Unelected Unions: Why Workers Should Be Allowed to Choose Their Representatives

Report Jobs and Labor

Unelected Unions: Why Workers Should Be Allowed to Choose Their Representatives

August 27, 2012 29 min read Download Report
James Sherk
James Sherk
Research Fellow, Labor Economics
As research fellow in labor economics at The Heritage Foundation, James Sherk researched ways to promote competition and mobility.

Abstract: Unions negotiate workers’ terms of pay, promotion, layoff, and retirement; union members may not bargain for themselves. However, very few union members voted for this representation. Once organized, unions remain certified indefinitely: Most union members join as a condition of employment. At the very least, unions should regularly stand for re-election. Even better, Congress and state legislatures should allow workers to choose whether they want union representation, and if so, who represents them—unionized or not. This would give workers the ability to negotiate terms that represent their needs, instead of being forced into a one-size-fits-all contract. The right to join, or refuse to join, a union rests on freedom of association. Union members should have the same rights as other citizens to decide who will represent them.

Under general union representation, employees relinquish their individual negotiating authority to a union. The union becomes the sole representative of the employees in negotiations with their employer. Unionized employers must negotiate employment terms with the union and the union alone. They may not bargain with individual workers. Whether employees like the terms their union has negotiated or not, they cannot negotiate them separately. The union decides how to trade off pay, benefits, and other working conditions at the bargaining table. Ostensibly, the union will negotiate better terms collectively than the workers could individually.

However, unions have institutional interests that conflict with their members’ interests. In these cases unions usually place their interests first. For example, unions want to increase their membership and dues income. In states without right-to-work laws, they bargain for “union security” clauses, which require employers to fire workers who do not pay union dues. Such clauses substantially boost union revenues—employees have no choice but to pay. Unions agree to lower compensation in exchange for forced dues.[1] While this benefits the union, it hurts employees by both cutting their pay and forcing them to pay union dues out of what remains.

Some unions will sacrifice their members’ compensation to facilitate organizing drives. Rian Wathen, the former Director of Collective Bargaining for United Food and Commercial Workers Local 700 in Indianapolis, described a situation he once witnessed. His union had organized several stores in a local grocery chain. During contract negotiations the union accepted substantial cuts in their existing members’ health benefits. In exchange, the employer agreed not to resist their organizing efforts at his non-union stores.[2] The contract meant hundreds of thousands more dollars for the union, but hurt the existing members. The Service Employees International Union expanded its ranks in California nursing homes by agreeing in advance to concessions in exchange for organizing assistance.[3] Unions can exercise their bargaining power in ways that harm their members.

Inherited Unions. Very few union members chose their union to represent them. Most accepted union representation as a condition of employment, but did not separately choose either general representation or the specific union that represents them.

This happens because the National Labor Relations Act (NLRA) does not require private-sector unions to stand for re-election. And only one state, Wisconsin, requires government unions to stand for re-election. A unionized workforce remains unionized until the employer goes bankrupt, or the workers decertify it (a prohibitively difficult undertaking). New employees are represented by the union for which previous employees voted. The overwhelming majority of workers in both the private sector and in government inherited collective representation in this manner.

Unions in the Private Sector. Just 7 percent of private-sector union members voted for their union.[4] The remaining 93 percent are automatically represented by a union they had no say in electing. This number may seem high, but the labor movement organized far more workers in the past than today. In 1964, for example, nearly 487,000 private-sector workers voted in certification elections.[5] In 2011, despite a significantly increased population and labor force, only 72,000 did so.[6] The vast majority of unions that exist today are inherited unions. Few current employees had a say in forming them.

The United Auto Workers (UAW), which organized General Motors’ Michigan factories in 1937, is a case in point. Michigan does not have a right-to-work law, so union-represented workers must pay the union's dues or get fired. General Motors’ current employees never had the chance to vote for or against the UAW. UAW representation was a non-negotiable condition of their employment.

Unions Represent Small Portion of Original Supports

Chart 1 shows the number of private-sector workers that voted for union representation each year, as well as the number of those employees still working for their original employer. Between retirements, layoffs, and job changes, unions represent only a small number of their original supporters.

Unions in Government. Inherited representation occurs even more often in government. Private-sector unions must constantly organize new businesses to replace members they lose at bankrupt firms. The government does not go out of business. Most state laws that give government employees collective bargaining powers were passed in the 1960s and 1970s. After they finished their initial organizing drives, unions had few school districts or cities left to organize. In most states, little recent organizing activity has taken place in government. As a result, the vast majority of unionized government workforces were organized at least 30 years ago. Few of today’s employees worked for the government then.

Proportion of Unionized Public-School Teachers in Six States

The situation of public-school teachers demonstrates how few government employees choose their union representation. Federal statistics track how long teachers have worked in the public-school system. Chart 2 shows these figures for several states. The median teacher has been on the job between nine and 14 years. In none of these states have more than 12 percent of teachers worked longer than 30 years. Very few of today’s teachers participated in the decades-old election that selected their current bargaining representatives.

In December 1975, New Hampshire passed a law giving government unions collective bargaining powers. Most of the unionizing elections took place within the next five years. Just 9 percent of New Hampshire’s teachers were hired before their school district voted to unionize. Only 9 percent of teachers were asked if they wanted their union to represent them. The other 91 percent had to accept union representation as a condition of teaching. (Appendix Table 1 displays the proportion of unionized public-school teachers that had the opportunity to vote on unionizing in each school district in New Hampshire.)

Table 1 shows the proportion of unionized public-school teachers that had the opportunity to vote on unionizing in several other states.[7] Florida passed legislation giving government unions collective bargaining powers in 1974, and by 1975 the state’s 10 largest school districts had unionized. Just 1 percent of current teachers were on the job in 1975. Fully 99 percent of the teachers in Florida’s largest school districts had no choice about being represented by their union.

