During the last 25 years, annual hours worked by prime aged men fell by the equivalent of six 40-hour workweeks. The reduction was more pronounced among those younger than among mid-age workers, among black men than among white men, and among those with less schooling. Thus hours worked not only fell but showed increased dispersion, increases that paralleled the growth in wage dispersion that has become so familiar to students of trends in wages. The argument advanced here is that the correspondence is not coincidental; the changes in hours worked are simply the labor supply responses that follow the changes in the structure of wages.

Trends in wages and employment of black men are contradictory. Wages are rising and employment is falling. The article first describes employment trends using data from the five decennial censuses, 1940-80, noting that education is becoming a more important determinant of participation. Next, it is shown that ratios of average wages have not been severely biased by falling participation of those who would earn the least if they worked. The final question asks whether falling participation of black men results from deteriorating job opportunities or from falling labor supply; falling supply appears to dominate.

Although increased wage inequality among men during the past three decades has received more attention, the growth in women's wages has been equally remarkable. In fact, by one measure of inequality, the ninth-decile/median ratio, the proportional growth in inequality has moved in exact proportion with the female/male wage ratio. It is suggested that both result from expansion in the value of brains relative to brawn. There is no way of knowing the full story of growth in women's relative wages, and it is important not to dismiss the import of changing career patterns. As is evident in the panel data, increasing labor market participation must be important. So, too, are the implications that follow the movement of women from the home to the job.

This is a perfect time to discuss inequality. After a period of three decades when wage inequality among men in the United States grew to approximate pre-World War II levels, measured inequality apparently has stabilized and may, in fact, be decreasing (Claudia Goldin and Robert A. Margo, 1992). It is, therefore, a time to assess the changes, to compare the gains and losses. I choose this topic, not to offend, but because I believe inequality is an economic good that has received too much bad press. I also think you will agree that it is a good, which like any other, can be scarce or overly abundant. I am neither trying to praise nor defend poverty, and I hope it is understood that the link between wages and income is not especially close, particularly at lower incomes where nonemployment dominates. Wages play many roles in our economy; along with time worked, they determine labor income, but they also signal relative scarcity and abundance, and with malleable skills, wages provide incentives to render the services that are most highly valued. Further, we all buy and sell labor either directly or indirectly as labor is embodied in products. To reverse of the themes of Milton Friedman's introductory theory class, one man's can be another's meat. Together with rising relative wages, the post-1950 increase in the job-market employment of women is of our major accomplishments, but it is an accomplishment that has been fueled in part by the low wages earned by those the Census classifies as private household workers, childcare workers, workers in laundries and cleaning establishments, food service workers, etc. As the population ages, the prospect of increasing dependence on personal assistance looms. For the most part, personal assistants and aides earn low wages, wages that extend the services we can afford. Friedman's actual observation is, however, that one man's meat is not necessarily another's poison (Friedman, 1976). The relatively high wages earned by physicians, scientists, and university professors have attracted many immigrants to the United States. Would we be better off if they had not immigrated? We could in fact reduce wage inequality by simply proscribing immigration, because the wage distribution of immigrants is U-shaped relative to that of natives (see J. P. Smith and B. Edmonston, 1997 p. 180 [table 5.4] ). My guess is that most economists, faced with such an option, would respond by directing attention to the services immigrants provide. My objective is to examine the recent period of increasing wage inequality with an eye toward the positive. I describe some of the changes that have occurred, speculate about the causes, and wrap up with illustrations of consequences. I contend that growing inequality has created opportunities that have been exploited by many and that the gains are not restricted to the traditional elite. Moreover, when we have adopted policies to mitigate the downside of increasing inequality, which is falling real wages in the lower parts of the distribution, a surprising number of individuals have also capitalized on those opportunities in ways that are not productive. Before turning to the data, I want to make a few general observations. * Department of Economics, Texas A&M University, College Station, TX 77843-4228. I am indebted to Martin Gritsch for assistance in extracting the data and to Barbara Charlton for secretarial support. I am also indebted to participants in the UCLA/RAND Labor Workshop and especially to Janet Currie and James P. Smith for comments and suggestions on an earlier draft. I have drawn heavily on my previous work, much of it in collaboration with Kevin M. Murphy.

The arrival of the post--World War II baby boom cohorts in the job market raises many questions of effects associated with a rapidly declining average age of the labor force. This paper first summarizes 1967-75 wage behavior, showing that relative wages between schooling groups have not changed for prime-aged workers, but there is some evidence, for new job-market entrants, that wages of more educated workers have fallen relative to wages of less educated workers. However, changes among schooling groups are small in comparison to those between new entrants and peak earners within schooling group. The evidence is very direct: as work-experience distributions shifted toward increased proportions of young workers, their relative wages fell. After examining a career-phase model in which workers at different phases are imperfect substitutes, estimates of empirical relationships between cohort size and wages are presented. The main result is that income-depressant effects of (own) cohort size decline over the career but do not vanish altogether. Initial effects include reductions in wage rates and in hours and weeks worked, while persistent effects extend only to wages.