Does Pretrial Detention Reduce or Increase Crime?

Two interesting papers came across my desk. First, The Economics of Rights: Does the Right to Counsel Increase Crime? by Ater, Givati and Rigbi:

We examine the broad consequences of the right to counsel by exploiting a legal reform in Israel that extended the right to publicly provided legal counsel to suspects in arrest proceedings. Using the staggered regional rollout of the reform, we find that the reform reduced arrest duration and the likelihood of arrestees being charged. We also find that the reform reduced the number of arrests made by the police. Lastly, we find that the reform increased crime. These findings indicate that the right to counsel improves suspects’ situation, but discourages the police from making arrests, which results in higher crime.

In other words, the right to counsel makes it easier for criminals to escape justice and since the price of crime falls the quantity of crime increases. Makes sense!

The second paper The Downstream Consequences of Misdemeanor Pretrial Detention is by Heaton, Mayson and Stevenson:

This Article uses detailed data on hundreds of thousands of misdemeanor cases resolved in Harris County, Texas — the third largest county in the U.S. — to measure the effects of pretrial detention on case outcomes and future crime. We find that detained defendants are 25% more likely than similarly situated releases to plead guilty, 43% more likely to be sentenced to jail, and receive jail sentences that are more than twice as long on average. Furthermore, those detained pretrial are more likely to commit future crime, suggesting that detention may have a criminogenic effect. These differences persist even after fully controlling for the initial bail amount as well as detailed offense, demographic, and criminal history characteristics. Use of more limited sets of controls, as in prior research, overstates the adverse impacts of detention. A quasi-experimental analysis based upon case timing confirms that these differences likely reflect the casual effect of detention. These results raise important constitutional questions, and suggest that Harris County could save millions of dollars a year, increase public safety, and reduce wrongful convictions with better pretrial release policy.

In other words, when you lock people up before trial they lose their jobs and are more likely to get a record so pretrial detention severs attachments to civil life and increases attachments to criminal life with the end result being increased crime. Makes sense!

What’s frustrating is that both of these papers are good–they have a plausible theory and sound research design–yet they reach opposite conclusions! To be sure, the time periods, places, people and exact experiment are different so both papers could be true. From the policy maker’s perspective, however, the fact that both papers could be true only adds to the difficulty of using academic evidence to make policy.


The first paper is on the effect of publicly provided legal counsel, the second on the effect of pretrial detention. These are different treatments. To say they "reach opposite conclusions" is a misrepresentation. Nonetheless very interesting.

And "...the difficulty of using academic evidence to make policy" arises because "academic evidence" is primarily theory & speculation -- not fact.

Why are academics supposed to guide the legal system, anyway ??

Accused Americans are supposed to be "innocent until proven guilty" ---pre-trial imprisonment for non-violent crime accusations contradicts the legal presumption of innocence.

Major problem is that individual policemen have far too much discretionary power to accuse & arrest people, clogging the system at the front end. Judicially issued arrest-warrants should be the norm, rather than the current cop-feels-like-arresting-you system. Police (and government prosecutors) are not supposed to have any judicial powers.

Ever since the US Sup Ct case of Gideon v Wainwright, as Rayward would know (I flunked out of law school), there's a right to counsel. The problem is: they're not that good if court appointed. I don't know about Israel. It could be that in Israel if you have a lawyer the police lay off you more, since you're rich.

So another conclusion might be: rich people who can afford lawyers don't go to jail, whereas poor people who end up in prison pre-trial, and cannot afford to get out, are more likely to go the jail since they are poor. Rich matters.

Basically that.

The problem is: they’re not that good if court appointed.

Very few private attorneys have criminal law as their main line of business. The public defenders are the experts and (if I'm not mistaken) supply about 85% of the manpower devoted to criminal defense where I grew up. The multi-departmental metropolitan firm my family uses has two attorneys (out of 20 or so) who have criminal law as one of their lines, and both have about three other lines.

