by Thomas Sowell

The presently proposed minimum wage legislation raises two major sets of issues: (1) the question of the real effects of minimum wages, as such, and (2) the effect of the new principle of automatically escalating minimum wage levels, tied to earnings elsewhere in the economy.
     Because of inflation, minimum wage levels have been periodically reviewed, and therefore the general issues and growing evidence on the actual effects of the Fair Labor Standards Act have also been reviewed. What an automatic escalation provision means is that we stop looking at the evidence. And we would stop at a time when a growing body of research by independent economists around the country is documenting the negative effects of the minimum wage law—and particularly its devastating impact on job opportunities for minority teenagers.1
     Minimum wage laws have been aptly described as “anathema to economists.”2 Even though 88 percent of academic economists supported the “war on poverty,” 61 percent of those same economists opposed the minimum wage law.3 In short, this is not opposition based on philosophy or political leanings, but on economic analysis and on the mounting factual evidence that the law increases unemployment among the very people intended to be benefited. Moreover, economic research has also revealed a disturbing correlation between teenage unemployment rates and teenage crime rates.4 In view of this, this seems like a particularly inappropriate time to stop looking at the evidence by putting in an escalator clause that will give the law a life of its own, independent of its effects on people.
     The minimum wage law might be discussed in terms of philosophical, political, or economic theory, but the real issues turn on facts. The crucial factual question is whether eliminating wage rates below some designated level also eliminates jobs. As a realistic matter, few people of any philosophical, political, or economic persuasion would want to see either (a) job-seekers unemployed with a hypothetical right to a “living wage,” or (b) people employed at very low wages when they could be equally fully employed at normal wages.
     While facts are crucial, they are not easy to get, for a number of reasons which will be discussed. It will also be necessary to consider a number of standard (or stereotyped) arguments about the minimum wage effects which have persisted over the years. Finally, I would like to venture a few suggestions about the continuing need to assess the impact of the law and some ways that assessment might be improved.

Factual Studies

There are serious problems inherent in trying to study the unemployment effects of minimum wages, as well as other problems that derive from the way the U. S. Department of Labor chooses to approach the issue.
     One problem that plagues minimum wage effects studies is getting statistical data for the specific workers directly affected by the minimum. Such workers are often only a small fraction of the total work force. Even where a substantial proportion of the directly affected workers lose their jobs as a result of a minimum wage increase, this effect can be lost statistically in the random fluctuations in employment of the much larger number of workers whose wages were always above the minimum. The statistical extraction of the relevant changes is analogous to trying to receive an electronic signal through a heavy background of static noise. Different economists use different methods and devices to mute the background statistical “noise” in order to read the signal. As a result of their different procedures for grappling with this problem, economists' numerical estimates of the unemployment effect of the law differ—a variation seized upon by proponents of minimum wages5—but it is increasingly clear that the consensus of these studies is that the law does cause substantial unemployment, and that is more fundamental than the question of exact numbers.
     One of the simplest ways of reducing the statistical “noise” in the data is by selecting some age-group which is known to receive very low wages, so that a relatively high percentage of the people in the category chosen are earning low enough wages to be directly affected by minimum wage changes. Teenagers are an obvious choice, and nonwhite teenagers even more so. Here the serious unemployment effect of minimum wage rates has been repeatedly demonstrated by economists operating independently of one another and using different statistical methods.6
     Extremely high unemployment rates among black teenagers have been so highly publicized in recent years, and so automatically attributed to employer discrimination, that certain historical facts must be noted. Large racial differences in teenage unemployment are of relatively recent vintage. In the late 1940s and early 1950s, there were no such large differences, and indeed, black youngsters 16 and 17 years old had consistently lower unemployment rates than whites in the same age brackets.7 Surely no one is going to claim that there was less employer discrimination then than now. We all know better. What was the difference, then? Minimum wages had not yet begun the rapid rise and spreading coverage which has been the dominant pattern since then.8
     The unemployment effect of minimum wages can also be seen in international comparisons of countries that do and do not exempt young people from the adult minimum wage. In countries where such exemptions are slight or nonexistent—such as the United States and Canada—youth unemployment is some multiple of adult unemployment. But where there are exemptions that are large and cover a number of working years—as in England, Germany and The Netherlands9—there are no significant differences between youth unemployment rates and adult unemployment rates.10
     These findings may reflect the special vulnerability of teenagers as an inexperienced and relatively unskilled group—or they may reflect the greater statistical ease of determining the facts for this group. A recent survey of minimum wage studies notes “the lack of acceptable continuing data on low-wage adults.”11  The same things known to be happening to teenagers may also be happening to other very low-wage people, who happen not to be grouped together statistically. There are some scattered clues that this is in fact the case. For example, an older study of domestic servants, before they were covered by the Fair Labor Standards Act, showed that their ranks tended to be increased in the wake of minimum wage increases, suggesting the displacement of low-skill women from other employment that was covered by the act.12
     Factual studies by independent (usually academic) economists must be sharply distinguished from studies by the U. S. Department of Labor. The Labor Department itself has recently been forced to acknowledge the gap between its perennially optimistic conclusions and the consensus of independent studies, the latter “using advanced economic and statistical analyses.”13 The crudity of the Labor Department studies has been scathingly criticized by academic economists.14 However, even so, the actual numbers appearing in Labor Department studies of minimum wage effects often show employment declines in the wake of minimum wage increases, even though the stated conclusions of these very same studies may be that the minimum wage did not cost people their jobs.15 Congressman Dent is correct only in the narrowest sense when he asserts that “Not once in the history of the minimum wage has there been an adverse report” from the Labor Department about “the lessening of job opportunities.”16 In this context, such a statement is far from reassuring. It will hardly be the first clean bill of health given by an agency evaluating itself or the legislation on which its own appropriations and staff depend. This is especially unsurprising to me, as one who worked inside the Labor Department on minimum wage research, and who personally experienced the pressures to reach conclusions consistent with the department's interests.

