Americans love to debate the size of government. The issue is the central cleavage between our two political parties these days, after all. But does size really matter? As we rediscover every time this misdirected argument arises—roughly every dozen years—the quality of government is different from and vastly more important than the size of government. One can find examples of large governments—in terms of the number of public employees relative to private-sector ones and the amount of money spent as a percentage of GDP—that are rather ineffective while others are paragons of effectiveness (think Brazil versus Denmark). On the other hand, some small governments, by the same metrics, are very effective and others very ineffective (think Hong Kong versus Guatemala). Therefore, a more fruitful line of inquiry is to analyze what government is actually able to do—and the cost of its doing it.

Figuring out what helps or hinders government quality is no easy task. As Francis Fukuyama recently argued in Political Order and Political Decay, in countries where democracy preceded the construction of state bureaucracies, government capacity is harder to develop. Egalitarianism in the United States arrived before the state, and democracy—beginning with white male suffrage—arrived before the existence of a professional civil service. The result was that from the 1820s onward a party patronage system developed in state and local government in which parties sought to capture public resources and redistribute them to their supporters in exchange for their votes. It took nearly a century of struggle, from the 1870s to the 1960s, to build a competent bureaucracy based on civil service rules and merit systems. Sequencing explains why democracy and bureaucratic performance are at odds in some countries but not so much in others.

Yet almost as soon as the American party bosses were defeated and the ink was dry on the civil service statutes that professionalized the bureaucracy, a new method of mobilizing the resources of the public sector for political purposes emerged. From 1959 to 1984 a majority of the fifty states adopted laws granting collective bargaining rights and union security provisions to various categories of state and local government employees. These provisions stipulated a process to determine legally binding contractual agreements for the terms and conditions of employment and provided for exclusive representation of a group of employees by a single union, dues collection by government, and “fair share” fees charged to non-union members for representation by the union.

Consequently, the number of public employees belonging to unions shot up from less than 10 percent of full-time workers in 1960 to 36 percent in 1980 (in some states more than 60 percent), where it has remained since then. These new labor organizations became a way of mobilizing the resources of public employees on behalf of candidates and parties. In exchange, the unions demanded more benefits for their members, including higher pay, better health and pension benefits, and rules that shield workers from managers. The result was a new, if more limited, spoils system. And just like the original spoils system of the Jacksonian era, it soon became an obstacle to government efficiency.

Defenders of public employee unions, who see them as a bulwark against efforts to shrink the size of government, often overlook the ways that powerful public sector unions actually weaken government’s ability to do things at a reasonable price. The overall effect of unionizing government employees has been to weaken the capacity of state and local government by distorting democratic processes, driving up government costs, skewing governing priorities, and depressing bureaucratic quality. These organizations have these effects because they constitute unique types of interest groups, distinct from both private-sector unions and other membership groups.

This difference is hugely important because, when we speak of the “state” in the United States, we are talking mainly about state and local government. Public employment counts for about a fifth of the total U.S. labor force today, but of that fifth state and local government employ nearly five times as many people (about 15 million) as does the Federal government (about 3.5 million). The vast majority of citizen interaction with and reliance on government occurs at the state and local level.

Government Unions and Democracy

The classic critique of public sector unions is that they undermine the sovereignty of democratic government. A stylized version of democracy holds that citizens vote in competitive elections for candidates offering distinct policy programs; the winners then take office, pass new statutes reflecting those programs, and implement them through the bureaucracy. By granting collective bargaining rights to unions representing government employees, elected representatives partially delegate control of the bureaucracy to unelected labor leaders. Elected officials negotiate with union leaders over public workers’ pay as well as the conditions under which they carry out their tasks. Many policy choices are thus the outcome of negotiations between officeholders and unions rather than the expression of the people’s will channeled through elections.

State courts objected to such delegation of legislative authority; indeed, American jurisprudence until the 1940s held that such negotiations were unconstitutional. The courts based their view on a simple observation: Since private corporations are partnerships with fiduciary obligations to shareholders but not to the general public, labor deserves the right to form a partnership (a union) to enable bargaining on theoretically equal terms; but governments are obligated to the general public in a democracy, so if government does not treat public employees properly, the proper recourse is to replace government decision-makers at election time. If the public vouchsafes its trust in elected officials to make good judgments about all issues of concern to the public weal, there is no sound basis upon which to place the salaries and benefits of bureaucrats in some other category.

