Professors Kreuger and Card have reviewed their New Jersey wage study, and the paper can be found at Working Papers no 393. We would be interested in publishing reviews of this work.
Economists, like Gilbert and Sullivan's Englishmen, are divided into two basic groups. On one side are those who believe that responses to price incentives are usually large -- Big Responders (BR). On the other side are those who believe that responses to price incentives are generally small -- Small Responders (SR).
Present a BR with an exogenous change in price or wage -- for instance a mandated increase in the minimum -- and his prior is that there will be a large response in quantities. Br's feel comfortable with perfect competition, Hecksher-Ohlin trade models; factor price equalisation; large responses in effort and hours to marginal taxes; welfare traps; arbitrage of financial opportunities across national lines; large employment losses to administered wages. Forced to choose between a first-approximation economic model with an infinite elasticity of response and one with zero elasticity, the BR economist opts for infinity: "in the long run, there are many substitutes, new competitors, suppliers, etc."
Present an SR with a change in price or wage, and his prior is that quantities will not change much. SR's feel comfortable with input-output analysis, imperfect competition, factor content analyses of the effects of trade on employment, the correlation of investment and savings across countries, backward-bending supply curves, and the persistence of economic rents. Forced to choose between a first-approximation model with infinite elasticity of response and one with zero elasticity, the SR economist opts for zero: "in real world, costs of adjustment are large, uncertainty slows responses, habits change gradually, etc."
What does the BR-SR divide have to do with the Card and Krueger volume? A lot. Along with its other valuable contributions, the book provides the most important evidence in recent years on the SR side of this recurrent debate. It does this through an exemplary empirical analysis of the effects of minimum wages, on which most economists tend to be Br's by inclination.
It is important to recognise at the outset that there is little in economic analysis to help us decide whether quantity responses to price incentives are likely to be large or small. Many economists expect larger responses in the long run; or when budget constraints are important. Logic tells us that massive changes in prices (say, tripling our wages) that turn balance sheets from the black to the red will have large effects on quantities (the Deans might close down our departments). But whether the BR or SR perspective applies to minimum wages in the range observed in the U.S. is a purely empirical question. It does not have to do with acceptance or rejection of neo-classical economic theory. There is no theoretic or a priori reason for assuming that minimum wages that average 40-50% of hourly earnings in manufacturing have large or small effects on employment, or even that minimum wages that average 70 or 80% of average earnings will necessarily do so. Indeed, economic theory is so "rich" that it offers us monopsony models that predict increases in employment in response to minimum wages. Only careful empirical analysis can determine the magnitude of employment responses to the minimum.
The BR-SR divide is an empirical one. This book offers the most careful and wide-ranging analysis of the empirical evidence on minimum wages in the U.S. that any social scientist could ask for. On many issues, the evidentiary base is sufficiently diffuse to allow economists with different priors to reach different conclusions -- to make a lawyers' case, as it were, for their preferred BR or SR world. (Labor economists think this is true of virtually all macro-economic studies, where conclusions seemingly hinge on assumptions -- called structural models -- that determine what one sees in limited time series data, because the evidence does not "speak for itself"). This is not the case for assessing the short or medium run effects of the minimum wage on employment. Minimum wages change in discrete exogenous jumps over time and across areas that creates "natural experiments".
Card and Krueger exploit this variation admirably. Indeed, they do so in such a way as to provide an exemplary example of how to do empirical economics. The authors examine not one, not two, but many such experiments to judge the minimum wage's employment effects. They (and colleagues) gathered new data or analysed available CPS files, before/after recent changes in minimum wages, with control groups, where feasible. They looked at evidence on the effects of the minimum on the distribution of earnings and, using event studies, of the effect of announced increases on the value of firms and on prices. Forget for the moment the results or possible data problems in particular parts of the work. If you want to find out what the world says about an issue, rather than to force the world into your prior structural model, this is the way to do it. Richard Lester, to whom the book is dedicated, has been greatly honoured by such a depthful analysis.
The only thing missing, which I would have liked to see, are some detailed case investigations of low wage employers' response to the minimum, or better yet, some ethnographic experience, to enrich the book. Such analysis would please Lester, also, if I am not mistaken. A few Ronald Reagan style anecdotes would go a long way to diffusing some of the criticisms the book has received from employer groups.
The chapter that I found weakest was on foreign experiences, which also is largely the case of limited information. Here, recent British work on the abolition of wages councils yields SR results consistent with those in this book: no adverse change in employment (Machin and Manning). And analysis of the employment consequences of comparable worth in Australia found little or no job losses for women from large mandated increases in their pay (Gregory and Duncan).
There is more support for the SR view in empirical analysis than most economists have recognised.
Methodology aside, Card and Krueger reach strong and in some cases, surprising conclusions, or, rather and more properly, they report empirical results that yield these conclusions: negligible employment effects in most cases, and positive employment effects in some well-designed natural experiments. Their summary table (table 12.1) shows 6 positive employment effects, of which 2 are significant, and one 0 effect.
This article is reprinted courtesy of Professor Richard Freeman from an article that first appeared in Industrial and Labor Relations Review.