Assumptions and Claims

The economic analysis which concludes that minimum wages increase the unemployment of low-wage workers rests essentially on the belief that labor is no exception to the general rule that less is demanded at a higher price than at a lower price. Attempts to overturn this basic economic principle usually reduce to one of four assumptions or assertions: (1) there is a fixed number of workers demanded, more or less without regard to wage rates; (2) low-wage workers are victims of employer monopoly power rather than low productivity, so that raising their wage rates will not price them beyond their value to the employer and therefore will not price them out of a job; (3) higher wage rates will cause employers to use labor more efficiently, so that workers will then become more valuable, and so will not lose their jobs; and (4) the increased “purchasing power” caused by higher minimum wages will lead to a greater demand for goods, and therefore a greater demand for labor, offsetting any tendency toward unemployment. These arguments will be examined in order.

Fixed Demand
     The idea that an employer “needs just so many men” is an old one, which dies hard. Factually, it can hardly stand up in the face of declining employment after wage increases, or the virtual elimination of such occupations as Western Union messenger and elevator operator (despite the continued existence of telegrams and elevators). As a theory, it implies that the substitution of capital—and of higher priced labor—is impossible. The problem arises not when the theory is stated directly and explicitly, but when it is implicitly assumed (and therefore insulated from critical scrutiny), as in the belief that more jobs for teenagers mean fewer jobs for adults.

Employer Monopoly Power
     Under special conditions, where there is only one employer in a labor market or a group of employers acting in concert, wages can be kept below what equally productive workers would earn otherwise. There is a special economic theory for such “exploitation” situations, and a minimum wage increase under those conditions would not produce unemployment.17 Unfortunately, low-wage workers are very unlikely to be in such situations. All sorts of firms, industries, and even households employ unskilled workers, and collusion under these conditions is out of the question. Even such a staunch advocate of minimum wages as the late Senator Paul H. Douglas noted that the market for unskilled labor was one of “almost perfect competition.”18 The sad fact is that low-wage workers are not so much underpaid as under-skilled, and there is no easy way around this problem without pricing them out of a job.

     Theories of offsetting rising wages by increasing efficiency have long been used to claim that minimum wage increases will not reduce employment. Unfortunately, those who argue this way have not distinguished real efficiency—larger output from given combinations of input—from a mere substitution of one input for another as their relative prices change. The examples they cite of “better” or “more efficient” methods of production after a minimum wage increase are methods well-known to employers before the imposed wage change, and were not used then simply because they were not the cheapest methods available under the previous input prices. If higher wage rates lead to the substitution of capital for labor, then by definition there will be more output per unit of labor; but it is mere word play to call this more “efficiency” if the product now costs more to produce and society has to support unemployed workers as well.

Purchasing Power
     The doctrine that workers' increased purchasing power after a minimum wage increase will sustain employment has many problems connected with it,19 but the most fundamental problem is that it assumes the very thing that is at issue: that the workers keep their jobs and work as many hours as before. If not, their hypothetical right to a higher wage rate will not buy anything. Workers can only spend real earnings, not hypothetical rights. Once this is realized, it is hardly necessary to go into the other deficiencies of the theory, such as the fact that inflationary increases mean that more spending power is not more purchasing power.


