Corps de l’article

Faced with financial pressures and neo-conservative ideologies, public managers have increasingly restructured public services in North America. Contracting out, the most studied form of privatization, has been at the centre of a debate over the role of the private sector in service provision (Savas, 2000; Sclar, 2000). Somewhat lost in this discourse is a careful examination of the impact of restructuring on employees and their unions. This study attempts to fill this gap by examining the impact of service contracting on unions and employees of cases at the local level of government in New York State.

The cases analyzed were drawn from a larger survey of townships and counties in upstate New York (outside New York City) by the author and Mildred Warner (Warner and Hebdon, 2001). The 1997 survey asked chief elected officials about restructuring decisions since 1990 in all townships and upstate counties. In the last part of the survey, respondents were asked to provide brief descriptive summaries of restructuring cases, together with a contact person and telephone number. The author employed a computer-assisted research team (CAST) at Cornell University to conduct a telephone survey using these contacts to examine various labour effects of the contracting for services process (discussed more fully below).

The paper begins with a brief overview of the New York state context at the time of the survey. The literature on the labour effects of privatization is then reviewed. Next, the data, methodology, and hypotheses are presented followed by the results and a conclusion.

The New York State Environment

It is useful to examine some of the factors that condition the privatization environment in New York at the time of this study. This context is important not only to understand the external factors in play, but also to examine limits on the applicability of the findings. Public employees in New York State are covered by the Taylor Law that provides for collective bargaining over a broad range of issues. The law, for example, requires negotiation over some aspects of contracting out (Donovan and Orr, 1982). New York ranked in the top two or three in unionization (as measured by union density) in both public and private sectors (BLS, 1998). For example, in 1998 New York’s public sector union density was over 70 percent while the average figure in the U.S. was 37.5 percent (BLS, 1998). Given the high levels of unionization, the existence of New York’s Taylor Law, and generally labour-friendly democratic governors, New York state was one of the least privatized states in the country in 1995 (Lauder Report, 1992).

In 1992, New York experienced a political change to a republican Governor, George Pataki, with an aggressive privatization agenda (see Lauder Report, 1992). Given this change and the fact that some 70 percent of the local politicians in the larger survey from which these cases were drawn were Republicans, the expectation was that there would be a new wave of privatization. The cases analyzed in this paper occurred in the three-year period immediately following the election of George Pataki as Governor.

Literature Review

Effects on Workers

The empirical question of the impact of privatization on workers has generated a vigorous debate in the literature. Proponents of privatization assert that workers do not lose job security, wages or conditions, or unionization as a result of contracting out (Savas, 2000). It is further argued that public employers provide generous adjustment policies for affected employees such as early retirement, job transfer, and no layoff policies (Savas, 2000). Others maintain, on the other hand, that public sector employees suffer lower wages and job loss (Kuttner, 1999). Also public sector unions have argued that contracting out adversely affects wages and benefits; often results in job loss; is a direct attack on unions and collective bargaining; and has a disproportionate effect on women and minorities (Dantico, 1987).

A survey of local government public works directors in the U.S. showed that unionized cities are less likely to privatize, but this opposition is not expressed through collective bargaining but by political action (Chandler and Feuille, 1991). Privatized public sector jobs shift from the public sector to the private service sector. Unionization in the private service sector in the U.S. has been between 5 percent and 6 percent for much of the past 25 years and has always been significantly lower than both the public and manufacturing sectors (BLS, 2004). Thus, the probability of remaining unionized after contracting out is extremely small. In terms of average union densities, for example, workers in the private service sector are seven times less likely to be unionized than in the public sector (BLS, 2004).

There is very little research on the post-privatization effects on labour (World Bank, 1999). An exception was a study of the U.S. Federal Government (GAO, 1985). It showed that 53 percent of the workers that took jobs with contractors had lower wages, 21 percent had higher wages, and in 25 percent of the cases, wages were the same. At the local level of government, it was also found that contractor wages were lower for bus drivers and mechanics even where the contractor was unionized (Peterson, Davis, and Walker, 1986).

In a national study of local government, Jackson (1997) examined the attempt by public employers to manage tensions between management and employees caused by private service delivery. She defined three categories of adjustment strategies: internal (retraining, early retirement, transfers); selective use plans (apply only to new services, tie to attrition); and private sector contracts (job offers to displaced workers, wage parity, affirmative action requirement). Two-thirds of cities in this national survey reported no labour adjustment strategy for privatization. Only 33 percent overall employed adjustment techniques in one of the three Jackson categories and only 15 percent of cities used at least one program in each of the three categories.

