1991. 158. Newt Gingrich, Republican Whip, letter to Betty Smith, June 27, 1991. Gingrich subsequently spoke out against any change in the law on replacement workers. 159. Gunderson, Morley, John Kervin, and Frank Reid, “The Effect of Labour Relations Legislation on Strike Incidence.” Canadian Journal of Economics, No. 4 (November 1989); Gunderson, Morley and Angelo Melino. 1990. “The effects of Public Policy on Strike Duration.” Journal of Labor Economics, 8:3 (1990); Alliance to Keep Americans Working: Reasons Against Strike Bill, 1994. The business lobby also rejected the notion that the U.S. was exceptional among developed nations by permitting the practice of permanent replacement. In an effort to refute the AFL-CIO’s contention during the campaign that only the U.S. and South Africa allowed permanent replacements, the U.S. Council for International Business issued a report stating that 18 out of the 23 countries surveyed prohibited the practice, but that five countries – Australia, Austria, Finland, Norway and the UK – had laws on replacements broadly similar to that of the U.S. Daily Labor Report, “Council Releases Strike Replacement Survey,” May 4, 1993, A-15. 160. A study by Cynthia Gramm of the University of Alabama concluded that disputes involving permanent replacements were five to six times longer than strikes in which permanent replacements were not hired. Cynthia Gramm, “Employers’ Decisions to Operate

242

JOHN LOGAN

During Strikes: Consequences and Policy Implications,” in William Spriggs, ed., Employee Rights and the Changing Economy (New York: M.E. Sharpe, 1991). See also General Accounting Office, Labor-Management Relations: Strikes and the Use of Permanent Replacements in the 1970s and 1980s, GOA/HRD-91-2, (January, 1991). 161. William E. Spriggs, Economist, Economic Policy Institute, Testimony Concerning H.R. 5 before the Subcommittee on Labor-Management Relations of the U.S. House of Representatives Committee on Education and Labor, March 13, 1991. 162. Michael Loitto, Jackson-Lewis, quoted in Lisa Ellis, “Will Eastern, Greyhound Battles Scare Other Firms,” Philadelphia Inquirer, June 10, 1990. 163. “The Striker Replacement Bill,” editorial, Washington Post, April 27, 1993, A16. 164. Peter G. Nash and Jonathan R. Mook, “Striker Replacement Legislation: If It Ain’t Broke, Don’t Fix It,” Employee Relations Law Journal, 16:3 (Winter 1990–1991), pp. 317–332. 165. Frank D. Martino, President, International Chemical Workers Union, letter to Howard Samuel, IUD, September 25, 1990; William D. Ford, Chairman, House Committee on Education and Labor, Letter to Howard Samuel, IUD, January 25, 1991. 166. Bureau National Affairs Special Reports, Employer Bargaining Objectives, 1991–1994 (BNA Press, Washington, DC). 167. Elliot Bredhoff, General Counsel, IUD, letter to Howard Samuel, October 12, 1990. 168. “Trends in the Number of Strikes and Use of Permanent Strike Replacements in the 1980s,” Statement of Franklin Frazier, Director of Education and Employment Issues Human Resource Division, General Accounting Office, Before the Subcommittee on Labor, Senate Committee on Labor and Human Resources, June 6, 1990. 169. Howard Samuel, IUD, letter to Elmer Chatak, IUD, October 5, 1990. 170. Peter Cramton and Joseph Tracy, “The Use of Replacement Workers in Union Contract Negotiations: The U.S. Experience, 1980-1989,” Journal of Labor Economics, 16:4 (1998), pp. 667–701. 171. Clifford Krauss, “Senate Won’t Ban the Tactic of Replacing Strikers,” New York Times, June 12, 1992; Marshall Hicks, President, Utility Workers Union of America, letter to Elmer Chatak, IUD, AFL-CIO, April 19, 1994; Wayne E. Glenn, President, United Paperworkers International Union, letter to Elmer Chatak, President, IUD, December 2, 1992; Joseph Uehlein, Director of Special Projects, AFL-CIO, letter to Howard Samuel, IUD, January 16, 1991. 172. Clark S. Herman, President, Clark S. Herman Associates, Inc., letter to Lane Kirkland, May 2, 1991. 173. American Mining Congress, “Issue Brief,” April 4, 1994. 174. Alliance to Keep Americans Working, “Strike Bill Hurts Non-Union Businesses Too!” (1994); Grassroots Action Information Network (Chamber of Commerce), “Senate to Vote Soon on the Strike Bill,” March 24, 1994; Statement of S. Jackson Faris, President of the NFIB Before the Senate Subcommittee on Labor on S.55, Striker Replacement Legislation, March 30, 1993. 175. James A. Burns, Jr., “Permanent Replacement of Striking Employees,” Employee Relations Law Journal, 19:1 (Summer 1993), pp. 169–173. 176. In March 1993, John Irving, former general counsel of the NLRB, used Yager’s study to attack Labor Secretary Robert Reich following his speech at the NAM, but Reich dismissed Yager’s study as “irrelevant at best and seriously misleading, at worst.” Daily Labor Report, “Reich Affirms Clinton Support for Striker Replacement Measure,” March

Labor’s “Last Stand” in National Politics?

243

31, 1993, AA-1; Daniel Yager, Loading the Scales: Is the Balance Between the Right to Strike and the Right to Operate in Need of Reform? (Economic Policy Foundation, Washington, DC, 1993.) 177. James Jeffords, U.S. Senate, letter to Paul Rohmann, June 14, 1994. Likewise, Fortune magazine, which had campaigned vigorously against the strike bill, asked, “why change [labor policy] now, after half a century – a period in which Democrat-controlled Congresses have repeatedly turned back efforts” to outlaw permanent replacements? “Strikers’ Rights,” Fortune, January 14, 1991, p. 112. 178. The one exception was New York State and the New York data on striker replacements was available only for the years since 1988. In 1991, the international offices of the Steelworkers, Machinists, and Graphic Communications unions started to keep detailed records on employers who had hired permanent replacements. 179. IUD, “A Preliminary Survey of the Use of Permanent Strike Replacements in Pennsylvania, 1980–1990,” January 31, 1991.

ACKNOWLEDGMENTS The author would like to thank David Brody, Adriana Craciun, Tom Donahue, Tony Giles, John Godard, Sandy Jacoby, Bruce Kaufman, and George Strauss for comments on previous drafts of this article.

THE TRANSFER OF IDEAS IN INDUSTRIAL RELATIONS: DUNLOP AND OXFORD IN THE DEVELOPMENT OF AUSTRALIAN INDUSTRIAL RELATIONS THOUGHT, 1960–1985 Diana Kelly ABSTRACT The primary objective of this paper is to understand the extent to which Australian industrial relations academics took up the different heuristic frameworks from USA and U.K. from the 1960s to the 1980s. A second objective is to begin to understand why, and in what ways ideas are transmitted in academic disciplines drawing on a “market model” for ideas. It is shown that in the years between 1960s and 1980s a modified U.S. (Dunlopian) model of interpreting industrial relations became more influential in Australia than that of U.K. scholarship, as exemplified by the British Oxford School. In part this reflects the breadth, flexibility and absence of an overt normative tenor in Dunlop’s model which thus offered lower transaction costs for scholars in an emergent discipline seeking recognition and approval from academia, practitioners and policy-makers. Despite frequent and wide-ranging criticism of Dunlop’s model, it proved a far more enduring transfer to Australian academic industrial relations than the British model, albeit in a distorted form. The market model for the diffusion of ideas illuminates the ways in

Advances in Industrial and Labor Relations Advances in Industrial and Labor Relations, Volume 13, 245–273 © 2004 Published by Elsevier Ltd. ISSN: 0742-6186/doi:10.1016/S0742-6186(04)13008-4

245

246

DIANA KELLY

which a variety of local contextual factors influenced the choices taken by Australian industrial relations academics.

INTRODUCTION Even before the age of cyber-communication, the demands of business inter alia have ensured that ideas have transcended cultural or national boundaries with notable speed.1 Despite the evident importance of understanding the diffusion of ideas, the processes of diffusion and transfer have only rarely been studied systematically outside of the rarefied world of the sociology of knowledge. The difficulties of a systematic study of the transfer of ideas are all the more important in an academic discipline such as industrial relations. This is because industrial relations is closely linked to other social sciences, humanities, behavioural sciences and the law. Moreover it is fraught with self-doubt and an apparent need to incorporate the immediacy, volatility and the inherent ideological traps of public phenomena and perceptions into its scholarship. Nevertheless, it is essential to come to an understanding of how ideas are transferred within the discipline, and also between industrial relations and other disciplines.2 This paper takes just one aspect of industrial relations thought – the influence on Australian industrial relations thought of two influential English language mainstream heuristic frameworks. The paper examines how the ideas of these were transferred to Australia. It raises general questions of how ideas are taken up, and what is taken up, and why certain ideas are accepted almost wholly, what ideas are transferred but are modified in the transfer, and what ideas are rejected. These issues raise further questions about why particular ideas or analytical concepts are accepted, modified or rejected. On one level, then this paper is about a particular epoch in Australian industrial relations thought – the 1970s and 1980s – but on another level it raises questions about the movement of ideas, an area of analysis which is important in industrial relations for a number of reasons. To explore these issues the paper draws on a simple market model of the diffusion of ideas in tracing the uptake of some international ideas in Australian industrial relations thought.

