Since the foundation of the American Federation of Labor(AFL) in 1886, most unions in the United States have displayed a pragmatic out look, largely compatible with that of business. The general purpose of unions has been to protect and advance the well being of workers, while that of business has been to promote the interests of stockholders. Higher wages and higher profits are compatible over the long run in a growing economy. Conflict does arise, however, from the fact that in the short run higher wages for workers imply lower profits for shareholders. Power, too, is a matter of dispute. In the absence of unions, managers have a monopoly of power over their employees. With unions on the scene, that power must be shared.

The standard economic analysis of what gave a particular union the power to raise the pay and benefits of its merabers was propounded by the eminent English economist Alfred Marshall toward the end of the 19th century. Marshall theorized that the strength of a union depended upon four factors. First, demand for the product should be inelastic, so that there is little, if any, decline in sales in response to price increases. Second, labor costs should be a small portion of the total costs of production, so that a rather large increase in wages would generate only a small increase in the price of the product. Third, the supply of factors that can be used as substitutes for union labor, such as nonunion labor or labor-saving machinery, should be inelastic, so that their price rises substantially as more units are employed. Fourth, the ability of these factors to substitute for union labor should be highly limited; it would be hard to substitute for workers with very high skil! ls or skills that are highly specific to a single employer.

Numerous studies have been made to estimate the extent to which unions in the United States have raised the wages of their merabers above what they would otherwise have been. These studies show substantial differences in the effectiveness of different unions, and that is in the spirit of Marshall\'s analysis. Substantial variation has also been found in the effectiveness of unions over the course of business cycles. On the average, unions have raised the wages of their merabers as compared to nonmerabers by about 15 percent, somewhat more during perioRAB of depression and somewhat less during perioRAB of prosperity. Empirical studies have also indicated that the productivity of union workers has been higher than that of nonunion workers, largely because union workers have tended to have more capital gooRAB at their disposal than nonunion workers. These studies also indicate that unionized workers have had lower turnover rates. This has lowered the costs of recruitment and trainin! g to employers. These cost savi ngs have materially diminished the wage disadvantage experienced by the employers of unionized labor.

Some participants in and observers of the U.S. labor movement have viewed unions as institutions with the potential to establish industrial democracy and socialism. Others have viewed unions as highly conservative institutions returning to workers the status lost in the transition from village societies to urban anonymity. In reality, their role has been more modest. In the early 1980\'s they enrolled in their ranks only one of five merabers of the labor force, down from one of four in the 1950\'s and 1960\'s. These workers had a somewhat greater say in their work lives and in the halls of Congress and state legislatures. They received somewhat greater pay and were more productive. They sometimes followed the political calls of their leaders and sometimes did not.