We've had this issue in our family, where a relation insisted on hiring a private attorney for his son rather than having his son make use of the public defender. The co-defendants who used the public defender got a better deal, because the private attorney had little experience with criminal law (no clue why my in-law insisted on hiring an insurance lawyer or why that lawyer took the case). That case is a bit cherry-picked, of course. There are general practitioners who take criminal cases that you can hire; I have a cousin who does that work, though he prefers real estate and personal injury. They're rank-and-file lawyers, however. They're not Brendan Sullivan.

Thanks for that point; I guess I was thinking of those celebrity criminal lawyers like OJ's defense team.

Assigned counsel plans are characteristic of non-metropolitan counties and are gradually being replaced by salaried public defenders.

It's simply not true that public defenders aren't that good. They are often the most experienced trial lawyers there are.

What IS true is that they are often overworked and underpaid, and they have to provide a vigorous defense to scumbags who are clearly guilty and stand little chance of acquittal.

"... the fact that both papers could be true only adds to the difficulty of using academic evidence to make policy." Can't we just flip a coin? Would we ever be able to tell the difference?

"Overall, GNP accounts turned out to be a powerful instrument to estimate militarization costs and calculate what speed of economic growth would be necessary ‘to pay for the war’, to paraphrase a well-known paper written by Keynes in 1940 during his tenure at the British Treasury. Thanks to this systematic approach, America managed to come very close to pure economic planning, and in 1944 war production goals alone surpassed the nation’s entire output just ten years earlier. According to economist John Kenneth Galbraith, celebrated author of The Affluent Society and adviser to US president Kennedy, Kuznets and his talented colleagues had been the equivalent of several infantry divisions in their contribution to the American war effort". - Gross Domestic Problem, Lorenzo Fioramonti, (c) 2013

"Do you know," said Henderson, "to what extent data concerning our production capacity, our resource potential, our trained manpower- everything of importance to the war effort, in fact-had become unreliable and untrustworthy during the last half of the war? Group leaders, both civilian and military, were intent on projecting their own improved image, so to speak, so they obscured the bad and magnified the good. Whatever the machines might do, the men who programmed them and interpreted the results had their own skins to think of and competitors to stab. “ – Isaac Asimov, “The Machine That Won The War”.

PS--I hear the opening ceremony in Rio was nice...thanks for the link.

"I hear the opening ceremony in Rio was nice…thanks for the link."

The opening ceremony was great, if I can say so myself, like the watch-fires of a hundred circling camps. Glad to be of service to you.

Substitution effect in the first case and income effect on the second?

'In other words, the right to counsel makes it easier for criminals to escape justice'

So, no presumption of innocence in Israel? Being arrested does not make one a criminal, and the right to counsel means that the police need to follow the law themselves - something that libertarians tend to be aware of, at least the sort that read Reason or Radley Balko. As any honest libertarian will note, the price of freedom involves risk - many well run police states (East Germany as an exemplar) tend to be notably free of such criminality as burglary or assault or robbery, not to mention unconcerned about anyone accused by the state having any rights.

'yet they reach opposite conclusions'

Almost as if Israel and the U.S. are two different societies.

'the fact that both papers could be true only adds to the difficulty of using academic evidence to make policy'

Some people might say that applying Israel or America's experience to the other society is the actual problem, one a bit more fundamental than the use of academic evidence.

In the US the government as prosecutor has unlimited funding while the defense does not. In fact, this is leads to plea bargaining, the defendant runs out of money and legal representation and is forced to admit guilt to a lesser crime. Public defenders are law school grads unable to find work in private practice.

"even after fully controlling for the initial ... as well as ..."

While I don't see an alternative to attempting to control for x, y, and z, I'm always doubtful of claims of "fully controlling".

It smacks too much of god-like infallibility. There are limits to what one can confidently tease out from mere observational data, and it would be wise to admit it.

I'd be suspicious of the controls too. What Tabarrok does not mention is that they are examining misdemeanor pre-trial detention. In New York, it's atypical to do any tiime - pre-trial or post-conviction - when the top count is a misdemeanor.