THE TRANSMISSION OF IDEAS An early notable systematic analysis of the transmission of ideas is the conference held in 1966 (Goodwin & Holley, 1968) at Duke University. In the resultant publication some emphasis was given to the model of vector analysis, in which the mathematical and physical concept of vectors offers the opportunity to examine

The Transfer of Ideas in Industrial Relations

247

the magnitude and direction of the transmission of ideas. This also fits with the disease model of the transmission of ideas postulated in 1989 by Colander, in which he proposes that one way in which ideas are spread is akin to a contagious disease. In this model vector analysis offers the opportunity, just as it does for epidemiologists to, monitor and therefore evaluate, the processes by which ideas spread (Colander, 1989). The difficulty with these kinds of models is that the recipient or consumer of ideas is passive, and the ideas spread without any respondent actions. In turn this does not illuminate why a theory, method or heuristic framework is taken up wholly or in a different form from that which was originally devised (see Coats & Colandere, 1989). By contrast a “market model” for the transmission of ideas posited by Colander (1989, pp. 12–14) considers the transmission of ideas, by which is meant not only concepts, but also theories, heuristic frameworks and methods, in terms of suppliers and consumers. The consumers respond actively to the available set of ideas in the market place and choose ideas or elements of ideas according to their preferences and perceived needs. These will reflect their perceptions of output requirements and the optimal inputs to achieve them. Even in an emerging discipline the scholars are subject to the same sorts of reputational systems as those existing in mature disciplines.3 What will motivate scholars in taking up a theory or heuristic framework will reflect their perception of what is deemed good scholarship in their discipline. Moreover, as with any factor inputs, scholars can choose to use frameworks in a variety of ways, so they may re-bundle or reshape ideas to meet their perceived needs. As Colander (1989, p. 12) has noted The diffusion process is inevitably selective, highlighting certain ideas or aspects . . . while obscuring or fragmenting others. Which ideas work their way through the system and how the ideas are changed as they proceed depends on the incentives of the individuals the particular institutions and how those incentives and institutions interact.

In the marketplace for ideas then, the tastes and preferences of the consumers will reflect the local “market,” and the ways in which ideas meet their needs. “Consumers” will not only take account of input factors which meet their perceived needs, but they will also consider the costs to them in taking up one idea, framework or theory over another. In this paper, a market model is utilised to illuminate salient elements in an examination of some developments in the nature and direction of Australian industrial relations scholarship from the formative years of the 1960s, when IR emerged as a field of study, until the mid 1980s after which other factors in business and social sciences began to influence research methodology in the field. The paper takes as its starting point the assertion that despite a tendency to avoid theory (Bray & Taylor, 1986; Woods, 1978), Australian industrial relations has been most influenced by British and American IR heuristic processes and

248

DIANA KELLY

modes of thought.4 In the paper, the pattern of influence of the predominant postwar “paradigms” from these two countries on industrial relations thought in Australia is examined. The process by which Australian scholars chose and modified a particular U.S. analytical framework over that developed by British scholars is then explored. Both frameworks were widely recognised in their host countries, and there is clear evidence that Australian scholars were familiar with each framework. In questioning why one was eventually preferred by mainstream scholars in Australia, the market model for ideas is used to explore the responses to the U.S. and U.K. heuristic frameworks.

METHODS AND METHODOLOGY IN AN EMERGENT SOCIAL SCIENCE5 For at least fifty years there has been a definable international community of scholars who have perceived themselves as expounding ideas of a discrete discipline, separate from the allied disciplines of law, economics, politics, sociology, psychology and personnel management. It has been a discipline characterised by a degree of uncertainty about its place.6 Industrial relations or employment relations (the terms are here used interchangeably) in Australia has been just as uncertain of its place, as in those Anglophone countries where it has developed as a separate academic discipline. As an open-bordered multimethod discipline, Australian academic industrial relations scholars are thus part of a borrower discipline in a nation with a long tradition as a borrower of intellectual ideas. After briefly considering definitional issues, this section discusses the heuristic frameworks offered by the two countries which most influenced Australian scholarly priorities, as well as the nature of influence on Australia scholarship in more general terms. As social science scholars, industrial relations researchers focus on all matters of work and employment. While the social sciences are more inexact than natural and physical sciences, all sciences nevertheless share a complexity which arises from multiple analytical frameworks. Such analytical frameworks may take the form of fundamental assumptions which determine epistemology, theoretical development and methodology. This is the case for example, considering the theoretical assumptions at the basis of debates over institutional vs. monetarist economics, or the impact of plate tectonics on geology (Laudan, 1983, pp. 79–104). By contrast, heuristic frameworks are rather less fundamental than theories. This is notably the case in new areas of intellectual activity and emergent disciplines where the area of study is sparsely populated and scholars are still identifying the nature of the terrain (Becher, 1989). Certainly this has been evident in academic industrial

The Transfer of Ideas in Industrial Relations

249

relations where strong theoretical developments are yet to influence analysis and research. Rather than develop theory, scholars have borrowed piecemeal from other disciplines or worked from heuristic frameworks which direct research questions, methodology and methods in particular directions. Following from British scholars Bain and Clegg (1974, p. 92) it can be argued that a heuristic framework can be defined as a construct for organising ideas, so that unlike a theory, it “must not be misunderstood as having predictive value in itself.” Despite the lack of predictive value, an heuristic framework provides a useful device for researching and communicating highly complex social phenomena such as industrial relations. For the purposes of comparison in this paper, three salient attributes of an heuristic framework can be identified. The first attribute is the substance of what is to be studied, the extent to which different facets are emphasised and the role ascribed to the context of the objects of study. As Hyman (1994, p. 167) has noted, “Any account of the ‘facts’ of industrial relations rests on the principles of exclusion and inclusion linked to (explicit or implicit) criteria of significance.” For example, scholars in the fields of labour economics, organisational behaviour, industrial psychology, human resource and critical management all take work and employment as their focus but they will include, exclude or emphasise different aspects. The second attribute of an heuristic framework can lead the first, for it is the assumptions and ideals which are the drivers for the researcher. This then can include ideological issues as well as the objectives of the researcher. Scholars who are investigating in order to generate particular kinds of reform will hold ideals and assumptions which will direct the focus of their research. In Australia in the 1980s numerous analysts sought to promote what was seen as American-style enterprise bargaining processes instead of the long-held commitment to state arbitration systems (see Niland, 1986). The assumptions apparent in their anathema to state arbitration thus directed their research. Similarly, Kaufman has demonstrated that the science-building goals of some industrial relations analysts led to different approaches to research than problem-solving imperatives of reformist scholars (see e.g. Kaufman, 1993, pp. 125–128). The third attribute of an heuristic framework flows from and feeds back into the other two attributes, and that is the research methods which are apt to the conceptualisation of the study and the objectives and ideals of the researcher. Many modern economists eschew qualitative research methods such as ethnography which they see as lacking in rigour, while critical management analysts in the U.K. see statistical analysis as arid and without insight. Heuristic frameworks then, can be considered in light of these three attributes – the objects of study, the imperatives of the research, and the research methods. Perhaps the most well known heuristic framework in Anglophone academic industrial relations is the “systems” approach. John Dunlop developed his systems

250

DIANA KELLY

framework in the 1940s and 1950s as a purposive response to his view that there was a lack of systematic investigation in IR analysis beyond collecting “mountains of facts” in all aspects of employment relations. In so doing he drew on his scholarly training as an economist. Less purposive but equally distinctive was the British heuristic framework, the “Oxford framework, which derived from and extended the foundation scholarship of the Webbs. These noted early sociologists were indefatigable inductivist researchers who investigated trade unions, cooperative societies and the Poor Laws from the turn of the century (Webb & Webb, 1898, 1902, 1975). As is explained below the Dunlop and Oxford heuristic frameworks are quite different in the assumptions and methods employed which in turn generated different kinds of research development. These two Anglophone frameworks appeared most apt as Australian academic industrial relations emerged and expanded in the 1960s and 1970s. While it is arguable that more than in most social science disciplines, scholars in employment relations/industrial relations operate within a real world context, the objective of scholarship is not simply to mirror or to reform current practices. This highlights the fact that industrial relations as a study or academic discipline is different from a national industrial relations system. Just as political science covers aspects of politics far deeper and broader than a national political system, so industrial relations scholarship is a field of study far deeper and broader than a national industrial relations system. Nevertheless, it is axiomatic that a significant influence on scholarship will be the current and historical context in which scholars are operating. The evident contextuality of industrial relations has implications in the market for ideas for how and which ideas are taken up by the scholarly community as factor inputs. At the level of practice, an industrial relations system can be described as the organisational, legal and economic framework which determines the coherent set of norms and processes evident in the formal institutions of industrial relations in a particular enterprise, region or nation-state. From around the turn of the century until recently, the conciliation and arbitration systems in Australia and New Zealand have meant that industrial relations processes have long been subject to government policy. This has contrasted with both the USA and U.K. for example where the control and administration of the employment relationship has been characterised by an enterprise focus and voluntarism. Thus in Australia and New Zealand with systems traditionally more in the “public” sphere than in the “private” sphere,7 the practice and analysis of industrial relations in Australia have perhaps been subjected to greater political and ideological considerations than overseas. The centrality of the “public” processes of the administration of the employment relationship has also been the focus of scholars from overseas, many of whom

The Transfer of Ideas in Industrial Relations

251

found the level of government involvement curious (see e.g. Perlman, 1954; de Vyver, 1956) Certainly, the “public” nature of industrial relations was more emphatic than elsewhere because national wage determination through the tribunal system gave governments in Australia greater access to this facet of economic policy making than in other countries.8 These national phenomena have led to the perception of the Australian system as unique. How unusual it has been is a moot point, but there is no doubt that the structure of the system has reinforced the conviction held by many practitioners and scholars that the control and administration of the employment relationship has been significantly different from that elsewhere. One effect of the perceived differences may have been to mediate or alter the ways in which international intellectual developments have influenced Australian industrial relations scholarship. The next section surveys the nature of influence on Australian scholarship, and then describes the two heuristic frameworks which have influenced academic industrial relations.