The enactment of public sector collective bargaining statutes also created a unique type of interest group that gets “two bites at the apple.” Unlike private sector unions and other interest groups, public sector unions get to bargain collectively with their employers (the first bite) and then they get to influence those on the other side of the bargaining table through electioneering, fundraising, and lobbying (the second bite). Private sector unions only really get one bite (at the bargaining table) as they try to directly improve the material conditions of their members. Other interest groups, such as the National Rifle Association or the Sierra Club, get to electioneer, fundraise, and lobby, but they don’t get to bargain collectively.

The basis of a public sector union’s power is that it is the exclusive representative of all workers in a bargaining unit, regardless of whether workers voted in favor of unionizing or whether individual workers decide to become union members or not. Indeed, more than 90 percent of public sector union members today belong to unions created long before they were hired. As the exclusive representative, these unions can collect “fair share” fees from non-members. The rationale for permitting unions to collect fees from non-members is to prevent freeriding on the union’s collective bargaining activities. (This argument is somewhat circular, insofar as these unions first asked to become the exclusive representative, and then argued that they could comply with that duty only by collecting fees from non-members.) Today, fair-share fees are often nearly identical amounts to union dues and provide a powerful incentive for workers to just join the union. These legal provisions inflate both the membership rolls and the bank accounts of public sector unions.

However, in two recent cases, Knox v. SEIU (2012) and Harris v. Quinn (2014), the Supreme Court questioned whether the requirement that non-members pay fair-share fees to a union violates the Constitution’s First Amendment protections of freedom of speech and association. The issue is the political character of public-sector collective bargaining, insofar as even negotiating something as standard as pay touches on the allocation of scarce public resources. Justice Samuel Alito explained the issue in Harris: “In the private sector, the line is easier to see. Collective bargaining concerns the union’s dealings with the employer; political advocacy and lobbying are directed at the government. But in the public sector, both collective bargaining and political advocacy and lobbying are directed at the government.” Taken to its logical conclusion—as it is in Friedrichs v. California Teachers Association, a case that may soon end up on the Court’s docket—this line of reasoning would declare fair-share fees an unconstitutional violation of workers’ rights, since there is no way to assure non-union workers that their monies do not underwrite political causes with which they do not agree. As Justice Alito put it in Knox: “Free-rider arguments . . . are generally insufficient to overcome First Amendment objections” of non-members.

The fulcrum of the Court’s reasoning is that collective bargaining in the public and private sector is a fundamentally different enterprise, the historic mean before the 1940s, as briefly described above. Almost all decisions relevant to public employees—such as pay, work rules, benefits, and hiring and firing—ultimately are political decisions. When public employees unionize, they are driven to try to manage legislators who are supposed to be managing them for the sake of the citizenry. Government unions make campaign contributions and organize get-out-the-vote drives to elect politicians who then act as “management” in negotiations. Politicians agree to generous contracts for public workers; those workers then pay their union dues—a portion of which is funneled back into those same politicians’ campaign war chests. This creates a pernicious cycle that is hard to break.

Furthermore, in the public sector the favorable impression created by the words “collective bargaining” is misleading. Those words convey a notion that the public interest is realized insofar as government elected officials fully press for the interests of the public (children in schools, citizens in need of various government services) and the unions press for the interests of workers (better salaries, benefits, and working conditions) and, through collegial negotiations, end up with a fair compromise. In reality, that happy result rarely happens because there are different and unequal incentive structures on either side of the bargaining table. Information asymmetries and distinct time horizons give unions an advantage. Elected officials, operating in areas with uncertain outcomes and harried staff, don’t always know the best interest of citizens. Politicians often are more concerned with the next election cycle than with the long-term budget impact of pay increases or pension formulas. On the other hand, unions are there for the long haul and have clear ideas about what their interests are and what their members want. Democratic accountability does little to constrain politicians or unions because the “rationally” ignorant public pays little attention to the technicalities of public sector labor relations. Newspapers devote little coverage to state and local budget issues. Consequently, over time collective bargaining works like a ratchet that slowly turns in favor of the union position.