The minimum wage law addresses a serious social problem, but creates no new options for dealing with it. In fact, it simply reduces the set of existing options available to the parties—employers and employees—who must voluntarily agree if there is to be a job. Trying to make people better off by reducing their options seems questionable even as a theory. In practice, what has happened is that fewer transactions (less employment) have taken place when there were fewer options open to the parties. It would be very surprising if it were otherwise.
     The great unsolved problem remains of what to do about the poor in general, or the low-wage workers in particular. Low-wage workers are not changed by calling them higher-wage workers, any more than students are improved by calling them B students instead of C students. The tragic educational results of the process of upgrading by fiat is hardly a recommendation for extending this practice into the economic sphere. In both cases, it is of course much harder, much slower—and more heartbreaking—to try to create real skills and real achievements. And yet nothing else will really do the job.
     Automatic escalation compounds the problems of the minimum wage law by making it possible to close our eyes to its effects hereafter. This seems unconscionable when those affected are poor, vulnerable, powerless, and inarticulate. If the Congress does not monitor what happens to them, there is no other powerful institution to do so. The set of incentives confronting the U. S. Department of Labor makes it unrealistic to expect it to critically evaluate minimum wage effects, and nearly forty years of history make it painfully apparent that it has no intention of doing so. Labor unions have their own imperatives and constraints. For them, the minimum wage law presents the same kind of opportunity that a tariff presents to a business firm. It is a way to price competitors out of the market. That this is accompanied by humanitarian statements may be a matter of rhetorical, or perhaps political, interest but it changes no economic fact. In the Union of South Africa, minimum wage laws were applied to native black Africans for the explicit purpose of stopping their competition with European workers.20 In the days of the British Empire, British unions and manufacturers attempted to get minimum wages applied to India for similar reasons, though with different rhetoric.21  American unions and businesses have been doing something very similar in Puerto Rico and other affiliated territories where minimum wages are set by tripartite boards of mainland Americans, often from competing firms.
     The point here is not to depict anyone as particularly evil. The point is that powerful institutional incentives exist to use the minimum wage laws for purposes very different from those announced in the Fair Labor Standards Act, and that these institutional incentives are likely to persist through turnovers of personnel in the future as in the past.
     Special interests, recognized as such, may be kept within bounds. For the special interests revolving around the minimum wage laws, Congressional oversight seems especially needed, and therefore automatic escalation seems especially dangerous.
     Finally, my hope would be that some way might be considered to have the statistical analysis of minimum wage effects performed by some organization other than the agency whose own fate is intertwined with that of the Fair Labor Standards Act.

Marvin Kosters and Finis Welch, “The Effects of Minimum Wages on the Distribution of Changes in Aggregate Employment,” American Economic Review, June 1972; Thomas G. Moore, “The Effect of Minimum Wages on Teenage Unemployment Rates,” Journal of Political Economy, July/August, 1971; Michael C. Lovell, “The Minimum Wage, Teenage Unemployment and the Business Cycle,” Western Economic Journal, December 1972.
Peter B. Doeringer and Michael J. Piore, Internal Labor Markets & Manpower Analysis (D. C. Heath & Go., 1971), p. 182.
Jacob J. Kaufman and Terry G. Foran, “The Minimum Wage and Poverty,” Readings in Labor Market Analysis, ed. J. F. Burton, Jr., et al. (Holt, Rinehart & Winston, Inc., 197 1), p. 508.
Belton M. Fleisher, The Economics of Delinquency (Quadrangle Books, 1966), Chapter 3.
“... the estimated magnitude of these employment impacts and the distribution of these impacts vary among the studies.” Statement of Assistant Secretary of Labor, quoted in Minimum Wage Legislation (American Enterprise Institute, 1977), p. 14.
See note 1 above.
Employment and Training Report of the President (Government Printing Office, 1976), pp. 242, 243.
Minimum Wage Legislation, p. 2.
U.S. Department of Labor, Bureau of Labor Statistics, Youth Unemployment and Minimum Wages, Bulletin 1657 (Government Printing Office, 1970),p.138.
Ibid., P. 149, Table 10.1.  It should be noted that The Netherlands did not have a minimum wage before 1966, so the entry in the table for 1960-64 is misleading.
Minimum Wage Legislation, p. 15.
Yale Brozen, “Minimum Wage Rates and Household Workers,” Journal of Law and Economics, October 1962.
Quoted in Minimum Wage Legislation, p. 14.
George Macesich and Charles T. Stewart, Jr., “Recent Department of Labor Studies of Minimum Wage Effects,” Southern Economic Journal, April 1960. See also Thomas Sowell, “The Shorter Work Week Controversy,” Industrial and Labor Relations Review, January 1965.
Numerous specific citations of Labor Department studies are listed in Sowell, op. cit. p. 243.
Representative John Dent quoted in Minimum Wage Legislation, p. 10.
Albert Rees, The Economics of Work and Pay (Harper & Row, 1973), pp. 75-78.
Paul H. Douglas, The Theory of Wages (Augustus M. Kelley, 1964), p. 78.
See Sowell, op. cit., pp. 241-242. 
P. T. Bauer, “Regulated Wages in Underdeveloped Countries,” The Public Stake in Union Power, ed. Philip D. Bradley (University of Virginia Press, 1959), p. 346.
Ibid., p. 332.