Minimum Wage Regulation

When workers who devote most of their energies to the service of their employers are nevertheless found living in poverty, there is a strong legislative urge to do something about it. The obvious thing to do is to require employers to pay a "living wage." Such legislation has now been in force in various countries for over a hundred years, and economists have been discussing the issue much longer than that. While recognizing the alleviation of poverty as a primary goal of policy, economists have almost uniformly opposed minimum wage legislation, at least until very recently. To some extent, this opposition reflects the division of labor. John Stuart Mill (1848) remarked on the many minimum wage plans current in his time, and went on to describe, at some length, the adverse consequences of such plans. As Marshall (1897) put it, "... there is popularity in the doctrine of a living wage; so we had better leave politicians to praise it and set ourselves to criticize it." More recently, the attempt by Card and Krueger (1995) to debunk the conventional economic wisdom on minimum wages has again generated much popular support, and much criticism by other economists.

Employment Effects

The main economic objection to minimum wage laws is that some workers will lose their jobs. As a theoretical proposition, this is easily demonstrated, under general conditions. Consider a profit-maximizing employer paying n* workers the minimum wage m, and making (maximal) profits . Suppose the minimum wage rises to M, (with no effect on other prices, or on the demand for the employer's product). If this increase is granted, and nothing else changes, profits fall by n*(M-m). Let N* be the number of workers employed after all profit-maximizing adjustments have been made, and note that the new profit level must be at least -n*(M-m). Now consider what would happen if the wage is restored to its old level: starting from -n*(M-m), the employer's profit rises by N*(M-m). But if N* is above n*, the profit would then be higher than its original level, contradicting the assumption that profit was maximal to begin with. So N* can't be above n*, and if the employer has any way to partially offset wage increases through adjustments in production methods, N* must be strictly below n*. This argument has nothing to do with the details of the firm's operations, or with whether the firm has some monopoly power. The only assumption (beyond profit maximization) is that if N* workers are available at the higher wage, then they are also available at the lower wage, so the result fails if the employer can only attract more workers by paying higher wages, that is, if the employer has some monopsony power.

The argument that a monopsony might find it profitable to increase employment in response to minimum wage regulation is exactly analogous to the argument that a monopoly might find it profitable to increase output in response to maximum price regulation. Let m now represent a wage freely chosen by a monopsony employer to maximize profit, where the profit-maximizing employment level n* is the total number of workers available at the wage m. Suppose that a minimum wage level M is set slightly above m, (so that the difference x = M-m is small), with N* - n* extra workers available at the higher wage. Observe that, before the minimum wage law, the employer did not find it profitable to increase the payroll by the amount MN* - mn*, in order to gain the extra revenue resulting from the addition of N* - n* workers. But this must have been a close call, because the employer did find it profitable to increase employment up to n*, and since x is small, the difference N* - n* must also be small, assuming that the number of workers available responds smoothly to increases in the wage offered. Now after the wage M is imposed, the employer is forced to increase the payroll by Mn* - mn*, with no offsetting increase in revenue, and must then consider whether a further payroll increase of MN* - Mn* is worth the gain in revenue generated by N* - n* extra workers. The answer to this must be yes, given that the employer was on the margin of hiring the extra workers anyway, when it would have been more expensive to do so. It follows that a minimum wage set slightly above the level that was freely chosen by the employer must necessarily increase employment. Further increases in the minimum must eventually cause employment to decline, however, since beyond some point the number of workers available is more than the employer wishes to hire, and the situation is then as it was in the first case discussed above, so the same analysis applies. Thus, as Stigler (1946) pointed out, even if monopsony is the relevant case, a national minimum wage cannot be imposed without fear of employment losses, because the appropriate wage level will vary from one employer to the next.


If the purpose of the minimum wage is to ensure that workers can achieve a decent standard of living, then, as Mill (1848) stressed, some provision has to be made for those who are laid off when the minimum constraint binds. There is then the familiar tension between alleviating poverty and providing incentives to work. If those who are laid off are covered by a welfare system that guarantees a minimum standard of living, then there is no incentive to work unless the minimum wage is substantially above this level. On the other hand, if the minimum is set high enough to provide an incentive to work, then some jobs that were initially viable will be eliminated for no good reason, which is clearly wasteful. This problem is endemic in all schemes designed to alleviate poverty, and the wasteful effects of minimum wage regulation should be judged in relation to the alternatives. In this context Stigler (1946) advocated a negative income tax as the best alternative.