Public sector unions also enjoy a number of advantages over traditional interest groups when it comes to political influence. They have easier access to policymakers through the collective bargaining process; they can more easily mobilize their members for protests and electoral participation; and they have their collective action problems solved for them by statute and enjoy a steady revenue stream, as union dues are deducted directly from members’ paychecks. Public sector unions are also deeply entwined with the Democratic Party and are consistently among the largest donors to candidates for office at all level of government as well as among the biggest spenders on independent campaign expenditures. In strong union states, public sector unions handily outspend other interest groups on campaigns and elections.

For example, in Federal races the nation’s two teacher unions (the National Education Association and the American Federation of Teachers) have been by far the largest contributors of hard-money donations to candidates and parties over the past twenty years, with over 90 percent of those funds going to Democrats. The American Federation of State, County and Municipal Employees (AFSCME) was the biggest spender in the 2010 Federal elections when one includes both hard and soft money. This is remarkable because most policies of concern to the AFSCME membership and the teachers unions are made at the state and local government level—and the unions often swamp the competition in political activity at those levels. According to the California Fair Political Practices Commission, the California Teachers Association spent more than $210 million on politics in the first decade of the 21^st century, which was more than the pharmaceutical, oil, and tobacco industries combined. In sum, few other interest groups possess such an arsenal of resources. Bringing them all to bear can actually reduce the quality of the democratic process.

Thus, unions representing government workers, elected officials, and agency managers can combine to create a uniquely powerful “iron triangle” in which discussion is technical and therefore confined to experienced insiders. To the extent that public sector unions succeed in insulating policy decisions from citizen pressures, they can to do so because they have the means and motive to pursue their goals tirelessly, whereas the attention of the general public is too episodic and under-informed to have much of an impact.

Fairness, Compensation, and Costs

As public employee unions became major political players in states and cities throughout the country, they contributed to growing differences between private and public sector work. In the former, cutthroat competition and rising inequality predominate, as constant evaluation, long hours, and pay-for-performance standards have become the norm. In the latter, job security and greater egalitarianism persist, as government work is characterized by limited evaluations, bureaucratic regularity, and slow advancement.

The differences between public and private sector work have raised questions about the fairness and efficacy of public sector compensation. Fairness demands that workers not be paid more for the same job just because they are in the public sector. And efficacy means that governments shouldn’t pay more than is necessary to attract qualified workers. Violation of these principles is unjust to both private sector workers and taxpayers. Therefore, many have asked whether public workers’ pay is linked to merit in measureable ways and whether they are overpaid in comparison to their private sector counterparts.

Many scholars find that, thanks to public sector unions’ collective bargaining and political activities, unionized government workers receive greater total compensation relative to non-unionized government workers and thereby increase the costs of government beyond what is actually needed to attain the same quality of public services. For example, political scientists Sarah Anzia and Terry Moe found that cities where firefighters are unionized spend 9 percent more on salary and 25 percent more on benefits than non-unionized cities.¹ Economist Henry Farber found that where collective bargaining is mandatory, government workers earned on average 10 percent more than where it is not.² Such findings are hardly surprising. Unions’ primary objectives are to secure greater pay and benefits for their members; it would be surprising if they didn’t succeed at least in some measure.

However, analysts also generally find that public sector unions boost the pay of their members slightly less than those in the private sector. Yet this finding can be explained easily given that so many public sector workers are white-collar employees with college degrees who start out at a salary level of relative affluence. According to political scientists Jan Leighley and Jonathan Nagler, as the percentage of public sector union members increased between 1971 and 2004, the fraction of union members in the top third of the nation’s income distribution increased by 24 percent, and the proportion of unionists in the bottom third of the distribution declined by 45 percent.³ A labor movement dominated by public sector unions does little for workers with only high school diplomas and limited skills, because it doesn’t represent them.⁴

Economists have extensively explored the issue of “comparative pay”: whether public employees earn more or less than private sector workers who do similar jobs with similar levels of education, experience, and other characteristics. Researchers on comparative compensation have arrived at a consensus on a few points. First, there is much more “wage compression” in the public than in the private sector, which means that the pay of those at the top and the bottom of the salary scale are closer together. Second, the public sector offers better compensation to workers with low levels of education and skills than the private sector does; the private sector is more lucrative than government for workers with high levels of education and skills. Third, in the aggregate, wages tend to be lower in the public sector than in the private sector, but benefits (in particular, pensions and retiree healthcare) are much more generous in public than in private employment.