A case study of privatization in Canada Post revealed that affected postal employees had post-privatization unemployment rates two and one-half times the national average despite the existence of a comprehensive employer adjustment policy (White and Janzen, 2000).

Privatization has been found to adversely impact minorities by disproportionately displacing them through contracting out, by reducing their employment mobility opportunities, and by diminishing their living standards, thereby increasing the discrimination wage gap with whites. There is ample evidence that governments employ minorities and women in higher numbers than in the private sector. According to data obtained from the Current Population Survey, women and blacks comprise 39.0 percent and 12.8 percent of all wage and salary earners in the U.S., respectively (BLS, 2000). But in public administration, women and blacks were currently 44.0 percent and 24.0 percent, respectively, of total wage and salary earners (BLS, 2000). Not only do governments employ more women and blacks, but women have higher salaries and are more likely to be in professional jobs than in the private sector (Dantico, 1987; Thurow, 1982). Similarly, blacks are also higher paid in government and are more likely to be in professional and managerial occupations than blacks in the private sector (Joint Center for Political Studies, 1985). Thurow (1982) found, for example, that women’s salaries were 20 percent higher and minorities’ salaries 30 percent higher in public employment than in the private sector. This result is due mainly to the higher levels of responsibility achieved by women and blacks in government employment.

The privatization policies of the Thatcher Government in the 1980s produced an extensive literature that examined various industrial relations effects. While the political will of Governor Pataki might have paralleled Thatcher’s a decade earlier, New York’s Governor lacked the legal authority to force privatization on local government. Despite this contextual difference, the U.K. experience is relevant to that of New York.

Pendleton (1999) compared conditions in 22 privatized firms with 25 public ones in the bus services industry in the United Kingdom. Several hypotheses were tested about the impact of privatization on industrial relations. Pendleton’s study is of interest for at least two reasons: the bus industry has an equal mix of private and public firms competing for contracts, and a clear policy goal of the Thatcher Government was to weaken union power and influence.

Pendleton found few or no differences between private and public firms on the extent of unionization and collective bargaining; however, wages were significantly lower in the private firms. The level of competition in the local market had no significant impact on any of these findings. Pendleton’s (1999) finding concerning the lack of impact on unionization is consistent with earlier U.K. studies. Colling (1991), for example, shows that privatization in the electricity sector did not have the predicted impact on unions. The attempts to weaken unions through decentralization of authority largely failed because of the inherent centralized nature of the service. Ferner and Colling (1991) further show that privatization does not necessarily improve industrial relations and that analysis should be carried out on a case-by-case basis. In a comprehensive review of over 75 studies in the U.K., Marsh (1991) found that a main impact of privatization on employees was deterioration in pay and conditions and a loss of jobs.

Union Opposition

Researchers have found employee opposition to be a significant negative factor in contracting out cases (National Commission, 1988; Ferris and Graddy, 1986). Studies provide support for the hypothesis that the labour-management relationship influences the decision to privatize (Chandler and Feuille, 1994). They found that cities in which there is a cooperative relationship are less likely to implement privatization than are non-union cities. However, cities with a more adversarial relationship with the union are more likely to privatize sanitation services. A problem with cross-sectional studies is their failure to answer the question of causality. One cannot rule out the reverse hypothesis that privatization consideration caused the adversarial relationship. It seems equally plausible, for example, that the consideration of privatization by a city could turn an otherwise cooperative relationship into a confrontational one. Only a longitudinal methodology can resolve this causality problem.

A broader view of privatization casts the restructuring decision not as a dichotomy between public and private but as a range of options, some of which may elicit employee opposition and others not (Warner and Hebdon, 2001; Warner and Hefetz, 2002; Brown and Potoski, 2003). For example, opposition may result from privatization or cessation of services where the size and scope of government is reduced, thus eliminating public service jobs. Restructuring that involves cooperation between governments to jointly administer services without affecting line employees may have a neutral impact on opposition. Finally, contracting back in services previously contracted out may increase the public provision of services and elicit support from public employees and their unions (Warner and Hebdon, 2001).

This model of local government sees managers operating in a very practical world where decisions are made in a non-ideological context. Rather than downsize government for its own sake, they will typically expand and diminish the role of the public sector at the same time through a complex mix of restructuring strategies.

Thus, union opposition to restructuring was found when privatization played a primary role in the mix of management strategies (Warner and Hebdon, 2001). Unions were less likely to oppose management strategies that expanded public employment. Unionization ceased to be significant when local governments employed a more complex mix of alternatives, which might expand public services through contracting services back in. Union opposition, therefore, was influenced by management restructuring choices.

