The Wrong Target: Public Sector Unions and State Budget Deficits

Sylvia Allegretto, Ken Jacobsand Laurel Lucia

Press Coverage

UC Berkeley Center on Wage and Employment Dynamics
UC Berkeley Labor Center

The economy fell off a cliff in 2008 as we experienced the most devastating downturn since the Great Depression. Unemployment hit double digits resulting in significant economic hardship and increasing demands on government safety nets and services to help those in need. At the same time, state revenues plummeted as workers lost their jobs and business activity contracted. As states struggled to balance their budgets in this weak economic climate, public officials in multiple states argued that the fiscal gaps were due to government workers and their unions.

The fallout from the political jostling around public workers has been that hundreds of bills related to public employees and unions were introduced in state legislatures—most of which sought to restrict public sector unions. At least twelve states have significantly restricted collective bargaining through new legislation in 2011 including: Wisconsin, Ohio, Indiana, Arizona, Idaho, Michigan, New Hampshire, Oklahoma, South Carolina, Tennessee, Utah and Wyoming.[1] A common rationale for these proposals is that growing costs associated with public sector workers, especially union-represented workers, are at the root of state budget deficits. Governor Scott Walker of Wisconsin said “we can no longer live in a society where the public employees are the haves and taxpayers who foot the bills are the have-nots.”[2] A conservative columnist in the Daily Journal noted “like Governor Christie, he [Governor Walker] decided to actually fix the problems that brought Wisconsin to this point. His budget limits government collective bargaining…”[3] But, were public sector workers and their unions to blame for state budget problems?

In this brief, we review the relevant research and analyze the relationship between public sector workers, their unions and state budget deficits. Our analysis finds that the size of the public sector workforce per thousand residents is not growing and previous studies have found that public sector compensation, as a share of public budgets, has not grown. Researchers have consistently found that public sector workers are not compensated more highly than their private sector counterparts after taking into account level of education, experience and other important factors. Finally, our regression analysis shows large state deficits were due, in large part, to the decline in house prices and not due to public sector workers and their unions. The bursting of the housing bubble was the precursor to the Great Recession and it is clear that the large drop in house prices (which capture much of the deteriorating economic climate) was central.

State Deficits Hit Record Levels

The Center on Budget and Policy Priorities analyzed trends in state budgets. They found that states have faced the largest budget shortfalls on record (Figure 1) during and following the Great Recession. State deficits totaled more than $100 billion each year in fiscal years (FY) 2008/09 through 2011/12, peaking at $191 billion in FY 2009/10. In FYs 2009/10 and 2010/11, all but two to three states faced shortfalls. [4] These record deficits have resulted in significant cuts to much needed programs such as: public health programs, the educational system including significant teacher layoffs, programs for the elderly and disabled, and massive layoffs of the public sector workforce.[5]

Figure 1: Largest State Budget Shortfalls on Record

State budget shortfalls have important economic consequences. Severe austerity measures implemented by states struggling to balance their budgets effectively reduce demand in an already underperforming economy suffering from weak demand. Along with direct layoffs due to reductions in spending, contracts are cancelled with vendors and benefits are cut for individuals. The impact is multiplied as vendors reduce purchases of supplies and services and individuals have less money to spend in the local economy. The impacts are further magnified for programs like Medicaid where the reduction in state spending also lowers federal matching funds. Tax increases also suppress demand, but, depending on the nature of the tax increase, they do so to a lesser extent.[6]

Public Sector Employment has been Steady

The large state deficits have frequently been blamed on a growing public sector. For example, Governor Scott Walker warned that Wisconsin “cannot grow if our people are weighed down paying for a larger and larger government.”[7] However, the size of the public sector has not grown in recent years, neither in terms of public sector employment levels nor public sector compensation.