Michigan gave government unions collective bargaining powers in 1965. Seven of the 10 largest school districts in the state had already unionized (even without full collective bargaining powers) before then or organized that year. One of the state's largest school districts unionized in 1971, and two others did so in the 1980s. Across Michigan’s 10 largest school districts, just 1 percent of teachers had the opportunity to decide who would represent them.[8]

In some states, the unionizing votes took place so long ago that the government has no records of the election. South Dakota passed legislation requiring government employers to meet and negotiate with union representatives in 1970.[9] The state Department of Labor no longer has the records showing when those elections took place.[10] Lawsuits decided by the South Dakota Supreme Court show that the Rapid City school district was organized by 1980, and the Aberdeen School district was organized by 1972.[11] At most, 9 percent of Rapid City public-school teachers and 1 percent of Aberdeen public-school teachers had the opportunity to vote on unionizing.

Proportion of Current Teachers Who Voted to Unionize

In other states, unionizing elections occurred so long ago that no current teachers voted for the union that negotiates their contracts. New York City public-school teachers voted to join the United Federation of Teachers (UFT) in 1961,[12] and the same union continues to represent teachers to this day. U.S. Department of Education data show that virtually every teacher who voted in that election has since retired.[13] Joel Klein, the former chancellor of New York City’s public schools, identified UFT opposition to reform as one of the greatest obstacles to improving New York City public schools.[14] The UFT hurts the education of hundreds of thousands of children, but no current teacher in New York voted to give them that power.

Similarly, no current teachers voted for their union in many school districts in Kansas. The state’s legislature passed legislation giving unions collective bargaining powers in 1970. By 1971, several of Kansas’s largest school districts had unionized. However, statistically speaking, none of the teachers in these districts have worked for more than 40 years.[15] They never voted to be represented by the Kansas National Education Association.

Mandatory Union Representation

Almost all union members in both the private sector and in government inherited union representation their predecessors voted for. The law did not give them the opportunity to decide for themselves whether general representation would benefit them and—if so—which union would provide the best representation. This leads to worse contracts for employees.

Union membership has costs and benefits. On the benefit side, unions try to raise pay and benefits. In a competitive economy, however, they have limited power to do so. Raising pay raises costs. If competition prevents a unionized company from also raising prices then such raises could bankrupt it. Unions know this and do not want to lose dues-paying members. Consequently, most union contracts at newly organized companies do not raise wages.[16] Unions only raise pay at companies with competitive advantages that can afford higher costs.[17]

Unionizing also has downsides. The most obvious are the union dues, which can run from several hundred to several thousand dollars a year.[18] Workers do not benefit from being charged hundreds of dollars a year to negotiate a contract that does not raise their pay.

General representation also has less obvious drawbacks. A collective bargaining agreement (CBA) is a collective contract. One contract sets the pay for every employee in the bargaining unit. Such a contract cannot reflect individual effort or contributions. Instead, unions precisely spell out pay with rigid job classifications and seniority schedules. Most unions do not allow companies to reward individual workers with raises or bonuses.[19]

Such one-size-fits-all CBAs were workable when all workers brought essentially the same skills to the bargaining table. But the nature of work in the economy is changing. Employers are automating many repetitive tasks. The fastest growing job sectors have been those requiring individual skills: professional specialty, executive and managerial, and technical and sales jobs.[20]

At the same time, employers are also flattening the job hierarchy. The line between management and workers is blurring. Employers increasingly expect workers to exercise independent judgment and take initiative on the job.[21] The unique skills of individual financial planners, web developers, or medical specialists do not lend themselves to general representation. Employers want to reward—and employees want to be rewarded for—individual contributions that no collective contract can reflect. Many workers feel that a contract prohibiting individual performance-based raises would hold them back.

Private-sector union membership has dropped sharply because workers’ demand for union representation has also fallen.[22] Employees should be allowed to choose whether they negotiate individually or collectively, instead of automatically inheriting general representation.

Unaccountable Unions. Workers who choose general representation should also get to choose which union represents them. Allowing unions to remain workers’ representatives indefinitely reduces their accountability. Currently, unions have few institutional incentives to represent their members well. Whether the union negotiates a good contract or a bad one it will stay in business. This can lead unions to ignore workers’ interests at the bargaining table.

Members of the United Healthcare Workers-West (UHW-West), a California chapter of the Service Employees International Union (SEIU), witnessed their union ignoring their interests firsthand. In order to expand its membership—and thus dues income—the national SEIU gave nursing homes that agreed to support their organizing drives a special deal.[23] The SEIU negotiated contracts that gave the owners of nursing homes who signed “alliance agreements” the “exclusive rights to run the company.” All employee relations (such as pay rates, incentives, promotions, and demotions) were left to the sole discretion of the employer, completely circumventing the union.[24] The union also agreed to give up its right to strike. It would, however, be allowed to charge employees mandatory dues.

Many companies were willing to support unionization if the union legally committed to doing little except collect dues. SEIU membership and income soared, but these new union members had no more influence over their working conditions than before.

Even when unions do not sell out their members, workers’ lack of choice makes unions less responsive to their members’ concerns. They do not have to work hard to serve their members to remain their exclusive bargaining representatives.

Wasted Dues. Inherited representation reduces the pressure on unions to be prudent with their members’ money. Workers cannot shop around for less expensive representatives. This lets unions spend dues wastefully.

Fully 60 percent of union members object to their dues being spent on political causes. [25] These objections do not prevent the AFL–CIO from spending one-sixth of its budget on lobbying and other political activities, often to support candidates its members oppose.[26]

Inherited representation also inflates union salaries. Mary Kay Henry, president of the SEIU, made $290,000 last year. Eliseo Medina, the SEIU’s secretary-treasurer, made $331,000.[27] Joseph Hansen made $361,000 as president of the United Food and Commercial Workers representing hourly workers at grocery stores.[28]

If workers could choose the union that represents them, union officers would have to earn their salaries. SEIU members upset over contracts in which the union agreed to do little would probably look for a new union. A full 57 percent of union members say they do not get enough value for their dues.[29]

Existing Options Limited. Existing laws do little to hold unions accountable to their members. Workers dissatisfied with their union representation can attempt to replace their union representatives in internal union elections, but local chapters of unions are subordinate to the national union. If the local’s policies clash with the policies of the national union, the national can place the local under trusteeship.