INTERNATIONAL INFLUENCES ON AUSTRALIAN ACADEMIC INDUSTRIAL RELATIONS Influences on Australian scholarship were traditionally derived from the nation’s historico-political links to Great Britain. White settlement by British government soldiers and convicts began at the end of the 18th century. As a colony of, and then a nation state within the British Empire and Commonwealth, Australian legal and political institutions have their source in Britain. Until the 1960s these were underpinned by dependence on important market links. Major firms in Australia were outposts of British firms, with particular similarities in management styles and employment relations. Until the middle of the 20th century the majority of immigrants came from Britain, so that it is not surprising that trade union structure also bore close similarities with that in Britain. Like Britain too, the industrial wing of the labour movement was complemented by a political party, with a concomitant broad set of principles beyond the economism which has tended to dominate labour in the USA. The surface features of the business, social, legal and political heritage have been replicated in academia. Australian universities were moulded on British lines, and Australian scholars aspired to British models of scholarship. For those reasons, until the 1960s, academics mainly took sabbaticals in the U.K. rather than elsewhere and used British texts in their teaching. One of Australia’s most influential and venerable scholars, J. E. Isaac took his Ph.D. in economics at London University, as did John Child. Their scholarly training and concomitant prestige

252

DIANA KELLY

influenced those whom they later taught or supervised, and in so doing affected the preferences of the incoming generation of scholars and practitioners. Yet despite the historical links Australian academic industrial relations was not only influenced by the “mother country.” In an overtly monolingual country such as Australia, only other Anglophone scholarship was ever under consideration as alternative exemplars to British research. Moreover, it is a paradox that while erstwhile colonial nations may replicate the home country mores and structures, they will simultaneously try to break from the tether of the imperial power. For that reason it is unremarkable that Australia has long looked to the largest of the “new world” countries, the USA, for leadership and ideas. After World War II this process gained pace as the USA gained international economic leadership. Not surprisingly American academic industrial relations was seen as the other significant supplier of ideas and models. In industrial relations, the pattern of following American ideas was bolstered by the fact that scholars from the USA had also had a long fascination with Australia. The scholar – judge who was the architect of much of the conciliation and arbitration machinery, Henry Bournes Higgins, lectured and published in the USA in the 1920s and there is evidence even earlier of American interest (Goodwin, 1965; Groenewegen & MacFarlane, 1990). Moreover from early this century U.S. scholars have visited Australia. The title of one of the first major publications on Australian industrial relations, Judges in Industry by Mark Perlman9 reflected precisely the nature of the American fascination with the institutional system, while visitors such as F. de Vyver also sought to analyse and understand the Australian system in the 1950s, and the lessons it might provide for the U.S. patterns of labour relations.10 Following similar patterns of research interest to Mark Perlman, other American writers such as Milton Derber, focussed on the processes of rulemaking in industrial relations in Australia. Their contributions were to be augmented and advanced in the 1970s with work of John Niland (Ph.D. Illinois) and Bill Howard (Cornell). In this respect Niland’s role in Australian academic industrial relations may be seen as crucial for the direction of Australian academic industrial relations since he was editor of the Journal of Industrial Relations for fifteen years from the mid-1970s. While the influence of the USA scholarship was to be greater than that of British scholarship, the close cultural and economic links between Australia and U.K., on the one hand, and the parallel structures of universities and of trade unions, on the other, meant that before the 1980s British influences on Australian intellectual life were very significant. To understand better the transmission of industrial relations research ideas from the U.K. and USA to Australia requires first a brief discussion of the notable heuristic frameworks from those two countries.

The Transfer of Ideas in Industrial Relations

253

DUNLOP AND SYSTEMS The widely used industrial relations systems framework was developed by the labour economist John Dunlop through the 1950s (see Dunlop 1948, 1950, 1958). It was devised as an analytical framework with the specific intent of integrating the study of industrial relations. Dunlop argued persuasively that industrial relations had been investigated simply as a subset of or several fields of study, notably labour relations, (“the labour problem”) varieties of management, (personnel management, scientific management, human relations), labour law, industrial psychology, public policy and labour economics. From Dunlop’s perspective, the outcome of study from all these perspectives had led to “mountains of facts” which needed an integrating analytical framework. He saw industrial relations as a separate system, wherein the actors and institutions could for purposes of research and policy making be studied within the separate entity of the industrial relations system, just as economists separate out the economic system from the rest of human activity. In this respect, Dunlop was adding the structural functional Parsonian model to the fundamental ideas derived from his economics training.11 Thus the archetypal industrial relations system was embedded with notions of equilibrium such that at any moment in time, where “an internal balance is likely to be restored if the system is displaced.” Dunlop’s industrial relations system comprises actors (employers (including managers), employees and government (including agencies) and the collective institutions, all bound together by a shared ideology to make the system work, and a shared set of contexts (market, technological, social, economic, political) (Olson, 1969). It is worth noting that in laying out his model Dunlop ascribes equal importance to each of the actors. However, in applying his model in his chapter on Bituminous Coal for example (Dunlop, 1993, pp. 131–178) there is virtually no reference to the roles, styles or functions of management. Nevertheless, in delineating his model Dunlop emphasises that the parties interact to produce, and then operate within, a “web of rules” which is the outcome of negotiation and bargaining. For Dunlop “the establishment of these procedures and rules – the procedures are themselves rules – is the centre of attention in the industrial relations system” (Dunlop, 1993, p. 53). What Dunlop neglects, despite his assertions to the contrary (Dunlop, 1993, pp. 51–53) is the ways in which “rules” can be developed. For Dunlop the mechanisms for rule making are narrowly defined and occur only within the system. He later notes (Dunlop, 1993, pp. 109–110) that “Differences in form, private or public, may be of concern to students of law but they cannot be primary interest of attempts to treat industrial relations more analytically.” Dunlop (1993, p. 8) also specifically rejects the notion that industrial relations is simply about relations between employers and employees.12

254

DIANA KELLY

Since the industrial relations system is indeed a systemic structure, attainment and maintenance of equilibrium is the appropriate outcome. Such assumptions of equilibrium are perhaps idealistic, but for Dunlop they were the logical outcome of the function of the systems approach which underpinned his framework. In this respect it is noteworthy that for Dunlop power was not a feature of the industrial relations system, but an external “context” like the technological or budgetary context (see Dunlop, 1993, pp. 107–130; Dunlop, 1977). In some countries Dunlop’s system was taken as a template for public policy, but more often his framework provided the basis for describing and evaluating the control and administration of the employment relationship, with analysts augmenting their analysis with their own personal and ideological perspectives. Overall though, it is worthwhile noting that Dunlop’s highly prescriptive approach to what were the accepted objects of study highlight his methodological singularity, While a pluralist from a political science perspective, insofar as the notions of different and competing interests at the enterprise or in the national industrial systems, Dunlop prescribed not only what were apt objects of study, but also how these should be researched. For Dunlop, the basis for investigation was agreements and laws which comprised the rules of the system, and statistical information which provided evidence of the working of the system. Notions of ethnography or other sociological methods were spurned in favour of “. . . deductive propositions, checked by empirical testing, relating specified changes in the system to specific changes in the rules” (Dunlop, 1993, p. 286). For Dunlop then (1) The industrial relations system is a complete and separate system which is influenced by other socio-economic systems or contexts. (2) There are three sets of actors employees, employers, and the State who each have different goals but are bound by shared perspectives of the system, and their shared occupation within the system.13 (3) The establishment of rules is the primary objective of the actors in the industrial relations system. (4) The objective of the researcher is value-free analysis of primary documents, notably policies, formal agreements and statistics.