I have to agree. Yes, the judge uses a "rap sheet" with a list of prior crimes and failures to appear for court that can to a degree be quantified and "controlled" so some sort of comparison can be done between potential bailees, but I doubt the probable cause statements can. The cops' descriptions of each potential bailee's crimes contain nuances that signal a person's threat level that I doubt can be controlled, like the degree to which they resisted arrest, seemed remorseful, and a multitude of factors that are often not explicit but emerge from "between the lines."

Most importantly, bail is a physical process of standing before a judge. Usually the judge is in some cinder-block room underneath a jail somewhere with a defendant a few yards away from him, not on a video screen. You are very close to the facial tattoos, disdain for their fellow inmates/defense counsel/judge/jailers, and overall demeanor. You take note when the "crazy eyes" one from the back of the room, comes up to the table with the defense attorney. And I don't think the judge is using some carefully-honed, hard-won expertise. My guess is that human beings have evolved such that we're very, very good at picking out the "problem children" from unspoken physical cues, such that combined with the "tingles" one gets from certain rap sheets and PC statements, that we tend to keep the real baddies locked up in ways that escape a "fully controlled" comparison.

I sometimes wonder if they have themselves convinced of this, or just feel obliged to make the rhetorical claim of the methodology.

In the US the government as prosecutor has unlimited funding while the defense does not.

The District Attorney in just about any county in the United States would be surprised to discover he has no budget or appropriation courtesy the county government and can just write hot checks.

The prosecution never says, "We don't have any more money in our budget to pursue this prostitution case so we'll just let the whore go."

Meanwhile, back in the real world, prosecutors decline to pursue cases as a matter of routine.

Thanks for removing my fingers from his throat.

Prosecutors are as limited in time, money and resources as public defenders. We often had to offer plea in mitigation to violent offenders just so we could handle the homicides, attempted homicides, aggravated batteries, and rapes. This is one reason I grind my teeth when libertarians talk about "non-violent" offenders being released. Many of these cases were bargained down from much more serious charges.

Does anyone listen to criminologists? And if they do, why?

I suspect were you to assemble a bibliography of apposite work by sociologists specializing in the study of crime and published between 1958 and 1994, you'd discover the vast majority incorporated the assumption or the conclusion that the technics of law enforcement and punishment are ineffectual. I suspect if you were to assemble a bibliography of apposite work assessing law enforcement in New York and published after 1997, most would conclude that the observable decline in crime rates was a mirage or had nothing to do with anything done by the NYPD (or that the NYPD was engaged in raaaaacist oppression).

Years ago, Karl Zinsmeister offered this observation about what he'd seen of soldiers and reporters during his time embedded in Iraq: "In junior high school, the reporters were being stuffed into lockers by the sort of guys they're covering". Barry Shain (a dissident academic) has also offered that if you want to understand how arts and sciences faculty look at students, just imagine what the faculty were like when they were 17 and imagine what their students were like.

If the studies are damaged by design flaws, you think the errors are going to be randomly distributed, or are they going to cut in a direction which you'd expect given the broken lenses with which the researchers look at the world?

"In junior high school, the reporters were being stuffed into lockers by the sort of guys they’re covering”

Again Brazilian moral superiority shines through. Brazilian schools have no lockers.

I enjoy the blog, so it was fun to see Alex picking up on this research this morning! It's worth noting that there are several other studies on bail (not all focused on misdemeanors) emerging for the U.S., so were beginning to build a better evidence base on this topic:

To respond to a few of Ray Midge's comments above, unfortunately, in Harris County the bail hearings actually are done via video (and almost always without representation for the defendant) so as to allow for more rapid processing of the large caseload, and a typical hearing lasts only a few minutes. It's not much time for a judge to learn a lot about someone to individualize bail.

And to deal with the problem of controlling for external factors, one thing we do in the paper is exploit what researchers call a natural experiment. People who have hearings on Thursdays look pretty much identical to those with hearings earlier in the week (including in their assigned bail amounts, suggesting that these groups of defendants are being viewed as equally "worthy" of being detained based on whatever information is available to the court). But the Thursday defendants are more likely to actually make bail, because it is easier to get someone to actually show up to bail you out on the weekend, and low income people tend to get paid on Fridays, so they are less liquidity constrained. We show that when you compare Thursday defendants to those from other days, even though they appear otherwise similar, they have better case outcomes, just as you'd expect if being detained causally impairs one's case.