This happened to the SEIU’s members in UHW-West. The local’s elected leaders believed that the national’s “growth” strategy of signing contracts that gave away the union’s bargaining powers betrayed their members. The leaders actively resisted the policy of organizing by agreeing in advance to do little.[30] So the national SEIU forced the UHW-West into trusteeship in 2009, replacing the local’s elected officers with appointees loyal to the national.[31] The rank-and-file members of UHW-West had no power to elect local officers who would put their interests ahead of the SEIU’s bottom line.

Alternatively, workers can petition for a decertification election to remove their union. However, both the law and unions make this prohibitively difficult: The National Labor Relations Act requires workers to collect signatures from 30 percent of the employees in the bargaining unit. These signatures may not be collected while employees are on the clock or in work areas, and employees have a one-month window every three years to collect them.[32] This makes gathering enough signatures to trigger an election extremely arduous, especially if the unit spans multiple cities.

Unions heavily pressure workers not to sign decertification petitions, and most unions expel workers who support decertification. The UAW constitution, for example, calls for the expulsion of any member “affirmatively engaged in efforts to decertify the International Union or any subordinate body thereof.”[33]

Being expelled from the union means that a worker cannot vote in elections for union officers or vote on ratifying a new contract. Signing a decertification petition means that workers lose the little say they have in how their union operateseven though it will still represent them if the decertification attempt fails.

As a result, decertification votes are extremely rare. Only 38,000 private-sector union members voted in a decertification election between 2009 and 2011.[34] That amounts to 0.5 percent of the 8 million private-sector workers unions represented during this time. Given the opportunity, however, many union members voted against their representatives: Unions lost 58 percent of these decertification elections.[35]

The difficulty of decertifying a union means that unions do not need to cultivate employee support. As long as dissatisfied workers grudgingly tolerate their union, it will continue to represent them.

Expanding Worker Choice. The law should allow workers to freely and easily choose whether they want collective representation and—if so—who that representative will be. Workers should not remain stuck with a union others elected decades ago. Nor should unions be allowed to automatically represent new members. They should have to earn employees’ support.

There are two good ways Congress and state legislatures can give workers that choice and make unions accountable to their members: They could either (1) require unions to run for re-election on a regular basis, or (2) let workers choose anyone as their bargaining representative, irrespective of who other employees have designated.

Union Re-Elections. Governors and legislators do not serve indefinitely. They must regularly stand for re-election. The most straightforward way to expand workplace choice is requiring unions also to run for re-election. Re-election votes every two to four years would allow employees to regularly assess their union’s performance as their representative. This would significantly increase unions’ accountability to their members—no matter how workers vote.

Workers who were dissatisfied with their representation could easily remove their union. Workers who were satisfied with the union could just as easily re-elect it. Many union members would probably re-elect their union. They should still get a vote on whether to do so.

Having to run for re-election would make unions more accountable and more responsive to their members’ concerns. Voters re-elect their congressmen about 95 percent of the time.[36] That does not make election campaigns superfluous. The fact that they have to prove themselves by running for re-election makes politicians better representatives than they would be if they served indefinitely.[37] How often would Members of Congress vote against pay increases if they did not face re-election? How concerned would they be with their constituent services or with the views of their constituents? Elections give voters an opportunity to evaluate their representatives’ performance. This forces often self-interested politicians to pay more attention to their constituents’ needs.

Re-election votes would similarly benefit union members. Wisconsin recently made membership in government unions voluntary and required unions to run for re-election. Many unions believed they would lose and did not file for re-election. Membership in the American Federation of State, County, and Municipal Employees fell by 55 percent.[38] Of the unions that ran for re-election 85 percent won.[39] To win the elections, however, many of them cut their dues. The Wisconsin Education Association, for example, reduced dues by 30 percent.[40]

Unions are not entitled to workers’ dues. They should have to earn them. If a union cannot convince employees that it has improved working conditions, they are probably better off without it. Most union members agree: 83 percent of union households believe that workers should have the right to regular votes on remaining unionized.[41] Union members understand that re-election votes benefit them.

Structuring Election Votes. Legislation before Congress, the Employee Rights Act, would require private-sector unions to run for re-election every three years.[42] As Wisconsin demonstrated, states can also require their government unions to stand for re-election. Congress and state legislatures should seriously consider requiring unions to do so. A crucial issue to consider in structuring such votes is whether the union should have to win the support of a majority of employees in the workplace, or just a majority of those who vote. Wisconsin required unions to win support from a majority of all employees. The Employee Rights Act requires unions to win a majority of those who choose to participate in the election, even if that majority represents a small portion of the overall workforce.

There is a good reason to require unions to obtain support from a majority of all workers: It prevents a small minority from imposing a bad contract on everyone else. This happened recently in Michigan when the SEIU petitioned for an election among 43,000 home health care aides paid by Medicaid. Most of these “employees” were family members taking care of disabled relatives. The state conducted a mail-in election, in which only 20 percent of homecare aides participated.[43] Most families threw away the ballots believing they were junk mail. The SEIU, however, mobilized its supporters and won the low turnout election. The union then negotiated a contract in which it provided virtually no benefits for these family members but required them to pay dues. Parents of disabled children soon found union dues being taken from their Medicaid payments when they had no idea an election had even taken place.[44]

Union leaders claim that requiring unions to get majority employee support unfairly counts non-voters as “no” votes. However, unions negotiate on behalf of all employees, and all workers are bound by the contract. Unions should need majority support to receive this power. The best policy would require unions to stand for re-election, and require that they win the support of a majority of all employees.[45]

If policymakers believe the argument against requiring majority support has merit, they should at least require unions to demonstrate substantial employee support. They could, for example, require the number of pro-union votes to constitute at least 40 percent of the workforce for the union to win certification. This would prevent a small minority of employees from imposing a union on an entire workplace. If a union can get the votes of only 20 percent of workers, it should not negotiate on behalf of everyone.

Representative Choice. Letting workers vote to re-elect their bargaining representatives is a straightforward reform that would benefit workers. A more comprehensive approach would provide workers with even greater benefits.

