Unions, economic freedom, and growth.

Holcombe, Randall G. ; Gwartney, James D.

The freedom to enter into contracts and to direct the use of economic resources one owns are essential to the operation of a market economy. Allowing employees to form unions to bargain collectively over wages and employment conditions is consistent with economic freedom, and any government intervention preventing unionization would be a violation of economic freedom. Nevertheless, American labor law, especially since the 1930s, has 'altered the terms mad conditions under which unions collectively bargain to heavily favor unions over the finns that hire union labor. Labor law has given unions the power to dictate to employees collective bargaining conditions, and has deprived employees of the right to bargain for themselves regarding their conditions of employment. While unions and economic freedom are conceptually compatible, labor law in the United States, and throughout the world, has restricted the freedom of contract between employees and employers.

The effect of unions on growth and prosperity can be examined at two levels. Narrowly, one can examine the effects that union contracts have had on unionized firms and industries. More broadly, one can look at the way that unions have affected labor law. Unions have successfully lobbied to increase the power of unions over firms, which in turn has allowed unions to impose more constraining conditions on employers. Union contracts likely would not contain some of the provisions they do were it not for the bargaining power labor law gives unions relative to the employers of union labor. Unions have 'also 'affected labor law by lobbying for conditions under which nonunion labor can be employed. Two notable examples are the minimum wage law and the Davis-Bacon Act, which require the federal government to pay the local prevailing wage rate--that is, the union wage rate--on all contracts.

If one takes the narrow view of simply evaluating the effects of union workers and union contracts on growth mad prosperity, the effects in the United States will be small and concentrated in a few industries. If one takes the broader view of examining the effects of union-promoted labor law that affects both union and nonunion workers, the effects will be larger.

The most visible effects of unionization in the U.S. economy are, first, the migration of the workforce away from unionized industries and professions toward nonunion employment, and second, the decline in those unionized firms and industries that have been unable to escape their unions. While market forces appear to generate a movement of labor away from unionized firms, those market forces are absent in public sector employment, so while unionization in the private sector has declined, it remains strong in the public sector. Looking ahead, perhaps the largest ramifications of unionized employment in the United States will be felt in the public sector, where unions have already imposed substantial costs on governments in the form of unfunded pension and retirement liabilities.

Economic Freedom and Prosperity

Freedom is something to be valued on its own merits, which provides a fundamental reason to question labor laws that impinge on workers' freedom of contract. Economic freedom also generates prosperity, so labor laws that reduce economic freedom also have a detrimental effect on a nation's standard of living. The evidence on the relationship between economic freedom and prosperity is substantial and persuasive. Mokyr (1990) and 'Landes (1998) provide excellent historical accounts showing that since the beginning of the industrial revolution, nations that have adopted the institutions of economic freedom have grown and prospered, while those that have not have languished in poverty.

Gwartney and Lawson (2005) have produced a quantitative measure of economic freedom in their Economic Freedom of the World (EFW) index, which is updated annually. While the idea that free markets and prosperity are linked has a long history, one might question what, exactly, constitutes a free market, and how one might evaluate whether some countries are more economically free than others. The EFW index was designed to address this question, providing a quantitative measure of economic freedom at a national level for 127 countries. The EFW index is composed of 38 different measures of economic freedom, aggregated into four major areas, which are then combined to get a single numerical index for each country. The four major areas are: (1) Size of Government: Expenditures, Taxes, and Enterprises; (2) Legal Structure and Security of Property Rights; (3) Freedom to Trade Internationally; and (4) Regulation of Credit, Labor, and Business.

Hundreds of studies have been undertaken using the EFW index as a measure of the degree to which the existence of market institutions affects outcomes in various dimensions. Berggren (2003) surveys a number of these studies, and Gwartney and Lawson (2005) outline additional results. Countries that have more economic freedom, as measured by the EFW index, have on average higher per capita incomes, and countries that increase their economic freedom exhibit higher rates of economic growth. In addition, countries with more economic freedom have lower unemployment rates, lower percentages of children in the labor force, higher life expectancies, and more political freedom.

Labor market regulations, including the laws outlining the rights and obligations of unions, are a component of economic freedom, but one can see from the EFW index that many other factors come into play. Thus, one would be hard-pressed to find a simple correlation between union activity, labor law, and prosperity. Regulation of labor markets is a small component of economic freedom, and while its effects on prosperity are important, they will be small compared with other aspects of economic freedom. In addition, there is not necessarily a close correlation between labor market regulations and unionization. As discussed in more detail below, some countries with low rates of unionization have very restrictive labor laws. Unionization tells only a small part of the story when looking at labor market restrictions on economic freedom.

