Pretrial detention

After someone is arrested, they might wait weeks or months for their case to be resolved, and they might have to wait in jail. On any given day in the United States, about half a million people are detained pretrial — that is, incarcerated before they are convicted or sentenced. In theory, pretrial detention ensures that defendants do not commit any additional offenses pre-disposition, including failing to appear for their legal proceedings. But does our use of pretrial detention make communities safer? If not, what should we do? Which policy reforms would meaningfully reduce the number of people detained pretrial?

In this survey, we asked the Criminal Justice Expert Panel to consider three statements about pretrial detention. Their responses are below.

Reducing the number of people detained pretrial will lead to a net increase in crime in the medium- to long-term.

View Individual Responses

Ending the use of cash bail will meaningfully reduce the number of people detained pretrial.

View Individual Responses

Using risk assessment to inform detention decisions will meaningfully reduce the number of people detained pretrial.

View Individual Responses

Responses

Reducing the number of people detained pretrial will lead to a net increase in crime in the medium- to long-term. - participant responses

Participant Vote Confidence Comment
John Donohue Stanford University Neutral/No Opinion 5 Pretrial detention incapacitates in the short run and ensures defendants show up for trial but I suspect it is not pursued in an effective way in that lots of low-risk individuals are locked up needlessly and the resources thereby wasted could be diverted to effective crime reduction strategies. I would need to know more about the costs and benefits of pretrial detention and the manner of inducing less detention to opine on whether a reduction will be benign or harmful.
David Kirk University of Oxford Strongly Disagree 8 Focus on medium to long-term is key here. Pretrial detention will have an incapacitation effect in the short-term, but detrimental consequences on employment and earnings in the future. Individuals detained pre-trial are more likely to plead guilty and face the collateral consequences of a conviction, increasing the risks for criminal activity long-term.
Alex Tabarrok George Mason University Neutral/No Opinion 6
Randi Hjalmarsson University of Gothenberg Disagree 8 Research suggests that pre-trial detention has significant labor market scarring and criminogenic effects in the medium term; these harmful effects of pretrial detention are likely to outweigh the potential short run increases (in relatively minor offenses) if individuals are not incapacitated in pre-trial detention.
Michael Makowsky Clemson University Strongly Disagree 9 The marginal individuals released from pre-trial detention from all but the most extreme policy shifts would have a negligible effect on crime.
Santiago Tobón Universidad EAFIT Disagree 9 Most evidence suggests harsh prison conditions increase recidivism. In most of the world prisons are overcrowded and over-use of pretrial detention only increases overcrowding. Hence, in most circumstances, reducing the number of people detained pre-trail might actually reduce crime. The exceptions could be those cases with a large excess capacity in prisons/jails
Paul Heaton University of Pennsylvania Disagree 8 Although the evidence from good causal studies is mixed, with some finding a zero net effect of pretrial detention in the medium to long term, and some finding that detention actually increases crime once we get beyond the short-run incapacitation stage, studies are consistent in showing the detention doesn't decrease crime.
Megan Stevenson University of Virginia Disagree 6 It may shift some crime from within jails to outside of jails, however.
David Weisburd George Mason University Neutral/No Opinion 1
Jamein Cunningham University of Memphis Neutral/No Opinion 1 This is tough. There is a cost/benefit analysis that I'm just not sure about. On one part, the alleged marginal offender faces huge cost from pretrial detention. This cost is borne before any guilt or innocence is established. And it is possible that the cost incentivizes future criminal activity. However, without any changes in the alleged offender circumstances, it is not clear that we should expect (on net) crime to decrease in the medium-to-long-run.
Charles Loeffler University of Pennsylvania Disagree 7 The best available evidence on this question is from studies of misdemeanor defendants. These studies suggest that crime will not increase as more affected defendants are released.
Michael Stoll University of California - Los Angeles Strongly Disagree 9
David Harding University of California - Berkeley Strongly Disagree 9
Giovanni Mastrobuoni Collegio Carlo Alberto Neutral/No Opinion 1 That depends on whether the marginal arrestee who is put in pretrial detention is eventually sentenced to prison time or not.
John MacDonald Disagree 10 If the policy is across the board release without monitoring in the community then I would expect crime to increase. If the policy is to release low level offenders and/or provide sufficient monitoring in the community then I think the evidence is that one would not expect crime to increase.
Monica Deza City University of New York Disagree 5 I believe the effects would be very short-term and potentially small.