The law currently forces workers in a bargaining unit to accept a single representative. If the majority votes for a certain union, then everyone must accept the same representative, and the contract that union negotiates. This leads to one-size-fits-all contracts that ignore the needs of individual employees. Merely requiring unions to run for re-election does not solve this problem. It still leaves every worker represented by the same agent and under one contract.

Congress and state legislatures should give workers representative choice: allow employees to designate any person as their bargaining representative. The law should allow individual workers to select different unions to represent them, or to represent themselves if they so choose.[46]

Representative choice would allow workers to negotiate contracts tailored to their situations. Workers who support their union and its priorities could continue to select it as their representative. Workers with different preferences—such as parents who value flexibility and paid leave over other benefits—could negotiate different contracts through another union or individually.

Allowing workers to switch bargaining representatives would force unions to respond to their members’ concerns. Like re-election votes, it would hold unions accountable to their members. Unlike re-election votes, it would fix one of the major downsides of unionizing: one-size-fits-all contracts.

Worker Voice vs. Monopoly Face. Unions perform two roles in the workplace: They express the collective employee voice, and they function as monopolists.[47] Most Americans are familiar with the “voice” role of unions. Unions communicate employees’ desires and concerns collectively to management. As workers’ representatives, unions inform employers about safety concerns, shortcomings in benefit structures, and other workplace concerns. While managers might discount individual employee complaints, they take concerns communicated collectively quite seriously. Economists believe this collective voice improves working conditions and productivity.

Unions also act as monopolists. Because unions exclusively represent all workers, they control a firms’ labor supply. They use this power to restrict entry and reduce output, raising their compensation. Unions are essentially labor cartels. This “monopoly face” imposes significant economic costs. If a city wants to hire employees it must hire them on the terms the union sets—even if some employees would accept different terms.

Free trade and deregulation have severely limited the monopoly power of private-sector unions—even if they control the labor of one company, consumers can buy elsewhere. This is why unions do not raise pay at most companies they organize. The government, however, has no competitors. Unions can substantially raise pay in government without risking jobs—the state or county simply has to raise taxes. Consequently, they do. In California, the average government employee earns 30 percent more than a comparable private-sector worker.[48] Their monopoly allows unions to force taxpayers to pay for government employees to retire at 55, and makes it impossible to reward excellent teachers or remove bad ones.

Representative choice would preserve the “voice” role of unions while removing their monopoly powers. Unions would still represent employees’ concerns to management. But if potential municipal employees were willing to work until age 65, one union could not stop them from separately agreeing to work for the city. Ineffective teachers could try to hold out for a contract that bases all raises on seniority, but they could not stop a school district from hiring better teachers unafraid of performance evaluations.

Equal Representation. Representative choice would also force unions to pay attention to the concerns of all their members. Unions currently do not represent all workers equally. Union contracts typically benefit workers with more influence in the union at the expense of those with less. One-size-fits-all contracts allow unions to force a bad contract on employees with less clout.

For example, unions usually negotiate seniority-based layoffs: When bad times come, the newest hires get laid off first. This is why Nevada laid off seven Teachers of the Year.[49] Senior union members strongly support the seniority-based layoffs. Newer workers would rather not be sacrificed to protect senior employees. They would prefer layoffs based on performance. However, junior employees have little clout in the union, so unions negotiate seniority-based layoffs.

Similarly, senior union members highly value retirement benefits—they are much closer to retirement. Having gained seniority-based raises, they care less about starting salaries. New graduates care much more about starting pay and less about retirement benefits.[50] Since senior employees have more clout in the union, union contracts reflect their preferences—not those of new hires. A disproportionate share of the compensation of public-school teachers, for example, goes toward retirement benefits.[51] This makes it harder for school districts to attract good potential teachers.

Unions can only benefit some workers at the expense of others because all employees must accept their representation. If junior teachers could negotiate separately with the school board, they would receive a contract that reflected their concerns. They could negotiate a contract with performance-based layoffs, a higher starting salary, and less expensive retirement benefits. Representative choice would prevent unions from making some workers more equal than others.

Freedom of Association. Wisconsin Governor Scott Walker closed a $3.6 billion budget deficit by sharply limiting the collective bargaining powers of government unions. Unions protested vehemently that Walker had stripped them of their “collective bargaining rights.” They organized recall elections against Walker and legislators who voted for the reforms. While Wisconsin voters re-elected Walker, the unions succeeded in recalling three state senators who voted for the reforms.

The union campaign was fueled by the moral argument that Walker had attacked their “collective bargaining rights.”[52] A long-standing moral tradition holds that workers have the right to form and join labor unions. In the 1891 Encyclical Rerum Novarum, Pope Leo XIII argued that “to enter into [a trade union] is a natural right of man; and the State has for its office to protect natural rights, not destroy them.”[53]

However, this right is founded upon the principle of freedom of association. For this reason the same moral tradition also holds that union membership should be voluntary. [54]

Representative choice allows workers to bargain collectively, and it enhances their freedom of association: It allows employees to join any union, irrespective of who their colleagues or predecessors selected. It fully preserves the “voice” and “association” functions of unions while removing their harmful monopoly. Unions would have difficulty arguing that representative choice takes away their rights. It simply reforms collective bargaining to let workers choose their own representatives. No union should have the right to force employees to accept its services.

State Initiatives. The law should not require all workers to accept representation from one and only one union. Neither should the law force workers to be represented by a union that actively ignores their concerns. Congress and state legislatures should revise collective-bargaining statutes to allow private-sector and government employees to choose anyone as their bargaining representative. Unions would still represent workers, but they would no longer exclusively represent every employee unless every employee voluntarily chose them.

Some state constitutions, such as Florida’s, expressly guarantee the ability of employees to bargain collectively. The Missouri Supreme Court has ruled that the Missouri constitution similarly gives government employees collective-bargaining powers. Unions are attempting to amend the Michigan constitution to include a similar guarantee. Wisconsin-type reforms are illegal in any state whose constitution guarantees collective-bargaining powers. These state constitutions still permit representative choice, however. Nothing in any state constitution mandates that employees must be forced to use the same bargaining representative.

