May 21, 2009
by James Sherk
Backgrounder #2275

What do unions do? The AFL-CIO argues that unions offer a pathway to higher wages and prosperity for the middle class. Critics point to the collapse of many highly unionized domestic industries and argue that unions harm the economy. To whom should policymakers listen? What unions do has been studied extensively by economists, and a broad survey of academic studies shows that while unions can sometimes achieve benefits for their members, they harm the overall economy.

Unions function as labor cartels. A labor cartel restricts the number of workers in a company or industry to drive up the remaining workers’ wages, just as the Organization of Petroleum Exporting Countries (OPEC) attempts to cut the supply of oil to raise its price. Companies pass on those higher wages to consumers through higher prices, and often they also earn lower profits. Economic research finds that unions benefit their members but hurt consumers generally, and especially workers who are denied job opportunities.

The average union member earns more than the average non-union worker. However, that does not mean that expanding union membership will raise wages: Few workers who join a union today get a pay raise. What explains these apparently contradictory findings? The economy has become more competitive over the past generation. Companies have less power to pass price increases on to consumers without going out of business. Consequently, unions do not negotiate higher wages for many newly organized workers. These days, unions win higher wages for employees only at companies with competitive advantages that allow them to pay higher wages, such as successful research and development (R&D) projects or capital investments.

Unions effectively tax these investments by negotiating higher wages for their members, thus lowering profits. Unionized companies respond to this union tax by reducing investment. Less investment makes unionized companies less competitive.

This, along with the fact that unions function as labor cartels that seek to reduce job opportunities, causes unionized companies to lose jobs. Economists consistently find that unions decrease the number of jobs available in the economy. The vast majority of manufacturing jobs lost over the past three decades have been among union members–non-union manufacturing employment has risen. Research also shows that widespread unionization delays recovery from economic downturns.

Some unions win higher wages for their members, though many do not. But with these higher wages, unions bring less investment, fewer jobs, higher prices, and smaller 401(k) plans for everyone else. On balance, labor cartels harm the economy, and enacting policies designed to force workers into unions will only prolong the recession.

Push for EFCA
Organized labor’s highest legislative priority is the misnamed Employee Free Choice Act (EFCA).[1] This legislation replaces traditional secret-ballot organizing elections with publicly signed cards, allowing union organizers to pressure and harass workers into joining a union. EFCA would also allow the government to impose contracts on newly organized workers and their employers. Both of these changes are highly controversial.

Supporters defend EFCA by sidestepping concerns about taking away workers’ right to vote. They argue that the bill will make it easier for unions to organize workers. They contend that unions are the path to the middle class and that expanding union membership will raise wages and help boost the economy out of the recession.[2] The official case for EFCA rests on the argument that greater union membership benefits the economy.

Opponents of EFCA largely confine their critique to the legislation itself: its undemocratic nature and the problems with giving government bureaucrats the power to dictate work assignments, benefit plans, business operations, and promotion policies. They also argue, however, that increasing union membership will harm the economy.[3]

Economists have exhaustively examined what unions do in the economy. When debating EFCA, Congress should look to the body of academic research to determine whether unions help or hurt the economy.

Unions in Theory
Unions argue that they can raise their members’ wages, but few Americans understand the economic theory explaining how they do this. Unions are labor cartels. Cartels work by restricting the supply of what they produce so that consumers will have to pay higher prices for it. OPEC, the best-known cartel, attempts to raise the price of oil by cutting oil production. As labor cartels, unions attempt to monopolize the labor supplied to a company or an industry in order to force employers to pay higher wages.[4] In this respect, they function like any other cartel and have the same effects on the economy. Cartels benefit their members in the short run and harm the overall economy.

Imagine that General Motors, Ford, and Chrysler jointly agreed to raise the price of the cars they sold by $2,000: Their profits would rise as every American who bought a car paid more. Some Americans would no longer be able to afford a car at the higher price, so the automakers would manufacture and sell fewer vehicles. Then they would need–and hire–fewer workers. The Detroit automakers’ stock prices would rise, but the overall economy would suffer. That is why federal anti-trust laws prohibit cartels and the automakers cannot collude to raise prices.

