Imprisonment and diminishing marginal returns

The rate of imprisonment in the United States has grown dramatically since the 1970's with a current prison population of over 1.4 million inmates. During this expansion, crime reduction served as a primary reason for passing sentencing legislation calling for longer prison sentences. The question is the extent to which this increase in incarceration and sentence length reduced crime. One aspect of this question and the debate that ensued deals with the limits of imprisonment as a crime reduction tool. Some argued that the returns of incarceration, in terms of crime reduction, inevitably decrease due to the imprisonment of less serious offenders as the prison population grows. In other words, the economic concept of diminishing marginal returns applies to imprisonment when used as a crime reduction tool.;Despite the debate, the question of whether or not diminishing marginal returns play a role in imprisonment has never been tested or resolved. Yet this question is as critical now as ever. Imprisonment is at an all time high and governments at the federal, state and local level struggle to balance budgets and decide which inmates to release from overcrowded facilities. This study outlines the importance of examining diminishing marginal returns in light of limitations for crime reduction through deterrence and incapacitation at the national and state level. Using annual data for the years 1971--2002, this study compares linear and non-linear models to determine whether diminishing marginal returns are present for time served and the incarceration rate at the national and state level. The findings indicate that while diminishing marginal returns are not always present, there is enough variation between the states and the national estimate to support a finding of variable marginal returns. The most consistent finding shows that diminishing marginal returns are likely when time served is moderate or high. The state estimates indicate that the incarceration rate tends to increase the crime rate. This is not evident in the national findings and may be due to an aggregation effect. Clearly, research needs to consider the potential for variable returns as well as state differences.