State and local government workers as a share of the workforce has been relatively steady since 1979, as shown in Figure 2. Using monthly Current Employment Statistics from the Bureau of Labor Statistics we look at state and local government jobs as a share of all jobs. We find a small increase during each of the five recessions over this time period, reflecting the disproportionate reduction in private sector employment during the recessionary periods—followed by a corresponding reduction in the share during the subsequent recovery and expansion periods. The high point for the share of total employees in state and local government is typically one or two years into the recovery, as employment tends to lag economic growth—especially in the jobless recoveries that followed the early 1990s and 2001 recessions. Overall, the share of workers in state and local employment averaged 14.2 percent over the thirty year period and ranged from a low of 13.6 percent at the height of the boom in 1999 to a high of 15.2 percent in the great recession in 2009 reflecting the greater loss in private sector employment—over 5 million private sector jobs were lost that year. [8] By midway through 2011, the share of workers employed by state and local governments had fallen back to 14.6 percent.

Figure 2: Public Sector employment has been steady share of workforce, 1979-2011

Not only has the share of state and local government jobs remained relatively steady as a percentage of all jobs, but state and local government employment per thousand residents has also remained steady. We analyze the number of public sector employees per thousand residents using Current Employment Statistics from the Bureau of Labor Statistics and population statistics from the U.S. Census Bureau. In 1990, the United States as a whole had an average of 17.2 state workers per thousand residents. In 2009, there were 16.8. The number of state and local workers per thousand residents rose slightly over the entire time period from 60.8 to 64.2, but remains unchanged since 2001 (Figure 3).

Figure 3: Public sector employment per 1,000 residents has been steady since 1990

Might union states be different? New York Times columnist David Brooks argued, “public sector unions can use political power to increase demand for their product.”[9] If he is correct, we should expect states with high public sector union density—the share of public sector workers in a union—to have more public sector workers per thousand residents, than states with lower public sector union density. In order to test this hypothesis, we examine the ten states with the highest share of public employees in unions and the ten states with the lowest share of public employees in unions. As illustrated in Figure 4, the lowest union density states averaged 69.1 state and local employees per thousand residents in 1990 and 74.6 in 2009. The highest union density states averaged 65.1 state and local employees per thousand residents in 1990 and 68.3 in 2009. The number of state and local employees per thousand actually fell in the high union density states between 2001 and 2009.[i]

Figure 4: Public sector employment per 1,000 residents higher in states with least public union density, 1990-2009

Finally, we test whether there was a significant correlation between the degree of union density in a state and the number of state and local employees per thousand residents in 2009. The plot of these data is in Figure 5. The two variables do not exhibit a defined pattern and there is no statistical correlation between them.[ii]

Figure 5: Higher public union density not correlated with greater public sector employment, by State, 2009

Public sector employment has remained steady, whether measured as a share of the workforce or in relation to state population. No correlation was found between public sector union density and the level of public sector employment in a state. Contrary to Brooks’ assertion, there is no evidence that public sector unionization has resulted in a growth of the public sector workforce.

Public Sector Compensation as a Share of State Budgets has Declined

Not only has the number of public sector workers per thousand residents remained steady, but public sector compensation as a share of state budget has actually declined. David Madland and Nick Bunker of the Center for American Progress Action Fund analyzed total compensation, including wages, salaries and benefits, for public sector workers as a share of state expenditures. They created a data set of ex- penditures on state employees between 1992 and 2009. The period is well timed to cover two complete business cycles. Madland and Bunker find that the share of state spending that went towards compensation fell steadily between 1992 and 2002 and remained stable from 2002 to 2009 (Figure 6).[10]

Figure 6: State budget woes not a problem of public sector worker compensation, 1992-2009

They further analyzed the results by the level of public sector unionization. They found that the average share of the budget spent on compensation over the time period for the ten most highly unionized states was 19.6 percent, compared to 18.7 percent for the ten least unionized states. By 2009, that gap between the two groups had narrowed to 0.5 percent (19.8 versus 19.3 percent). They conclude that state budget deficits were not caused by an increase in funding going to compensation for public sector workers.