Now consider how the United Auto Workers (UAW)–the union representing the autoworkers in Detroit–functions. Before the current downturn, the UAW routinely went on strike unless the Detroit automakers paid what they demanded– until recently, $70 an hour in wages and benefits. Gold-plated UAW health benefits for retirees and active workers added $1,200 to the cost of each vehicle that GM produced in 2007.[5] Other benefits, such as full retirement after 30 years of employment and the recently eliminated JOBS bank (which paid workers for not working), added more.

Some of these costs come out of profits, and some get passed to consumers through higher prices. UAW members earn higher wages, but every American who buys a car pays more, stock owners’ wealth falls, and some Americans can no longer afford to buy a new car. The automakers also hire fewer workers because they now make and sell fewer cars.

Unions raise the wages of their members both by forcing consumers to pay more for what they buy or do without and by costing some workers their jobs. They have the same harmful effect on the economy as other cartels, despite benefiting some workers instead of stock owners. That is why the federal anti-trust laws exempt labor unions; otherwise, anti-monopoly statutes would also prohibit union activity.

Unions’ role as monopoly cartels explains their opposition to trade and competition. A cartel can charge higher prices only as long as it remains a monopoly. If consumers can buy elsewhere, a company must cut its prices or go out of business.

This has happened to the UAW. Non-union workers at Honda and Toyota plants now produce high-quality cars at lower prices than are possible in Detroit. As consumers have voted with their feet, the Detroit automakers have been brought to the brink of bankruptcy. The UAW has now agreed to significant concessions that will eliminate a sizeable portion of the gap between UAW and non-union wages. With competition, the union cartel breaks down, and unions cannot force consumers to pay higher prices or capture higher wages for their members.

Unions in Practice
Economic theory consequently suggests that unions raise the wages of their members at the cost of lower profits and fewer jobs, that lower profits cause businesses to invest less, and that unions have a smaller effect in competitive markets (where a union cannot obtain a monopoly). Dozens of economic studies have examined how unions affect the economy, and empirical research largely confirms the results of economic theory.

What follows is a summary of the state of economic research on labor unions. The Appendix summarizes the papers referenced in the main body of this paper.

Unions in the Workplace.
Unionizing significantly changes the workplace in addition to its effects on wages or jobs. Employers are prohibited from negotiating directly with unionized employees. Certified unions become employees’ exclusive collective bargaining representatives. All discussions about pay, performance, promotions, or any other working conditions must occur between the union and the employer. An employer may not change working conditions–including raising salaries–without negotiations.

Unionized employers must pay thousands of dollars in attorney’s fees and spend months negotiating before making any changes in the workplace. Unionized companies often avoid making changes because the benefits are not worth the time and cost of negotiations. Both of these effects make unionized businesses less flexible and less competitive.[6]

Final union contracts typically give workers group identities instead of treating them as individuals. Unions do not have the resources to monitor each worker’s performance and tailor the contract accordingly. Even if they could, they would not want to do so. Unions want employees to view the union–not their individual achievements–as the source of their economic gains. As a result, union contracts typically base pay and promotions on seniority or detailed union job classifications. Unions rarely allow employers to base pay on individual performance or promote workers on the basis of individual ability.[7]

Consequently, union contracts compress wages: They suppress the wages of more productive workers and raise the wages of the less competent. Unions redistribute wealth between workers. Everyone gets the same seniority-based raise regardless of how much or little he contributes, and this reduces wage inequality in unionized companies.[8] But this increased equality comes at a cost to employers. Often, the best workers will not work under union contracts that put a cap on their wages, so union firms have difficulty attracting and retaining top employees.[9]

Effect on Wages.
Unions organize workers by promising higher wages for all workers. Economists have studied the effects of unions on wages exhaustively and have come to mixed conclusions.

Numerous economic studies compare the average earnings of union and non-union workers, holding other measurable factors–age, gender, education, and industry–constant. These studies typically find that the average union member earns roughly 15 percent more than comparable non-union workers.[10] More recent research shows that errors in the data used to estimate wages caused these estimates to understate the true difference. Estimates that correct these errors show that the average union member earns between 20 percent and 25 percent more than similar non-union workers.[11]

Correlation Is Not Causation.
But these studies do not show that unionizing would give the typical worker 20 percent higher wages: Correlation does not imply causation. Controlling for factors like age and education, the average worker in Silicon Valley earns more than the average worker in Memphis, but moving every worker in Memphis to Silicon Valley would not raise his or her wages. Workers in Silicon Valley earn more than elsewhere because they have specialized skills and work for high-paying technology companies, not because they picked the right place to live.