Unionization itself does not constitute a reduction ha economic freedom, nor does it necessarily reduce prosperity. That depends on the labor law that governs the rights of workers and the power of unions. Unionization of labor is completely consistent with economic freedom. People have property rights to their labor, and a right to control under what conditions they exchange their labor for income. The right to freedom of contract means that individuals have the right to join with other individuals to bargain collectively over their terms of employment. Similarly, freedom of contract also implies that employers have no obligation to contract with a union, or to bargain collectively if they choose not to. However, labor law has eroded those freedoms rather than supported them, so in fact, unionization has compromised economic freedom.

Labor Law in the United States

Throughout the 20th century, labor law ha the United States has reduced the ability of individuals to contract for the terms of their employment, and has reduced the ability of employers to contract with the individuals they employ. Labor law has not only solidified the rights of unions to bargain collectively for their employees; it has compelled employees to be party to collective bargaining whether they want to or not. Meanwhile, employers have no freedom to avoid entering into collective bargaining with unions. They are required to bargain "in good faith," which essentially means arriving at an outcome satisfactory to union leaders. Reynolds (1987: chap. 2) gives a good summary of the changes in U.S. labor law over the 20th century.

The first significant piece of union legislation in the 20th century was the 1914 Clayton Act. It gave unions an exemption from the provisions of the 1890 Sherman Antitrust Act, declared acts such as picketing by unions to be lawful, and limited the use of court injunctions in labor disputes. The biggest changes in labor law, though, took place in the 1930s. The Davis-Bacon Act of 1931 required government-financed construction projects to pay local prevailing wages for labor, which essentially meant union wage rates. The Davis-Bacon Act eliminated the ability of nonunion labor to compete by offering to work for less. The 1932 Norris-LaGuardia Act made nonunion employment agreements unenforceable in federal courts, sheltered unions from liability from wrongful actions under antitrust law, and gave unions immunity from damages in private lawsuits. The 1935 Wagner Act, the most significant piece of union legislation, specified as unfair a number of labor practices that businesses had used to resist unions, and created the federally appointed National Labor Relations Board to enforce the Act. The creation of the NLRB allowed unions to avoid relying on court decisions, which were more likely to be anti-union than those of a politically appointed board. The NLRB enforced union elections, decided who had the right to vote in union elections, and enforced monopoly bargaining for all employees in a bargaining unit. The NLRB also enforced union pay rates for all employees represented in the unit, regardless of whether they were union members. This provision prevents nonunion workers from competing for jobs by undercutting the collectively bargained union wages.

The Fair Labor Standards Act of 1938 set a federal minimum wage for many nonagricultural workers. Originally 25 cents an hour, the federal minimum wage is now $7.25 an hour, and many states mandate minimum wages higher than the federal minimum. While the minimum wage law does not directly address union workers, it raises the cost of nonunion labor, providing a benefit to unions by limiting nonunion price competition.

As a reaction to the Wagner Act, the Taft-Hartley Act was passed in 1947, which created a list of unfair labor practices to balance the Wagner Act's list of unfair practices for employers. One of the provisions of Taft-Hartley was to restrict union shops--where all employees are required to be union members--to states that did not pass right-to-work laws. A total of 22 states have passed right-to-work laws, meaning that in those right-to-work states employees do not have to join a union, even if the union undertakes collective bargaining for the employees as a requirement of the Wagner Act. The Landrum-Griffin Act of 1959 enacted regulations on the internal affairs of labor unions and established rides governing the relationships of unions with both employers and union members.

The rhetoric in labor law tends to be couched in terms of workers' rights, but review of U.S. labor law shows that in fact workers have lost a substantial amount of their freedom to contract for the terms of their employment, because labor law has given union leaders the right to dictate conditions of employment. Of course, unions must reach an agreement with employers, but here too employers have lost the ability to negotiate with employees, being required by law to negotiate "in good faith" with the union leadership. Also, employers are covered by antitrust law, making employment agreements among employers illegal, while unions are exempt from antitrust law, so there is no limit to the scope of their bargaining power. For example, the United Auto Workers is the bargaining agent for all unionized auto workers at Ford, General Motors, and Chrysler, while those employers must bargain individually with the union that controls the employees for 'all three.

Baird (1984, chap. 3) gives a list of individual freedoms that have been compromised by U.S. labor law. It gives unions the right to be "exclusive representatives of all the employees" in a bargaining unit, taking away individuals' freedom of contract; makes it illegal for employers to refuse to bargain with unions, taking away employers' freedom of contract; and requires employers to bargain "in good faith," which has been interpreted to mean arriving at an outcome that is satisfactory to the union.

The developments in 20th century U.S. labor law clearly show the erosion of individual rights and economic freedom, especially through the middle of the century. The 1947 Taft-Hartley Act and 1959 Landrum-Griffin Act did give some control of employment conditions back to workers, but the restrictions on freedom of contract imposed by labor law clearly reduce economic freedom for employees whose conditions of employment are governed by a union. There can be no doubt that the result of 20th century labor law was to give economic power to union leaders while reducing the economic freedom of both employers and employees.