Public Sector Workers are not Overcompensated

In recent years, public officials have claimed that public sector worker compensation is more generous than that of private sector workers. Senator Scott Brown of Massachusetts said in 2010, “it’s not right that lesser-paid private sector workers suffering through a recession have to pay for expensive government salaries.”[11] In his March 2011 budget address, Governor Scott Walker argued that Wisconsin’s economic growth was being hindered by a “government that pays its workers unsustainable benefits that are out of line with the private sector.”[12] New Jersey Governor Christie said “if you were paying attention the problems here that are created on the state budget—sure we have a deficit problem that was helped by the economic downturn but what we also have are benefits and costs that are out of control.”[13] These misperceptions about compensation inequity have been reinforced by analyses that simply compare average compensation of public sector workers to those of private sector workers, without taking into account differences in staffing between the two sectors with regard to education, experience and other relevant factors.[14]

However, research has consistently shown that public sector workers are paid the same or less than similar private sector workers when taking into account relevant factors that affect compensation:

  • A study by Jeffrey Keefe of the Economic Policy Institute (EPI and Rutgers University) analyzed wage and demographic data from the Current Population Survey and benefits data from the Employer Costs for Employee Compensation Survey by the Bureau of Labor Statistics. After accounting for education, experience, hours of work, organizational size, gender, race, ethnicity and disability, Keefe found that full-time state and local employees receive 3.7 percent less in total compensation (wages and benefits) compared to similar private sector workers.[15]
  • John Schmitt of the Center for Economic and Policy Research (CEPR) examined wage and demographic data from the Current Population Survey and reported similar findings: on average, state and local employees are paid 4 percent less in wages than similar private-sector workers after adjusting for education, age, gender, and other factors.[16]
  • Sylvia Allegretto of the UC Berkeley Center on Wage and Employment Dynamics (with Jeffrey Keefe) analyzed compensation in California and found that “an apples-to-apples comparison, or one that controls for education, experience, and other factors that may influence pay, reveals no significant difference in the level of employee compensation costs on an annual or per hour basis between private and public sector workers.” The study analyzed total compensation, including wages and benefits, based on data from the Current Population Survey and the Employer Costs for Employee Compensation Survey by the Bureau of Labor Statistics.[17]

These are a few of many examples that illustrate why simply comparing average pay between the two sectors, without taking into account workforce differences, is highly misleading when, in general, better educated and older, more experienced workers earn more than less educated and younger workers.

Journalists Michael Fletcher and Brady Dennis of the Washington Post acknowledged the research demonstrating comparable pay, but instead blamed state budget deficits on workers’ benefits. “Though studies show that state employees are generally paid less than comparably educated private workers, public employees often enjoy more generous pension and health-care benefits, and these are at the root of the long-term budget problems confronting many states.” [18] However, the studies by Keefe and Allegretto illustrate that even after taking benefits into account, public sector workers are compensated the same or less than comparable private sector workers.

Not only is compensation comparable for similar private and public sector workers, but private sector employment costs have actually increased at a faster rate than those of public sector workers in recent years. The Bureau of Labor Statistics measures the average cost to employers for wages, salaries and benefits per employee hour worked, published in the Employer Costs for Employee Compensation (ECEC) series. Our analysis of the ECEC data shows that from 2004 to 2009 the cost of employing a worker in the private sector increased 14.7 percent while it increased by 12.6 percent in the public sector.

Budget Woes Due to Bursting of the Housing Bubble and the Great Recession Not Public Sector Workers and their Unions

As we stated up front, the recession that officially lasted from December 2007 to June 2009 was the most severe downturn since the Great Depression—and the bursting of the housing bubble and the subsequent financial collapse caused it. Regardless, some elected officials and media have falsely suggested that collective bargaining for public sector workers caused state and local budget deficits. In a television interview, Governor Scott Walker of Wisconsin said “what stood in the way [of balancing local budgets] time and time again was collective bargaining.”[19] David Brooks of the New York Times wrote “states with public sector unions tend to run into fiscal crises.”[20] However, states experienced record deficits at a time when public sector unions did not grow and collective bargaining rights were not strengthened. What were the real underlying causes of state deficits?

