Search

The observed gap in average wages between black men and white men inadequately reflects the relative economic standing of blacks, who suffer from a high rate of joblessness. The authors estimate the black‐white gap in hourly wages from 1980 to 1999 adjusting for the sample selection effect of labor inactivity. Among working‐age men in 1999, accounting for labor inactivity—including prison and jail incarceration—leads to an increase of 7%–20% in the black‐white wage gap. Adjusting for sample selectivity among men ages 22–30 in 1999 increases the wage gap by as much as 58%. Increasing selection bias, which can be attributed to incarceration and conventional joblessness, explains about two‐thirds of the rise in black relative wages among young men between 1985 and 1998. Apparent improvement in the economic position of young black men is thus largely an artifact of rising joblessness fueled by the growth in incarceration during the 1990s.

One question dominates methodological debate in macrosociology: In explaining large‐scale social processes should we seek simple theories that apply under a range of conditions or complex theories that are tailored to specific cases? Complex explanations are justified by their explanatory power and theoretical appeal. Simple explanations are justified by their falsifiability and their capacity to sustain strong conclusions. I weigh the merits of simple and complex explanations, applying some basic ideas from Bayesian statistics. The Bayesian analysis provides support for a compromise position that is rigorously grounded in the fundamental rules of probability. These ideas are illustrated in an analysis of labor union growth in Sweden.

Comparative research contrasts the corporatist welfare states of Europe with the unregulated U.S. labor market to explain low rates of U.S. unemployment in the 1980s and 1990s. In contrast, this article argues that the U.S. state made a large and coercive intervention into the labor market through the expansion of the penal system. The impact of incarceration on unemployment has two conflicting dynamics. In the short run, U.S. incarceration lowers conventional unemployment measures by removing able‐bodied, working‐age men from labor force counts. In the long run, social survey data show that incarceration raises unemployment by reducing the job prospects of ex‐convicts. Strong U.S. employment performance in the 1980s and 1990s has thus depended in part on a high and increasing incarceration rate.

This article examines the depedence of unionization on labor market institutions and long-run trends in inflation, unemployment, and the sectoral composition of the labor force in 18 OECD countries. The large number of contending explanations highlights the difficulty in comparative sociology of drawing confident conclusions when there are few cases to compare. This problem is addressed through diverse evidence-a review of the comparative historical record and a time-series analysis-to form a Bayesian prior distribution, which is combined with a cross-sectional data set to sharpen conventional estimates of economic and institutional effects. This analysis indicates working-class organization in trade unions expanded in the postwar period where labor acquired an institutional influence over capitalist labor markets.