This paper explores a model where private charity and public transfers are determined simultaneously. In political equilibrium, the government "overprovides" public transfers, transferring more to the poor than altruistic taxpayers prefer. At this equilibrium, private charity is zero. Evidence for this result is found by examining various types of data from the 1920s to the present. While private charity currently exceeds $50 billion, very little of it goes to the poor. I provide evidence that this phenomenon of zero private charity began, as the model predicts, in the 1930s, the beginning of federal intervention in the charity market.

This paper explores a model where private charity and public transfers are determined simultaneously. In political equilibrium, the government "overprovides" public transfers, transferring more to the poor than altruistic taxpayers prefer. At this equilibrium, private charity is zero. Evidence for this result is found by examining various types of data from the 1920s to the present. While private charity currently exceeds $50 billion, very little of it goes to the poor. I provide evidence that this phenomenon of zero private charity began, as the model predicts, in the 1930s, the beginning of federal intervention in the charity market.

We analyze the variance of log earnings within labor force cohorts of U.S. males with particular attention to the influence of labor force growth rates. Estimates with data from the 1968-79 current population surveys reveal increases in earnings inequality within labor force cohorts even after we control for the level of education, experience, and unemployment. A model of human capital investment is used to analyze differences among cohorts in the life-cycle profile of the variance of log earnings. It is shown that a pattern of sharply increasing and then declining labor force growth rates will tend to raise the return on human capital investment and thereby increase the variance of postschooling investment and log earnings early in the life cycle. This positive effect on relative earnings inequality should decline with experience and possibly become negative. Because of the post-World War II baby boom and baby bust, recent labor force entrants have faced just such a pattern of actual and projected labor force growth rates during their work life. An extended model is tested and generally confirmed using estimates and forecasts of labor force growth rates for the period 1920-2000.

We analyze the variance of log earnings within labor force cohorts of U.S. males with particular attention to the influence of labor force growth rates. Estimates with data from the 1968-79 current population surveys reveal increases in earnings inequality within labor force cohorts even after we control for the level of education, experience, and unemployment. A model of human capital investment is used to analyze differences among cohorts in the life-cycle profile of the variance of log earnings. It is shown that a pattern of sharply increasing and then declining labor force growth rates will tend to raise the return on human capital investment and thereby increase the variance of postschooling investment and log earnings early in the life cycle. This positive effect on relative earnings inequality should decline with experience and possibly become negative. Because of the post-World War II baby boom and baby bust, recent labor force entrants have faced just such a pattern of actual and projected labor force growth rates during their work life. An extended model is tested and generally confirmed using estimates and forecasts of labor force growth rates for the period 1920-2000.

This paper addresses the question, What happens to the arrest rate when the number of law enforcers increases? One implication of the analysis is that arrest statistics are a poor instrumental variable for judging the quality of law enforcement. Increasing the number of police can increase of decrease the number of arrests. An increased probability of arrest induces fewer criminal acts; hence the ambiguity. Because of this result, we apply the theory in the setting of college basketball. We find a large reduction, 34 percent, in the number of fouls committed during a basketball game when the number of referees increases from two to three. Additional empirical evidence is presented which suggests that this elastic supply of basketball crime is due to more competent officiating and cleaner play.

This paper addresses the question, What happens to the arrest rate when the number of law enforcers increases? One implication of the analysis is that arrest statistics are a poor instrumental variable for judging the quality of law enforcement. Increasing the number of police can increase of decrease the number of arrests. An increased probability of arrest induces fewer criminal acts; hence the ambiguity. Because of this result, we apply the theory in the setting of college basketball. We find a large reduction, 34 percent, in the number of fouls committed during a basketball game when the number of referees increases from two to three. Additional empirical evidence is presented which suggests that this elastic supply of basketball crime is due to more competent officiating and cleaner play.