Given that the bursting of the eight trillion dollar housing bubble was the precursor to the Great Recession, it is necessary to assess the role of falling house prices in state budget woes. Using a simple regression without controls, John Sides found a small correlation between union density and state budget deficits.[21] We employ a similar regression analysis to investigate the impact that public sector union density and collective bargaining rights had on the size of state (50 states plus D.C.) deficits in FYs 2009 and 2010 adding in a number of controls. Table 1 presents the regression results.

Following the methodology of Sides, the first specification simply uses public sector union density as the lone independent or control variable.[22] The coefficient on union density of 0.245 (significant at the 5 percent level) implies that a one-percentage point increase in union density would increase the average budget shortfall by approximately 0.2 percentage points. In the second specification we add a variable for the change in house prices, by state, over the five-year period from 2005-10. The coefficient on union density becomes much smaller economically (0.114) and is rendered not statistically significant. The coefficient or effect of a one-percentage point decline in house prices is attributed, on average, to a 0.56 percentage point increase a state’s budget shortfall. Furthermore, the ‘Adjusted R2 ’, which explains the variation in state budget shortfalls, increases from less than ten percent in the first regression to over forty percent in the second one.

In the third specification, we add additional controls to capture union strength. Specifically, we add a set of dummy, or dichotomous, variables to account for whether state employees, police, fire, teachers and local employees are covered under collective bargaining by law, constitution, charter, executive order or court decision. Another dummy variable indicates whether a state has a right-to-work law. A strike variable is used if at least one public sector occupation is allowed to strike.[23] Again, the results show that the coefficient on union density is not statistically significant while the house price effect is approximately the same as in the second specification. We have not listed the coefficients of the union strength variables because they were all very small (absolute value less than .09) and not statistically significant. Thus, going beyond union density to account for union strength does not affect the bottom line finding that house price declines were, to a large extent, a central reason why state budgets are in such dire straits. It should be noted that the outcome of this analysis is not substantially different whether changes in state house prices, state GDP or unemployment rate changes by state are used as explanatory variables. They are all capturing to a great extent the economic collapse.

Table 1: State budget deficits--a problem of unions or a bursting housing bubble?
See endnote iii [iii]for details of union variable that measure union strength.

To many, this regression analysis may seem like proving the obvious given the state of the economy and the enormity of the collapse. This has nonetheless not stopped the attack on public workers nor their unions.



Following the 2010 elections, multiple states took action to curtail collective bargaining rights arguing that public sector unions were a major cause of state budget deficits. A close examination of the available evidence finds that the claim that public sector unionization leads to greater deficits does not withstand scrutiny.

The public sector workforce has not been growing relative to the population; this is true in union and non-union states alike. There is no correlation between the share of public workers in unions and the size of the public sector workforce. This belies the notion that public sector unions are increasing the demand for their product.

Compensation has fallen as a share of state expenditures over the last twenty years; this is true for both high and low-union states. Controlling for education, experience and other relevant factors, public sector workers are not more highly paid than their private sector counterparts. Public sector unions provide workers with a voice on the job and enable members to choose their form of compensation. This has generally led to a greater share of compensation paid in health and retirement benefits than in cash wages.

Budget deficits were primarily caused by the housing crisis and subsequent economic downturn which resulted in a decline in revenues as the economy contracted. Finally, controlling for the decline in housing prices, we find no statistically significant correlation between union density, union strength and the size of state budget deficits.

For states to address their budget deficits, the most important factors are national economic growth and a resolution to the housing crisis. Solutions that focus on cutting state and local budgets can be expected to further weaken the economy. Federal aid to the states is essential to maintain the public infrastructure while the economy rebounds. Federal action is also needed to address the housing crisis, which continues to provide a drag on the economy and on state and local revenues.