Analysts disagree over whether there is a compensation “premium” for government work. Looking at salary alone most labor economists find that public workers make less. However, if we factor in pension and health benefits, public employees tend to make more than their private sector counterparts do—on average as much as 15 percent more.⁵ It’s hard to deny that public sector unions do indeed drive up the costs of government.

Public sector costs are largely driven by the generosity of employee benefits. Whereas the average public sector retirement pension is only about $30,000 a year, most people in the private sector have little retirement savings at all. Yet that pension figure doesn’t capture what is really going on. It includes people who worked for government a long time ago, when pensions were less generous, and people who worked for government for a short time. In California today, state workers often retire at 55 years of age with pensions that exceed what they were paid during most of their working years. New York City now pays out more for retired than for active police officers. In addition, under the “defined benefit” pension plans that prevail in the public sector, recipients do not have to shoulder any of the financial risk of investments, which is borne by taxpayers.

A study by the Center for Retirement Research at Boston College found that workers who spent nearly their entire career in public employment ended up better off financially in retirement than similarly situated workers—people with similar jobs and skills—who spent their entire careers in the private sector or worked for government only for a short time. Workers who remain in government employ for many years are the ones who really reap the benefits of generous pension and health plans. Therefore, such munificence is unlikely to serve as a powerful instrument for retaining younger workers. Teachers’ compensation structure, which heavily favors older teachers, is a case in point.

The takeaway point beyond the details, again, is that unionization has increased the costs of government. In addition, in some places in America today, citizens are paying more than citizens elsewhere for the same or inferior public services. And public union bargaining has led to violations of fairness, as some public sector workers are making more than their private sector counterparts. Finally, government work has become especially attractive for those with limited skills, whereas it often repels ambitious strivers at the top end of the labor market who can do better in the private sector. Such a one-two punch is hardly good for government quality.

Reduced Productivity

Public sector unions clearly shape how well American government operates and at what cost. Besides salary and benefits, they affect government performance through the negotiation of work rules that often emasculate management. Although such rules clearly protect workers from capricious managers, they also tend to reduce productivity by eliminating discretion and squelching innovation. Many unionized government agencies cannot structure pay incentives to reward excellence; they cannot easily transfer, reassign, or dismiss poor performers; and they cannot easily hold union-member managers accountable. Over time, many parts of American government have come to resemble Gulliver in Lilliput: tied down by hundreds of rules lobbied for and negotiated by public sector unions. Since the 1970s, public sector unions have consistently resisted organizational enhancements embraced by the private sector, blocked efforts to privatize and streamline services, and sought to make government more rule bound and risk averse. There is a good argument to be made that the United States needs fewer courts and parties hamstringing what bureaucratic professionals can do, but the impact of public sector unions does little to inspire hope that a more autonomous Executive Branch bureaucracy would deliver better quality government.

A main reason comes down to the motivational uses of money. Money is certainly a motivator, but union-negotiated public sector contracts tend to eliminate the connection between pay and performance. Contracts stipulate a salary schedule based on longevity: If a worker remains on the job, his or her salary will slowly but surely rise over time. Exerting greater effort on the job won’t win workers a raise or a bonus. For instance, teacher salaries are determined by only two factors: seniority and educational credentials accumulated. Monetary incentives and performance assessment aren’t part of the public manager’s toolkit.