FLANDERS, CLEGG AND THE OXFORD SCHOOL The term Oxford School reflects the reformist pragmatism of British industrial relations scholars of the 1960s and beyond. Although it was more informally developed than Dunlop’s framework, the patterns of assumptions in the

The Transfer of Ideas in Industrial Relations

255

epistemology and methodology of the Oxford School provide the basis for an heuristic framework, with identifiable regularities and differences from other frameworks. The Oxford School had its antecedents in Sidney and Beatrice Webb’s particular approaches to inductive case study research (Webb, 1938a, b), and as is evident in the breadth and consistency with which the Webbs are cited in Oxford school publications, had a fundamental impact on their approaches to research. The most notable exponents were Flanders (1965, 1970), Clegg (1954) and Fox (1966, 1973) (see also e.g. Boraston et al., 1975; Brown, 1978; Rimmer, 1973). It was perhaps most well-known in the research for the Donovan Commission which some see as the grand exposition of the Oxford school in the 1960s. Of interest here however, is not the findings or report, but rather the research and assumptions which informed investigation. The major recommendations of the Donovan Commission (1966) aimed not only to maintain the degree of participation that had developed in the informal shop steward movement, but also to contain it within new formality. These recommendations were thus the policy outcomes of a deep commitment to pluralism, which, by the 1960s was a descriptor given to an industrial relations ideology.14 Pluralism, a concept borrowed from political philosophy, rested on questions about the nature of power in society. For industrial relations scholars, pluralism drew on normative assumptions of the ideal society as one in which collectivities of interest should be sufficiently strong and diverse as to ensure that no single interest group could predominate (Clegg, 1972, 1990; Fox, 1966; Fox & Flanders, 1969; for an outstanding scholarly critique, see Hyman & Brough, 197515 ). In their analysis, the pluralist industrial relations scholars began with the assumption that there existed a system of industrial relations which focussed on the making of substantive and procedural employment rules by managers, employers, and workers as represented by their collectivities. In the British system of industrial relations “Each employee is likely to be affected by a considerable number of rules and the complex of rules within a particular plant can be regarded as a system” (Clegg & Flanders, 1961). Thus in Britain in the 1960s, the system of what was called job regulation in each plant was portrayed as a sub-system of the complex of rules in each industry.16 In this respect it seemed only a little different from the Dunlopian framework, except that British pluralists in the 1960s and 1970s focused close attention on the British system, and their research drew deeply on the British historical context. Apart from slightly different terminology and emphases, it appeared as if the objects of study of both the Dunlopian and Oxford frameworks were internal and external job regulation. The primary difference in their approach to job regulation reflected the Oxford School interpretation of the business enterprise as rather more complicated than had Dunlop. Drawing on management analyst Peter Drucker, Flanders (1975, p. 89) asserted that internal job regulation reflected the “triple personality” of the business

256

DIANA KELLY

enterprise which is “at once an economic, a political and a social institution.” It was the same at the level of the state – economic, political and social factors were all important. To investigate job regulation, therefore required recognition of all three of these “personalities,” not as external and discrete contexts but as part of a whole, integrated by normative notions of what is a good industrial relations system.17 And whatever was the ideal of good industrial relations, this complexity of personalities, levels of study, and sense of integration were important. This leads to the second set of shared attributes of the Oxford School framework, which, is of much more interest to the disciplinary historian. These relate to the assumptions and research methods of the Oxford School framework. They are: (1) the centrality of power relations in employment – equity and fairness; (2) close integration of macro (social, economic, and institutional contexts) and micro (workplace); (3) the inclusion of employers and managerial employees as well as nonmanagerial employees; and (4) inductive approaches to research and reliance on the case study as central method. Of central importance to the framework was the notion of employment relationships as power relations. By inserting the concept of power into their analysis, these scholars were not only acknowledging the allied normative concepts of social responsibility, equity and fairness, but also promoting ideals of “industrial democracy.” The goal of industrial democracy, the sharing of power, was a key theme through much research of Oxford scholars, with others’ studies highlighting the forms of power and its misuse (see also Beynon, 1973). In discussing power at the workplace meant that the Oxford researchers also gave attention to management roles. It is not surprising for example that Flanders’ 1966 analysis of the “The Internal Social Responsibilities of Industry” (Flanders, 1975) discusses the power of management at length, because not only was the employment relationship a power relation, but also it was fundamental for Flanders that “responsibility is a function of power.” Thus reflecting on his well-known ethnographic study, of the Fawley productivity agreements (Flanders, 1964, 1975) “uncovering the facts and trying to make sense of them forced me to realise how crucial was the notion of managerial responsibility, especially the responsibility towards the managed . . .” (Flanders, 1975, p. 64) (see also Clegg, 1960; Fox, 1973, 1985). Immediately these issues become embedded into the research framework, other attributes follow. From the acknowledgement of power as a central variable, flowed a recognition that all parties to the analysis of work and employment regulation are

The Transfer of Ideas in Industrial Relations

257

the subjects of analysis. Thus integrated into analysis of the public and institutional aspects of industrial relations analysis were aspects of management, the law, enterprise and workplace culture as well as political economy and economic policy. Moreover, as a normative, complex, and particularistic framework, the research was overtly and necessarily multidisciplinary. Finally claim Brown and Wight (1994) in their analysis of the Oxford framework, all of these conceptual and theoretical assumptions required particular research techniques, most notably the case study involving observation and interview. Even Ackers and Wilkinson (2003), who somewhat scathingly identify the Oxford School research as close to “casual journalism,” acknowledge the School’s emphasis on ethnography for capturing the complexity of organisations. What they perhaps omit from their analysis is that scholars such as Clegg and Flanders it was important not to oversimplify such complexity. Thus while Clegg was assiduous in seeking theoretical approaches, he sought to uphold the traditions of rigour and thoroughness, It is not surprising then that Clegg asserted in evaluating theories of collective bargaining “Justice is done to the subject’s complexity by not trying to force it into a single theoretical mould” (Clegg, 1968, p. 19). At the same time, as Ackers and Wilkinson adduce, the Oxford School analysts were notable in endorsing the need for industrial relations research, as well as policies and processes to uphold access by all to the political and economic gains of British capitalism. The egalitarian pluralism of the British school gave their analysis and studies a clearly normative element, with their writings providing clear evidence of their objective to see a shift in the distribution of wealth towards those with lower incomes, and a shift of power over the conduct of their working lives and environment towards working men and women, and . . . emphasising the importance of trade unions in industry, in the economy and in society (Clegg, 1990).

Much of their analysis therefore was on the rules and rulemaking in employment but always utilising an institutional perspective, rather than an individualist focus. For these scholars the processes of understanding the practice and operation of rulemaking, had the purpose of finding ways of empowering workers, so that consideration of ideas of fairness underpin much of their analysis (see e.g. Clegg, 1972; Hyman & Brough, 1975, Chap. 4; Kahn Freund, 1969, see also Hyman, 1995). This is not to say that they sought radical economic reform. Neither did their analysis elide management. While Flanders was a practitioner claiming socialist ideals who moved back and forth between academia and union or activist roles, he saw management as having an essential role in making things better (Flanders, 1964, 1965, 1970). He was not alone in this view; British

258

DIANA KELLY

scholars usually investigated the role of management in their industrial relations analyses. Another British scholar, Hyman (Hyman & Brough, 1975, pp. 157–183) took the pluralists to task for taking the economic system as a given. In doing so it was argued they were upholding an economic system in which unfairness was unavoidable, indeed essential. Hyman argued that the contradictions inherent in such approaches diminished the work of these scholars. Nevertheless, rather more than in the USA, the objective of the British pluralists was to work towards a fairer society. To this end they gave considerable weight to achieving order through increasing effective participation, as is apparent in their emphasis on joint regulation. Thus, while the contradictions highlighted by Hyman hold true, the underlying, and sometimes overt, objective of the Oxford analysts was a free society which would give weight to egalitarian values. By contrast the American scholarship tended toward logical positivism, given the deductivist approach to research and the emphatic insistence that value-free research was essential. The mainstream approaches developed in the U.S. and the U.K. were thus similar in that they were based on strong empirically bound frameworks. However, they differed in their methodologies. Where the British framework was particularistic, inductive, and had strong normative overtones, Dunlop’s framework, based as it was on a model industrial relations system, was deductive, broad and positivist.18 Both were in a position to influence Australian industrial relations thought.

AUSTRALIAN INDUSTRIAL RELATIONS THOUGHT 1960–1985 In the market for ideas the contexts in which scholars work will be central to the ideas or theories which they choose. For example, in an area such as academic industrial relations the contexts for the academics include the nature of the university system, the scholarly imperatives and traditions of neighbouring (and “senior”) disciplines, the norms and values extant and the public institutions and structures which affect or determine the objects of research. In choosing a theory or analytical framework as a factor input, scholars will, perhaps not purposively, consider the transaction costs of each framework against their assessment of the significance of these central facets of their environment. This section will briefly examine the contextual factors and then delineate and analyse the development of ideas in Australian academic industrial relations. In the postwar decades the publicly acknowledged industrial relations system in Australia continued to be collective and institutional. These were decades of relatively high profits, solid levels of growth and high union density. The high

The Transfer of Ideas in Industrial Relations

259

level of union consciousness meant continuing attention was paid to “the labour problem” which was accentuated by a relatively large number of strikes, albeit of much briefer duration than elsewhere. In the main management tended to be ad hoc (with low levels of management training) and production-focussed, accepting unions, but generally ensuring their roles remained limited to those issues acceptable by the tribunal system (see e.g. Byrt & Masters, 1974; Hall, 1971; Tsokhas, 1981). The media, together with Coalition Liberal (conservative) governments (until 1972) and activist conservative political pressure identified the high numbers of strikes as being linked to leftist normative values, and sometimes to the spectre of communism. Differing perspectives on Communism had split the Australian labour movement in the 1950s, a split which business lobbies used to good effect.19 Nevertheless, employment relations processes in these decades were characterised by optimism about economic growth offsetting concerns of managers and workers, interspersed with short periods of numerous industrial disputes and considerable recourse to tribunals. Throughout these years notions of collectivism were widely accepted in Australia. Not only was union density over 60% but the unions intersected and sometimes led movements for equal rights for women, and against war, racism and unfettered building construction. During the 1950s industrial relations analysis in Australia had been the province of the practitioners – the industrial psychologists at workplace level, and the lawyers and economists, as well as trade union officials, employer advocates and industrial tribunal practitioners at industry and national level. Initially Australian scholars were slow to adopt any analytical framework, perhaps because there were almost no academics trained in industrial relations. With a few relatively isolated exceptions, analysis of Australian industrial relations was the province of the practitioner, not the academic (Niland, 1978b, see also Walker, 1955). Thus despite the publication of Dunlop’s Industrial Relations System in 1958, and those by Flanders and Clegg from the mid-1950s (Clegg, 1972; Flanders, 1975; Flanders & Clegg, 1954; Fox, 1966), there was only occasional utilisation of an identifiable analytical framework in industrial relations studies in Australia until the 1970s (Journal of Industrial Relations, 1959–1975; Child, 1962, 1963). This is perhaps because institutional industrial relations operated within the public sphere in Australia, together with the fact that the academic community did not develop until the 1970s. The proto-community of industrial relations analysts in the 1960s grew out of the Industrial Relations Society which formed in 1959 and which was not an academic society, but one in which the discussions and differences of practical men can take place under circumstances where the disciplines of the academician, the administrator, and the lawyer may all help to control and direct the course of the debate to the mutual benefit of all concerned (Kerr, 1961).20