It may prove difficult to modify private-sector labor law—Congress has not substantively modified the National Labor Relations Act since 1959. However, many states have revised the collective-bargaining powers they give unions.[55] State legislatures should lead the way in giving workers representative choice. If several states passed and demonstrated the benefits of representative choice, they could serve as a model for labor reforms at the federal level.

The Freedom Foundation has developed model legislation to establish representative choice in government workforces.[56] State legislatures should seriously evaluate this proposal. If they opt not to enact representative choice, they should at least require unions to stand for re-election.


Unions represent employees in the workplace, but most workers never chose their union. The overwhelming majority of workers accepted union representation as a condition of being hired at a unionized firm. They inherited the union that their predecessors voted for decades earlier. Just 7 percent of private-sector workers voted for their union. An even smaller portion of government employees chose their union. For instance, virtually all of the teachers who voted to unionize Kansas’s largest school districts in 1971 have since retired. The current teachers did not choose their representative.

Inherited representation encourages unions to put their interests first—at the expense of the workers they ostensibly represent. Congress and state legislatures should require unions to run for re-election, or allow workers to designate their own bargaining representative. Workers should not be forced to accept a union’s services.

James Sherk is Senior Policy Analyst in Labor Economics in the Center for Data Analysis at The Heritage Foundation. The author extends his gratitude to Heritage Foundation interns Michael Quinn, Stephanie Jaczkowski, and Alex Entz for their invaluable research assistance with this paper.




To determine the proportion of unionized public-school teachers that had an opportunity to vote on unionizing, The Heritage Foundation combined data on statewide teacher tenure collected by the U.S. Department of Education’s National Center for Education Statistics with data on unionizing elections in various school districts.

The teacher tenure data came from the 2007–2008 Schools and Staffing Survey (SASS).[57] One of the survey instruments consists of interviews with a representative sample of public-school teachers in each state. One of the interview questions identifies teachers as union members or non-union members. Two other questions identify the number of years since the individual began teaching full-time in public schools, and the number of years since the individual began teaching part-time in public schools. Heritage Foundation analysts constructed a new variable—public-school tenure—defined as the maximum of these two variables.

This constructed tenure variable will slightly overstate the portion of teachers who voted in the unionizing election. It measures the tenure of teachers in all public schools, not just the school district in which they now teach. If a veteran teacher moved to a district after it unionized, but that individual started teaching before that time, this methodology counts that teacher as having voted in the election. These results should thus be considered an upper bound on the portion of teachers who had the opportunity to vote.

Chart 2 displays the distribution of public-school tenure for unionized teachers in several states. The sample sizes[58] for these surveys are:

Florida: 510 respondents

Kansas: 420 respondents

Michigan: 680 respondents

New Hampshire: 480 respondents

New York City: 650 respondents

South Dakota: 370 respondents

The Heritage Foundation used multiple data sources to determine the date of the most recent unionizing elections.

Florida: Florida passed legislation creating collective bargaining in government in 1974. The Heritage Foundation made public-records requests by phone for four district unions (Polk, Lee, Palm Beach, and Miami–Dade) between June 6 and June 8, 2012. Five other unions (Broward, Hillsborough, Duval, Pinellas, and Orange) published the dates on which they were certified on their websites. The teacher association in Brevard responded to a public-records request for their unionization timeline via e-mail. In all but two of these districts, the teachers voted to unionize before 1974, even though they did not have full collective-bargaining powers. It is unclear whether they voted again under the 1974 law. Consequently, 1974 was used as the latest possible election date for unions formed before then.

Kansas: Kansas instituted collective bargaining in government in 1970. The starting dates of three district unions (Shawnee Mission, Kansas City, and Topeka) were taken from the National Directory of Nonprofit Organizations.[59] The organization date for the union in the Olathe district was cited in a court case: Unified School District No. 233 Johnson County v. Kansas Association of American Educators (2003). The President of the Wichita Teacher Association provided the certification date of his union via a telephone interview.[60]

Michigan: Michigan gave government unions collective-bargaining powers in 1965. The 10 largest school districts in the state are unionized. Secretaries in the union’s main office gave the dates of unionization for the Detroit City School District, Dearborn City School District, and Walled Lake Consolidated Schools.[61] The dates for the formation of the other seven school districts’ teachers unions were obtained from the National Directory of Nonprofit Organizations.[62] In six of these districts, the teachers voted to unionize before 1965, even though they did not have full collective-bargaining powers. It is unclear whether they voted again under the 1965 law. Consequently, 1965 was used as the starting date for unions formed before then.

New Hampshire: The New Hampshire legislature instituted collective bargaining in government with the passage of the Public Employee Labor Relations Act on December 21, 1975. Unlike most other states, the New Hampshire Public Employees Labor Relations Board (PELRB) maintains records of most union certifications going back to the 1970s. However, no central database records the holding of representative elections.[63] The only way to obtain this information is to review each representation file since 1976. The state PELRB suggested using the date of certification to approximate the date of the election—as certification forms are issued after a representative election.

On the union certification forms, if a representation proceeding had been conducted in accordance with “RSA 274-A, Section 10,” it was assumed an election had occurred—as this refers to the section of the New Hampshire Public Employee Labor Relations Act that relates to the election process.

A reference, instead, to Chapter 490:3 in a PELRB certification indicates that it went through the “grandfathering” process—that is, no election was conducted by the PELRB. While this means an election may have occurred earlier, it was not supervised by the PELRB, as such an election would have occurred before the Public Employee Labor Relations Act. Unions grandfathered under the act were assumed to have voted in 1976, the latest date possible under the law. This represents a lower bound on the length of time since the election.

If a modification petition was used to create a new bargaining unit, the board may conduct elections for the “modification proceedings.” Elections are also held if the modification petition seeks to create a new unit. However, if the union lost the election, the bargaining unit would not be amended, and thus the certification would not record each election conducted. Elections may also occur during “challenging” petitions of modification (in the hopes of changing representation) and such failed attempts may also not be recorded. If the union lost the election to challengers or to a “no representation” vote, it would either lose certification or the PELRB would issue a new certification.