Jeffrey Grogger University of Chicago Neutral/No Opinion 10 It depends on how it's done. Recent research suggests that there is a subset of defendants who could be released with no adverse effect on offending. However, blanket release policies would probably release others as well.
Anna Bindler University of Cologne Disagree 8 The existing evidence suggests that the rates of pretrial detention may not be justified by the average reduction in the risk of crime. However, it is possible that pretrial detention leads to general deterrence effects that should be taken into account.
Jennifer Doleac Texas A&M University Disagree 7 There's solid evidence that scaling back pretrial detention would reduce future offending by those who are released, over the medium-term.
Aaron Chalfin University of Pennsylvania Disagree 6 Two important papers (Dobbie, Goldin & Yang; Leslie & Pope) find that pre-trial detention incapacitates offenders and increases public safety in the short run but increases re-offending in the longer-run with offsetting effects over a medium-term horizon (2 years). A recent paper on non-prosecution by Agan, Doleac and Harvey also suggests that a net increase in crime isn't likely. That said, there may be subpopulations (eg, gun arrestees) for whom effects could be quite different.
Manisha Shah University of California - Los Angeles Disagree 6
Stephen Billings University of Colorado - Boulder Disagree 5
John Pepper University of Virginia Disagree 9 There is very little credible evidence.
Morgan Williams, Jr. New York University Neutral/No Opinion 7
Anna Harvey New York University Disagree 8 Several studies report that pretrial detention (mechanically) decreases reoffending pre-disposition, but increases reoffending post-disposition for marginal defendants, with no net effect on offending (Gupta et al 2016, Leslie and Pope 2017, Heaton et al 2017, Dobbie et al 2018). Reducing pretrial detention would not increase crime in the medium- to long-term if the latter effect remains at least as large as the former effect.
Patrick Sharkey Princeton University Neutral/No Opinion 5 I haven't seen strong evidence on this question. I consider the legal and moral questions about holding people in jail without being convicted, and the logistical challenge of reducing the number of people held pretrial, to be more important questions.
Rosanna Smart RAND Corporation Strongly Disagree 8 Several rigorous studies have found pretrial detention to have negative consequences for those detained, and little evidence that pretrial release leads to higher likelihood of re-arrest, subsequent crime, etc. – with the exception of increased pretrial crime and failure to appear (e.g., Dobbie et al., 2018), which I am considering short-run effects.
Robynn Cox University of Southern California Strongly Disagree 9 Detaining people pretrial, especially for non-violent offenses, has many costs and may lead to greater crime in the long run due to the costs of being detained (e.g., loss of employment, loss of housing, family separation, etc.).
Paolo Pinotti Bocconi University Disagree 5
J.J. Prescott University of Michigan Disagree 8
Peter Reuter University of Maryland Disagree 5
Robert Apel Rutgers University Disagree 8 The claim that pretrial detention as used in practice has mostly (or anything) to do with public safety is laughable. If there is any medium- to long-term impact of a reduction in pretrial detention, I predict it might actually reduce crime considering all the negative collateral effects of even short custody spells. And evidence is growing that offending worsens after pretrial detention and outweighs any incapacitation benefit.
Beau Kilmer RAND Corporation Disagree 7
Kevin Schnepel Simon Fraser University Disagree 8 While there may be more crimes in between arrest and trial without pretrial detention, a medium- to long-term decrease in crime among those no longer detained pretrial (through less disruption in employment, less exposure to the carceral environment) likely outweighs any increase during the pretrial period.
Felipe Goncalves University of California, Los Angeles Disagree 2 The crime effect of reducing pre-trial detention will depend on which defendants are affected, and even then it is hard to know. While some individuals released pre-trial commit offenses while awaiting their trial, the experience of pre-trial detention may itself lead to future criminal behavior (in addition to its other damaging effects). So it is unclear what the net impact on crime would be.
Benjamin Hansen University of Oregon Neutral/No Opinion 5 There are several pathways to consider. Of course it's worth noting pre-trial detention is justified legally on the likelihood someone will appear for trial, not on their future disposition to be commit crime (at this point they are presumed innocent). That said, if fewer people are detained, likely fewer of them will be convicted which may reduce their recidivism (there's some evidence on this). However there are also incapacitation effects and spillovers onto peers/siblings/kids.
Ariel White Massachusetts Institute of Technology Disagree 8
Emily Owens University of California - Irvine Disagree 8 I don't think there is good evidence that this happens- rather than shift some offending around in a temporal sense.
David Abrams University of Pennsylvania Disagree 6
Greg Midgette University of Maryland Disagree 4 Some research suggests pretrial detention is associated with higher rates of recidivism later, though I'm not aware of a big enough evidence base to suggest this finding generalizes broadly.
Manudeep Bhuller University of Oslo Disagree 5
Shawn Bushway RAND Corporation Disagree 7