[1] David Schaper, “Collective Bargaining Curbs Spread Across The U.S.,” National Public Radio, May 24, 2011,

[2] Steven Greenhouse, “Strained States Turning to Laws to Curb Labor Unions,” New York Times, January 4, 2011, p. A1,

[3] James Sherk, The National Review Online, “Wisconsin’s Governor Walker takes on Public Sector Unions,” February 15, 2011.

[4] Elizabeth McNichol, Phil Oliff, and Nicholas Johnson, Center on Budget and Policy Priorities. “States Continue to Feel Recession’s Impact,” June 17, 2011,

[5] Nicholas Johnson, Phil Oliff and Erica Williams, Center on Budget and Policy Priorities, “An Update on State Budget Cuts: At Least 46 States Have Imposed Cuts That Hurt Vulnerable Residents and the Economy,” February 9, 2011,

[6] Ken Jacobs, Laurel Lucia and T. William Lester, “The Economic Consequences of Proposed California Budget Cuts,” UC Berkeley Center for Labor Research and Education Policy Brief, May 2010,

[7] State of Wisconsin Governor Scott Walker, “Text of Governor Walker’s Budget Address,” March 1, 2011,

[8] U.S. Bureau of Labor Statistics, Current Employment Statistics, 2009.

[9] David Brooks, “The Paralysis of the State,” New York Times, October 12, 2010,

[10] David Madland and Nick Bunker, “State Budget Deficits are Not and Employee Compensation Problem: The Great Recession is to Blame,” Center for American Progress Action Fund, March 10, 2011,

[11] Ed O’Keefe, “Scott Brown to join Senate homeland security, military committees,” Washington Post, March 3, 2010,

[12] State of Wisconsin Governor Scott Walker, “Text of Governor Walker’s Budget Address,” March 1, 2011,

[13] Transcript: New Jersey Governor Chris Christie on NBC’s The Today Show with Ann Curry on Wednesday, February 23, 2011,

[14] See for example: Dennis Cauchon, “Federal pay ahead of private industry,” USA Today, March 8, 2010,

[15] Jeffrey Keefe, Economic Policy Institute, “Debunking the Myth of the Overcompensated Public Employee: The Evidence,” September 15, 2010,

[16] John Schmitt, Center for Economic and Policy Research, “The Wage Penalty for State and Local Gov- ernment Employees,” May 2010,

[17] Sylvia A. Allegretto and Jeffrey Keefe, “The Truth About Public Sector Workers in California: They are Neither Overpaid nor Overcompensated,” UC Berkeley Center on Wage and Employment Dynamics Policy Brief, October 2010,

[18] Michael A. Fletcher and Brady Dennis, “State budget clashes spread to Indiana, Ohio,” The Washington Post, February 23, 2010,

[19] Transcript of Fox News Sunday, February 20, 2011,

[20] David Brooks, “Make Everybody Hurt,” New York Times, February 22, 2011, p. A25,

[21] John Sides, “The Relationship between Union Membership and State Budget Deficits,” The Monkey Cage, February 21, 2011,

[22] Barry Hirsch and David Macpherson, Union Membership and Coverage Database from the CPS,

[23] Union strength variables based on data from American Federation of State County and Municipal Employees (AFSCME) Department of Research and Collective Bargaining, compiled from state bargaining laws.


[i] The state and national employment numbers are from two different BLS sources; caution should be taken in comparing the data in Figure 3 with the data in Figures 4 and 5.

[ii] The union density data is only available for public sector workers as a whole and includes: state, local and federal workers. Federal workers make up a small fraction of public sector workers as a whole.

[iii] Set of union dummy variables are dichotomus variables defined as follows:
State EE = 1 if state employees are covered under collective bargaining by law, constitution, charter, executive order or court decision (COVERED).
Police = 1 if police officers are COVERED.
Fire = 1 if firefighters are COVERED.
Teacher = 1 if teachers are COVERED.
Local = 1 if public employees are COVERED.
RTW = 1 if state has right to work law.
Strike = 1 if at least one type of public workers are permitted to strike.