It is often tempting to attempt to infer the welfare effects of actual minimum wage changes from empirical observations on pre and post change employment and unemployment levels and wage or earnings distributions. For example, minimum wage increases are often explicitly or implicitly taken to be beneficial for the impacted population if reductions in group employment rates are found to be small or positive. As another example, although a very small percentage of U.S. workers are paid the minimum wage, larger impacts on welfare on the general population are often taken to result from a sort of ripple effect from the bottom up, which in terms of the wage distribution itself is often referred to as spillover. Empirical evidence that demonstrates a change in the shape of the wage distribution above the minimum wage is often interpreted as resulting from these spillover effects, and by their very nature these are assumed to be beneficial for individuals on the supply side of the market.
In this paper we shall rigorously attempt to define, characterize, and explain the phenomenon of spillover. To accomplish this task, we will utilize a simple model of search and bargaining in a stationary environment. While the model is admittedly highly stylized, in a companion paper (Flinn 1999) we show that it can be estimated using Current Population Survey (CPS) data and fits observed wage and unemployment duration distributions quite well. Thus the model at a minimum provides a parsimonious and readily interpretable view of the labor market as it is reflected in CPS data, and for this reason can be given some credibility. The model is readily adapted to allow for the existence of binding minimum wages rates. The equilibrium which results from the imposition of a minimum wage [or an increase in value of an already binding minimum wage] is roughly in accordance with empirical work on this subject based on disaggregated data. In particular, a few of the implications of the model are: (1) the existence of a probability mass at the minimum wage m with an absolutely continuous distribution to the right of m> (2) decreases in employment rates with increases in the minimum wage; and (3) the possible existence of spillover effects in response to a change in the minimum wage rate. Using the model, we will rigorously define a particular welfare measure as well as spillovers. We show that welfare in the population can increase even though employment rates decrease and with or without spillovers in the wage distribution. Conversely, spillover effects of minimum wage changes do not indicate that the change was necessarily welfare?enhancing. The main point to which we want to draw attention is that the welfare effects of minimum wage changes can only be inferred by using empirical evidence on employment rates, wage distributions, and a formal model within which welfare can be rigorously defined and evaluated.
We will not be directly concerned with the impacts of minimum wage levels on unemployment or employment levels in this paper. Instead, we will use our model to attempt to understand the impact of minimum wage levels on accepted wage offer distributions, both in terms of truncation and shape changing. Several authors have attempted to adapt standard econometric models for truncated and limited dependent variables to incorporate minimum wages into standard wage function estimation schemes. Some of the more important research efforts in this area include Meyer and Wise (1983a,1983b), Dinardo et al (1996), and Dickens et al (1997). Meyer and Wise estimated a model of minimum wage effects using individual?level data which allowed them to infer what the wage distribution and employment level would have been in the absence of a minimum wage. The basic idea behind the econometric specification is to assume the form of the wage dis tribution in the absence of a minimum wage, and then to allow the minimum wage to alter this distribution by essentially aggregating probability mass around the minimum wage to that exact value.
This results in a wage distribution which has a continuous and discrete component to it. While their model specification has been criticized by a number of researchers [e.g., Card and Krueger (pp.232 236) and Dickens et al. (1997)], primarily for relying on functional form assumptions for identification and for choosing a parameterization which rules out the possibility of employment increases in response to a minimum wage increase, it remains one of the better econometric attempts to identify minimum wage effects using individual level data in the literature. From our perspective, the main weakness of their model is the arbitrary specification of the manner in which a minimum wage distorts the preexisting distribution wage distribution. In the model developed here and in our companion paper (Flinn 1999), optimizing behavior by searchers and firms determines the nature of this distortion, and it is roughly consistent both with the Meyer and Wise econometric specification and with the empirical evidence cited in Card and Krueger.
Contents
1. Introduction
2. Labor Market Search with Bargaining
- 2.1 Labor Market Decisions without Minimum Wages
2.2 Labor Market Decisions in the Precence of Minimum Wages
3. Minimum Wage Effects on Welfare and Wage Distributions
4. Data and Empirical Exercises
- 4.1 Cross-Sectional Wage Distributions
4.2 Empirical Analysis using the Matched CPS Sample
5. Lessons from Simulations
6. Conclusion
Table 1
References
Figures
Download
PDF Ebook Interpreting Minimum Wage Effects on Wage Distributions: A Cautionary Tale