Those few studies which have acknowledged the reverse privatization (i.e., contracting back in) phenomenon have not included it in their analysis (Chandler and Feuille, 1991, Kodrazycki, 1998). An exception is the paper by Lopez-de-Silanes, Shleifer and Vishny (1995) that identified a process of “switching” services back and forth between public and private sectors.

Data and Methodology

In the Warner and Hebdon (2001) mailed survey of towns and counties in upstate New York, recipients were asked to provide written cases of restructuring and telephone numbers of key management contact persons. This open-ended question produced 70 contacts.[1]

Questions were asked about the fate of the affected workers (if any) and conditions of employment in the privatized firm. The survey resulted in 54 usable responses of which, 39 were cases of privatization, nine were cases of contracting back in to the public sector, and the remaining six cases were contracting out to other government or non-profit employers. There were 242 workers affected in the 39 cases of privatization (Table 3).[2] A broad range of services was affected (e.g., payroll, school bus drivers, solid waste, snow plowing, ambulance, airport management, and human services).

Because it is unlikely that managers would volunteer privatization failures, there is a potential selection bias in the reporting of these 70 cases. To verify the information provided by managers, an attempt was made to contact the union representative where a union was involved in the case. This process had limited success due to the lack of knowledge of the case on the part of the union representative or their unavailability. Thus, it was not possible to rigorously assess the impact of the bias except to point out that the reported cases probably understate the negative labour effects of restructuring.

In order to examine union opposition, employee morale, and the effect of adjustment policies, it was necessary to build a model to control for various factors. As discussed above, the early public administration literature examined privatization in the form of contracting out as a dichotomy between private and public provision. The decision to contract out was found to be more likely when there are greater cost savings from wages, more fiscal pressure to reduce public service costs, and less opposition from such public service constituencies as the elderly, lower income groups, and public employees (Ferris, 1986). Privatization models typically included the cost of service provision, local economic conditions, complexity and size of government, and unionization (Kodrazycki, 1994, 1998; Stein, 1990; Lopez-de-Silanes, Shleifer and Vishny, 1995).

To provide a more rigorous test of various hypotheses about the role of adjustment policies and important various factors in the privatization decision-making process, a model is constructed to perform multiple regression analysis. The two dependent variables that are employed are employee opposition and employee morale. Opposition is a dummy variable taking a value of 1 if there was employee or union opposition to the privatization, 0 otherwise. Employee morale is a variable with a value of 2 indicating a positive atmosphere, 1 neutral, and 0 a negative one. Given these distributions, logistic regression is used in the case of opposition and tobit analysis for morale.

The regression sample is restricted to the 39 cases of privatization for profit.[3] The model includes independent variables that record the presence of various adjustment policies (moved employees to another county MOVEDC and offered early retirement EARLYRET), the existence of a union (UNION), whether the employees are skilled or not (SKILLED), and whether the following factors were important (legal/political, community values, and contract monitoring).[4]

Because of the small size of the sample, it was necessary to omit variables where there was no strong theoretical reason for including them or where they were highly insignificant. The number of employees affected was dropped from the analysis because it was not significant in any of the regressions. This was not the expected result since we anticipated greater opposition where a large number of employees was affected. The following factors were also not included in the analysis: economic, budgetary, management, local employment, service quality, availability of information and past experience with restructuring. Note that reasons of parsimony, the legal and political factors were combined into one variable.

Hypotheses

Hypothesis on Wages and Benefits. Given the mixed findings in the literature, the effect on wages and benefits could not be predicted.

Hypothesis on Employment Effects. Similarly, the employment effects are indeterminate.

Hypothesis on Race and Gender. Given that public employment has higher than average levels of women and visible minorities than societal averages, it was expected that there will be disproportionately more women and minorities affected by privatization. A more rigorous test of this proposition would be to compare race and gender proportions in the sample with the expected rate based on the average rate for women and visible minorities in public employment in the areas affected.

Hypothesis on Union Effects: Unionization. Given the findings in the literature and the higher union density in the public sector (especially in New York), privatization is expected to reduce unionization levels.

Hypothesis on Union Effects: Union Opposition and Employee Morale. It was expected that union opposition would be reduced and morale increased where an adjustment policy existed or if the case involved the contracting back in of previously contracted out work. If, for example, the town guaranteed no layoffs and provided for early retirement, less opposition would be expected. On the other hand, the presence of such a strong policy may have been a response to opposition. It is further anticipated that a case of contracting in of work would generate less opposition from employees and higher morale since jobs are not threatened and in some cases, unions may be gaining members.

Results

Wages and Benefits. The data in Table 1 reveal that private contractor wages and benefits were not substantially different from those of the public employers. Slightly more employees received lower wages and benefits than those receiving higher compensation but for the largest category of contractors, wage and benefits were about the same.