Labor Law in Other Countries

Other developed economies have seen even more bias in favor of collective bargaining than the United States. Sometimes this has come in the form of legal powers given to unions, but in other cases labor law covers all workers in a nation, regardless of their union status. Botero et al. (2004) note that countries with left-leaning politics tend to have more stringent labor market regulations, and that countries with French, Scandinavian, and socialist legal origins have higher levels of labor regulation than common law countries.

Labor market regulation abridges freedom of contract, so national laws erode economic freedom for employees and employers. In countries with strong government labor regulations, union contracts have less scope for influence. Botero et al. (2004) find that countries with more regulated labor markets have higher unemployment and lower rates of labor force participation, so the reduction in economic freedom is associated with a reduction in employment. Unions influence labor law, beyond a doubt, but unionization is less important precisely because of the restrictive government labor laws that apply to all workers regardless of their union status.

France is a good example of a nation with substantial restrictions on economic freedom imposed by national labor laws. Siebert (1997) noted that the French minimum wage was at about 60 percent of the median wage in 1997, compared with 34 percent ha the United States. France attempted to lower the minimum wage for workers under 25 in 1993, Saint-Paul (1996: 297) reports. However, that attempt was extensively protested, largely by people paid more than the minimum wage who viewed young workers as threats to their employment, and the reform was withdrawn. The political power of some reduced the economic freedom of others. France also has high unemployment benefits, and labor law requires substantial severance pay to employees who are terminated. France also mandated a 35-hour work week for many workers in 2000. The mandate was repealed in 2005, but in exchange for higher pay for extra hours of work.

Faced with labor laws that made it very costly to terminate employees, employers attempted to avoid those costs by hiring workers on "determined duration contracts" (DDCs), which specify a temporary term of employment. This creates a two-tier labor market where some workers have permanent jobs with substantial protections while others are in DDCs. However, Saint-Paul (1996) reports that France has also limited the ability of employers to hire on DDCs, restricting them only to work that is temporary in nature, and that more generally, European governments have tried to restrict two-tier labor markets. France restricted DDCs to a maximum of 12 months of employment in 1989 and limited the ability to use DCCs for newly created jobs.

Saint-Paul (1996) concludes that labor market laws in Europe are designed to preserve the interests of those who are employed, reducing labor market flexibility and increasing unemployment. Labor market regulations that make it more costly to terminate employees make it more costly to hire them in the first place. When there are high termination costs imposed on the employer, making a hiring mistake can be very costly, so employers are more reluctant to hire, which reduces total employment. However, if workers can be terminated anytime at the discretion of the employer, the employer can take a chance on hiring someone if it might pay off.

Siebert (1997) notes stringent labor market regulations in other European countries as well. Italy passed regulations on firing procedures in 1966, and following strikes made firing costs prohibitive by 1970. Regulations throughout the 1970s and 1980s did loosen firing regulations and make it easier to hire DCCs, although restrictive labor practices remain in place. Germany passed a host of regulations in the 1970s. Mandatory social plans were required in 1972 to close a firm, unemployment benefits were raised, and in 1976 a codetermination law was passed requiring that half the members of the supervisory boards of large firms had to be labor representatives. By the 1970s most European countries required dismissals to be approved by work councils, which considered factors like marital status, the number of children, and the worker's health in deciding whether the dismissal would be allowed. The stated goal was to make employment more secure, but as Siebert notes, the actual effect was to make it more costly to hire workers, resulting in increased unemployment. Nickell (1997) finds that high unionization, with collective bargaining for wages and no coordination among employers leads to higher unemployment when looking at a cross-section of developed economies, including those in Europe and North America.

Labor market restrictions are not limited to mature economies. Besley and Burgess (2004) present evidence from India that when labor law gives workers more collective bargaining power relative to employers, output, employment, and productivity tend to fall. As de Soto (1990) notes, when regulatory restrictions make employment costly in less-developed economies, people find employment in the underground economy. However, this type of work places people outside many of the protections of the legal system and the tax system, and makes long-term contracts difficult to undertake, resulting in a negative effect on productivity.

The literature shows that, as Gwartney and Lawson (2005) note, labor market restrictions not only reduce the economic freedom of employers and employees but also result in higher unemployment and slower economic growth. Labor market restrictions are not synonymous with union activity, however, and government-imposed labor law has even more potential than union contracts to restrict economic freedom.

To help illustrate the tenuous relationship between labor market restrictions and union activity, Table 1 shows the level of private sector union density in a number of developed economies for 2005. Union density, a common measure of the strength of unions, is defined as the proportion of the labor force that is unionized, including union coverage of workers not belonging to a union. Union density thus includes workers who are in a collective bargaining unit but who are not union members. Comparing across countries, the United States has a relatively low private sector union density, but some countries with very restrictive labor laws 'also have low union density.