In order to ascertain the nature of the modification proceedings and whether they involved a new vote, Heritage Foundation analysts examined the board decisions that dealt with each of the relevant unions between 1976 and 2012, as well as noting the original certification date (a certain representative election). The Heritage Foundation used the date of the most recent election conducted.

The election date information was combined with the teacher tenure distribution to estimate the portion of teachers working in that school district on or before the election date. For example, Manchester Public Schools in New Hampshire voted to unionize in 1977. Three percent of teachers in New Hampshire have been working for 35 or more years. Thus, Table 1 indicates that at most 3 percent of Manchester Public Schools' teachers voted on whether they wanted to be represented by the Manchester Education Association. This does not mean that 3 percent of teachers voted for the union, only that at most 3 percent of teachers could have participated in the election.

New York City: The United Federation of Teachers published a book celebrating its 50th anniversary as collective bargaining representatives in New York City.[64] That book identified the original certification election as occurring in 1961.

South Dakota: South Dakota required local governments to meet and negotiate with government unions in 1970. However, the state Department of Labor does not maintain records of unionizing votes from before the 1990s, so it is not possible to determine exactly when the states’ school districts voted.[65] South Dakota Supreme Court decisions reference negotiations with the Aberdeen and Rapid City teacher associations, and the earliest dates mentioned in these decisions were used as the start date of collective bargaining. These dates are lower bounds on the length of time these school districts have bargained.[66]

The analysis embodied in Table 1 and Appendix Table 1 assumes that teacher tenure is distributed normally throughout each state. It also assumes that the distribution of teacher tenure has not changed substantially since the 2007–2008 survey. In other words, if 3 percent of teachers in Manchester Public Schools had been on the job for 35 years in 2007–2008, the same would be true in 2012.

This assumption will be violated if unionized teachers delayed their retirement during the recession. However, even assuming that no teacher retired between 2007 and 2012 would increase the estimated proportion of teachers with the opportunity to select their union by at most a few percentage points. In some districts, such as New York City, it would not increase the proportion of potential voters at all.

Few Teachers in New Hampshire Participated in Unionization Elections

Few Teachers in New Hampshire Participated in Unionization Elections

 Few Teachers in New Hampshire Participated in Unionization Elections

 Few Teachers in New Hampshire Participated in Unionization Elections

[1]In most cases newly organized unions do not raise their members’ pay. See footnote 16 for details. However, virtually all new union contracts in non-right-to-work states require all employees to pay dues or lose their job.

[2]Rian Wathen, “Protecting Employee Rights,” Heritage Foundation panel discussion, March 20, 2012, http://www.heritage.org/events/2012/03/protecting-employee-rights.

[3]Matt Smith, “Union Disunity,” San Francisco Weekly, April 11, 2007, http://www.sfweekly.com/2007-04-11/news/union-disunity/ (accessed August 15, 2012).

[4]Analysis of data on job tenure of union members released by the Bureau of Labor Statistics, and data on union organizing from the National Labor Relations Board, conducted by J. Justin Wilson, “Job Tenure and Union Elections: Non-voting Union Membership in the Private Sector, 1964 to 2009,” Center for Union Facts, March 2012, http://www.unionfacts.com/downloads/Union_Tenure_Elections.pdf (accessed August 15, 2012). (accessed August 15, 2012).

[5]These figures are for workers covered under the National Labor Relations Act, which applies to most private-sector workers. The Railway Labor Act covers workers on railways and on airlines.

[6]National Labor Relations Board, Election Reports, FY 2011, “Election Data, FY 2011,” http://www.nlrb.gov/sites/default/files/documents/3429/election_data_fy_2011_.xls (accessed August 15, 2012), and Twenty-Ninth Annual Report of the National Labor Relations Board for the Fiscal Year Ended June 30, 1964 (Washington, DC: U.S. Government Printing Office, 1965), Table 14, http://www.nlrb.gov/sites/default/files/documents/119/nlrb1964.pdf (accessed August 15, 2012).

[7]Heritage Foundation calculations using data on the tenure of unionized public-school teachers from the Schools and Staffing Survey (SASS) 2007–2008, and separately collected data on union organizing by school district. See the appendix of this Backgrounder for details.

[8]Ibid. Average across school districts weighted by the number of teachers in each district.

[9]In 1973, the requirement was expanded to meet and negotiate “in good faith.”

[10]The Heritage Foundation requested information on the dates of unionizing elections in South Dakota and was informed by the state Department of Labor that it did not have records for votes before the1990s. E-mail correspondence on June 29, 2012.

[11] Aberdeen Ed. Ass’n v. Aberdeen Bd. of Ed., Aberdeen Independent School Dist., 88 S.D. 127 (1974), and Rapid City Educ. Ass’n v. Rapid City Area School Dist. No. 51-4, 376 N.W.2d 562 (1985).

[12]United Federation of Teachers, “50 Years: United Federation of Teachers 1960–2010,” http://www.uft.org/files/attachments/uft-50-years-book.pdf (accessed August 15, 2012).

[13]Heritage Foundation calculations using data from the 2007–2008 Schools and Staffing Survey. Based on a sample of 650 unionized full-time and part-time teachers in New York City public schools. See the appendix of this Backgrounder for details. The maximum tenure of unionized public-school teachers in New York in this sample was 41 years. Note: It is possible that a few teachers not sampled have worked longer than this, but the number of such veteran teachers is not statistically distinguishable from zero.

[14]Joel Klein, “The Failure of American Schools,” The Atlantic, June 2011, http://www.theatlantic.com/magazine/archive/2011/06/the-failure-of-american-schools/8497/4/?single_page=true (accessed August 15, 2012).

[15]Heritage Foundation calculations using data from the 2007–2008 Schools and Staffing Survey. Based on a sample of 420 unionized full-time and part-time teachers in Kansas Public Schools. See the appendix of this Backgrounder for details. Note: This means that no teachers in the representative sample have been working for more than 40 years. It is possible that some teachers not sampled have worked longer, but the number of such veteran teachers is not statistically distinguishable from zero.

