Abstract
This paper argues that public sector labor relations is best understood in a framework that focuses on unions’ ability to shift demand curves rather than to raise wages, as is the case in the private sector. It reviews the public sector labor relations literature and finds that: (i) public sector unionism has flourished as a result of changes in laws; the effects of public sector unions on wages are likely to have been underestimated; public sector unions have a somewhat different effect on wage structures than do private sector nions; compulsory arbitration reduces strikes with no clearcut impact on the level of wage settlements; public sector unions have diverse effects on non—wage outcomes as do private sector unions. In terms of evaluating public sector unionism, the paper argues that by raising both the cost of’ public services (taxes) and the amount of services public sector unionism involves a different welfare calculus than private sector unionism.
Introduction and Summary
In he 1950s only a small minority of public sector workers were organized and even AFL—CIO President George Meany believed “It is impossible to bargain collectively with the government. “1vjrtually no states had laws permitting collective bargaining for public employees. Strikes were prohibited. Analysts of unionism regarded most public sector employees as unorganizable. In 1980 about 13 per cent of government employees were represented by labor organizations and 32 per cent were covered by contractual agreements. ‘ Over three—quarters of the states had legislation authorizing collective argaining by public employees. Eight states permitted strikes by some public workers. According to the Current Population Survey (oPs) 3% of government employees compared to 22 % of private sector employees were “represented by labor organizations” in May 198O. What caused the sudden surge in public sector unionism in the United States? What are its economic consequences? How have various states and localities dealt with the organization of their employees? Which laws and procedures have proven more/less successful in coping with the unionism of public employees?
In what ways does labor relations in the public sector parallel labor relations in the private sector? In what ways has the public sector been unique? To answer these questions, I review the results of the past two or so decades of research on public sector labor relations. The research represents a significant and expanding effort by labor specialists. In 1960 relatively few labor relations articles in the leading journals dealt with the public sector; in 1983 not only were there numerous public sector articles in the labor jour— 2 nals but the field had grown sufficiently to produce two specialized journals, s well as numerous books and research treatises. The review shows that while we have amassed considerable knowledge about the nature of public sector labor relations, there are noticeable gaps in our understanding of what public sector unions actually do. The following seven propositions provide a general overview of the major findings and issues in public sector labor relations:
A fundamental difference between public sector and private sector collective bargaining is that public sector unions can affect the demand for labor through the political process, as well as affect wages and work conditions hrough collective bargaining. However, the intrinsic political aspect of public sector bargaining does not necessarily make the demand for public sector workers more inelastic than the demand for private sector workers, and thus does not necessarily give public sector unions great economic power. While it is true that public employers do not face competition in their locality, they are subject to the discipline of a budget and, in the long run, to exit and entry of residents and businesses (the Tiebout adjustment). Similarly, while public sector unions may have an advantage in bargaining because they can help elect the olitical leaders against whom they negotiate, “legislative vetoes” of neo— tiated settlements and taxpayers’ referenda which limit taxes create potential weaknesses for unions. Finally, in contrast to private sector unions which are free to wield the strike threat as a weapon, union power in most public sector jurisdictions is limited by restrictions on striking.
The growth of public sector unionism in the past two decades can be traced, in large part, to the passage of laws (executive orders) which have 3 sought to bring the private sector industrial relations model to the public eec— tor.
In states with laws favorable to unionism, public sector unionism has flourished; in states without such laws, it has not. General lack of management opposition to public sector unionism, possibly due to the political power of unions, has been important for union victories in representation elections in the public sector. There is some indication that the spurt in public sector unionism has ended, leaving the U. S. with about a 30—140% union coverage of public sector labor. Much of the literature concludes that public sector unions have had relatively modest effects on wages compared to the effects of private sector nions on wages. This result, if true, would resolve the debate over the relative strength of unions in public and private employment. A careful reading of the evidence, however, shows the finding to be questionable on several grounds: First, many public sector union wage studies looked at unions when they were Just establishing themselves during a period of public sector expansion. If, as seems reasonable, the wage effects in first contracts are smaller than in later contracts and if union wage effects are smaller in booming than in declining markets, much of the early literature has understated the long term equilibrium” impact of public sector unions on compensation. A second problem with the generalization is that few studies have distinguished between the effects of public employment se versus effects from the type of occupations organized in the public sector. Public sector unions include a relatively large number of white collar workers, whose wages appear to be only moderately affected by unionism in the private sector. They include protective service workers, for whom it is difficult to find comparable occupational groups in the 4 private sector.