The 12 percent union density in the United States is at the low end of the range. France, at 8 percent, is the lowest, despite very restrictive labor laws. Korea's union density is about 10 percent, but most countries have union densities much greater than the United States. Sweden tops out Table 1 with a union density of 76.5 percent, and Finland's union density exceeds 72 percent. One can see that comparing union density across countries, union penetration in the U.S. private sector is relatively low by international standards.

Many more workers are covered by collective bargaining agreements than just unionized workers. In some countries--France again provides an example--wages for all workers in many professions are set by law nationwide, so even nonunionized workers employed in nonunionized firms are subject to the wage and other labor agreements determined by collective bargaining. The impact of collective bargaining then extends much further than the unionized workforce. Table 2 shows the union density figures from Table 1 for selected countries along with the share of the labor force that is subject to collective bargaining. Countries are listed in order from lowest to highest union density, and one can see that there is little correlation between union density and the percentage of workers covered by collective bargaining agreements. France, with the lowest union density, has the highest share of workers covered by collective bargaining. Table 2 illustrates the difference between unionization and the extent to which labor law influences wages and working conditions. Ultimately, a country's labor law will have more influence than the extent of unionization.

In addition to the impact of collective bargaining and labor law on wages, labor law also influences the cost of hiring and firing workers, which in turn 'affects their employment prospects. Mandated benefits for employees obviously make hiring more costly, but the cost of firing employees can 'also be substantial, which also makes hiring more costly. If termination costs are low, employers will find it more worthwhile to take a chance on an employee, because if the employee does not work out, he or she can be let go. If there are substantial severance costs, hiring employees is more risky and employers will be more reluctant to hire. The hiring and firing costs mandated by governments are aspects of labor law that introduce frictions into the labor market, increasing inefficiencies, reducing job mobility, and increasing unemployment.

Table 3 shows the cost of hiring and firing workers as a percentage of worker pay for the same countries listed in Table 2. The figures show that these costs vary substantially. The United States is at the low end for both hiring and firing costs, with hiring costs being 8.5 percent of wages and firing costs at zero. At the other end of the scale France has the highest hiring costs, which are 47.4 percent of wages, while Germany has the highest dismissal cost at 66.7 percent.

The countries in Table 3 are divided into three groups: those with relatively low hiring and firing costs, those in the middle, and those with high costs. The column HC+DC adds together the hiring and dismissal costs, which gives an indicator of the costs of taking on an employee over and above wages paid. This indicator is approximate, because hiring and firing costs may have different effects. The hiring costs apply to 'all hires, whereas the dismissal costs apply only in the event the employer wants to dismiss an employee, so different firms in different circumstances may weight these differently. So, while HC+DC is clearly a back-of-the-envelope calculation, it is useful in that it measures real costs of employing workers in addition to the wages they are paid.

Japan, Australia, and the United States are substantially lower in HC+DC than the other countries, with Japan the highest of the three at 33.9 percent of worker pay. The United Kingdom mad Canada fall in the middle, at around 40 percent. Italy, France, and Germany are all around 80 percent or more. The countries are listed in Table 3 in order of their 2007 unemployment rates, and the countries with the low hiring and dismissal costs all have 2007 unemployment rates lower than those in the middle group, which in turn all have unemployment rates lower than those in the high group. The rightmost column shows average unemployment rates for 1999-2006, which tells 'almost the same story. Tiffs provides some evidence of the impact that labor law has on one easily measurable indicator of labor market efficiency: the unemployment rate.

The point of looking at international labor law is to illustrate that there is much more to labor market restrictions than ration activity, and to illustrate that there is not a close correlation between unionization and labor market restrictiveness. Table 3 provides some evidence on the effect of labor market restrictiveness on unemployment. Labor law restricts economic freedom when it reduces the ability of employees and employers to bargain over the terms of employment. Labor law compromises economic freedom regardless of the union status of an employee.

An Example of the Productivity Effects of Unions: Right-to-Work Laws

The effects of unions on productivity can be seen by examining right-to-work laws. In the United States, the 1947 Taft-Hartley Act gave states the right to pass right-to-work laws, which outlaw the requirement that workers join and financially support a union as a condition of employment. As such, right-to-work laws enhance individual freedom. Moore (1998) reviews a substantial empirical literature showing that right-to-work laws do affect productivity and unionization on a number of dimensions.

Right-to-work laws support economic freedom because they ensure that workers are not coerced into joining a union as a prerequisite for taking a particular job. They do not 'allow a worker to bargain independently of the ration with firms that are covered by union contracts, however. While it may be that right-to-work laws allow those who do not join to free-ride off the collective bargaining provided by the union, as Sobel (1995) notes, laws that force individuals to join a union as a condition of employment clearly compromise those individuals' freedom of contract. The freedom-enhancing solution to the free-rider problem would be to exempt nonunion workers from being bound by the union contract, not forcing them to join the union.

