There are two competing explanations of why women workers earn less than men with equivalent education, work experience, and job tenure: the human capital explanation and the discrimination explanation. The human capital explanation argues that sex differences in human capital investment which arise from sex differences in expectations surrounding labor force participation account for the wage differential. Women workers are expected to invest less in jobspecific human capital than otherwise comparable men workers because women expect to spend less time on the job. Furthermore, even for men and women workers with equal ex post levels of job tenure and/or work experience, women have invested less in onthe-job training because their a priori expectations of job tenure and/or work experience were less than those of men who now have the same tenure and/or experience. Therefore, in this view, women workers earn less than comparable men because they have invested less in specific human capital. Women earn less because they are less productive; the sex-wage differential is economically efficient. The discrimination explanation argues that sex differences in labor market opportunities, that is, sex discrimination in the labor market, account for the sex-wage differential. In this view, women workers earn less than comparable men because they are the victims of sex discrimination in the labor market. Women do not earn less because they are less productive; the sex-wage differential is economically inefficient. While the economic implications of these two explanations of the sex-wage differential are enormously different, both explanations are consistent with empirical studies simply because both resort to nonmeasurables to explain the sex-wage differential: empirical studies cannot measure directly either discrimination or job-specific human capital. Therefore, the problem with these two competing explanations of the sex-wage differential is that neither has been empirically sorted from the other. Both explanations are consistent with data which show a wage differential by sex after controlling for education, work experience, and job tenure. Newly available data on displaced workers provide an opportunity to empirically disentangle these two competing explanations of the sex-wage differential. Displaced workers are workers who have lost their jobs either because their workplaces have closed or because they were permanently laid off due to slack demand for the outputs of their firms. Displaced workers represent a special case of worker mobility. Unlike voluntary job movers, that is, workers who have voluntarily quit their prior jobs, the job mobility of displaced workers is not the result of their own expectations that better jobs are available. A worker who voluntarily changes jobs does so because there is another job which offers higher wages (or other improvements in the conditions of employment). The worker moves precisely because his or her productivity (and wages) is higher on the subsequent job. The wage change is endogenous. Unlike workers who are fired or involuntarily laid off because their personal productivity is lower than that of other tDiscussants: Rebecca Blank, Princeton University; Elyce Rotella, Indiana University.

Two-earner households differ from oneearner, husband-wife households in other ways than the number of earners. They have higher money incomes; they are smaller and also less likely to include young children; they have less adult time. Urban residential location choices have traditionally been explained in terms of tradeoffs between housing consumption and job accessibility. These differences in income, household composition, number of workplaces, and available leisure time thus suggest that the residential and work location choices, and housing demands of two-earner households differ from those of other households. Because two-earner households have higher money incomes and because recent empirical studies have found the income elasticity of housing demand in the United States to be greater than one (see Edwin Mills), two-earner households are expected to consume more housing. Because the housing expenditure savings from a more suburban location increase as housing consumption increases, households who consume larger amounts of housing (i.e., twoearner) are expected to find suburban locations more attractive, ceteris paribus. On the other hand, two-earner households have members commuting to two jobs. This suggests that commuting costs are doubly incurred and that access to jobs plays a larger role in their residential location decision. If employment of both spouses tends to be centrally located, the total commuting costs associated with more suburban locations are higher for two-earner than for one-earner households. Alternately, if one spouse is employed at a suburban location (see Michelle White), some suburban locations may offer proportionately lower commuting costs for two-earner than for oneearner households. To the extent that: 1) employment location is determined by residential location and commuting preferences of employees; and 2) either two-earner households locate differently than oneearner households, or employed wives have different commuting preferences than other employees, the increasing labor force participation of married women will also affect the rate of suburbanization of employment. Finally, besides having higher money incomes and more job locations to access, two-earner households may differ systematically from one-earner households in their preferences for particular housing characteristics such as neighborhood and density. This paper analyzes the net effect of income, commuting costs, and housing composition on the household's choices of house size, house location relative to employment location, and other housing characteristics for two-earner and one-earner households. In the first section a household decisionmaking model of location and housing consumption is developed; the second section discusses the data used to estimate the model; the third section discusses the estimation procedure and presents the empirical results; the final section presents the conclusions.