The limited studies of union effects on blue collar workers in the public sector show wage effects similar to those in the private sector. A third reason for questioning the conclusion that public sector unions have weak wage effects is the likelihood that, because of “comparability” of wages across cities, there are greater “spillovers” of wage settlements in the public sector than in the private sector, biasing downward standard cross—section estimates of the union effect. Finally, for some groups, public sector compensation packages differ greatly in dimensions not captured in standard data sets.
While these considerations do not necessarily prove the conventional generalization to be wrong, they suggest it should not be accepted without further investigation. With respect to wage structure, public sector unions have quite different effects on some aspects of wage structure than private sector unions do, and quite similar effects on other aspects. The teachers’ unions appears to widen educational wage differentials, while policemen and firefighters’ unions appear to have little effect on the range of salaries for their members. This contrasts with the general equalizing effect of unionism on such differentials n the private sector. On the other hand, nest studies find that in both the public sector and the private sector the union impact on fringe benefits is greater than the union effect on wages, and our analysis of overall inequality of earnings shows less inequality among union than among nonunion public sector workers. 5. Despite being illegal in all but eight states, strikes are part of public sector labor relations. From the 1960s to the l980s, the number of strikes increased drastically until 1 of every 8 strikes occurred in the public sector. Public sector strikes are generally of short duration.
While injunctions 5 often fail to halt strikes, there is evidence that the nature of strike laws affects their frequency. Public sector alternatives to the strike, in the form of various types of compulsory arbitration, have been successful in reducing strikes with no clearcut impact on the level of wage settlements. Arbitrators’ decisions show no apparent bias toward one side or the other, despite frequent claims to the contrary. Final offer arbitration appears to have worked reasonably well, with less of a ‘chilling effect’ on negotiations than conventional arbitration, consistent with theory.
The limited evidence available on the effects of public sector unions on productivity shows that unionism is not inimical to productivity. Some studies find positive union effects, some find essentially nO effects, but few find negative effects. However, problems of measuring output of public services make any firm generalization difficult. With respect to public budgets, studies suggest that unionization does, indeed, increase the share of a municipality’s budget going to the workers in an organized function. Several aspects of market performance have not, as yet, been exten- sively explored by public sector researchers.
These include: the effect of public sector unions on turnover, which is known to be lower, in general, in public employment; the net effect of unions on the dispersion of earnings; the effect of unions on employment and wage responses to cyclical and other economic swings; and, most important of all, the effect of unions on the “price” of output, taxes, property values, and municipal finances in total. Succeeding sections of this essay provide detailed evaluation of the research that underlies these seven propositions.
The New Unionism
Unions typically grow in sudden spurts after years of stagnation. As can 1960’s — e seen in Figure 1, this is true of public sector unions in the United States. During the mid early 1970’s, public sector union membership sore than in that it uadrup1ed. 6 The sudden growth of public sector unions was remarkable occurred during a period of noticeable decline in the organization of the private sector. It came as a surprise to most observers, who had cited such factors as the security of public sector work, the high proportion of female, black, and white collar workers, adverse public attitudes, civil service laws and prohibi— tions 7 on strikes as factors inhibiting organization. Indeed, as late as
February 1959, the AFL—CIO executive council stated that ‘in terms of accepted collective bargaining procedures, government workers have no right beyond the authority to petition Congress——a right available to every citizen. ‘8 Table 1 shows that, while some of the newly organized public sector workers are in unions that are predominantly in the private sector, the majority of workers are in newly emergent public sector unions. These unions have certain distinct characteristics. First, in jurisdictional terms, many are more like craft than like industrial unions, being organized along occupational lines (i. . , the postal workers’ unions, the teachers’ unions, the police, the firefighters, the sanitation workers, and the transit workers). Second, the new unions have organized an exceptionally large number of “white collar” workers, even if one excludes the teachers.