Businesses like right-to-work laws because they make it more difficult for unions to organize their employees, and unions dislike right-to-work laws for the same reason. Thus, the debate tends to be centered on pro- versus anti-union lines, not on the issue of economic freedom. Right-to-work laws also appear to help economic development, as Palomba and Palomba (1971) and Moore and Thomas (1974) note, which can factor into the debate. Calzonetti and Walker (1991) present survey data showing that firms do consider right-to-work laws in their location decisions.

Holmes (1998) presents some interesting evidence on the effect of right-to-work laws on manufacturing activity. He measures manufacturing activity along the borders of states with and without right-to-work laws, comparing manufacturing activity in border counties in right-to-work states with counties directly adjacent in non-right-to-work states. Holmes finds that manufacturing employment is about one-third higher in border counties of right-to-work states than in the adjacent counties in non-right-to-work states. A number of other studies, such as Newman (1983) and Woodward mad Glickman (1991), support the conclusion that states without right-to-work laws have lower manufacturing employment and employment growth. Moore's (1998) survey cites many studies showing the positive effect of right-to-work laws on employment and business activity, but no studies with the opposite results.

Right-to-work laws make a good case study on unions, economic freedom, and growth, because right-to-work laws clearly preserve economic freedom by preserving employees' rights of contract, and decades of empirical research in economics shows that the absence of right-to-work laws hinders economic development.

The Decline in Private Sector Unionization

One bit of evidence on the effect of unionization on productivity is the decline in private sector unionization in the United States over the past half century. By 1945, partly due to the strength of pro-union legislation in the 1930s, 22.4 percent of the civilian labor force was unionized. Table 4 gives figures for union density in the United States for various years. It shows that private sector union density has declined from more than 30 percent of the labor force in 1960 to less than 12 percent by 2007.

There is a substantial variation in union density among industries, with transportation and warehousing, utilities, and telecommunications having union densities above 20 percent while finance, business, and professional services have union densities below 3 percent. Table 5 gives some figures on union densities across industries. With the problems facing the auto industry, the effects of unionization in manufacturing have been highlighted, but manufacturing overall has a union density of about 12 percent, close to the U.S. average for the private sector workforce. Table 5 shows that even in the most unionized industries, under a quarter of the labor force is covered by union contracts, with the exception of utilities. Utilities are a special case in at least one respect, however, because they tend to be regulated monopolies and therefore sheltered to a degree from the effects of competition that firms in other industries face. Taken together, Tables 4 and 5 show that union labor is a small and declining part of the private sector U.S. labor force.

Meanwhile, union density among public sector employees is about 40 percent, and has been relatively stable for several decades. Table 6 shows that union density among government workers rises for lower levels of government. Union density is 33 percent at the federal level, 35 percent at the state level, and 46 percent at the local level. While private sector unionization in the United States is relatively small and declining, public sector unionization remains much stronger. Like the utilities industry mentioned in the previous paragraph, governments are shielded from competition. High union density remains in areas where competition is limited, and it appears that market competition erodes the penetration of union labor. As the global economy has become more competitive, unionization of the U.S. workforce has fallen.

When considering the decline in private sector unionization in the United States, one must recognize that the higher labor costs created by union rules and contracts price union workers out of markets. When industries cannot escape union labor, high costs eventually eliminate those industries. The U.S. railroad industry has seen significant decline over the past half-century. Passenger rail has been taken over by the federal government with the creation of Amtrak and freight shipments have shifted toward trucks. Reynolds (1987: 111) states that without union work rules in the railroad industry the number of rail workers could be cut approximately in half. Work rules have been modified since Reynolds wrote, but by then private passenger service had already disappeared and freight traffic was already in substantial decline. Reynolds (1987: 118) notes that from 1959 to 1985 railroad employment fell from 900,000 to 301,000, showing that the union was not effective at saving railworkers' jobs. Nobel Laureate Paul A. Samuelson (quoted in Reynolds 1987: 147) has said, "The whole history of unionism has been ... in determining how industries in decline are accelerated toward their extinction."

The auto industry provides a good ease study to illustrate Samuelson's point. One would be hard-pressed to blame the 2009 bankruptcies of Chrysler and GM solely on the UAW. Nevertheless, high labor costs mad reduced labor flexibility imposed by the UAW on the auto companies dearly raised their costs and limited their ability to innovate to keep up with a changing market.