260

DIANA KELLY

One of the initial outcomes of the IRS was the Journal of Industrial Relations (JIR), the first issue of which was published in 1959. In its first decade or so the JIR articles reflected both the nature of the public aspects of the industrial relations system in Australia and the intellectual training of the authors. The majority of articles in the journal dealt with the aspects of the employment relationship from a managerial, legal or economic perspective, or with collective bargaining or strikes as economic or legal processes. There were also articles on industrial psychology and personnel management in JIR articles in the 1960s, unlike in later years when the focus narrowed to institutional parties and processes. Nevertheless, the method was essentially pragmatic and oriented toward dealing with the labour problem.21 By the 1970s however, there was clear evidence that scholars were using and evaluating both the British and American models, and considering the costs and benefits of each.22 The economist turned industrial relations scholar, Kingsley Laffer, took the opportunity in 1972 to evaluate Dunlop’s framework in his review of Kenneth Walker’s new book The Industrial Relations System, which was published at the same time as a reprint of Dunlop’s Industrial Relations System.23 At the Industrial Relations section of ANZAAS in 1973, Laffer compared and analysed the definitions of industrial relations of both Dunlop and the British scholars, and found them both wanting, preferring his own definition which focussed on “bargaining relations” at the workplace (the individual effort bargain) and at national level (incomes policy) (Laffer, 1974).24 Scholars claimed to prefer the British model (Howard, 1978), but there was plenty of evidence of the American model, particularly by Dufty (1972, 1975, 1979) and Walker (1956, 1970), who even in the 1950s used Dunlop’s earlier models as the basis for his analysis. Walker’s (1956) book was published by Harvard University, and included a foreword by Dunlop. It did not seek to replicate Dunlop’s model however, but rather altered Dunlop’s notion of the parties, shifting emphasis towards the processes in the formal system. As with later Australian writers and unlike the British writers such as Clegg, Walker’s analysis paid little attention to managers and much more to the State and to government than Dunlop had indicated. Once again these modifications reflect the view held by many that the tribunal system was a central dependent variable in Australian industrial relations analysis. Given their primacy, the value to the Australian academics of the Dunlopian framework, in whatever form, was that the tribunals could be readily incorporated into their research (see e.g. Moth, 1972; Dufty, 1975). While Howard was perhaps exaggerating when he talked of “Dunlopians v Flandersites small scale warfare” (Howard, 1978, p. 34), there is no doubt that those scholars who were concerned with methodology (and they were very few) were ambivalent about the British and American models. For some (W. A. Howard, G. W. Ford, K. W. Hince, Interviews 1998) it appears that Dunlop’s system was

The Transfer of Ideas in Industrial Relations

261

a useful teaching tool, but that for analytical or ideological reasons the British model was preferable because it seemed to allow for closer attention to social and political factors. This ambivalence, allied with a trend towards the American approach, is apparent in the text Australian Industrial Relations (Plowman et al., 1980), which was to become the primary text for a decade or more.25 Introductory texts, particularly those which are widely used, are an important measure of the direction of thought in a discipline, since they signify the foundational assumptions of the mainstream scholars, or those scholars who aim to capture the mainstream. Moreover, and this is particularly true of methodology, textbooks also direct the next generation of scholars, policy-makers and practitioners to particular interpretative approaches. In the first edition of their text, David Plowman, Stephen Deery and Chris Fisher deal with the competing demands of the British and American approaches in two ways. First they seek to conflate the two as pluralist “systems” approaches, providing highly selective descriptions and discussions of both. Second while they claim that the text follows closely a Dunlopian framework, almost all of their suggestions for further reading derive from the British tradition. Nevertheless what the Australian students of industrial relations learned from the book was an approach to industrial relations analysis which used a modified systems framework and which omitted the important aspects of the Oxford model such as the role of power. It is also notable however, that the Dunlopian framework was modified particularly in terms of its actors – it was not until the third edition in the 1990s that Deery and Plowman dealt in any depth with managers. This reflected their perception, and one which is especially evident in much Australian industrial relations research from the latter 1960s, that the locus of industrial relations was the public sphere of rulemaking and necessarily incorporating the tribunal system. There were other notable differences, such as the weighting given to the environmental factors (particularly economic) which was also ascribed rather less importance by Deery and Plowman than Dunlop. In other words, a modified version of Dunlop’s framework was chosen as a device for structuring investigation and teaching in Australian industrial relations. This framework, implicitly or explicitly underpinned much mainstream academic Australian industrial relations for the rest of the 1980s (see e.g. Dufty, 1980; Hill et al., 1982; Lansbury, 1978)

AUSTRALIAN ACADEMIC IR THOUGHT: COSTS AND BENEFITS OF THE TWO MODELS Understanding why Dunlop’s model transferred more readily to Australia requires attention away from the nature of the models, and towards the consumers, the

262

DIANA KELLY

Australian academic industrial relations community. This section examines the nature of the take-up of the Dunlopian framework and the rationales for the choices made by the scholarly community. The reasons for making a choice within a discipline to take up one paradigm (interpretive framework) rather than another, is multifactorial. As Kindleberger (1989) has noted neither plausibility nor logical excellence is of itself sufficient. In the market for ideas, there are both supply side (the plausibility, explanatory value, generalisability) and demand side factors (state of the discipline, manifestations of the invisible college, and the nature of the predominant culture within the discipline). That the Australian industrial relations community took up a modified Dunlop system rather than the British framework reflects not only the differences in the two models of industrial relations, but also the nature of the small community of Australian industrial relations scholars. In analysing the transfer of ideas understanding the reasons for the consumers’ choice is at least as important as the capacity, fitness and generalisability of the source ideas. Thus the features and self perceptions of the Australian IR community, the historical context and the politico-economic climate of the 1970s, all clarify why Dunlop became preferred choice of research base for those mainstream scholars seeking a “theoretical” stance. In this respect there are three elements which deserve consideration. The first two are both to do with legitimacy – attaining legitimacy within the academic community and achieving legitimacy with the large and articulate practitioner base, while the third relates to the nature of public opinion during the Cold War years. The acceptance of external groups was important for the Australian academic industrial relations community in the small, elite and conservative world of Australian academia. Indeed, despite a relatively high growth rate in terms of students and researchers, the academic industrial relations community before the mid-1980s was very small, and stood at the fringes of labour law, labour economics, industrial psychology and politics.26 For many, including some of its leading exponents, industrial relations did not have disciplinary status, in fair part because of the absence of “grand theory.” Industrial relations scholars saw themselves in need of a definable and ideally readily presentable “grand theory,” which would give them unquestioned standing as a “real” discipline (Howard, 1978; Laffer, 1974; Turkington, 1978). In this respect the influence of the nearby discipline of economics, which served as the early home of many academic industrial relations programmes was important. Unlike in the U.K., industrial sociology was not welldeveloped in Australia, and industrial relations was generally linked to economics and law. From the perspectives of those in the mainstream, economics and law are both “tight” disciplines which have highly defined theoretical and methodological bases and little room for heterodoxy.27 Thus scholars were not necessarily

The Transfer of Ideas in Industrial Relations

263

responding to high ideals of scholarship but simply to a need to convince their economist colleagues of the fitness of their proto-discipline to take its place according to the imperatives of the economic discipline of the time. That Dunlop’s analytical framework was developed by an American economist at a prestigious institution was an important aspect for academic economists in the 1970s and 1980s. Yet given the traditional, historical and Commonwealth links, transaction costs associated with take-up of ideas from U.K. would have been lower than those from the U.S. Despite this fact, those in the mainstream of academic industrial relations in Australia chose the Dunlop model as reinforcing what seemed to them to be the greater benefits accruing to the to their choice of this framework. Moreover, because of the significant role of public institutions in Australian industrial relations, there was, as in economics and law, a robust practitioner base. As a consequence, despite its small size, the Australian industrial relations academics also felt that they had to work to be relevant to current practitioner and policy concerns. As noted earlier, the Journal of Industrial Relations, the sole “academic” journal until the latter 1980s, had a strong practitioner bias. It was, in effect, the practitioners who set the research agenda.28 At the same time, the reformist bent of the Oxford researchers which might have seemed useful was not compelling for the Australian practitioners. This was in part because the particularistic British system with its shop stewards and voluntarism was perhaps too different. Moreover, the marginal status of industrial relations as an academic discipline was exacerbated by the nature of research funding processes in Australia. In general, funding was more easily accessible for recognised disciplines or those with a professional basis. It seems likely that academic industrial relations was borderline and the directions of research needed to recognise what was most accepted by the funding agencies (Niland, 1978b). Thus from the 1960s until the 1980s (and beyond?) academic industrial relations was at the margin of the social sciences within universities, and yet also viewed with some misgivings by the industrial relations practitioners.29 Some industrial relations academics believed that to gain legitimacy with those scholars in traditional disciplines, a clearly defined and tangible theoretical framework was essential. On the other hand, to convince the practitioners of the legitimacy of academic industrial relations analysis, two things were needed – relevance for contemporary imperatives and a surety that theories were “practical,” or to use the term of Kaufman (1993), problem-solving. Thus in evaluating the choice between the Oxford and Dunlop heuristic frameworks, the Australian scholars had to take account of the psychic costs and benefits of each model. What exaggerated the marginality of academic industrial relations was the public fear of communism in Australia and the portrayal of trade unions as potential hotbeds of radicalism. While Australian trade unions had long achieved legal