While the various laws differ among types of employees (i. e. police and firefighters have different regulations than the teachers), states that have favorable collective bargaining laws in one area also have them in others, permitting the broad classification of states in Table 3. As an example of the rapid change in state laws, consider Massachusetts. In 1958 it passed a law allowing public employees to join unions but not to bargain over economic matters. In 196)4, it passed a law requiring bargaining over work conditions, which was extended in 1965 to include wages, but did not 11
Then, responding to niinicipalities unhappy with final offer arbitration, the legisla- ture developed a labor—management committee to resolve impasses through a variety of mechanisms. In the span of a decade, the state moved from no legal bargaining to bargaining with a dispute resolution mechanism guaranteeing a 10 contract. In addition to passing laws which encouraged collective bargaining, the majority of states instituted public employee relations boards (PERB’s) to hold representation elections among workers and to resolve charges of unfair labor 11 practices.
Studies of the spurt in public sector unionism (see Table 4) uniformly show that these laws were a major factor in the growth of public sector unioni- zation. States that enacted laws had rapid increases in unionization in ensuing years. States that did not had no such growth. The more favorable the laws were to unions the greater the growth of unionization. For instance, in states where laws permit strong union security clauses, teachers unions appear to have done better than in other states. This is not to say that public sector bargaining does not exist in the absence of a law.
Indeed, Ohio is a good counter example. But one counter example does not disprove a social science generalization. For the most part, the spurt in public sector unionism was associated with changes in state laws regulating collective bargaining. Finally, it should be noted that there is some indication in the growth curves for public sector union organization that this spurt has come to an end. From 1972 to 1980 the share of state and local full—time employees with employee 13 Table 4: Studies of Impact of Legal Environment on Public Sector Union Growth and Determinants of Favorable Laws Study Group
Union Growth
Result Moore Teachers, cross—state and time series, 1919— 1910 Public employee unionism, 1972—1980 Mandatory bargaining laws help AFT not NEA Reid and Kurth Progressive state labor law raises iini cni sm Dalton Government employees, 1976 Mandatory bargaining laws greatly raise union density Teachers union growth spurred by provisions allowing exclusive representation, checkoff of dues, deterred by right— to—strike laws Comprehensive collective bargaining laws raise unionism Collective bargaining laws are a key factor in police organization Reid and Kurth Teacher unionism, 1972—1980
Moore Public sector unionism, 1968 Ichniowski; Lauer Police unionism, 1960—1980 in DETERMINANTS OF LAWS Faber and Martin Laws on teacher collective bargaining Urbanized states and those scoring high (low) on ADA (ACA) ratings passed laws earlier Per capita change in income 1960—TO, state per capita expenditures of government, and irinovativeness of state leads to public sector collective bargaining laws Kochan State public employee bargaining laws 14 organization representation was virtually unchanged in Public at almost 50 per cent. 12 Why Did the New Laws Induce the Spurt Sector Unionism?
The various state public employee labor laws of the mid 1960’s — established 1970’s mechanisms for workers to vote for/against collective representation these and required public sector employees to bargain with them. Both of innovations legal were already part of the National Labor Relations Act governing private sector unionism. They can be viewed as making public employees more like private employees. Yet unionism in the private sector declined during the period. One possible reason for the different results is that there was “pent—up” demand for unionization in the public sector, but not in the private sector.
Another reason, which I believe to be more important, is that public sector employers have not fought union organization of their workers to the extent that private sector employers have. As public officials, they cannot break the spirit or letter of the law, as management can in the private sector. Hence, the same nominal election procedures produce different results in the two settings. Finally, an important issue which must be addressed in any impact analysis of the of laws on society is whether the legal changes have a true independent effect on social outcomes or whether they are simply an intervening or mediating actor for more fundamental forces. Studies of the variables associated with passage of laws favorable to public sector collective bargaining show that the laws have come first in states with high income per capita and a history of innovations in governmental activity. Some show that these laws are more likely in states that are highly unionized in the private sector, but others do not. While detailed analysis of’ legislative votes and of the history of key legisla— 15 tion is required to determine full causal routes the available evidence suggests that the laws had an independent effect.
Without legislation favorable to collective bargaining we would not have observed the extensive unionization of the public sector in United States, at least not to the extent which actually occurred.
The New Setting
The public sector differs in several important ways from the private sec- tor (see Figure 2). Public employers have the sovereign powers of the state; they generally are monopolies in their local area; they are ultimately responsible to the voting public, including public sector workers; and some produce essential services (police, fire, and defense).