The UAW's hold on the auto industry began with what Asher et al. (2001:161) called the union's greatest victory: the unionization of the GM workforce in 1937. Pushed by the UAW, GM began paying some employee health insurance benefits in 1950, and by 1973 was paying all costs for employee health insurance, pensions, and all health insurance for retirees and their survivors. GM also agreed to a "30 and out" retirement program that provided full pensions to employees after 30 years of employment, regardless of the employee's age. When these benefit packages were agreed to they were much less costly, but as costs (especially health care costs) have risen, the UAW has stood firm in insisting that those benefits remain. Other costly programs included the Jobs Bank Program, which paid laid-off workers 95 percent of their previous wage once they had exhausted unemployment benefits, regardless of whether they worked, and job classification rules that prevented workers in one classification from doing work outside that classification, even when workers were idle.

As long as America's "big three" controlled the auto industry as an oligopoly, the UAW contracts that hampered 'all three companies did not put one at a competitive disadvantage over another. The "big three's" hold on the industry began to crumble in the 1970s when higher gasoline prices tilted the market toward more fuel-efficient imports. In addition, the impression of poor quality control at the "big three" contrasted with an increasing reputation for quality in Japanese autos, further eroding the "big three's" hold on the market. UAW membership exceeded 1.5 million in 1979, but had fallen to 465,000 in 2007. Now, a number of foreign automobile manufacturers produce cars in the United States, including BMW, Honda, Mercedes Benz, Nissan, Toyota, and Volkswagen. They have located mostly in Southern states where labor costs are lower, and perhaps as significant, in right-to-work states where auto workers are not unionized.

Despite many factors working against the "big three," in response to increasing competition for U.S. auto sales and jobs, the UAW held firm in fighting for conditions of employment, and continued to strike when the terms of employment were not met. From 1996 to 1998 the UAW had seven local strikes against GM, and its most recent strike in 2007 pulled workers off the job in 82 facilities across the nation. Even as the industry was collapsing, the UAW was unwilling to offer concessions to the auto companies. Only when the companies declared bankruptcy in 2009 did the UAW offer concessions, which it would have had to have made anyway in a bankruptcy proceeding. The UAW provides an excellent case study illustrating Samuelson's observation that unions accelerate declining industries toward their extinction.

However, this is only part of the story. Research indicates that unions push up wages and reduce the profits of unionized firms (Hirsch 2004). But this is a two-edged sword. While the unionized workers will enjoy higher wages in the short run, the higher costs and lower profits will make it more difficult for the unionized firms to compete effectively. Capital can be exploited in the short run, but dais will not be the case in the long run. Therefore, to the extent that the profits of unionized firms are lower, investment expenditures on fixed structures, research, and development will flow into the nonunion sector and away from unionized firms. As a result, the growth of both productivity and employment will tend to lag in the unionized sector. Thus, unions will not only hasten the demise of declining industries as Samuelson states, they will also hasten the shift of unionized firms mad industries from expansion toward decline.

The primary impact of private sector unionism in the United States has been to shift output and employment away from heavily unionized firms and industries toward the nonunion and less heavily unionized sectors. These shifts have been gradual, but they have reduced both the size and impact of unions on the private sector of the U.S. economy. However, unionization remains strong in the public sector, which is sheltered from competition because it is financed by compulsory taxation.

The Effect of Unions on U.S. Productivity and Growth

Vedder and Gallaway (1993: 141) estimate that in 1939 the unemployment rate was more than 6 percent higher than it would have been in the absence of the growth of unionized labor in the 1930s. They are looking at a period in which labor law greatly expanded the power of unions, and before mitigating legislation was passed in the 1940s and 1950s. Reynolds (1987: 61) reviews the evidence and concludes that there is "no obvious association between the degree of unionism and aggregate productivity growth in the historical data," and reviewing the literature finds that a consensus is that unionization may be responsible for a 0.33 percent reduction in aggregate income. He notes that these estimates do not take account of the effects of unionization on investment and entrepreneurship, where unions may have a larger effect. The estimates show a small impact of unionization, and private sector union density has fallen substantially since he drew those conclusions.

In the United States, where private sector union density is below 12 percent and falling, the direct effects of unions on productivity and growth fall into two primary areas: the impact of unions in retarding productivity in heavily unionized industries, pushing those industries into decline; and in the public sector, where union density remains higher and stable, where there are few alternatives to those public sector services, and where they can be paid for by compulsory taxation. The auto and railroad industries provide two examples where union contracts substantially raised the costs of unionized firms, resulting in their declines and shifting employment toward nonunionized industries. In the private sector, market forces are reducing the impact of unions on productivity, hut public sector unionization is more problematic.