Table 1

Comparison of Wages and Benefits, Public Employers and Private Contractors, New York State, 1990–1995 (39 cases)

 

Cases

Employees

Wages higher in private contractor

8

18

Wages lower in private contractor

8

28

Wages about the same in private contractor

21

138

Missing cases

2

 

Total

39

184

Benefits higher in private contractor

8

18

Benefits lower in private contractor

7

39

Benefits about the same in private contractor

15

109

Missing cases

9

 

Total

39

166

-> Voir la liste des tableaux

Employment Effects. Retraining was the most common adjustment policy, applying to 36 percent of employees affected by privatization (Table 2). The next highest category covering 32.2 percent of employees was to have no policy. Only 23 percent of employees found jobs with the private sector firm and an even smaller number (3.7 percent) obtained another job in the local town or county. The picture presented here is not as bleak as the national survey discussed above that showed that two-thirds of cities reported no labour adjustment strategy for privatization (Jackson, 1997). On the other hand, the adjustment policies that were found here do not support the rosy picture of privatization that some proponents portray (see Savas, 2000).

Table 2

How Employees Were Affected by Contracting Out to Private Sector, Towns and Counties in New York State, 1990–1995 (26 cases)*

Category

Privatization

 

Cases

% of total cases (26)

Employees

% of total emps. (242)

Retrained

11

42.3

87

35.9

Found other job in town or county

5

19.2

9

3.7

Moved to private firm

9

34.6

56

23.1

Early retirement

4

15.4

6

2.5

Formal layoff

8

30.8

11

4.5

None

7

26.9

78

32.2

* In 13 of the 39 cases there were no employees affected by privatization. Percentages are expressed as a proportion of 26 cases and 242 affected employees. Please note that columns cannot be totalled because in some cases more than one policy applied to one employee.

-> Voir la liste des tableaux

Race and Gender. The question that was examined was this: are disproportionately more women and minorities contracted out in relation to the current employment distribution of government employers in the area?

The worker characteristics shown in Table 3 of those employees affected by privatization were compared with New York county averages using New York 1990 census of government data. When we assumed that the contracting out of 128 blue collar, 30 white collar, 5 technical, and 34 professional employees would be made based on the census of government as to the proportion of women in each of these occupational categories for actual counties surveyed, the results predicted that from 61–68 women would be affected.[5] Since the actual number affected was 98, it is concluded that women were adversely impacted by privatization. This result occurred despite the relatively high number of typically male blue-collar employees affected.

Table 3

Characteristics of Employees Affected by Contracting Out to Private Sector, Towns and Counties in New York State, 1990–1995 (26 cases)*

Category

Privatization

 

Cases

% of cases (26)

Employees

% of emps. (242)

Skill level

 

 

 

 

– Unskilled

9

34.6

32

13.2

– Partly skilled

9

34.6

58

24.0

– Skilled

13

50.0

104

43.0

– Highly skilled

7

26.9

43

17.8

Occupation type

 

 

 

 

– Blue collar

18

69.2

128

52.9

– White collar

6

23.1

30

12.4

– Technical

7

26.9

5

2.1

– Professional

2

3.8

34

14.0

Women

19

73.1

98

40.5

Visible minorities

4

15.4

9

3.7

Part time

13

50

52

21.5

Union

9

34.6

134

55.4

* In 13 of the 39 cases there were no employees affected by privatization. Percentages are expressed as a proportion of 26 cases and 242 affected employees.

-> Voir la liste des tableaux

A similar calculation can be made with respect to visible minorities. According to the census of government data, the proportion of visible minorities in these occupations in upstate New York local government averages about 1 percent. Thus, even the relatively small proportion of affected visible minorities of 3.7 percent is substantially more than the predicted number of 1 percent based on existing minority proportions in the four occupational groups.

Union Effects: Unionization. Table 4 shows that out of 134 unionized workers before contracting out, 98 lost unionization status as a result of privatization. This means that 73 percent of all unionized jobs in the survey lost their union status after contracting out. This de-unionizing effect of privatization supports the hypothesis of Dantico (1987) and was similar to the U.K. experience.

Table 4

Unionization Losses Due to Privatization, New York State Survey, 1990–1995 (39 cases)*

 

Before privatization

After privatization

 

Cases

Employees

Cases

Change

Employees

Change

Non-union cases

30

108

36

+ 6

206

+ 98

Unionized cases

9

134

3

– 6

36

– 98

Total

39

242

39

 

242

 

* Unions actually lost eight out of the nine bargaining units but gained two because two of the private contractors were unionized. Hence the “after privatization” total of three unionized units.