Do these differences make the private sector industrial relations system essentially inapplicable to the public sector, or are these differences substantially no greater than the differences between various private industries such as steel and construction, for example? Do the unique features of the public sector give unions of governmental employees particularly strong economic power, as some have alleged? The Sovereign Power Issue Among the first objections to public sector unionism was the claim that such unionism infringes on the sovereign power of the state in determining levels f services and rules for the behavior of public employees. Surely, it was argued, one cannot allow public employees, particularly the military, the police, or fire fighters to have power independently of the elected government. 1-3 The counter argument is that there is nothing special about employee—employer relations for most public sector workers. Why should a worker employed in a private firm under city contract operate under a drastically different labor law than the equivalent worker in a city owned service?
The sovereignty issue has, for the most part, been resolved by allowing public sector workers to organize, but by forbidding most, particularly those in “essential services,” from striking. For workers in “nonessential services,” 17 Figure 2: Industrial Relations System in Public Sector Compared to Private Sector Actors
- Management Public Sector Elected officials Legislative/City Council voters, through referendum Often white collar Private Sector Managers
- Unions Mostly blue collar Some mediation some arbitration
- Third Party Compulsory arbitration Police, military power of state . Workers Usually have some job security, more likely to be worse on blacks Workers Technolor /Market
- Output Produced Unpriced goods, hard—to-. measure public goods Priced goods and services for private consumption/investment Competitive markets in nDst cases
- Competition Monopoly in local market T. Entry & exit Residents and businesses can move across locales New firms enter! leave given locale Profits “buffer” changes in wages Raise/lower prices to change revenues 8. Budget Conditions Budget constraint Can change taxes Intergovernmental grants Short—term debt Power
- Influence on other side Unions can help elect political leaders Unions have no say in company policy Free to strike with only modest Taft—Hartley restrictions
Conflict tools Strikes often prohibited
Illegal Acts
Management unfair labor practices unlikely Little penalty for management unfair labor practices. Illegal acts frequent and rising 18 how-ever, actual penalties for strikes are often uxderate, effectively allowing short strikes (see section v). The federal government stands in a unique position with respect to sovereignty.
While federal employees are allowed to form unions and negotiate over working conditions, they do not negotiate over wages and are not allowed to strike. The high unionization in the federal sector thus provides evidence for worker desire for representation in a large bureaucratic organization, exclusive of the “monopoly” power of unions to raise wages through collective bargaining. The Monopoly! Inelastic Demand Issue The argument that the nnopoly power which governments have in their jurisdiction creates such inelastic demand for public sector employees as to give ublic employee unions great economic power was stressed by Wellington and Winter in their 19T1 book: • . to the extent union power is delimited by market or other forces in the public sector, these constraints do not come into play nearly as quickly as in the private. “… soine of these services are such that any prolonged disruption would entail an actual danger to health and safety. the demand for numerous governmental services is relatively Inelastic, that is relatively insensitive to changes In price. Indeed, the lack of close substitutes is typical of many governmental endeavors. I reject this claim for three reasons:
In the short run governments face tax and budget constraints that create a potentially nre rather than less elastic demand for labor. This is because an employer operating under a budget constraint has no profit “residual” from which to pay higher wages or in which to put savings from lower wages. All of the adjustment to changes in wages take the form of an adjustment in quantities. More broadly, the budget serves as a ‘disciplinary’ device in the 19 public sector as does market demand in the private sector, forcing a quantity price radeoff on the employer. In the long run, cities and states are not really monopolies, since residents and businesses can move from one jurisdiction to another. Indeed, by analogy with the factor equalization theorem of trade theory, mobility should compensate fully for the monopoly power governments have in their jurisdiction. Citizens unhappy with level of public services can move elsewhere, reducing the taxable population and thus the ability to pay public sector wages.
Mobility places great constraints on public sector union bargaining power as has been stressed by Courant, Gram. ich, and Rubinfeld. Moreover, in jurisdictions where taxation of property raises funds for operation of governments, the capitalization of taxes in property values is an additional constraint on public sector budgets and thus on union ability to raise wages. Where public sector workers are “essential” they are almost a1wars forbidden to strike which greatly reduces their economic power, as noted earlier. The Political Context The political dimension clearly creates a distinct environment for labor relations which leads collective bargaining down different paths than in the private sector.
