Race, Prediction, and Pretrial Detention

• Published date: 01 December 2013
• Author: Shima Baradaran,Frank McIntyre
• DOI: http://doi.org/10.1111/jels.12026
• Date: 01 December 2013

Race, Prediction, and Pretrial Detention

Frank McIntyre and Shima Baradaran*

This article uses the nationally representative State Court Processing Statistics on felony

defendants to analyze how judges decide if defendants should be held pretrial. We find a

large (11 percentage points) racial gap in hold rates within a county. Judicial decisions are

significantly influenced by the probability that the defendant will be rearrested pretrial for

a violent felony. Controlling for this probability causes the racial gap in hold rates to

disappear. Bail amounts follow the same pattern. The most plausible sources of bias—

mismeasurement of the control probabilities or selection bias—likely either do not matter or

cause an upward bias.

I. Introduction

Most scholars, political leaders, and experts would agree that racial bias exists in the U.S.

criminal justice system (Johnson 1987; Lopez 1999).1Despite widespread efforts to create

equality in criminal laws, it is still true that blacks are imprisoned at far higher rates than

whites. In 2006, blacks comprised 37.5 percent of all prisoners, though they only consti-

tuted 13 percent of the national population. About one in 33 black men was in prison in

2006, compared with one in 205 white men. Although some blame this massive gap in

outcomes on unobserved differences across groups, many claim that racism enters the

criminal justice system at every step, from arrest through conviction and sentencing. One

can readily find work casting blame on all actors involved, from police and media (McCain

1992)2to lawyers, juries, and judges (see, e.g., Wardle 2007; Anderson 2002).3

*Address correspondence to Frank McIntyre, Assistant Professor of Economics, Rutgers Business School Department

of Finance and Economics; email: frank.mcintyre@rutgers.edu. Baradaran is Associate Professor of Law, University of

Utah College of Law.

Many thanks to colleagues who reviewed and commented on this draft as well as earlier related work.

1Johnson (1987) explained that unconscious racism plays a part with the Supreme Court’s analysis of equal protection

clause and cruel and unusual punishment cases, that ignorance of racism blinds judges, and that indirect, covert, and

unconscious racism often goes unnoticed in the criminal justice system. Lopez (1999) discussed judicial institutional

racism in criminal justice cases resulting from unconscious bias in society, even though judges lack intent to

discriminate.

2McCain (1992) discussed racism pretrial with media and prosecutorial discretion that perpetuates racism in criminal

justice and the importance of eliminating racism early on because later stages do not have mechanisms to filter out

pretrial prejudice.

3Other scholars have explained that family relationships may be at the root of the link between race and crime (see,

e.g., Wardle 2007; Anderson 2002).

bs_bs_banner

Journal of Empirical Legal Studies

Volume 10, Issue 4, 741–770, December 2013

741

In the past, data limitations have made it impossible to say much about the extent to

which these gaps in incarceration are due to racial bias. The available data simply did not

make it possible to convincingly tease out the extent to which outcomes might be due to

unobserved differences between the inmates, a necessary precondition to isolating racial

bias. Thankfully, both data and statistical sophistication have been growing at such a

phenomenal rate that we can, if not end the debate, at least inform it. This article takes

advantage of a large, nationally representative data set on felony defendants who were

exhaustively followed from arrest to sentencing. We use these data to focus on a very specific

decision: Should the defendant be released or held pretrial? We find that the racial gap

disappears on average once one flexibly accounts for differing probabilities of rearrest for

violent crime. These results are robust to controlling for drug crime rearrests or nonviolent

felonies or when looking at bail amounts. We find that the most plausible sources of

bias—mismeasurement of the probability of violent crime, selection bias, and missing

income data—likely either do not matter or cause an upward bias, making our results a

viable upper bound.

Shortly after arrest, the judge determines whether a defendant should be released or

detained pretrial. The bail determination is important both in its own right and as a specific

point at which judges make a quick decision based on limited information, much of which

we have available to us, about how to treat a defendant standing before them. This is an

obvious potential source of bias as, statistically, 43 percent of black defendants are detained

pretrial and only 34 percent of white defendants. This deprivation of liberty for months

prior to having a trial is in and of itself a serious issue because we find that about one in six

of those released will be rearrested for a second crime while awaiting trial. Some of those

released, about one in 50, will be rearrested for a violent felony. Thus the decision is

important both to the defendant desiring freedom and to the community that bears the

burden of “bail crime.”

In the pretrial decision, the judge quickly examines the crime the defendant is

charged with, his or her prior record, and the evidence against the defendant and deter-

mines whether the defendant should be released or detained pretrial. This decision is

highly indicative of whether the defendant receives a custodial sentence or parole and may

impact whether a defendant is convicted at trial. Given this information, the judge is also

charged (in most jurisdictions) to detain defendants who are likely to be a “danger” to

society or are likely to commit a crime if released. This pretrial decision, then, is a good

place to begin the inquiry into race and the judiciary.

Beyond pretrial detention’s inherent importance, it is an excellent opportunity to see

how judges make quick decisions. Price and Wolfers (2010) showed that judges in sporting

events may exhibit racial biases in snap judgments, as do many citizens and potential jurists

in implicit association tests (Greenwald et al. 2009). This pretrial release decision sheds

light on racial bias in a more important context, but still one based on a limited and

contained set of information. Econometricians have little hope of recovering all the details

of a trial, but the data brought forward in a bail hearing are not nearly so extensive. Given

our data on arrest charge and an extensive criminal history, we can come far closer to

recovering the pertinent data the judge uses in making a decision. Furthermore, we have

very good data on the outcomes of interest in a bail hearing—such as will the defendant flee

742 McIntyre and Baradaran