[16]Robert J. Lalonde, Gerard Marschke, and Kenneth Troske, “Using Longitudinal Data on Establishments to Analyze the Effects of Union Organizing Campaigns in the United States,” Annales d' Economie et de Statistique, Vol. 41–42 (January–June 1996), pp. 155–185; Richard B. Freeman and Morris M. Kleiner, “The Impact of New Unionization on Wages and Working Conditions,” Journal of Labor Economics, Vol. 8, No. 1 (January 1990), pp. S8–25; and John DiNardo and David S. Lee, “Economic Impacts of New Unionization on Private Sector Employers: 1984–2001,” The Quarterly Journal of Economics, Vol. 119, No. 4 (November 2004), pp. 1383–1441.

[17]James Sherk, “What Unions Do: How Labor Unions Affect Jobs and the Economy,” Heritage Foundation Backgrounder No. 2275, May 21, 2009, http://www.heritage.org/research/reports/2009/05/what-unions-do-how-labor-unions-affect-jobs-and-the-economy.

[18]The Indiana Local of the International Union of Operating Engineers charges average dues of $2,000 to $2,500 a year. See “Union Seeks Restraining Order on Right-to-Work Law,” Associated Press, February 29, 2012, http://www.bsudailynews.com/union-seeks-restraining-order-on-right-to-work-law-1.2708444 (accessed August 15, 2012).

[19]James Sherk and Ryan O’Donnell, “RAISE Act Lifts Pay Cap on Millions of American Workers,” Heritage Foundation Backgrounder No. 2702, June 19, 2012, http://www.heritage.org/research/reports/2012/06/raise-act-lifts-pay-cap-on-millions-of-american-workers.

[20]James Sherk, “A Good Job Is Not So Hard to Find,” Heritage Foundation Center for Data AnalysisReport No. 08-04, June 17, 2008, http://www.heritage.org/research/Labor/cda08-04.cfm.

[21]Assar Lindbeck and Dennis Snower, “Centralized Bargaining and Reorganized Work: Are They Compatible?” European Economic Review, Vol. 45, No. 10 (December 2001), pp. 1851–1875.

[22]Henry S. Farber and Alan B. Krueger, “Union Membership in the United States: The Decline Continues,” NBER Working Paper No. 4216, 1992.

[23]Smith, “Union Disunity.”


[25]“Benchmark Study of Union Employee Election Year Attitudes,” The Word Doctors, October 2010, Question 41, http://www.theworddoctors.com/docs/Benchmark%20Study%20of%20Union%20Employee%20Election%20Year%20Attitudes.pdf (accessed August 15, 2012). Representative sample of 760 union members; margin of error + or – 3.7 percent. The Word Doctors is Frank Luntz’s polling and focus group firm.

[26]U.S. Department of Labor, Office of Labor-Management Standards (OLMS), “Form LM-2 Labor Organization Annual Report,” filed by the American Federation of Labor–Congress of Industrial Organizations, 2010, File No. 000-106, http://www.unionreports.gov (accessed August 15, 2012). These figures cover cash compensation. They do not include benefits or deferred compensation.

[27]U.S. Department of Labor, Office of Labor-Management Standards (OLMS), “Form LM-2 Labor Organization Annual Report,” filed by the Service Employees International Union, 2011, File No. 000-137, http://www.unionreports.gov (accessed August 15, 2012). These figures cover cash compensation. They do not include benefits or deferred compensation.

[28]U.S. Department of Labor, Office of Labor-Management Standards (OLMS), “Form LM-2 Labor Organization Annual Report,” filed by the United Food and Commercial Workers, 2011, File No. 000-056, http://www.unionreports.gov (accessed August 15, 2012). These figures cover cash compensation. They do not include benefits or deferred compensation.

[29]“Benchmark Study of Union Employee Election Year Attitudes,” The Word Doctors, Question 34.

[30]United Healthcare Workers-West, “The California Alliance Agreement: Lessons Learned in Moving Forward in Organizing California’s Nursing Home Industry,” January 4, 2007, http://media.sfweekly.com/789842.0.pdf (accessed August 15, 2012).

[31]Evelyn Larrubia, “Bay Area Health Union Seeks Vote on Membership,” Los Angeles Times, February 3, 2009, http://www.latimes.com/news/local/la-me-nuhw3-2009feb03,0,4633353.story (accessed August 15, 2012).

[32]National Labor Relations Board, “Basic Guide to the National Labor Relations Act,” 1997, pp. 8–10, http://www.shrm.org/Publications/hrmagazine/EditorialContent/Documents/basicguide.pdf (accessed August 15, 2012).

[33]Article XXXI, Section 24 of the constitution of the United Automobile, Aerospace, and Agricultural Implement Workers of America.

[34]Heritage Foundation calculations using data from the National Labor Relations Board, “Election Reports,” FY 2009 to FY 2011, http://www.nlrb.gov/election-reports (accessed August 15, 2012). During this time 214,000 workers voted in union certification elections.

[35]Ibid. Note: Since decertification elections occur only in workplaces where workers are sufficiently upset with their union to navigate the obstacles to calling an election, unions are more likely to lose these votes than they would in a randomly selected bargaining unit.

[36]John N. Friedman and Richard T. Holden, “The Rising Incumbent Reelection Rate: What’s Gerrymandering Got to Do With It?” The Journal of Politics, Vol. 71 (April 2009), pp. 593–611.

[37]See, for example, Claudio Ferraz and Frederico Finan, “Electoral Accountability and Corruption: Evidence from the Audits of Local Governments,” American Economic Review, Vol. 101, No. 4, June 2011, pp. 1274–1311.

[38]Douglas Belkin and Kris Maher, “Wisconsin Unions See Ranks Drop Ahead of Recall Vote,” The Wall Street Journal, May 30, 2012, http://online.wsj.com/article/SB10001424052702304821304577436462413999718.html (accessed August 15, 2012).

[39]Tom Tolan and Erin Richards, “Majority of Education Unions Vote to Recertify,” Journal Sentinel (Milwaukee), December 8, 2011, http://www.jsonline.com/news/education/education-unions-vote-to-recertify-sd3caui-135285458.html (accessed August 15, 2012).

