Prison Crowding and Violent Misconduct
56 Pages Posted: 29 Jun 2017 Last revised: 8 Dec 2017
Date Written: November 15, 2017
In recent years justice reform has been a popular bipartisan topic in U.S. politics, with reducing the burgeoning U.S. prison population as one of the primary goals. The first objective of this research is to estimate the causal relationship between prison crowdedness and prison violence that is essential to understanding the impacts of having severely overcrowded prisons as well as efforts to reduce such crowding. I exploit exogenous variation in California prison populations, resulting from a Supreme Court mandate to reduce prison crowding, to estimate this relationship. Using both difference-in-difference and instrumental variable identification strategies, I identify a significant positive effect robust to a variety of model specifications. The estimates suggest that reducing prison crowding by 10 percentage points leads to a reduction in the rates of assaults and batteries by 15% or more. These estimates of the relationship between prison crowding and violent misconduct are, to my knowledge, the first in the literature with a justifiable argument for causality.
The second objective of this research is to explain the paucity of empirical evidence to support the widely held belief among correctional policy makers and practitioners that there is a positive causal relationship between prison crowding and violence. A simple reduced form model is presented which recognizes that population shocks inevitably change both crowdedness and the composition of the prison population, assuming some heterogeneity in inmates’ baseline propensities toward violence. Failure to account for this can bias estimates towards zero. Although the estimation strategy used in this paper does not directly control for compositional changes, I provide persuasive evidence towards the presence of a compositional effect. The estimates provided are therefore a lower bound on the true effect of crowding on violence.
Keywords: Crowding, Population Density, Violence, Misconduct
JEL Classification: K14, K42, D03, A12
Suggested Citation: Suggested Citation