Similarly, it is not necessarily unions that raise wages. They may simply organize workers who would naturally earn higher wages anyway. Unions do not organize random companies. They target large and profitable firms that tend to pay higher wages. Union contracts also make firing underperforming workers difficult, so unionized companies try to avoid hiring workers who might prove to be underperformers. High-earning workers do not want seniority schedules to hold them back and therefore avoid unionized companies.

Estimates from the Same Worker.
Economists have attempted to correct this problem by examining how workers’ wages change when they take or leave union jobs. This controls for unobservable worker qualities such as initiative or diligence that raise wages and may be correlated with union membership–the worker has the same skills whether he belongs to a union or not. These studies typically show that workers’ wages rise roughly 10 percent when they take union jobs and fall by a similar amount when they leave those jobs.[12]

Data errors become particularly problematic when following workers over time instead of comparing averages across groups. Some economists argue that these errors artificially diminish the union effect.[13] More recent research explicitly correcting for measurement errors has found that taking union jobs causes workers’ wages to rise between 8 percent and 12 percent.[14] One Canadian study expressly examined how much of the difference between union and non-union wages was caused by unions and how much came from unmeasured individual skills. Over three-fifths of the higher wages earned by union members came from having more valuable skills, not from union membership itself.[15] Just as the land surrounding Silicon Valley does not itself raise wages, most of the difference between union and non-union wages has little or nothing to do with unions themselves.

Wage Changes After Unionization.
Studies tracking individual workers also do not prove that unionizing necessarily raises wages. Individual data do not account for firm-specific factors, such as large firms both paying higher wages and being targeted more commonly for organizing drives.

To discover the causal affect of organizing on wages, researchers compare wage changes at newly organized plants with wage changes at plants where organizing drives failed. Such studies look at the same workers and same plants over time, thereby controlling for many unmeasured effects. These studies come to the surprising conclusion that forming a union does not raise workers’ wages.[16] Wages do not rise in plants that unionize relative to plants that vote against unionizing.

Several of the authors of these studies have endorsed EFCA, but their research argues that expanding union membership will not raise wages. This should not come as a complete surprise. Unions in competitive markets have little power to raise wages because companies cannot raise prices without losing customers. Additionally, some unions– such as the Service Employees International Union–have expanded by striking deals promising not to seek wage increases for workers if the employer agrees not to campaign against the union.

Total Wage Effects.
While economic research as a whole does not conclusively disprove that unions raise wages, some studies do come to this conclusion. It is difficult to reconcile these studies with the large body of other research showing that union members earn more than non-union members, or with the strong evidence that unions reduce profits.

A better summary of the economic research is that unions do not increase workers’ wages by nearly as much as they claim and that, at a number of companies, they do not raise wages at all. Once researchers control for individual ability, unions raise wages between 0 percent and 10 percent, depending on the circumstances of the particular companies and workers.

Effect on Businesses.
Union wage gains do not materialize out of thin air. They come out of business earnings. Other union policies, such as union work rules designed to increase the number of workers needed to do a job and stringent job classifications, also raise costs. Often, unionized companies must raise prices to cover these costs, losing customers in the process. Fewer customers and higher costs would be expected to cut businesses’ earnings, and economists find that unions have exactly this effect. Unionized companies earn lower profits than are earned by non-union businesses.

Studies typically find that unionized companies earn profits between 10 percent and 15 percent lower than those of comparable non-union firms.[17] Unlike the findings with respect to wage effects, the research shows unambiguously that unions directly cause lower profits. Profits drop at companies whose unions win certification elections but remain at normal levels for non-union firms. One recent study found that shareholder returns fall by 10 percent over two years at companies where unions win certification.[18]

These studies do not create controversy, because both unions and businesses agree that unions cut profits. They merely disagree over whether this represents a feature or a problem. Unions argue that they get workers their “fair share,” while employers complain that union contracts make them uncompetitive.

Which Profits Fall?
Unions do not have the same effect at all companies. In competitive markets, unions have very little power to raise wages and reduce profits. Companies cannot raise prices without losing business, but if union wage increases come out of normal operating profits, investors take their money elsewhere. However, not all markets are perfectly competitive. Unions can redistribute from profits to wages when firms have competitive advantages.

