Silverman ’13: Minimum wage column flunks on all counts

Guest Columnist

Recently, Oliver Hudson ’14 argued that raising the minimum wage will increase unemployment (“The $9.00 minimum wage: A policy to increase unemployment,” March 7). His claim is empirically unfounded. A 1994 paper by Princeton economists David Card and Alan Krueger found no employment impact of an increase in the New Jersey minimum wage. The study used a cross-sectional method to compare changes in employment outcomes between New Jersey and eastern Pennsylvania, which have similar seasonal patterns of employment. Using information gathered from employers before and after the increase, the authors found no statistically significant change in employment.

Card published two other minimum wage studies in 1992. One used regional variations in wages to measure the impact of an increase in the federal minimum wage. The other compared employment changes in California, which increased its minimum wage in 1987, with outcomes in similar states that did not increase their wages. Both found raising the minimum wage transfers wealth from employers to workers, but has no effects on overall employment rates. Numerous studies since 1992 have shown similar results. One 1994 paper that replicated the Card and Krueger study using administrative data rather than survey responses did find a negative relationship between employment and the minimum wage. However, it drops certain data included in the original paper. When these are restored, no employment impact is found.

Hudson candidly acknowledges the Card and Krueger study, although he incorrectly states that it found a positive relationship between employment and the minimum wage — the correlation wasn’t statistically significant. Many studies showing an adverse impact on employment use flawed time-series analysis examining all cross-state variation in minimum wages to estimate employment effects. Perhaps Hudson understands that this approach fails to account for numerous unobserved variables that can yield false conclusions. Moreover, when updated with more recent data, the negative link between employment and the minimum wage often disappears. Hudson cites no evidence to support his premise, only that which debunks it. His is the rare column that refutes itself.

Hudson also notes that “the economics profession leans strongly toward endorsing the textbook view that a minimum wage increases unemployment.” This in and of itself is neither revelatory nor interesting. Diverse groups possess superior wisdom to individuals when they benefit from the aggregated contributions of each member’s unique knowledge. However, crowd wisdom requires that discrete opinions be unique and independent. Many economists base their opinion not on firsthand research or independent observation but on signals from others — i.e., conventional wisdom. The result of individuals basing opinions on others’ views rather than private information is an information cascade. When these occur, crowds are likely to make incorrect decisions. Furthermore, there is significant evidence of a publication bias in favor of studies showing a significant negative link between employment and the minimum wage.

More disturbingly, Hudson asserts that “empirical studies do not provide enough evidence to reject the theoretical model.” Hudson’s conception of scientific inquiry is backwards. Theories are only credible to the degree that they accurately explain real-world phenomena. Until supported by empirical evidence, theory by itself lacks explanative authority. The dominant explanation for the cholera epidemic in Europe in the early 1800s wasn’t germs or infection but the “miasmas” theory that illness is caused by tiny airborne poison particles. Theories can be conjured to support anything. This doesn’t make them true.

The neoclassical theory of supply and demand rests on three assumptions: perfectly competitive input markets, perfectly competitive output markets and diminishing marginal utility of labor and capital. If any of these assumptions are violated, the model is invalidated. In real life, all three are suspect.

There is a theory that explains how increasing the minimum wage can have no effect on, or increase, employment. The monopsony model, an alternative to neoclassical theory, disregards the unrealistic assumption of perfectly competitive labor markets.

It holds that employers are price-setters, meaning they have market power in determining wage levels. Employers set workers’ wages rather than accepting wages given by the market. This means they cannot hire as many workers as they want at the prevailing wage, as neoclassical theory assumes. If they want to hire more, they must raise wages. Because raising wages entails paying more to all existing workers as well, the marginal cost of labor exceeds labor supply, allowing employers to maximize profits by paying workers less than the “equilibrium” intersection of marginal cost and marginal revenue, accruing rents.

When a minimum wage is set, however, the marginal cost and labor supply curves partially converge, which can cause profit-maximizing employers to both raise wages and hire more workers. So long as the increase is small, raising the minimum wage constitutes a pure wealth transfer from employers to workers, reducing deadweight loss and increasing efficiency.

Karl Popper wrote that “the scientific status of a theory is its falsifiability.” When a theory persists despite its failure to withstand empirical scrutiny, adherence is based in ideology, not fact. Writing on religion, Richard Dawkins criticized the attitude that the “book is true, and if evidence seems to contradict it, it is the evidence that must be thrown out not the book.” Hudson’s opposition to increasing the minimum wage is itself religious dogma, not scientific belief.



Bradley Silverman ’13 is a concentrator in economics, public policy and political science. He can be reached at

  • heyholetsgo

    Hmmm, I think it’s fair to criticize the deadweight loss mentality of economists. However, it is a simple fact that when wages increase in certain (not all) areas, it will be more profitable to outsource those low-skill jobs to China. Do you really want that?

    This is also the reason why I think your move to introduce the Monopsony model is self-refuting. Why would you spend a great deal criticizing neoclassical economics, just to introduce a model that is even further from reality? The monopsony model assumes a closed labor market, which is much further from reality than the neoclassical model’s assumptions….