There are several reasons to favour other means of reducing poverty. First, if the objective is to provide a decent living standard for each household, then it is unnecessary to intervene in a situation where the sub-standard earnings of one household member are offset by the high earnings of other members. Second, it is a mistake to think that all workers found earning less than a living wage at a point in time are trapped in poverty: most low-wage workers move on to higher-paying jobs.


By far the best-known example of minimum wage legislation is the U.S. Fair Labor Standards Act of 1938. This was preceded in the U.S. by more limited legislation at the State level (setting minimum wages for women and minors only), beginning in Massachusetts in 1912, and in eight other states in 1913 (see U.S. Department of Labor, 1915). The early U.S. legislation was partly stimulated by the living wage movement in New Zealand and in Australia, which led to the New Zealand Industrial Conciliation and Arbitration Act of 1894, and the Factories and Shops Act of 1896 in Victoria. While the stated purpose of the New Zealand legislation was the peaceful settlement (through arbitration) of wage disputes that might otherwise lead to strikes or lockouts, the Victoria measure established wage boards empowered to raise wages in a few "sweated" trades. Within a few years, the wage board system spread throughout Australia and New Zealand, with more and more trades being covered, and a similar system was established in Great Britain by the Trade Boards Act of 1909.

Prior to the passage of the State minimum wage laws in 1913, there were extensive feasibility studies by various ad hoc Commissions. These studies drew on a massive report by the U.S. Bureau of Labor, authorized by an act of Congress in 1907, detailing (in nineteen volumes covering about 1200 pages) the conditions of employment of women and children (see Bureau of Labor Statistics 1915b). The quality of the data on women's wages is remarkable: for example, the Massachusetts commission in 1911 collected individual wage schedules for 6,900 people, and supplemented this with wage data from the Bureau of Labor study for some 8,000 women.

One result of particular importance from these investigations was that wages were found to vary greatly from one establishment to another, for no apparent reason. For example, the Bureau of Labor study tabulated women's wages in 13 establishments in the glass industry, and found that for four distinct categories of relatively unskilled work, the wage paid by the establishment at the top of the distribution was twice the wage paid at the bottom. The conclusion drawn from this was that if one establishment can pay high wages and compete successfully with others paying low wages, there should be little reason to fear that minimum wage laws would cause workers to lose their jobs. Indeed, this kind of evidence seriously undermines the theoretical argument for employment losses given earlier, since it appears from the data that employers do not take the wage as given, as the theory supposes. There is no suggestion that monopsony is the source of these wage variations; the interpretation is rather that some employers are not managing their businesses efficiently, or that they are making excessive profits at the expense of their workers. Whatever the reason, the finding of large unexplained wage variations is characteristic of many field studies of labour markets, and it is echoed in studies of wage variation across industries, such as Krueger and Summers (1988).

The U.S. Fair Labour Standards Act of 1938 largely resolved the long series of political and legal struggles over state minimum wage laws (see Grossman 1978 for an accessible summary of the legislative history of this Act). Amendments have subsequently expanded the coverage of the Act, and the minimum wage has been raised 19 times since it was originally set at 25 cents an hour in 1938; the most recent amendment, passed in August 1996, raised the minimum to $5.15 in September 1997. Yet the struggle continues: the amendments have been preceded by long series of Congressional hearings, sometimes extending over several Congresses, and often calling forth passionate testimony on both sides.

Evidence on Employment Effects

The bulk of the evidence on the effects of minimum wages is concerned with the effects of the Fair Labour Standards Act, as amended.  Figure 1 shows the monthly history of the real minimum wage (deflated by the Consumer Price Index), in relation to real average hourly earnings of production workers in manufacturing. The vertical lines show the months in which increases in the nominal minimum were implemented, and the downward drift between each pair of lines shows the declining value of a dollar. One might expect to find that each jump in the minimum would be accompanied by a reduction in employment, particularly for workers employed in low-wage jobs, such as teenagers, or workers in retail industries. Brown, Gilroy and Kohen (1982) surveyed many regression studies using time-series data to test this hypothesis, leading to the conclusion that "In summary, our survey indicates a reduction of between one and three percent in teenage employment as a result of a 10 percent increase in the federal minimum wage." But as Card and Krueger (1995) pointed out, these are fragile results that are unlikely to convince those who are sceptical about the practical relevance of economic theory.