264

DIANA KELLY

legitimacy, and trade union density was high by world standards in the postwar era, social legitimacy was less readily acceded. Since the 1950s media and government opinion alike had scorned any evident leftist views. Although government actions were never as overt nor as specific as that of the McCarthy era in the USA, the experiences of the Split (see Note 19) and the strong overtones of antiCommunism had a clear effect on Australian universities. They also probably influenced the directions of Australian academic industrial relations, which sought to remove all apparent political signals. The preferences for the tentative, emerging discipline thus lay with the need to convince practitioners and policy-makers of detached analysis. This is evident for example in Dufty’s (1975) claims that Australian industrial relations were “non-ideological.” Unlike in the U.K. evidence of overt pro-trade union views, much less left-radical bias, was strongly resisted in mainstream academic industrial relations in Australia. As a consequence, industrial relations researchers in the 1960s–1980s were faced with the difficult task of investigating a field where the activities of one of the parties was viewed with scepticism, and yet where the effects of trade unions were perhaps greater than at any other time in Australian history.30 Not surprisingly, many industrial relations scholars took up the issue of “the labour problem,” in their quest to be “relevant.” This placed trade unions at the centre of their investigations throughout the 1970s and 1980s. As noted above the attention directed to researching the labour problem was partly related to the rising strike rates. Among the academics this turned the debates towards strike activity, and that old Australian chestnut of arbitration vs. collective bargaining (Cupper, 1976; Hancock, 1962; Howard, 1983; Kirby, 1965; Isaac, 1966, 1974; Niland, 1978a, 1986). The vigour of this debate was a strong theme in industrial relations thought until the mid-1980s, narrowing the focus of scholars who also felt the need to deal with this complex topic as if it should be value free. As exponents of a small and uncertain discipline, seeking legitimation from governments, from funding bodies and from the other social sciences, industrial relations academics took up an analytical framework which was seemingly without sympathy to any ideology. Evaluating the merits of such a (non?) position is not relevant here, although it has been noted that this characteristic was also apparent in the USA where ideological positions were rejected and the “field became both more unified in outlook and more neutrally professional in approach” (Kerr, 1978, p. 133).31 In order to promote the relevance and apparent rigour of their field, the Australian industrial relations analysts thus sought to draw on a model which appeared to sustain their “value free” stance, while offering a positivist, if not hypothetico-deductivist methodology. That these scholars achieved their goals and met the challenges of the many and conflicting problems of marginality of their discipline in the two decades

The Transfer of Ideas in Industrial Relations

265

from 1960 was evident in the growth of academic industrial relations (Cupper & Hearn, 1978). By the mid-1980s there were half a dozen chairs and full degree programmes at most Australian universities where there had been virtually none in the 1960s. There were new journals, major increases in postgraduate courses, and a successful academic association, alongside the Industrial Relations Society. All of these attest to the growth of the discipline and the achievement of some measure of status as a “discipline.” What had assisted them was their preference for use of Dunlop’s model which unlike the idealistic U.K. model provided an ideologically malleable analytical framework. It provided a means to analyse which seemed to be able stand outside moral or ideological bounds, yet also sufficiently resembled a classical economic model as to convey to social scientists that empiricism and rigour were hallmarks of industrial relations. Such considerations were important to industrial relations scholars, who sought acknowledgment from those in economics and personnel management/industrial psychology which stood close to industrial relations. Coincidentally these latter disciplines were themselves influenced by American patterns of thought (see e.g. Coats, 1985, 1993). There is no doubt that the Dunlop model could easily incorporate the actual Australian industrial relations system, given proximity between Dunlop’s specification of government agencies and the view of Australian scholars that tribunals were central actors. Thus while both the Dunlop and Oxford models focused on rulemaking and regulation of employment, Dunlop’s model was more readily adapted to Australian concerns over strikes and the effectiveness of the tribunal system in preventing strikes. In this respect the underpinning assumptions of the Dunlopian model appeared to enable both policy problem-solving and demands for an intellectually acceptable analytical framework. This was because, deriving as it did from Parsons’ social systems theory, the Dunlopian model was premised on the belief that such approaches should both analyse the “real world” and identify ways to achieve order or equilibrium or harmony. On the other hand it is important not to overstate the extent to which Dunlop’s systems approach was taken up in mainstream academic industrial relations in Australia. First, much Australian industrial relations thought remained atheoretical, and more closely aligned to history insofar as analysis was based on observation or investigation of events and processes, utilising small interpretative models, but without clear reference to any wider theory. Other scholars found Dunlop’s framework was inadequate in some way, especially those scholars who had come from economics (see e.g. Laffer, Dabscheck and Niland, Isaac). At the other end of the spectrum were those who claimed to use Dunlop’s analytical model, but who omitted or greatly modified sections of it. Most notably, from earliest days, Australian scholars excluded managers and management approaches, and were

266

DIANA KELLY

content to examine use those elements of the industrial relations system which most suited the concerns extant, strikes and the validity of the tribunal system. It was the same with Dunlop’s contexts – technological and budgetary which were excluded or viewed in ways somewhat different from that laid down originally (Dufty, 1975; Moth, 1972; Plowman et al., 1981; Walker, 1956, 1970). A purist may even argue that what was transferred to Australian industrial relations thought was not Dunlop’s framework at all, since so many elements were removed or altered. Nevertheless, the extent to which Dunlop’s model has been conveyed to successive generations of students, many of whom become practitioners, is evident in the apparent high levels of borrowing of Dunlop’s book at almost any Australian university library, the number of times it appears on reading lists and the frequency with which it is still cited. By early 1980s, before the situation for Australian industrial relations scholars became further complicated by new issues, the preference for Dunlop’s model over the Oxford approach was clearly evident.

CONCLUSION A study of the transfer of the ideas from U.K. and USA to Australian industrial relations thought suggests that a modified U.S. model was eventually more influential than that of U.K. In part this reflects the lower transaction costs inherent in the breadth of Dunlop’s model, its flexibility and its simplicity as a useful pedagogical device. However, credence needs also be give to the absence of a normative element in the U.S.-derived model, both in terms of the search for legitimacy and as a pedagogical tool. Students begin industrial relations with a great deal of ideological baggage and often very definite, if highly simplistic, perspectives on the surface aspects of industrial relations. In order to develop their capacity to examine the actors, processes and underlying motivations in industrial relations, academics, especially teachers, sought a framework which did not appear to advance an ideological perspective. Situated as they were in or near economics departments whose exponents eschewed any form of apparent ideology, the Australian industrial relations scholars selected a model they saw as having the most benefits in terms of legitimacy and acceptance from the other social scientists and practitioners. Dunlop’s model proved to be a safe and simple method of research which incorporated all the objects of study and research imperatives evident to academics, policy-makers and practitioners of the time. The framework simplified and systematised the immense complexity of the elements of industrial relations to students – the macro as well as the micro elements of the control and administration of the employment relationship. In the same way, Australian industrial relations scholars and analysts have sought to make their studies appear

The Transfer of Ideas in Industrial Relations

267

as “scientific” as other disciplines and ideologically amenable to governments and state agencies in order to be relevant for public policy. As a consequence, despite frequent and wide-ranging criticism of Dunlop’s model, it proved, albeit with some twists, a far more enduring transfer for mainstream analysis, than the British model. In terms of the transfer of ideas, it is worth noting firstly that because of the monolingual nature of Australian education, the choices for the industrial relations academics were first limited to Anglophone heuristic frameworks. Moreover, the tradition of looking to the example of U.S. and U.K. scholarship was very strong in Australia. Given the choice of the two Anglophone frameworks, then, the Australian scholars chose not the take up the Oxford model in fair part because of its psychic costs. Despite the belief of some that the Oxford model engendered more thorough-going research, they rejected the model not only because of its more particularistic approach, but also because it was more normative, an attribute under increasing suspicion in the economics departments of the 1960s and 1970s. By contrast through ease of use, perceived positivism and appreciation by policymakers and other, the Dunlop framework proved more desirable. There were compromises and offsets for the Australian scholars but the choice nevertheless was made for a modified Dunlopian framework. Clearly in considering the market for ideas consideration must be given for contextual determinants that may not seem immediately apparent. In the choice between Dunlop and Oxford these local factors proved important. In these times when ideologues are working to marginalise employment studies, reflective scholarship on the ways and modes in which ideas are transmitted within industrial relations seems a useful approach.