Ending the use of cash bail will meaningfully reduce the number of people detained pretrial. - participant responses

Participant Vote Confidence Comment
John Donohue Stanford University Agree 8
David Kirk University of Oxford Agree 8
Alex Tabarrok George Mason University Strongly Disagree 7
Randi Hjalmarsson University of Gothenberg Agree 6
Michael Makowsky Clemson University Strongly Agree 8
Santiago Tobón Universidad EAFIT Neutral/No Opinion 5
Paul Heaton University of Pennsylvania Agree 5 This really depends on what process to determine who gets detained is used in place of cash bail, but the evidence from places like New Jersey suggests that eliminating bail in real-world settings would reduce detention.
Megan Stevenson University of Virginia Agree 6 Likely, but it depends on what replaces it…
David Weisburd George Mason University Neutral/No Opinion 1
Jamein Cunningham University of Memphis Agree 9
Charles Loeffler University of Pennsylvania Disagree 7 For the often affected population of misdemeanor defendants, decreases in cash bail have mostly led to increases in release without the need to make financial payments at time of release. These changes often impact individuals required to pay relatively small amounts. Since many of these individuals eventually made payment under the previous cash bail system, the overall population detained does not experience a large change in size (e.g., Philadelphia).
Michael Stoll University of California - Los Angeles Agree 8
David Harding University of California - Berkeley Agree 7
Giovanni Mastrobuoni Collegio Carlo Alberto Neutral/No Opinion 1 I know little about the US cash bail system, but if judges use cash bails to discipline defendants the number may also increase.
John MacDonald Strongly Agree 10 Way too many people are held in jail because they are poor and cannot afford cash bail.
Monica Deza City University of New York Disagree 5
Jeffrey Grogger University of Chicago Agree 9 This is consistent with an abundance of evidence.
Anna Bindler University of Cologne Neutral/No Opinion 5 The response may very much depend on the counterfactual - what is cash bail substituted with.
Jennifer Doleac Texas A&M University Neutral/No Opinion 7 If judges are setting bail in order to detain people, they might simply detain the same people when cash bail is not an intermediate step. There is some recent evidence from Philadelphia that reducing the use of cash bail had on effect on detention rates (Ouss & Stevenson, 2021).
Aaron Chalfin University of Pennsylvania Agree 5 Recent work by Ouss and Stevenson suggests that when prosecutors stop asking for bail, it can make a big difference. But the answer to this question is likely to be informed by the changing nature of politics and the institutional details of CJ reform. What options does the judge have? What replaces cash bail? Is bail reform accompanied by other reforms? Cash bail is a tool that judges often use to detain defendants but, depending on what is allowed, judges could simply remand more defendants.
Manisha Shah University of California - Los Angeles Agree 5 Also depends on policy that replaces cash bail.
Stephen Billings University of Colorado - Boulder Agree 9
John Pepper University of Virginia Agree 7
Morgan Williams, Jr. New York University Agree 5
Anna Harvey New York University Neutral/No Opinion 5 Ouss and Stevenson (2021) report that ending low cash bail requests had no effect on pretrial detention in Philadelphia, because those affected by the policy were generally able to make bail prior to the policy's implementation. Judges may also offset a broader prohibition on cash bail by increasing their use of detention orders, with unclear effects on the number of people detained pretrial.
Patrick Sharkey Princeton University Agree 7
Rosanna Smart RAND Corporation Agree 7 This would depend on what mechanism, if any, is put in places as a replacement to cash bail. However, looking at states and jurisdictions that have reduced or ended the use of cash bail (e.g., New Jersey, Philadelphia), these policies do seem to reduce the number of people detained pretrial (although I have some uncertainty around the magnitude that can be causally attributed to the end of cash bail).
Robynn Cox University of Southern California Agree 7 In theory it could but it depends on implementation and what the cash system is replaced with. DC was an early reformer and has had lots of success.
Paolo Pinotti Bocconi University Disagree 5
J.J. Prescott University of Michigan Agree 8
Peter Reuter University of Maryland Agree 5
Robert Apel Rutgers University Strongly Agree 8 Elimination of cash bail is a much-needed reform since the system amounts to punishment for poverty, and informational nudges can be used effectively to ensure appearance at future court proceedings. Carve outs for certain offenses are still overly broad, in my view, and should be restricted. An even more impactful reform would be to eliminate bookings altogether for a large class of offenses.
Beau Kilmer RAND Corporation Agree 6
Kevin Schnepel Simon Fraser University Neutral/No Opinion 5 I think it all depends on what, if anything, replaces cash bail. Would there be other requirements that may be hard for individuals to meet?
Felipe Goncalves University of California, Los Angeles Disagree 5
Benjamin Hansen University of Oregon Agree 7 Many people detained ahead of trial lack the resources to pay bail. Moving away from cash bail and instead to a different risk based system could reduce how many are detained. At the same time, it just change who is detained.
Ariel White Massachusetts Institute of Technology Agree 4 I'm interpreting this question as being about a New Jersey- or Illinois-style legislative change to the bail system, not simply changes in local prosecutorial policies for requesting bail.
Emily Owens University of California - Irvine Neutral/No Opinion 8 This depends on what would be replacing cash bail. There are many possible options.
David Abrams University of Pennsylvania Strongly Agree 9
Greg Midgette University of Maryland Agree 7 Among people with identical cash bail amounts--so ostensibly similar perceived restaurant risk--those from poorer areas are more likely to be detained. It's not crystal clear to me that replacing cash bail with another policy will lead to lower rates of detention or will change the characteristics of those who are detained based on the new heuristic a judge would adopt.
Manudeep Bhuller University of Oslo Disagree 8
Shawn Bushway RAND Corporation Agree 8 The answer depends in part on how aggressively cash bail is currently being used, and the exact structure of the new system replacing cash bail.