Economic research shows that union wage gains come from redistributing abnormal profits that firms earn from competitive advantages such as limited foreign competition or from growing demand for the company’s products. Unions also redistribute the profits that stem from investments in successful R&D projects and long-lasting capital investments.[19]

Consider a manufacturing company that invests in productivity-enhancing machines. Workers’ output increases, and the company earns higher profits years after the initial investment. It has an advantage in the marketplace over companies that did not make that same investment. Unions redistribute the higher profits from this investment–not the normal return from operating a business–to their members.

Unions Reduce Investment.
In essence, unions “tax” investments that corporations make, redistributing part of the return from these investments to their members. This makes undertaking a new investment less worthwhile. Companies respond to the union tax in the same way they respond to government taxes on investment–by investing less. By cutting profits, unions also reduce the money that firms have available for new investments, so they also indirectly reduce investment.

Consider General Motors, now on the verge of bankruptcy. The UAW agreed to concessions in the 2007 contracts and has made more concessions since then. If General Motors had invested successfully in producing an inexpensive electric car, and if sales of that new vehicle had made GM profitable, then the UAW would not have agreed to any concessions. The UAW would be demanding higher wages. After the union tax, R&D investments earn lower returns for GM than for its non-union competitors such as Toyota and Honda.

Economic research demonstrates overwhelmingly that unionized firms invest less in both physical capital and intangible R&D than non-union firms do.[20] One study found that unions directly reduce capital investment by 6 percent and indirectly reduce capital investment through lower profits by another 7 percent. This same study also found that unions reduce R&D activity by 15 percent to 20 percent.[21] Other studies find that unions reduce R&D spending by even larger amounts.[22]

Research shows that unions directly cause firms to reduce their investments. In fact, investment drops sharply after unions organize a company. One study found that unionizing reduces capital investment by 30 percent–the same effect as a 33 percentage point increase in the corporate tax rate.[23]

Unions Reduce Jobs.
Lower investment obviously hinders the competitiveness of unionized firms. The Detroit automakers have done so poorly in the recent economic downturn in part because they invested far less than their non-union competitors in researching and developing fuel-efficient vehicles. When the price of gas jumped to $4 a gallon, consumers shifted away from SUVs to hybrids, leaving the Detroit carmakers unable to compete and costing many UAW members their jobs.

Economists would expect reduced investment, coupled with the intentional effort of the union cartel to reduce employment, to cause unions to reduce jobs in the companies they organize. Economic research shows exactly this: Over the long term, unionized jobs disappear.

Consider the manufacturing industry. Most Americans take it as fact that manufacturing jobs have decreased over the past 30 years. However, that is not fully accurate. Chart 1 shows manufacturing employment for union and non-union workers. Unionized manufacturing jobs fell by 75 percent between 1977 and 2008. Non-union manufacturing employment increased by 6 percent over that time. In the aggregate, only unionized manufacturing jobs have disappeared from the economy. As a result, collective bargaining coverage fell from 38 percent of manufacturing workers to 12 percent over those years.


Manufacturing jobs have fallen in both sectors since 2000, but non-union workers have fared much better: 38 percent of unionized manufacturing jobs have disappeared since 2000, compared to 18 percent of non-union jobs.[24]

Other industries experienced similar shifts. Chart 2 shows union and non-union employment in the construction industry. Unlike the manufacturing sector, the construction industry has grown considerably since the late 1970s. However, in the aggregate, that growth has occurred exclusively in non-union jobs, expanding 159 percent since 1977. Unionized construction jobs fell by 17 percent. As a result, union coverage fell from 38 percent to 16 percent of all construction workers between 1977 and 2008.[25]


This pattern holds across many industries: Between new companies starting up and existing companies expanding, non-union jobs grow by roughly 3 percent each year, while 3 percent of union jobs disappear.[26] In the long term, unionized jobs disappear and unions need to replenish their membership by organizing new firms. Union jobs have disappeared especially quickly in industries where unions win the highest relative wages.[27] Widespread unionization reduces employment opportunities.