Then there is the inability of managers to threaten workers credibly with dismissal. In strong union jurisdictions teachers are almost never dismissed by school administrators, as “due process” procedures render the task so costly and time-consuming that many principals don’t even try. For example, in the first decade of the 21^st century only 19 of nearly 295,000 California teachers were fired for poor performance. During that period, Los Angeles spent $3.5 million trying to dismiss seven teachers for poor performance. It managed to get rid of three, two of whom received large severance packages. Government managers also lack control over their workforces in other ways. Police unions have negotiated extensive procedures that shield workers accused of misconduct from disciplinary action.

This dense network of rules—enforced by union lawyers ready to ensure that managers follow them to the letter—reduce the performance of many public bureaucracies. American public schools provide a welter of examples. Stanford economist Caroline Hoxby has shown that when school districts begin collective bargaining with teacher unions student performance declines and some outputs, such as the student dropout rate, get worse.⁶ Hoxby’s colleague in political science, Terry Moe, found that as teacher contracts become denser with rules, student achievement falls, especially among minority students. The result is a major loss of human capital, especially among the least fortunate.⁷

Take, for instance, work rules such as seniority transfer rights for teachers and “last in, first out” layoff policies. Under the former, senior teachers in schools that are being downsized can transfer to other schools within a district and bump junior teachers out of their jobs. According to the latter, the most junior teachers must be laid off first in the event of a budget crunch, regardless of performance. Such provisions leave principals with limited control over the quality of teachers they are putting into classrooms.

In the Golden State, the lawsuit Vergara v. California challenged work rules that granted new teachers tenure after only 16 months on the job, due process to dismiss tenured teachers, and state laws that stipulated that during downsizing teachers could be dismissed only on the basis of seniority and irrespective of performance. Testimony about the effects of these policies, Los Angeles County Superior Court Judge Rolf Treu wrote, “shocks the conscience” and suggests that the organization of the schools “disproportionately affect[s] poor and/or minority students.”

Besides schools, consider that in California’s prison system, the prison guards’ 1999–2001 union contract stipulated that 70 percent of prison positions would be doled out on the basis of seniority—irrespective of whether officers’ skills matched particular jobs. Prison wardens and managers could not assign the most effective guards where they were most needed. One wonders what California was paying the wardens to do. In his book The Toughest Beat (2013) sociologist Joshua Page documents how other rules negotiated by the union make it nearly impossible to discipline corrections officers accused of misconduct.

Finally, consider trash collection, one service that economists of all stripes generally agree is more efficient when done by private contractors than by government employees. Timothy Chandler and Peter Feuille examined 740 cities and found that when sanitation workers were unionized, cities were less likely to consider contracting out the service.⁸ Public officials simply feared union wrath at the polls, and in other ways too. Those cities with strong unions thus were forced to pay more for carting and hauling. In Chicago, ward politics interacted with sanitation union contracts to drive up the costs of trash collection to the point that the city was paying $231 per ton in 2011 compared to $129 per ton in Los Angeles—the next biggest spender.

The protected aspects of the public union world have reduced efficiency in the delivery of public sector services. Ultimately, the poor and the middle class who send their children to public school, rely on cities to collect their trash, commute to work using public transportation, and depend on various forms of state support almost invariably get the short end of the stick. They are paying more in taxes and receiving less in services. All sensible people should oppose policies that lead overall costs to outweigh benefits. The problem is that public sector unionism and collective bargaining too often are at odds with common-sense policy and effective government.

Skewing Government Purpose

By increasing the costs of government, public sector unions also skew its priorities. For example, economists Richard Freeman and Robert Valletta found that in urban governments where select categories of workers are unionized and others are not, the government tends to feed the unionized sectors and starve the non-unionized.⁹ Such distortions hardly comport with allowing elected officials to balance competing priorities with scarce public resources.

To the extent that public employee unions are the primary organized opposition to addressing the soaring costs of state and local pension and health care obligations, they restrict the range of democratic choice. Today, states and localities are budgeting more but deciding less, as large slices of public budgets must be devoted to the elements of public employee compensation that are back-loaded into retirement. So while budgets grow in size, what government can do with those funds is more limited. This should hardly be congenial to either liberals or conservatives.