Using risk assessment to inform detention decisions will meaningfully reduce the number of people detained pretrial. - participant responses

Participant Vote Confidence Comment
John Donohue Stanford University Agree 8
David Kirk University of Oxford Neutral/No Opinion 3
Alex Tabarrok George Mason University Strongly Disagree 7
Randi Hjalmarsson University of Gothenberg Neutral/No Opinion 5 This likely depends on (i) the goal of the risk assessment (to reduce number detained? to reduce disparities in who is detained?) and (ii) whether judges or bail officers use significant discretion in deviating from the guidelines provided in the risk assessment tool.
Michael Makowsky Clemson University Strongly Agree 8 Given the risk aversion of judges, any exogenous assessment of risk would reduce their personal and reputational risk of releasing individuals pre-trial.
Santiago Tobón Universidad EAFIT Neutral/No Opinion 8 It depends on how the assessment tools are implemented and which are the cutoffs that prosecutors and judges use to decide whether or not to send someone to prison
Paul Heaton University of Pennsylvania Agree 3 Much depends on implementation--some versions of risk assessment could increase detention, and others judges would ignore. However, one advantage of risk assessment may be political rather than technical--by addressing fears about public safety, it can allow people otherwise unwilling to support reductions in detention to do so
Megan Stevenson University of Virginia Disagree 9 Little evidence that it has done so so far.
David Weisburd George Mason University Agree 7
Jamein Cunningham University of Memphis Agree 7 My concern is related to how these risk assessment may incorporate preexisting biases and further exacerbate racial disparities in the criminal justice systems. In my opinion, risk assessments should be part of a multifaceted approach to reduce the number of people detained pretrial.
Charles Loeffler University of Pennsylvania Neutral/No Opinion 8 The utility of risk assessment for reducing the pretrial detention population depends on how release thresholds are set and how any such tool is integrated into existing pretrial release procedures. In practice, the implementation of risk assessment often serves as an opportunity to reconsider the utilization of pretrial detention. However, the use of risk assessment itself does not guarantee any specific impact on the size of the pretrial population.
Michael Stoll University of California - Los Angeles Agree 8
David Harding University of California - Berkeley Neutral/No Opinion 5
Giovanni Mastrobuoni Collegio Carlo Alberto Neutral/No Opinion 1 More information will certainly be useful to pinpoint the marginal arrestee who should be put in pretrail detention, but changing the threshold may also lead to an increase in people detained pretrial.
John MacDonald Strongly Agree 8 This agreement is contingent on simple risk assessments that focus on low-risk offenders and magistrates or judges that actually follow the guidelines.
Monica Deza City University of New York Agree 5
Jeffrey Grogger University of Chicago Neutral/No Opinion 5
Anna Bindler University of Cologne Disagree 6 The existing research suggests that using risk assessment to inform detention may overcome errors in pretrial decisions and reduce the number of people detained pretrial. However, this will depend on the extent to which risk assessments feature into judicial decisions; so far, the evidence suggests that judges may override the suggestions of algorithms.
Jennifer Doleac Texas A&M University Neutral/No Opinion 7 While risk assessments could push judges in this direction, there is little evidence that it's having this effect in practice. This is likely because risk assessments simply provide information and a recommended action; judges are not required to follow the recommendation (and often don't).
Aaron Chalfin University of Pennsylvania Neutral/No Opinion 1 Tough question. Kleinberg et al shows us that algorithms could be deployed to either shrink pre-trial detention or improve public safety, depending upon the normative goal of the social planner. Stevenson and Doleac finds that judges' positions are shifted by risk scores, though maybe less so over time. So a lot depends on (a) how actuarially fair a risk tool is and (b) whether the ultimate policy goal is to shrink jail populations or improve safety. I don't know how this will play out.