One reason why it is difficult to find evidence on the employment effects of minimum wages is that there are many ways to circumvent the effect of an apparently binding minimum wage, even if we ignore the real problem of outright noncompliance. A simple example is the tradeoff between wages and fringe benefits: if the minimum is increased, benefits can be cut. Employers also spend money to make the workplace more pleasant, or give discounts on goods and services, or adopt flexible rules on work scheduling, and so forth. These things are easily adjusted, especially in the long run, and from the worker's point of view, they just transfer money from one pocket to another. If a minimum wage regulation merely changes the composition of the worker's pay packet, it should not have any effect on the employment decision.

Another reason for the lack of strong empirical evidence on employment effects is that, as Figure 1 suggests, the minimum wage has never been high enough to affect more than a small minority of workers, and these are the very workers whose employment patterns are most volatile in any case, so the signal to noise ratio in any time-series study is not favourable. Moreover, although the minimum wage series depicted in Figure 1 has about 700 observations, the amount of relevant information in this series is effectively the number of times the minimum was increased by a substantial amount, and the increase was not quickly erased by inflation.

The most recent innovation in the empirical analysis of minimum wage effects is the work of Katz and Krueger (1992) and of Card and Krueger (1995), using original survey data to examine the effects of changes in state minimum wage laws on fast food restaurants in Texas and New Jersey. This research returns to the painstaking methods of the older Bureau of Labour Statistics studies (although the authors of the older studies had many more research workers at their disposal). The most controversial finding was that when New Jersey raised the minimum wage to $5.05 in 1992, while the minimum in neighbouring Pennsylvania remained at the Federal level of $4.25, employment appeared to actually increase in New Jersey relative to Pennsylvania, after the change took effect. As discussed in Kennan (1995), this result becomes less impressive when it is recognized that the relative increase in employment resulted, not from an increase in New Jersey, where the minimum wage had increased, but rather from a decrease in Pennsylvania, where nothing had apparently changed. Without tracing the source of the decrease in Pennsylvania, it is difficult to have confidence that the same decrease would have occurred in New Jersey if the minimum wage law had not been put into effect.

As is illustrated by the remarkable reaction to the work of Card and Krueger (1995), the political economy of minimum wages is at least as vital now as it was in John Stuart Mill's time, even though the economic importance of this kind of regulation does not seem large, after a century of experience.

John Kennan


Brown, C. Gilroy, C. and Kohen, A. 1982. The effect of the minimum wage law on employment and unemployment. Journal of Economic Literature 20: 487-528.

Card, D. and Krueger, A. 1995. Myth and Measurement: The New Economics of the Minimum Wage. Princeton: Princeton University Press.

Grossman, J. 1978. Fair labour standards act of 1938: maximum struggle for a minimum wage Monthly Labour Review, June.

Katz, L. and Krueger, A. 1992. The effect of the minimum wage on the fast food industry. Industrial and Labour Relations Review 46:6-21.

Kennan, J. 1995. The elusive effects of minimum wages. Journal of Economic Literature 33: 1949-1965.

Krueger, A. B. and Summers, L. H. Efficiency wages and the inter-industry wage structure. Econometrica 56: 259-293.

Marshall, A. 1897. The old generation of economists and the new. Quarterly Journal of Economics 11: 128-129.

Mill, J. S. 1848. Principles of Political Economy. New York: Co-operative Publication Society; revised edn, 1900.

Stigler, G. 1946. The economics of minimum wage legislation. American Economic Review 36: 358-365.

U.S. Department of Labour 1915a. Minimum Wage Legislation in the United States and Foreign Countries. Bureau of Labour Statistics, Bulletin 167. Washington: Government Printing Office.

U.S. Department of Labour 1915b. Summary of the Report on Condition of Woman and Child Wage earners in the United States. Bureau of Labour Statistics, Bulletin 175. Washington: Government Printing Office.