NOTES 1. See e.g. the growth of scientific management in Europe prior to World War II (Nyland, 1996) or the modes in which some of the ideas from the Hawthorne experiments were accepted (Gillespie, 1993). 2. See especially Chap. 5 and 6 of Kaufman (1993) for discussion on some these issues. 3. This raises the question of whether the “invisible college” performs the same function as the “invisible hand.” 4. Since the 1980s the practice and ideas of other countries, particularly Sweden and Germany, have come to influence ideas about employment relations in Australia, but USA and U.K. were unquestionably predominant until recently. 5. As I have noted elsewhere (Kelly, 1997), the debate over whether IR (or indeed any other area of specialist study) is a discipline or a field of study or a “specialism” is in most respects a fruitless exercise which does not advance or enhance our knowledge-getting processes. The terms “field of study” and “discipline” are used interchangeably in this paper. Graham et al. (1983) provide useful insights.

268

DIANA KELLY

6. Much of this self-doubt is self-imposed, and not limited to industrial relations. However, IR is rarely found in European universities although academic analysis of the work, management and employment is widespread. Moreover, industrial relations/employment relations scholarship is rarely cited in other disciplines, despite the considerable analytical insights that industrial relations analysts have developed. 7. These terms which have gained more common parlance in the feminist literature are clearly germane to the Australian situation – especially at present when governments and business are working to return the employment relationship to the private sphere. 8. In recent years governments have radically reduced national wage determination. An examination of the March issue of the Journal of Industrial Relations in the 1970s and 1980s demonstrates the extent and effect of public sphere industrial relations. 9. See also M. Perlman, “An Industrial Problem: Australia’s Longshoremen,” Labor Law Journal, Vol. 4, 1953, pp. 462–473, which more than his larger work focuses on management. 10. See especially, F. T. de Vyver, “The Weakening of Managerial Rights” Business Horizons, 2, 1, 1959, pp. 38–48. What is notable about the work of both de Vyver and Perlman is that like Dunlop, they gave far greater credence to the role of management than Australian scholars. 11. Dunlop studied economics initially in California, gaining his doctorate in 1937. In 1937–1938 he studied wage theories at Cambridge, before returning to take up his post at Harvard, with which he has remained associated continuously since. His orientation and epistemological norms are therefore those of an economist. 12. As is widely recognised, Dunlop (1993) comprises Dunlop (1958) with an extensive commentary of responses to critics from the previous forty-five years. 13. Some writers have argued that Dunlop’s focus was only on unionised environments although in his “Commentary” which prefaced the 1993 printing of Industrial Relations Systems, the latter strongly disagrees (Dunlop, 1993, pp. 14, 15). 14. These days pluralism is an overused term, frequently utilised (not quite appropriately) as a catch-all term to indicate suggest acknowledgment of the conflictual aspects of the employment relationship. The term has caught the imagination of business liberals who seek to discredit ideas of competing interests in employment. It is also a preferred term for those analysts who seek to emphasise the antagonistic aspects of the employment relationship over the cooperative attributes. 15. See also Hyman (1994). Hyman, a notable and insightful Marxist scholar was supervised by Clegg. 16. The notion of job regulation is different from the Webbs’ notion of the Device of the Common Rule which they saw as a core trade union strategy which contrasted with the less equal outcomes of individual bargaining. 17. In this respect it is worth noting that while Flanders at one point specifically eschews “unstructured relationships” as being part of industrial relations (Flanders, 1975, p. 86) much of the rest of his writing identifies the importance of all relationships in the workplace (see e.g. his work on Fawley Productivity agreements, and “The Internal Social Responsibilities of Industry.”) 18. Dunlop is adamant that scholarly analysis must remain positivist and separate from “values.” He retains his long-held view that it is the role of the academic practitioner, the problem-solver’ first to analyse, then if they wish, to add recommendations (ideals) as

The Transfer of Ideas in Industrial Relations

269

addendum. In this way the normative may be inserted into the regulation, but only after, and separate from positivist analysis. Interview John T. Dunlop, 14 October, 1997. 19. “The Split” was a major rupture in Australian labour politics. Throughout the 1950s there was considerable rumour and turmoil over the extent to which communism had permeated the Australian labour movement. The large anti-communist faction, especially in Melbourne, split from the Australian Labor Party to form the Democratic Labor Party (DLP) in 1955. The DLP, which had considerable effect on Australian political events from the 1950s, enabled the Liberal Coalition to remain in government until 1972 when the Whitlam Labor government was elected. The DLP ceased to have any parliamentary presence from that time, but the schismatic effects of the Split have had long term impacts on the Australian labour movement (see Ellem, 1998, see also Kuhn, 1955). 20. Kerr, a foundation member of the Industrial Relations Society was a widely respected lawyer, particularly in areas of labour and employment law. In 1975, as Governor General of Australia he oversaw the controversial dismissal of the Whitlam Labour Government. 21. This is not uncommon in an emerging discipline. For example, Jensen (1969) notes the same tensions between practitioner orientation and scholarship orientation in the emergence of political science. See also Ash (1983). 22. Of course Dunlop’s influence reached further back than the 1970s. For example, in his thesis submitted at London University in 1947, Isaac cited five of Dunlop’s publications, albeit all focussing on the economics of wages. This raises the question, not dealt with in this paper, of how far U.S. thought influenced U.K. thought. For excellent discussion the relationship between U.S. economics and the British profession, see Coats (1993), Chap. 8 and 21. 23. Laffer (1972, p. 68), see also Laffer (1976). Laffer was central to the emergence of academic industrial relations in Australia. 24. In the same J. I. R., there were inter alia articles or research notes by Isaac, Niland, Groenewegen, Gordijiew, and Beaumont, all economists, and similar numbers of articles by scholars from organisational behaviour/psychology and law. This disciplinary array highlights the influence of different disciplines on industrial relations, an influence which to some extent transects the cultural-geographical patterns of influence, even if at times one discipline appears predominant. 25. This is not to reject the significance of the text by Dabscheck and Niland in the same year which gave immense emphasis to state agencies and processes but specifically rejected Dunlop (pp. 27, 28). This text was undoubtedly quite important for some years, but there were no further editions. Also important as texts in this period several anthologies of Labour Relations Readings (see e.g. Ford, Hearn & Lansbury, 1980). 26. Beyond a few small programmes in Colleges of Advanced Education, business education in the broad sense, only began in Australian universities in the 1980s when the Australian Graduate School of Management was founded in 1984. 27. In this respect, it is worth noting that major IR scholars of the time, – Niland, Isaac, Hancock, Laffer, and Howard, to name a few – for example all trained as economists, and exhibited economists’ discomfort with intangible values and other loose ends. For discussion on loose and tight disciplines, see e.g. Becher (1989), and Coats (1993). 28. As Kaufman (1993) has clearly highlighted in his science-building vs. problemsolving dichotomy, this was not a problem unique to Australia. 29. Keenoy (1984, pp. 255–257) describes industrial relations as the Cinderella of the social sciences. See also, Ford (1980).

270

DIANA KELLY

30. The rapid decline of union density in fifteen years from 1980, from over 55 to 30% perhaps attests to the level of social legitimacy of unionism. 31. It is perhaps no coincidence that explicit rejection of the left in the postwar years was rather more evident in the USA and Australia than in the U.K.

ACKNOWLEDGMENTS The author wishes to thank the readers who have commented assiduously on previous drafts of this paper.

REFERENCES Ash, M. G. (1983). The self presentation of a discipline: History of psychology in the United States between pedagogy and scholarship. In: Graham, Lepenies & Weingart (Eds), Functions and Uses of Disciplinary Histories. Dordrecht: P. D. Reidel. Bain, G. S., & Clegg, H. (1974). A strategy for industrial relations research in Great Britain. British Journal of Industrial Relations, 12(1), 91–113. Becher, T. (1989). Academic tribes and territories: Intellectual enquiry and the culture of disciplines. Milton Keynes: SRHE & Open University Press. Bray, M., & Taylor, V. (1986). Introduction: Directions in industrial relations research. In: M. Bray & V. Taylor (Eds), Managing Labour: Essays in the Political Economy of Australian Industrial Relations. Sydney: McGraw-Hill. Boraston, I., Clegg, H., & Rimmer, M. (1975). Workplace and union. London: Heinneman. Brown, W. (1978). Transforming the miraculous into the natural: Current trends in industrial relations research in Britain. In: D. Turkington (Ed.), Industrial Relations Research and Teaching in Australia and New Zealand. Wellington, New Zealand: IRC, Victoria University. Byrt, W., & Masters, P. R. (1974). The Australian manager. Melbourne: Sun Books. Child, J. (1962). Recent developments in labour relations. Proceedings of the 4th Convention of the Industrial Relations Society of NSW, Sydney: IRS of NSW. Child, J. (1963). Recent problems in labour relations. Proceedings of the 5th Convention of the Industrial Relations Society of NSW. Sydney: IRS of NSW. Clegg, H. (1960). A new approach to industrial democracy. London: Basil Blackwell. Clegg, H. (1972). The British system of industrial relations. Oxford: Basil Blackwell. Clegg, H. (1990). The Oxford school of industrial relations. Warwick Papers in Industrial Relations, No. 31, January. Warwick University, UK: School of IR and Business Studies/IRRU. Coats, A. W. (1985). The American economic association and the economics profession. Journal of Economic Literature, 23(December), 433–473. Coats, A. W. (1993). The Sociology and Professionalization of Economics: British and American Economic Essays. London: Routledge. Coats, A. W., & Colander, D. C. (1989). An introduction to the spread of economic ideas. In: D. C. Colander & A. W. Coats (Eds), The Spread of Economic Ideas. Cambridge: Cambridge University Press. Colander, D. C. (1989). The invisible hand of truth. In: D. C. Colander & A. W. Coats (Eds), The Spread of Economic Ideas. Cambridge: Cambridge University Press.