Crowding out, or redirection of government spending, is most evident in the skyrocketing costs of public employees’ pensions. Moody’s recently estimated the unfunded pension liability of the states to be $3.2 trillion, or 21 percent of total U.S. GDP. Indeed, in 2009, New Jersey’s unfunded liability was $130 billion, which was more than four times the state’s 2008 fiscal year budget. In 2012, California paid pensions of $100,000 or greater to more than 12,000 state and municipal retirees. In California the share of general fund revenues used to pay pensions and public employee health care has doubled over the past decade. In addition, state and local governments now carry a health care liability of $1 trillion, or 33 percent of all state and local government revenue.

Consequently, the cost to local government of employing a worker, including benefits, has soared. In San Jose, the average cost of a full-time worker is $142,000 a year, up 85 percent in ten years. A sanitation worker in New York City now costs $144,000 annually, up from $79,000 a decade ago. In New York City, combined public employee health, pension, and fringe benefits went from consuming 15 percent of Gotham’s budget in 2002 to 34 percent in 2014. The problem is especially grave because the number of retiring government employees is growing faster than the number of retirees in general.

Reordering government priorities is always painful. Governmental institutions are conservative by nature. Around the country, governors and mayors have been forced to propose cuts to education, policing, fire protection, health care programs, and other social services. Philadelphia recently confronted the agonizing choice whether to lay off 2,350 teachers in a school system with 131,000 students or to address the city’s $5 billion unfunded pension liability with $120 million in new revenues by extending an 8 percent sales tax. These sorts of tough choices face elected officials in many parts of the country, and are only made more difficult by union opposition. Sacrificing a swath of states’ and cities’ spending on education, health, and poverty programs to pay for the benefits of a small slice of people who have worked for government for their entire careers is hardly a wise allocation of scarce resources.

Because unionized public services cost more and deliver poorer services, many states with strong unions have resorted to borrowing to finance them. Such borrowing is yet another impediment to government productivity and to a sensible allocation of resources since it requires government to spend more on debt service. Harvard economists Richard Freeman and Eunice Han found that “states with public sector bargaining laws have higher debt-to-state gross domestic product (GDP) ratios and slightly higher deficits than states without such laws.”¹⁰ High debt means higher interest payments and lower bond ratings, which is not good for anyone except bankers, as governments are unable to use those tax dollars for other purposes—or return them to citizens through tax cuts. The costs of that debt are clearly constricting states’ ability to do other things.

The Future of Unionized Democracy

The conflict between the interests of government unions and the public interest is profound. The long-standing hope that the two could be reconciled rested on a misguided notion that these organizations are functionally the same as unions in the private sector, which is plainly not the case. Government workers’ unions have increased the costs of government, transformed the labor movement into a tool of the relatively affluent, decreased government productivity, and skewed the allocation of public resources. Ultimately, they have shaped the way that state and local American democracy works. This situation is or ought to be a serious concern for liberals and conservatives, Republicans and Democrats, and private and even public sector workers themselves.

Republicans have every incentive to attack public sector unions, as they are major underwriters of the Democrats and champion higher taxes and government growth. Of course, Republicans have principled objections to public sector unionism alongside their political considerations. On the other side, Democrats face a challenging situation: They rely on the unions for money and manpower during elections but, as the party of government, they also want affordable and effective public services. Therefore, the Democratic Party is divided between centrists who have openly confronted the unions and liberals who have favored union interests. How this battle plays out will have major implications for the quality of American governance.

To set the United States on a path to compete globally, to adjust to rapid technological change, and to assist those most in need of government support in the 21^st century, American government has to become more flexible and responsive. It needs to improve its bureaucratic capacity. Moving in that direction will be contentious, since the unions have a huge stake in preserving the status quo. Other countries face similar challenges in dealing with government unions. It’s not just California’s prison guards who pose a problem for fiscal sanity and effective services, but also unionized British police officers, French rail workers, and Mexican teachers.

Yet reconfiguring relations between government and labor will contribute to the long-term project of creating a government that can balance American society’s aspirations, economic dynamism, and social compassion. As Alexander Hamilton remarked long ago, “A government ill executed, whatever it may be in theory, must be, in practice, a bad government.”