While private sector unions can occasionally alter the demand for labor curve (through union label campaigns, or by bargaining over employment as well as wages) the usual assumption is that they alter the cost of labor, with firms responding by changing the level of employment (see Figure 3A). Indeed, the standard model of private sector unionism evaluates the monopoly loss due to unionism in terms of the lower national output because of the reduc— 20 Figure 3: Differences Between Public Sector and Private Sector Union Effects A. Private Sector Union Wage Effects Wages Effect of Unionism is to raise wage W0 to W1
Wi Demand Curve Employment B. Public Sector Shift in Demand Effects Case I: Perfectly elastic supply curve Wages ,, Effect of Unionism is to shift demand D to D’ and raise wages above wo Supply ‘NN E1 Case II: Upward—sloping Supply Curve Employment Wages // Wo Effect of Unionism is to shift and demand D to D’ raise wage above Employment 21 tion in employment and increased marginal product in the union sector. By contrast, public sector unions can be viewed as using their political power to raise demand for public services, as well as using their bargaining power to fight for higher wages (Figure 3B and 3C).
A possible criticism of public sector unions is that they increase public goods production beyond the social optimum, however, is a two—edged sword, and it is by no means clear raising rather than lowering employment. Politics, whether collective bargaining in a political context increases or reduces union power. On the one hand, public sector workers are an identifiable voting bloc and potential activist group in local election campaigns. But so too are tax- payers. In the private market consumers affect wages of workers only indirectly lead firms to raise prices. through shifts in purchases when increased wages
In the public sector ‘consumers’ can affect wages directly by electing officials, passing referendum and the like which restrict settlements. In virtually every state, so-called “legislative vetos” can vitiate bargains, as legislatures! councils refuse to raise the iney to fund signed contracts. For example, in the l9TOs, despite signed contracts college professors in the University of Massachusetts system did not receive salary increases for several years because the legislature did not allocate the funds)5Taxpayer revolts, as evidenced in proposition 13 (California) or J. Massachusetts), have also been used by opponents of public spending to limit potential union wage gains by capping tax revenues or budgets. In other circumstances unions have effectively used legislature or voter support to win terms they could not gain at the negotiating table, as the following case cited by the Labor Management Relations Service indicates: 22 A prime example of union use of the referendum to bypass a municipal employer unreceptive to union demands Is that of the St. Louis Fire Fighters Local 73, who sought by special election to obtain equal pay with policemen.
They rang doorbells and conducted an intensive campaign for votes in homes and taverns and at barbecues and labor and political meetings. The firemen also appealed for votes in newspaper, television, and radio advertisements. It was estimated that the entire campaign, financed by assessments on members of Local 73, cost between $35,000 and $50,000. This compared to the $6,000 spent by the firemen’s principal opponent, Mayor Alfonso J. Cervantes, who campaigned against the raise claiming it would cost the city $1. 7 million for a full year and probably cause a reduction in the number of fire companies.
The firemen, claiming that these were “scare tactics,” prevailed, as the voters gave the proposal a 61. 7 percent majority; it needed sixty percent to pass. ]-The term ‘multilateral bargaining’ is commonly used to refer to the situation in which public sector unions bargain not simply with those across the table from them but with other interested public parties as well. In such bargaining, need for public services, public expenditures, quality of services, as well as wage packages are often at stake. Neither in theory nor in practice does multilateral bargaining necessarily improve the union’s ability to win wage gains.
In sum, the unique features of the public sector do, indeed, make it different from the private sector, producing a different industrial relations system. Careful examination of how these features affect the bargaining power of the two sides suggests, however, that the relative strength of unions in the public versus the private sector cannot be resolved by a priori logic. The issue requires empirical analysis, to which we turn next.
Union Compensation Effects
The general tone of studies dealing with the effect of public sector unionism on compensation is that the effects tend to be small.