More Contractions but Not More Bankruptcies. Counterintuitively, research shows that unions do not make companies more likely to go bankrupt. Unionized firms do not go out of business at higher rates than non-union firms.[28] Unionized firms do, however, shed jobs more frequently and expand less frequently than non-union firms.[29] Most studies show that jobs contract or grow more slowly, by between 3 and 4 percentage points a year, in unionized businesses than they do in non-unionized businesses.[30]

How can union firms both lose jobs at faster rates than non-union firms and have no greater likelihood of going out of business? Unions try not to ruin the companies they organize. They agree to concessions at distressed firms to keep them afloat. However, unions prefer layoffs over pay cuts when a firm does not face imminent liquidation. Layoffs at most union firms occur on the basis of seniority: Newer hires lose their jobs before workers with more tenure lose theirs. Senior members with the greatest influence in the union know that they will keep their jobs in the event of layoffs but that they will also suffer pay reductions. Consequently, unions negotiate contracts that allow firms to lay off newer hires and keep pay high for senior members instead of contracts that lower wages for all workers and preserve jobs.[31]

Economists expect unions to behave like this. They are cartels that work by keeping employment down to raise wages for their members.

Consider General Motors. GM shed tens of thousands of jobs over the past decade, but the UAW steadfastly refused to any concessions that would have improved GM’s competitive standing. Only in 2007–with the company on the brink of bankruptcy–did the UAW agree to lower wages, and then only for new hires. The UAW accepted steep job losses as the price of keeping wages high for senior members. If GM does file for bankruptcy, it will likely emerge as a smaller but more competitive firm. It will still exist and employ union members–but tens of thousands of UAW members have lost their jobs.

Unions Cause Job Losses. The balance of economic research shows that unions do not just happen to organize firms with more layoffs and less job growth: They cause job losses. Most studies find that jobs drop at newly organized companies, with employment falling between 5 percent and 10 percent.[32]

One prominent study comparing workers who voted narrowly for unionizing with those who voted narrowly against unionizing came to the opposite conclusion, finding that newly organized companies were no more likely to shed jobs or go out of business.[33] That study, however–prominently cited by labor advocates–essentially found that unions have no effect on the workplace. Jobs did not disappear, but wages did not rise either. Unless the labor movement wants to concede that unions do not raise wages, it cannot use this research to argue that unions do not cost jobs.

Slower Economic Recovery. Labor cartels attempt to reduce the number of jobs in an industry in order to raise the wages of their members. Unions cut into corporate profitability, also reducing business investment and employment over the long term.

These effects do not help the job market during normal economic circumstances, and they cause particular harm during recessions. Economists have found that unions delay economic recoveries. States with more union members took considerably longer than those with fewer union members to recover from the 1982 and 1991 recessions.[34]

Policies designed to expand union membership whether workers want it or not–such as the misnamed Employee Free Choice Act–will delay the recovery. Economic research has demonstrated that policies adopted to encourage union membership in the 1930s deepened and prolonged the Great Depression. President Franklin Roosevelt signed the National Labor Relations Act. He also permitted industries to collude to reduce output and raise prices–but only if the companies in that industry unionized and paid above-market wages.

This policy of cartelizing both labor and businesses caused over half of the economic losses that occurred in the 1930s.[35] Encouraging labor cartels will also lengthen the current recession.


Unions simply do not provide the economic benefits that their supporters claim they provide. They are labor cartels, intentionally reducing the number of jobs to drive up wages for their members.

In competitive markets, unions cannot cartelize labor and raise wages. Companies with higher labor costs go out of business. Consequently, unions do not raise wages in many newly organized companies. Unions can raise wages only at companies that have competitive advantages that permit them to pay higher wages, such as successful R&D projects or long-lasting capital investments.

On balance, unionizing raises wages between 0 percent and 10 percent, but these wage increases come at a steep economic cost. They cut into profits and reduce the returns on investments. Businesses respond predictably by investing significantly less in capital and R&D projects. Unions have the same effect on business investment as does a 33 percentage point corporate income tax increase.

Less investment makes unionized companies less competitive, and they gradually shrink. Combined with the intentional efforts of a labor cartel to restrict labor, unions cut jobs. Unionized firms are no more likely than non-union firms to go out of business–unions make concessions to avoid bankruptcy–but jobs grow at a 4 percent slower rate at unionized businesses than at other companies. Over time, unions destroy jobs in the companies they organize. In manufacturing, three-quarters of all union jobs have disappeared over the past three decades, while the number of non-union jobs has increased.

No economic theory posits that cartels improve economic efficiency. Nor has reality ever shown them to do so. Union cartels retard economic growth and delay recovery from recession. Congress should remember this when considering legislation, such as EFCA, that would abolish secret-ballot elections and force workers to join unions.

James Sherk is Bradley Fellow in Labor Policy in the Center for Data Analysis at The Heritage Foundation.
Appendix and Resources