Optimal Bail and the Value of Freedom: Evidence from the Philadelphia Bail Experiment

Abstract

This paper performs a cost-benefit analysis to determine socially optimal bail levels that balance the costs to defendants against the costs to other members of society. We consider jailing costs, the cost of lost freedom to incarcerated defendants, and the social costs of flight and new crimes committed by released defendants. We estimate the effects of bail amounts on the fraction of defendants posting bail, fleeing, and committing crimes during pre-trial release, using data from a randomized experiment. We also use defendants' bail posting decisions to measure their subjective values of freedom. We find that the typical defendant in our sample would be willing to pay roughly $1,000 for 90 days of freedom. While imprecise, our optimal bail estimates are similar to the observed levels of bail prior to bail reform.

Testing for Racial Discrimination in Bail Setting Using Nonparametric Estimation of a Parametric Model

“…To put these numbers in perspective, we adopt figures used in Abrams and Rohlfs (2006, henceforth, AR) to measure components of R; doing so lets us back out an implied estimate of the value of lost freedom, V . Recalling that J = P + V , solving for V yields…”

“…Perhaps its earliest treatment in the economics literature was by Landes (1973), who models bail setting as an optimizing procedure. More recently, Abrams and Rohlfs (2006) use data from the 1981…”

“…We find that there is consistent, though often imprecisely estimated, evidence of discrimination against blacks in bail setting, especially for Dallas county; Los Angeles pretty clearly can be absolved of any charge of discriminatory bail setting. Using figures from Abrams and Rohlfs (2006) to put our estimates in practical context, we find that the magnitude of discrimination against blacks implied by our point estimates is substantial. Back-of-theconditioning on the level of bail and x, while the second set involves conditioning as well on u.…”

Hedonic Estimation Under Very General Conditions Using Experimental and Quasi-Experimental Designs

“…In another example, Abrams and Rohlfs (2011) examine criminal defendants who are deciding whether to post bail or to remain in jail until trial. The market of interest is one's activities over the 90 days until one's trial.…”

“…Variations on these estimators have been applied in recent studies to estimate the value of freedom from jail, the demand for avoiding the Vietnam draft, the value of a statistical life, and the demand for class size reductions in elementary school (Abrams and Rohlfs, 2011;Rohlfs, 2012;Rohlfs, Sullivan, and Kniesner, 2013;Rohlfs and Zilora, 2013). Section VII applies one of the estimators of aggregate demand for an attribute to measure the value that military recruits place on funds for higher education.…”