One of the looming issues regarding public sector union contracts is retirement benefits, which threaten to overwhelm governments, especially at the state and local level. The two problems are the ability of public sector employees to retire young coupled with very generous retirement benefits. A key factor is that because public sector employees are funded by compulsory tax payments, governments do not face the same market discipline as private sector firms, and the tax burden that will be applied to finance generous pension benefits will impose a cost on private sector productivity. Chrysler and GM were able to renegotiate generous benefits to union retirees in bankruptcy. That burden will be more difficult for governments to overcome. Looking ahead, the biggest impact unions will have on prosperity will come from public sector unions, not those in the private sector.

Conclusion

While the right of workers to unionize and bargain collectively is completely consistent with freedom of contract and individual rights, 20th century labor law has created an environment in which unions have the power to compromise the freedom of contract by compelling workers to bargain collectively, in some eases to compel them to join a union and pay union dues, and to compel employers to negotiate with unions for labor contracts even when individuals may prefer to bargain themselves, independent of other workers. The concept of collective bargaining is consistent with economic freedom, but the developments of 20th century labor law have compromised economic freedom, and the powers given to unions have limited the rights of workers and employers.

Unions have consistently bargained for higher wages and other benefits for their employees, and in the short run, because labor law has given to unions an advantage in the bargaining process, union contracts have had the effect of increasing the wages and benefits of union workers. In the long run, the higher cost of union labor brought on by those union contracts has resulted in a steady decline in private sector unionism, and has eroded U.S. manufacturing in unionized industries--most visibly, the railroad and auto industries.

With private sector union density in the United States at about 12 percent, the overall effect of unionization on economic growth is not substantial. In a few industries such as the railroad industry and the auto industry, it has been devastating. Those two industries illustrate the larger effect of unionization, which has been the shift of employment away from unionized firms.

While private sector union density is relatively low and declining, public sector union density is higher and stable. Local government employees have a union density of 46 percent, and many of the same factors that applied to the UAW's effect on the auto industry also apply to local public sector employees. Benefits are very generous, imposing a less visible future cost that will have to be borne by taxpayers unless those benefits are restructured. The effect of unions on overall economic growth in the United States has been minor, because market forces have shifted private sector employment from unionized toward nonunionized industries, but international comparisons show that more restrictive labor law does place a measurable burden on the economy.

In the future, the largest impact of unionization in the United States will come from public sector unionization. The burden of generous retirement benefits will crowd out other government expenditures, will be a force for higher taxes, and will impose an increasing burden on the private sector of the economy that pays those taxes.

References

Asher, H. B.; Heberlig, E. S.; Ripley, R. B.; and Snyder, K. (2001) American Labor Unions in the Electoral Arena. Lanham, Md: Rowman and Littlefield.

Baird, C. W. (1984) Opportunity or Privilege: Labor Legislation in America. Bowling Green, Ky: Social Philosophy and Policy Center.

Berggren, N. (2003) "The Benefits of Economic Freedom: A Survey." Independent Review 8 (2): 193-211.

Besley, T., and Burgess, R. (2004) "Can Labor Regulation Hinder Economic Performance? Evidence from India." Quarterly Journal of Economics 119 (1): 91-134.

Botero, J. C.; Djankov, S.; La Porta, R.; Lopez-de-Silanes, F.; and Shleifer, A. (2004) "The Regulation of Labor." Quarterly Journal of Economics 119 (4): 1339-82.

Calzonetti, F. J., and Walker, R. T. (1991) "Factors Affecting Industrial Location Decisions: A Survey Approach." In H. W. Herzog, Jr., and A. M. Schlottman (eds.) Industry Location and Public Policy, 221-40. Knoxville: University of Tennessee Press.

de Soto, H. (1990) The Other Path: The Invisible Revolution in the Third World. New York: Harper and Row.

Gwartney, J., and Lawson, R. (2005) Economic Freedom of the World: 2005 Annual Report. Vancouver, B.C.: Fraser Institute.

Hirsch, B. T. (2004) "What Do Unions Do for Economic Performance?" Journal of Labor Research 25 (3): 415--55.

Holmes, T. J. (1998) "The Effect of State Policies on the Location of Manufacturing: Evidence from State Borders." Journal of Political Economy 106 (4): 667-705.

Landes, D. S. (1998) The Wealth and Poverty of Nations: Why Some Are So Rich, and Some So Poor. New York: W. W. Norton.

Mokyr, J. (1990) The Lever of Riches. Oxford: Oxford University Press.

Moore, W. J. (1998) "The Effects of Right-to-Work Laws: A Review of the Recent Literature." Journal of Labor Research 19 (3): 445-69.

Moore, W. J., and Thomas, R. W. (1974) "Determinants of the Passage of Right-to-Work Laws: An Alternative Explanation." Journal of Law and Economics 17 (1): 197-211.

Newman, R. J. (1983) "Industry Migration and Growth in the South." Review of Economics and Statistics 65: 76-86.

Nickell, S. (1997) "Unemployment and Labor Market Rigidities: Europe versus North America." Journal of Economic Perspectives 11 (3): 55-74.