The Transfer of Ideas in Industrial Relations

271

Cupper, L. (1976). Legalism in the Australian conciliation and arbitration commission: The gradual transition. Journal of Industrial Relations, 18(4), 337–364. Cupper, L., & Hearn, J. M. (1978). Industrial relations teaching in Australian universities: The industrial relations programme, University of Melbourne. In: D. Turkington (Ed.), Industrial Relations Research and Teaching in Australia and New Zealand. Wellington, New Zealand: IRC, Victoria University of Wellington, New Zealand. Dufty, N. (1972). Industrial relations in the Australian metal industry. Sydney: West Publishing. Dufty, N. (1975). Changes in Labour Management Relations At The Enterprise. Paris: OECD. Dufty, N. (1979). Industrial relations in the public sector: The firemen. St Lucia, Queensland, Australia: University of Queensland Press. Dufty, N. (1980). Employer associations in Australia. Australian Bulletin of Labour, 6(4), 209–244. Dunlop, J. T. (1958). Industrial relations systems. New York: Holt. Dunlop, J. T. (1977). Policy decisions and research in economics and industrial relations. ILRR (April), 275–282. Dunlop, J. T. (1993). Industrial relations systems. Boston, MA: Harvard Business School. Ellem, B. (1998). Preface and introduction: The split revisited. In: B. Ellem (Ed.), The Great Labour Movement Split in New South Wales: Inside Stories. Sydney: Australian Society for the Study of Labour History. Flanders, A. (1964). The Fawley productivity agreements: A case study of management and collective bargaining. London: Faber. Flanders, A. (1965). Industrial relations: What is wrong with the system. London: Faber. Flanders, A. (1975). The internal social responsibilities of industry. Management and Unions: The Theory and Reform of Industrial Relations. London: Faber. Flanders, A., & Clegg, H. (Eds) (1954). The system of industrial relations in Great Britain. Oxford: Basil Blackwell and Mott. Ford, G. W. (1980). Industrial relations research in Australia: The question of responsibility and the importance of the theory of ignorance. In: G. W. Ford, J. Hearn & R. D. Lansbury (Eds), Australian Labour Relations Reading (3rd ed.). South Melbourne: Macmillan. Ford, G. W., Hearn, J., & Lansbury, R. D. (Eds) (1980). Australian labour relations reading (3rd ed.). South Melbourne: Macmillan. Fox, A. (1966). Industrial sociology and industrial relations. London: HMSO. Fox, A. (1973). Industrial relations: A social critique of pluralist ideology. In: J. Child (Ed.), Man and Organization. London: Allen & Unwin. Fox, A. (1985). Man mismanagement. Hutchinson, London 2nd ed. – 1st ed. 1975. Fox, A., & Flanders, A. (1969). The reform of collective bargaining: From Donovan to Durkheim. British Journal of Industrial Relations, 7(2), 151–174. Gillespie, R. (1993). Manufacturing Knowledge: A History of the Hawthorne Experiments. Cambridge: Cambridge University Press. Goodwin, C. D. W., & Holley, I. B. (Eds) (1968). Towards and theory of the intercultural transfer of ideas. In: C. D. W. Goodwin & I. B. Holley (Eds), The Transfer of Ideas: Historical Essays. Durham, NC: Duke University Press. Graham, L., Lepenies, W., & Weingart, P. (Eds) (1983). Functions and uses of disciplinary histories. Dordrecht: D. Reidel. Groenewegen, P., & MacFarlane, B. A. (1990). A history of Australian economic thought. London: Routledge. Hall, C. R. (1971). Australian Manufacturing Achievements to 1960. Sydney: Angus and Robertson.

272

DIANA KELLY

Hancock, K. (1962). Compulsory arbitration vs. collective bargaining: Three recent assessments. Journal of Industrial Relations, 4(1), 20–31. Hill, J., Howard, W. A., & Lansbury, R. (1982). Industrial relations: An Australian introduction. Melbourne: Australian Management Studies series, Longman Cheshire. Howard, W. A. (1978). Doctrine, theory, and teaching in industrial relations. In: D. Turkington (Ed.), Industrial Relations Research and Teaching in Australia and New Zealand. Wellington, New Zealand: IRC, Victoria University of Wellington, New Zealand. Howard, W. A. (1983). Centralism and perceptions of Australian industrial relations. Journal of Industrial Relations, 25(1), 3–25. Hyman, R. (1994). Theory and industrial relations. British Journal of Industrial Relations, 32(2), 165–180. Hyman, R. (1995). The historical evolution of British industrial relations. In: P. K. Edwards (Ed.), Industrial Relations: Theory and Practice in Britain. Oxford: Basil Blackwell. Hyman, R., & Brough, I. (1975). Social values and industrial relations: A study of fairness and inequality. Oxford: Basil Blackwell. Isaac, J. E. (1966). The prospects for collective bargaining in Australia. In: J. E. Isaac & G. W. Ford (Eds), Australian Labour Relations: Readings. Melbourne: Sun Books. Isaac, J. E. (1974). Compulsory arbitration and collective bargaining reconsidered. Journal of Industrial Relations, 16(1), 1–12. Jensen, R. (1969). History and the political scientist. In: S. M. Lipset (Ed.), Politics and the Social Sciences. New York: Oxford University Press. Kahn-Freund, O. (1969). Industrial relations and the law – retrospect and prospect. British Journal of Industrial Relations, 7(3), 301–316. Kaufman, B. E. (1993). The origins and evaluation of the field of industrial relations in the United States. Ithaca: ILR Press. Keenoy, T. (1984). Invitation to Industrial Relations. Oxford: Basil Blackwell (reprint). Kerr, C. (1978). Industrial relations research: A personal retrospective. Industrial Relations, 17(2), 131–143. Kindleberger, C. P. (1989). How ideas spread among economists: Examples from international economics. In: D. C. Colander & A. W. Coats (Eds), The Spread of Economic Ideas. Cambridge: Cambridge University Press. Kirby, R. (1965). Some comparisons between compulsory arbitration and collective bargaining. Journal of Industrial Relations, 7(1), 1–17. Kuhn, J. W. (1955). A note on communists and strikes in Australia. Political Science Quarterly, 70(March), 103–106. Laffer, K. (1972). Dr Walker on industrial relations. Journal of Industrial Relations, 14(1), 67–70. Laffer, K. (1974). Is industrial relations an academic discipline? Journal of Industrial Relations, 16(1), 62–73. Laffer, K. (1976). Where are we going in industrial relations? Journal of Industrial Relations, 18(3), 283–293. Lansbury, R. (1978). Industrial relations in the post-Dunlop era: An Australian view of teaching and research in the United States. In: D. Turkington (Ed.), Industrial Relations Research and Teaching in Australia and New Zealand. Wellington, New Zealand: IRC, Victoria University of Wellington, New Zealand. Laudan, R. (1983). Redefinitions of a Discipline: Histrories of Geology and Geological History. In L. Graham, W. Lepenies & P. Weingart (Eds.), Functions and Uses of Disciplinary Histories. Dordrecht: Reidel.

The Transfer of Ideas in Industrial Relations

273

Moth, L. G. (1972). Industrial relations in the plant: A study of evolution within the system. Journal of Industrial Relations, 14(3), 282–293. Niland, J. (1978a). Collective bargaining and compulsory arbitration in Australia. Kensington: NSW University Press. Niland, J. (1978b). Funding aspects of university research in industrial relations: An Australian perspective. In: D. Turkington (Ed.) Industrial Relations Research and Teaching in Australia and New Zealand. Wellington, New Zealand: IRC, Victoria University of Wellington, New Zealand. Nyland, C. (1996). Taylorism, John R. Commons, and the Hoxie Report. Journal of Economic Issues, 30(4), 985–1017. Olson, M. (1969). The relationship between economics and the other social sciences. In: S. M. Lipset (Ed.), Politics and the Social Sciences. New York: Oxford University Press. Plowman, D., Deery, S., & Fisher, C. (1980). Australia Industrial Relations. Sydney: McGraw-Hill. (Revised edition, 1981.) Report of the Royal Commission on Trades Unions and Employer Associations (1966). London: HMSO. (Donovan Report). Rimmer, M. R. (1973). Review of The system of industrial relations in Britain. Journal of Industrial Relations, 15(2), 221–222. Turkington, D. (Ed.) (1978). Industrial relations research and teaching in Australia and New Zealand. Wellington, New Zealand: IRC, Victoria University of Wellington, New Zealand. Walker, K. F. (1956). Industrial relations in Australia. Cambridge, MA: Harvard University Press. Walker, K. F. (1955). The scientific study of industrial relations. Australian Quarterly, 27(September), 55–66. Walker, K. F. (1970). Australian industrial relations systems. Cambridge, MA: Harvard University Press. Webb, B. (1938a). My apprenticeship, Vol. 1. Harmondsworth: Penguin. Webb, B. (1938b). My apprenticeship, Vol. 2. Harmondsworth: Penguin. Webb, S., & Webb, B. (1902). Industrial democracy. London: Longmans Green. Webb, S., & Webb, B. (1975). Methods of social study. Cambridge, UK: C. U. P., first published 1932. Woods, N. S. (1978). The future of teaching and research in industrial relations. In: D. Turkington (Ed.), Industrial Relations Research and Teaching in Australia and New Zealand. Wellington, New Zealand: IRC, Victoria University of Wellington, New Zealand.