-> Voir la liste des tableaux

Union Effects: Union Opposition and Employee Morale. The regression results in Table 5, column 1, show that unionized towns were 68.5 percent more likely to oppose privatization. Union presence, however, had no significant impact on employee morale (column 2).

It was anticipated that union opposition would be substantially less in the cases where work is brought back in house. While multiple regression analysis was not possible, a bivariate correlation coefficient of –.316 with union opposition (significant at the 2 percent level), indicated union support in these contracting back in cases. In addition, in the six cases of contracting out to other governments, there were no cases of employee or union opposition. These results contrast sharply with the significance of the union variable in the regression analysis and provide support for the Warner and Hebdon (2001) result showing that towns and counties that employ a mix of restructuring strategies that include contracting in may mute union opposition.

The only adjustment policy that was significantly associated with both higher opposition and lower morale was moving employees to another county. While the policy may seem progressive, it is believed that the negative result was obtained because of the conflict that preceded the policy. In this case, the policy may reflect a conflict-laden process between employers and the towns that employed this approach.

Other factors. Where skilled employees were involved, there were more negative effects on morale when compared to less-skilled occupations. Perhaps skilled employees have a greater expectation that their work will not be contracted out.

Table 5

Regression Coefficients, Employee Opposition to Privatization, New York Towns and Counties, 1990–1995 [marginal effects] (“t” statistics in parentheses)

 

Employee Opposition (logit)

Employee Morale (tobit)

Union exists

6.611**

.267

 

[.685]

[.256]

 

(2.16)

(.94)

Adjustment policies

 

 

– Moved employees to other county

11.527**

–.160*

 

[1.195]

[–.585]

 

(2.06)

–(1.68)

– Early retirement

4.245

–.556

 

[.440]

[–.534]

 

(1.40)

(–1.56)

Skilled employees

–4.559

–.769**

 

[–.427]

[–.738]

 

(–1.59)

(–2.88)

Important factors

 

 

– Legal/political

–3.882**

–.019

 

[–.402]

[–.018]

 

(–2.30)

(–.17)

– Community values

–7.445**

.077

 

[–.772]

[.074]

 

(–2.23)

(.32)

– Monitoring

4.869*

–.855**

 

[.505]

[–.820]

 

(1.67)

(–3.28)

Constant

12.097**

2.456**

 

(2.40)

(4.89)

Equation Statistics

Log. Like. = –9.079

Log. Like. = –34.098

N

39

39

* p < .10 (critical t = 1.645, two-tailed test) ** p < .05 (critical t = 1.960, two-tailed test)

-> Voir la liste des tableaux

Only the important factors of community values, legal/political, and contract monitoring were significant in the regressions. Towns where community values were more prevalent were significantly less likely to encounter employee opposition. This result may reflect a privatization project that is not so much driven by ideology as by community concerns about service quality. In such cases, downsizing government is less likely to be a goal and employee concerns are more likely to be taken into account. The interpretation of lower opposition where legal and political issues were important indicates the strength of a strong legal or political mandate. Conversely, when there is weak political support for restructuring, there is a greater possibility of employee opposition. Neither community values nor legal/political factors significantly affected morale.

Finally, where the monitoring of the contract was important, opposition was greater and morale was lower. More complex contracts where specification is difficult may generate employee opposition because existing employees find their role diminished to one of contract monitors (see Kettl, 2002: 206).

Conclusion

Local government privatization in New York State was found to have some harmful effects on workers. Surprisingly, few local employers had adjustment policies to protect affected employees and disproportionate negative impacts were found on women and minorities. Privatization was found to have significant de-unionizing effects with substantially fewer workers in unions after restructuring. On the other hand, privatization had no pronounced impact on wages and benefits. In the majority of cases, wages and benefits were about the same after privatization.

In terms of employee opposition, where community values were an important factor, towns experienced less opposition. Employee morale was significantly lower for skilled employees and where contract monitoring was important. These findings may be attributed to the perceived watering down of the role of skilled and professional workers. Privatization may reduce the practice of their profession or craft to the role of contract monitor.

Evidence was found that the role of unions in the restructuring process is not one of automatic opposition. Unions were the catalyst for opposition actions, but only in for-profit restructuring cases. In the nine cases that involved contracting work back into the public sector, unions supported restructuring changes, and in the six cases of contracting out to another government, there was no union opposition.

There is a need for more research into the post-privatization effects on workers and their organizations. Greater efforts ought to be made to match management and union accounts of privatization. A wider research agenda would also include national surveys with longitudinal data to resolve causality questions.