Manisha Shah University of California - Los Angeles Agree 5 This depends on how good of a job the risk assessment does predicting risk!
Stephen Billings University of Colorado - Boulder Neutral/No Opinion 3
John Pepper University of Virginia Neutral/No Opinion 1 I think this is a question for policymakers.
Morgan Williams, Jr. New York University Agree 4
Anna Harvey New York University Neutral/No Opinion 5 A really well-designed risk assessment tool that was implemented faithfully might be able to reduce the numbers of people detained pretrial (e.g., Kleinberg et al 2018 QJE), but it's unlikely that most risk assessment tools are either that well-designed or are implemented faithfully (Stevenson and Doleac 2021).
Patrick Sharkey Princeton University Neutral/No Opinion 5
Rosanna Smart RAND Corporation Agree 4 While the use of risk assessment tools to make pretrial detention decisions certainly has the potential to reduce the number of people detained, again these effects would depend on the algorithm/tool used (e.g., purpose, inputs, etc.), the extent to which directives for classifying risk levels are designed to reduce detention, the extent to which the application of such a tool actually affects pretrial decisions (eg, whether tool recommendations are consistently followed [Stevenson, 2018]), etc.
Robynn Cox University of Southern California Strongly Disagree 9 There are lots of problems with risk assessments. They have baked in racial and class bias. It’s not clear that they have meaningful reduced the number of people detained pre-trial and chances are they will increase racial disparities in pre-trial detention. Of course, criminal legal system actors like it because they let them off the hook politically for bail decisions.
Paolo Pinotti Bocconi University Agree 5
J.J. Prescott University of Michigan Neutral/No Opinion 10
Peter Reuter University of Maryland Disagree 5
Robert Apel Rutgers University Agree 6 In the short term, yes, and the reduction can be sizable. But whether this can be sustained is unclear, as in some jurisdictions, pretrial detentions exhibit mild reversion to trend following implementation of a risk assessment regime. This suggests a need for recalibration over time. Reduction in racial disparity also needs to be a stated goal of these efforts.
Beau Kilmer RAND Corporation Neutral/No Opinion 5
Kevin Schnepel Simon Fraser University Disagree 6 It could go in either direction depending on how risk assessment is done and where thresholds for detention decisions are set.
Felipe Goncalves University of California, Los Angeles Disagree 5
Benjamin Hansen University of Oregon Neutral/No Opinion 5 It's not clear how courts, judges or states are willing to risk assessments. Is this just a piece of information that can use? Or does this actually replace judges? Moreover, the judicial system actually could just change it's variability, where instead of judges making decisions, algorithms could simply decide to keep the riskiest people detained, subject to bed availability. This would ensure the pretrial jails are always full and by so doing increase pretrial detention.
Ariel White Massachusetts Institute of Technology Neutral/No Opinion 5 There is simulation work that suggests risk assessment tools could theoretically reduce the number of people detained pretrial, but that doesn't mean that as actually implemented they will do so. See Stevenson 2018, for example, on Kentucky's use of these tools.
Emily Owens University of California - Irvine Strongly Disagree 9 Just because a jurisdiction is using "a risk assessment tool" doesn't tell me what that assessment is based on, or how it is being implemented.
David Abrams University of Pennsylvania Agree 4
Greg Midgette University of Maryland Neutral/No Opinion 6 Judges and community corrections agents deviate from actions recommended by assessments pretty often, so it's tough to know the degree to which detention rates are driven by risk scores versus other inputs to judges' decisions.
Manudeep Bhuller University of Oslo Agree 8
Shawn Bushway RAND Corporation Strongly Disagree 9 1) Research shows quite clearly that risk assessment alone doesn't change anything if magistrates still retain ultimate discretion. 2)The risk assessment tool only rank orders risk. Policy makers must still decide where to put the cutoffs for detention.