Unemployment, Choice and Inequality

This monograph analyzes how unemployment and job search affect the distribution of labor earnings and contribute to inequality. Unemployment and job search generate inequality both through unequal time spent employed and through the unequal wage offers facing unemployed workers. This monograph studies the choices that workers make between earnings and unemployment, and calculates how much inequality this choice generates. It further studies the contribution of random outcomes to inequality and relates this inequality to the labor market assignment problem solved by search.

This chapter describes the job search model that will be used to study the distribution of earnings. The Markov process describing movements between employment and unemployment will be the basis of the distribution of unemployment in Chapter 5. The wage offer distribution will be used to explain the distribution of wage rates in Chapter 6. The impacts of labor market conditions will yield the valuations of unemployment in terms of earnings in Chapter 4. The behavior of workers will explain the joint distribution of unemployment and wage rates in Chapter 6.

This chapter develops methods of measuring the value workers set upon a period of unemployment. In existing studies of the distribution of earnings, there is already an implicit valuation of unemployment. If a given worker is employed one week less, his or her earnings are reduced by the weekly wage rate. If only earnings are examined in studying inequality, then a week’s unemployment is valued at a week’s less earnings.

The intention of this chapter is to examine the distribution of employment and unemployment among individual workers. By and large, the literature has dealt with differences in aggregate unemployment rates by demographic groups and with the distribution of unemployed workers by duration of unemployment. In contrast with these ways of looking at unemployment, this chapter concentrates on the size distribution of employment. This distribution is described by a cumulative distribution function for the proportion of a particular group that is unemployed less than or equal to an amount of time x in a period of length t. While the duration of unemployment is simple to model mathematically (at least when transition rates are constant), the proportion of the time spent employed is complicated by the possibility of multiple transitions between employment and unemployment. The appropriate probability density functions are developed in section 4. Related to the description of the size distribution of employment is the question of how unequally unemployment is distributed. That is, how unequally is the burden of unemployment distributed among the population, and how much of this inequality is the result of random outcomes?

The previous chapter develops one ingredient in the final distribution of earnings: the distribution of employment. This chapter develops the other ingredient, the distribution of wage offers facing the individual and the resulting distribution of accepted wage rates. Several important questions are introduced into the study of the distribution of earnings at this point. In very simple theories, the distribution of wage rates is related directly to the distribution of some individual characteristic, such as education, ability or experience. In more complex models, it arises from the assignment of workers to jobs in a deterministic model with full employment, in which differences in wage rates are related to trade-offs between product outputs or preferences in equilibrium. With job search, however, there are several intermediate steps. Even for a group facing identical labor market conditions, the distribution of accepted wage rates will resemble neither the distribution of wage offers nor the distribution of reservation wages for the group but will depend upon an interaction between the two.

The main theme of this monograph is the way in which unemployment and job search generate inequality and the nature of that inequality. We have discussed all the essential elements in the determination of inequality and are now in a position to combine those elements.

Labor markets are characterized by a number of complexities that are absent in simple markets. Because of job search, unemployment arises which may or may not be efficient. Supply and demand behavior are no longer represented by simple curves but instead depend both on the expected wage prevailing in the market for individual workers and on the level of unemployment or likelihood of getting a job. Also, there is not a single market; instead there are many overlapping markets, so that one worker’s prospects depend on the behavior of participants in the related markets. This chapter examines the qualitative conclusions concerning the operation of labor markets that arise from the theoretical and empirical results of this monograph. In particular, the analysis undertaken here suggests that labor markets operate very differently for higher grade workers versus lower grade workers. These differences are described in the next three sections.

The standard presumption regarding unemployment generated by search is that it is efficient (Robert Hall, 1979a; E. Prescott, 1975). But this conclusion is based on an incomplete notion of the role of search. Previous work on the subject, by not recognizing that search assigns workers to jobs, has neglected to study the assignment brought about by search and has ruled out the phenomenon of regression towards the mean. This phenomenon distorts the assignment of workers to jobs in a systematic way and produces chronic underemployment (excessive unemployment) of less skilled workers. Although the job matching literature (Boyan Jovanovic, 1979; Dale Mortensen, 1976) has incorporated some of the allocative role of search, it has not specified the heterogeneous nature of the workers and firms to be assigned to each other, with the result that regression towards the mean cannot arise.