Palomba, N. A., and Palomba, C. (1971) "Right-to-Work Laws: A Suggested Economic Rationale." Journal of Law and Economics 14 (2): 475-83.

Reynolds, M. (1987) Making America Poorer: The Cost of Labor Law. Washington: Cato Institute.

Saint-Paul, G. (1996) "Exploring the Political Economy of Labour Market Institutions." Economic Policy 11 (23): 265-300.

Siebert, H. (1997) "Labor Market Rigidities: At the Root of Unemployment in Europe." Journal of Economic Perspectives 11 (3): 37-54.

Sobel, R. S. (1995) "Empirical Evidence on the Union Free-Rider Problem: Do Right-to-Work Laws Matter?" Journal of Labor Research 16: 347-65.

Vedder, R., and Gallaway, L. (1993) Chit of Work: Unemployment and Government in Twentieth-Century America. New York: Holmes and Meier.

Visser, J. (2006) "Union Membership Statistics in 24 Countries." Monthly Labor Review 129: 38-49.

Woodward, D. P., and Glickman, N. J. (1991) "Regional and Local Determinants of Foreign Finn Location in the United States." In H. W. Herzog, Jr., and A. M. Schlottman (eds.) Industry Location and Public Policy, 190-217. Knoxville: University of Tennessee Press.

Randall G. Holcombe is DeVoe Moore Professor of Economies at Florida State University, and James D. Gwartney is Gus A. Stavros Eminent Scholar at Florida State University.

TABLE 1
PRIVATE SECTOR UNION DENSITY
IN SELECTED COUNTRIES, 2005

Country          Union Density (%)

Australia                    22.1
Canada                       29.9
Finland                      72.4
France                        8.0
Germany                      21.6
Italy                        33.8
Japan                        18.8
Korea                         9.9
Netherlands                  21.0
Sweden                       76.5
United Kingdom               28.8
United States                12.0

SOURCE: OECD (www.oecd.org/dataoeed/25/42/39891561.xls).

TABLE 2
UNION DENSITY AND SHARE OF EMPLOYEES WHOSE WAGES
ARE SET BY COLLECTIVE BARGAINING

Country          Union Density   Collective Bargaining
                      (%)                 (%)

France                8.0                 93
United States        12.0                 15
Japan                18.8                 20
Germany              21.6                 67
Australia            22.1                 80
United Kingdom       28.8                 36
Canada               29.9                 32
Italy                33.8                 90

SOURCES: Union Density from Table 1. Collective bargaining from
www.abs.gov.au,
http://unionstats.com, http://www.statcan.ca,
and Visser (2006). Collective bargaining figures are for years
2000-02. These data are updated infrequently, but collective
bargaining coverage changes slowly.

TABLE 3
UNEMPLOYMENT RATES AND THE COST OF HIRING AND FIRING EMPLOYEES

Country          Hiring Cost (%)   Dismissal Cost (%)   HC+DC (%)

Japan                 12.7                21.2            33.9
Australia             21.0                 4.0            25.0
United States          8.5                 0.0             8.5
United Kingdom         8.7                33.5            42.2
Canada                12.0                28.0            40.0
Italy                 32.6                47.0            79.6
France                47.4                31.9            79.3
Germany               21.3                66.7            88.0

                 Unemployment Rate (%)

Country          2007   1999-2006

Japan            3.8       4.8
Australia        4.4       5.9
United States    4.6       5.0
United Kingdom   5.3       5.1
Canada           6.0       7.2
Italy            6.2       8.7
France           7.9       9.4
Germany          9.0       8.4

SOURCES: World Bank, Doing Business in 2006: Creating Jobs,
and CIA World Factbook, 2008.

TABLE 4
PRIVATE SECTOR UNION DENSITY IN THE UNITED STATES

Year   Union Density (%)

1960         30.9
1970         27.4
1980         22.3
1990         15.5
2000         12.8
2007         11.6

SOURCE: OECD (ww.oeed.org/dataoecd/25/42/39891561.xls).

TABLE 5
UNION DENSITY IN SELECTED U.S. INDUSTRIES, 2009

Industry                           Union Density (%)

Transportation and Warehousing           22.4
Utilities                                28.3
Telecommunications                       20.4
Construction                             16.2
Manufacturing                            12.3
Educational Services                     16.0
Health Care and Social Assistance         8.9
Financial Activities                      2.3
Professional and Business Services        2.7

SOURCE: www.bls.gov/news.release/union2.t03.htm, Table 3.

TABLE 6
U.S. PUBLIC SECTOR UNION DENSITY, 2009

Level of Government   Union Density (%)

All Public Sector           40.7
Federal                     33.0
State                       35.1
Local                       46.1

SOURCE: www.bls.gov/news.release/union2.t03.htm, Table 3.