Alternative Models of Earnings Determination and Labor Market Structures

There are three distinct research traditions in the analysis of individual earnings determination: human capital, or earnings function, analyses; aggregate wage analyses; and labor demand analyses. An important and incongruous aspect of each is the treatment of geographical differences in labor markets. This paper first investigates the magnitude and character of geographical wage differentials. The sizable differences discovered there are then related to the existing, and highly simplified, models of labor market differences. While the two major classes of models (compensating differentials and labor demand) differ signficantly in assumptions and implications, it is impossible to distinguish adequately between them. There appears to be a clear need for more structural analyses of labor market operations. Research into the structure of individual earnings has been voluminous. Most recent analyses, under the heading "human capital" analysis, concentrate on "quality" differences among workers. Nevertheless, other research traditions addressing essentially the same questions have taken quite different, and conflicting, views of wage determination. These include investigations of aggregate differences in earnings patterns arising from differences in employing industries, in occupations, and in employment location and analyses based upon production relationships and the derived demand for labor. A major difference among the alternatives relates to labor market definition and the modeling of how labor market structure affects indiThe author is Professor of Economics and Political Science and Director, Public Policy Analysis, University of Rochester. *Helpful comments on previous versions were provided by Robert Evenson, Gary Fields, Richard Nelson, Randall Olsen, Walter Oi, John Quigley, Paul Schultz, and Kenneth Wolpin [Manuscript received April 1979; accepted September 1979.] The Journal of Human Resources * XVI 2