VACANCIES, UNEMPLOYMENT AND WAGE GROWTH

Most previous empirical work on unemployment has essentially ignored the demand side of the labor market. This is a serious shortcoming; many important policy questions cannot be answered without information on both job vacancies (the unsatisfied demand for labor) and unemployment (the unutilized supply of labor). This dissertation represents an effort to better integrate the demand side of the labor market into the empirical investigation of unemployment.

Chapter 1 pulls together all of the existing U.S. and Canadian data on the job vacancy rate and combines it with unemployment rate figures for the purpose of assessing how large the number of vacant jobs has been relative to the number of unemployed persons. The chapter's central finding is that the number of unemployed persons has typically far exceeded the number of vacant jobs; this result has important implications for the selection of an effective unemployment policy.

Time series data relevant for assessing whether the relationship between the U.S. vacancy rate and the U.S. unemployment rate has been stable over time are presented in Chapter 2. All the evidence from a large number of different sources seems to indicate that the U.S. Beveridge curve moved sharply away from the origin during the 1970's. While the number of unemployed persons has exceeded the number of vacant jobs throughout the 1970's, the vacancy rate associated with any given unemployment rate was substantially larger at the end of that decade than would have been true ten years earlier. Aggregate time series data and pooled cross section data are explored in an effort to determine what produced this shift. While only some of the potentially important factors could be considered in the formal analysis, the results do suggest that both the changing age structure of the labor force and the evolving unemployment insurance system played important roles.

Chapter 3 explores the relationship between the aggregate unemployment rate and the rate of wage growth. The chapter's central conclusion is that proxies for the aggregate vacancy rate perform better in wage growth regressions than do either the official or the prime age male unemployment rate. This result should be of considerable interest to macroeconomic model builders.