Indeed, most survey articles have concluded that “The ‘average’ wage effect of unionism in government.. , is roughly on the order of five per cent… sinaller than the average union wages impact in private industry. “17”The general effects (of 18 public sector unions) which have been measured are not huge. ” This generalization rests on extensive analysis of teachers unions, police and firefighters organizations, and on studies of the Current Population Surveys. The studies vary substantially. Some analyze wages of public sector workers by state, city, or district at a moment in time, using the frac- ion in a collective organization or having contracts as the “union variable. ” Others analyze rates of change in wages over time, looking for an acceleration in the rate of change after unionization. Some look at wages of individuals on large data tapes. Others perform before/after comparisons of wage rates. For the most part, the studies relate to the period from the mid 1960s to the early or mid 1970s. In this section I review briefly the relevant studies and examine the evidence and arguments that I believe question the existing generalization. The Studies of Teachers Table 5 eports the results of diverse estimates of the effect of school time teachers unionization on wages, organized by period, unit of analysis, and the approach taken. The results for 1965—1968, which cover the beginnings of teacher unionism, support the view that at that time teachers unionism Table 5: Finding The Wage Effect of Teachers Unionization by Year, Study and Type of Experiment (Cs = Cross—Section; B/A = Before/After)
The most recent Baugh and Stone analysis of individuals for 19TT finds very large effects, of 21 per cent in a cross section and 12 per cent in a before/after framework. Comparing 19Th—19T5 CPS results with 19T8—1979 results they find a tripling in the union wage effect over time. As a check on this finding, I estimated their cross—sectional model for 19T3 and 1981 and obtained a similar result. As a further check, I have also examined the pay of teachers across states by degree of unionization and also find evidence of a rising union impact. There are several possible causes for the increased wage effect of the eachers union. One likely cause is changed economic conditions. In the early period the market for teachers was strong, with consequent good wages for non—union as well as union teachers; in the latter period the market was weak with the teachers union using its power to offset some downward pressure on wages. Another hypothesis is that modestly higher differentials in rates of change of pay per settlement cumulated over time to a sizeable union effect.
A third possibility is that, in fact, the power of the teachers union has risen because of increased willingness to strike and changes in state laws regulating eachers collective bargaining. Which, if any, of these possibilities is correct requires not only a careful analysis of the teachers union wage effect using comparable data over time (along lines of Baugh-Stone) but also better measures of union economic power than has been common in studies. Because of differences in the legal treatment of teachers unions across states and over 26 time, simple dichotomous or percentage organized variables are potentially misleading: two areas with the same outcomes because in one area the law “unionization” gives ay have different economic the union of teachers greater power than in the other, for instance, by requiring final offer arbitration or allowing strikes. Protective Service Workers The second most extensively studied group of public sector unions are the police and firefighters, for whom the standard study has found moderate effects in the area of 5—10 percent (see Table 6). For both occupations, there is a general finding that effects are larger on total compensation than on wages; that effects are larger for larger cities; and that effects may vary over time.
The Feuille, Hendricks and Delaney estimates of police union effects over time show a definite rise from the early 1970s to the rnid—l9TOs, followed by a decline at the turn of the decade. There are two possible problems with the studies of union effects among protective service workers which suggest the 5—10 percent wage effects may underestimate the actual impact of unionism. The first problem is the not—so simple matter of measures for pay. All of the studies use reported salary rates or average salary compensation from the Census of Governments or the International City Management Association (ICMA) or comparable sources.
The second potential problem with estimates of the effects of protective service worker unionism is the possibility of sizeable “spillovers” from union to nonunion cities. Such spillovers are more likely for these occupations than for others because of the lack of equivalent occupations in the private sector and the consequent tendency for police and fire fighters to stress pay com- parability across cities. One admittedly crude way to examine this idea is to relate wages to unionism in a greater geographic area. Studies which have done this suggest union wage effects are roughly twice as large s the 5—10 percent the standard cross—section analysis. 19 In Additional Groups addition to the studies of teachers and protective service unions, there have been a few studies of union wage effects in particular blue—collar occupations which are directly comparable to those in the private sector. As Table S shows, these studies present a very different picture of the relative economic impact of unionism in the public and private sectors than is indicated in the earlier literature: even in the early 60s and TOs these studies show no noticable difference in the wage impact of unions in the two sectors.
If public sector unions have smaller wage effects, it could be because of the occupations covered. The Total Work Force An alternative to the analysis of detailed occupations which occupies much of public sector labor relations research is to examine the pay of all govern- ment employees on large data sets. The most important such study, by Sharon Smith, found that in 1975 government employees in general were paid more than private sector employees, but that unionism raised wages less for public sector than for private sector workers, controlling for a wide variety of additional age determinants. Using a slightly different model, with fewer controls, I report in Table 9 similar union results for 1973 and 1981 with, however, evidence of increases in the union impact in some government sectors and decreases in others. Moore and Raisian’s analysis of the Michigan Panel Survey of Income Dynamics shows greater variation in the public sector union effect over time, with a drop from the late 1960s (when few were organized) to the early 1970s and a rising effect thereafter.
In the absence of a comprehensive study of various data sets, definitions of government employees, and different models, the safest conclusions are that the union effects differ significantly over time and are 32.