[40]Steven Verburg, “Recall Aftermath: Public Unions Face Uncertain Future,” Wisconsin State Journal, June 10, 2012, http://m.host.madison.com/wsj/news/local/govt-and-politics/recall-aftermath-public-unions-face-uncertain-future/article_82fa0a7e-b25f-11e1-9c50-001a4bcf887a.html (accessed August 15, 2012).

[41]Poll of 3,021 adults conducted by CARAVAN Surveys between August 5 and August 15, 2011, “Poll of Union Members Regarding the Employee Rights Act,” http://www.employeerightsact.com/poll.html (accessed August 15, 2012).

[42]S. 1507, The Employee Rights Act, §2(b).

[43]Suzanne Lowe, “PERA Bargaining Unit Exclusion, S.B. 1018 (S-1): Analysis as Passed by the Senate,” Michigan State Senate Fiscal Agency Bill Analysis, March 28, 2012, http://www.legislature.mi.gov/documents/2011-2012/billanalysis/Senate/pdf/2011-SFA-1018-B.pdf (accessed August 15, 2012).

[44]The Mackinac Center for Public Policy, “The SEIU in Michigan: Home Health Aides,” July 7, 2011, http://www.youtube.com/watch?v=G4g9ILNUr_8 (accessed August 15, 2012).

[45]A further concern about union re-elections is their cost. The Employee Rights Act requires private employers and unions to split the cost of conducting union re-elections. A similar provision in such a bill applying to government unions would add slightly to strained government budgets. States could solve the problem of these higher costs by requiring government unions to pay for the cost of holding the elections. This is how internal union elections are paid for under the Labor Management Reporting and Disclosure Act (LMRDA). The LMRDA requires unions to conduct regular elections for union officers, and unions—not taxpayers—cover the cost of holding these elections. Government unions spend only a portion of their budgets representing their members. The rest they spend on other causes, such as lobbying. Unions can easily afford to pay for certification elections without negatively impacting the services they provide.

[46]Under this proposal, workers could also select non-union agents as their bargaining representative, i.e., a co-worker, lawyer, or a representative from a professional association.

[47]Bruce E. Kaufman, “What Unions Do: Insights from Economic Theory,” Journal of Labor Research, Vol. 25, No. 3 (July 2004), pp. 351–382.

[48]Jason Richwine and Andrew Biggs, “Are California Public Employees Overpaid?” Heritage Foundation Center for Data Analysis Report No. 11-01, March 17, 2011, http://www.heritage.org/research/reports/2011/03/are-california-public-employees-overpaid.

[49]Paul Takahashi, “Teachers New to District—Even the Outstanding Ones—Face Uncertain Future,” Las Vegas Sun, June 9, 2012, http://www.lasvegassun.com/news/2012/jun/09/teachers-new-district---even-outstanding-ones---fa/ (accessed August 15, 2012).

[50]This is especially true for those who want to teach for a few years, not their whole lives.

[51]Jason Richwine and Andrew G. Biggs, “Assessing the Compensation of Public-School Teachers,” Heritage Foundation Center for Data Analysis Report No. 11-03, November 1, 2011, http://www.heritage.org/research/reports/2011/10/assessing-the-compensation-of-public-school-teachers (accessed August 15, 2012).

[52]Walker did not in fact prevent state employees from joining unions. He only removed those unions’ monopoly control over the state of Wisconsin’s labor force.

[53]Pope Leo XIII, Rerum Novarum, 1891, §51, http://www.vatican.va/holy_father/leo_xiii/encyclicals/documents/hf_l-xiii_enc_15051891_rerum-novarum_en.html (accessed August 15, 2012).

[54]For example, Pope Pius XI argued that “just as inhabitants of a town are wont to found associations with the widest diversity of purposes, which each is quite free to join or not, so those engaged in the same industry or profession will combine with one another into associations equally free for purposes connected in some manner with the pursuit of the calling itself.” (Emphasis added.) Quadragesimo Anno,1931, §87, http://www.vatican.va/holy_father/pius_xi/encyclicals/documents/hf_p-xi_enc_19310515_quadragesimo-anno_en.html (accessed August 15, 2012).

[55]Recent examples include Ohio (though the law was blocked by a referendum), Oklahoma, Tennessee, and Wisconsin. Other states like Idaho, Nevada, and South Dakota have passed education reform laws restricting the ability of unions to negotiate provisions like tenure in their contracts. Other states like Michigan and New Jersey have required government employees to pay more for health benefits.

[56]Evergreen Freedom Foundation, Sweeping the Shop Floor: A New Labor Model for America,” 2010, “Appendix B: Public Employee Choice in Representation Act,” http://www.myfreedomfoundation.com/files/pdf/SWEEPINGTHESHOPFLOOR.pdf (accessed August 15, 2012).

[57]U.S. Department of Education National Center for Education Statistics, “Schools and Staffing Survey,” 2007–2008, http://nces.ed.gov/surveys/sass/ (accessed August 15, 2012).

[58]Federal guidelines require researchers who publish data generated from the Schools and Staffing Survey to round their sample sizes to the nearest 10. These sample sizes have been rounded accordingly.

[59] National Directory of Nonprofit Organizations, ed. by Bohdan Romaniuk, 27th ed. (Detroit: Gale, 2012).

[60]Interview on August 8, 2012.

[61]Telephone interviews conducted June 6–8, 2012.

[62] National Directory of Nonprofit Organizations, ed. by Bohdan Romaniuk.

[63]New Hampshire Public Employee Labor Relations Board, “Board Decisions,” 1976–2012, http://www.nh.gov/pelrb/decisions/board/index.htm (accessed August 15, 2012).

[64]United Federation of Teachers, “50 Years.”

[65]E-mail correspondence with the state Department of Labor, June 29, 2012.

[66] Rapid City Educ. Ass’n v. Rapid City Area School Dist. No. 51-4, 376 N.W.2d 562 (1985).


James Sherk
James Sherk

Research Fellow, Labor Economics