Essays on the marriageable men hypothesis and the cyclicality of relative wages.

Abstract

This dissertation consists of two empirical studies. In the first, I examine William Wilson's "marriageable men" hypothesis of the decline in black marriage rates. I estimate a fixed-effects model of marriage rates, using SMSA-level data from the 1970 and 1980 censuses to estimate the effect of the availability of high-earning black men on black marriage rates. My initial results indicate a weak relationship, suggesting that the decline in the pool of high-earning young black men was responsible for 7 to 10 percent of the decline in black marriage rates during the 1970s. When I instrument changes in the pool of high-earning men with changes in the SMSA's industrial mix, to correct for the bias caused by the endogeneity of marriage rates and male earnings levels, my estimates are substantially reduced. These results suggest that the decline in the pool of high-earning young black men was responsible for only 1 to 4 percent of the decline in the marriage rates of young black women during the 1970s. In the second study, I examine the movement of interindustry wage differentials over the business cycle. I estimate the cyclicality of these wage gaps using data from the BLS establishment survey and the Current Population Survey. Using BLS data, I find little evidence of any cyclicality in interindustry wage gaps. The inclusion of overtime pay in the BLS average hourly earnings data, however, creates a procyclical bias in my estimates. When I examine CPS wage data, which exclude overtime, my estimates suggest somewhat more countercyclicality in interindustry wage differences. The CPS data also allow me to control for worker characteristics, making it possible to control partially for cyclical changes in the composition of the labor force. In general, when I add controls for worker observables my estimates suggest less countercyclicality. This reduction is consistent with a pattern of low-skill workers leaving low-wage industries for high-wage industries during expansions. My estimates suggest a small degree of countercyclicality in interindustry wage differences. Since I am unable to control completely for worker quality, the actual degree of countercyclicality may be smaller.