#1. When you receive lower benefits than you are used to, for the same price you normally pay, you are less willing to pay the price.

The benefits to police work are convictions. The price is the effort associated with said police work.

Legal representation means more people are set free. The mix of freed innocents versus freed criminals - though a case can be made that it is primarily innocents who are now set free (low hanging fruit for lawyers) - may not matter as much as the fact that police now receive lower benefits for the same price, making them less willing to pay the price.

The benefits to police work are high wages, medical coverage, uniform allowance, fabulous retirement, easy, clean duties with minimum heavy lifting and the occasional opportunity to fill somebody full of bullet holes. There's a price for any line of work. That's why it's called work.

Alternate mechanism for second study result: Judges are actually somewhat accurate when determining who is dangerous to society, and are less likely to release them on bail or recognizance. And they're not applying purely demographic data to their decisions.

"... We show that when you compare Thursday defendants to those from other days, even though they appear otherwise similar, they have better case outcomes, just as you’d expect if being detained causally impairs one’s case."

I take it what you are getting at here is the notion that this shows that if you are detained, you are more likely going to be found guilty than similarly situated people who are released (at least that's how I interpret "causally impairs one's case"). I don't imagine there's much controversy to the notion that whether one is in custody or out of custody affects whether one takes a plea deal or not. I've heard it is a truism among prosecutors that a case is more likely to resolve in a plea agreement if the defendant is in custody. I think what's assumed here, though, is not that a person's ability to defend themselves at trial is damaged by pre-trial custody, but the result of the psychological fact that people are more likely to accustom their minds to the notion of a stretch of prison/jail, if the have been in jail for a spell already, and those are the people who are thus more likely to take a plea.

If the further claim is being made that of two similarly-situated groups with similar bail imposed, divided only in respect to detained/not-detained (Can't post-bail/can post bail), that the detained get longer jail sentences for the same offenses, well that's a troubling outcome. But here I think again, we get into whether the controlled groups are actually as controlled as we hope. Those that can post bail, I suspect, are more likely to be the sort of people who have a job, save money, and/or have a bail-paying "social support network" (an amorphous phrase often uttered by judges at sentencing) that all play favorably to a sentencing judge. In short, they might possess a lot of factors that directly or indirectly result in a life flight path that plays more favorably to the guy in the black robe deciding how much of a social threat the guy standing in front of him is.

I'll check out your links though.

It seems to me a more interesting subject might be whether a high-pretrial-detention system leads to more or less crime than a low-pre-trial detention system. One might be cheaper than the other, but in the criminal justice system, that question often misses the point. We could save a lot of taxdollars by doing away with the system as a whole (nothing is illegal, and never have to worry about someone's constitutional rights.

"In other words, the right to counsel makes it easier for criminals to escape justice and since the price of crime falls the quantity of crime increases. Makes sense!"

Alternative interpretation: crime has not increased; until defendants were given a basic right to counsel, the police made excessive arrests, probably coerced confessions from innocent people, and prosecutors were more eager to prosecute weak and trumped up cases. Now that they are held accountable, they are more circumspect before they act. The increase in crime is an illusion.

Police tend to sulk when the rules change.

It seems to me that not just the crime rate but how many innocent people are jailed because of the policies should be considered

Both links go to the same article. Could you please correct? Thanks

Here's the correct link

First paper intuitive. Second one not so much.

I suppose it follows from a standard Becker model that reducing the probability of conviction increases crime. This comes from two effects: 1) ineffective deterrent and 2) the criminal is on the streets instead of in prison.

Clearly jail space is costly and limited so there are usually never enough cells for Pre-trial confinement as we would like and are justified by circumstances. As in all things, this is a balance of costs and benefits with the added wrinkles of prohibitions against excessive bail and overcrowding.

In the latter paper there is some reverse causation. Misdemeanants seldom get pre-trial confinement so those who do are the worst of the worst, often a plea agreement from felony charges. Not all of these people have jobs to speak of, so the mechanism of confinement to crime is speculative.

Comments for this post are closed