Skip to main content

Attorney General Eric Holder at the Vera Institute of Justice’s Third Annual Justice Address


United States


Remarks as prepared for delivery.


Laurie, thank you for that wonderful introduction. When I asked Laurie to come back to the Justice Department to lead our Office of Justice Programs, I was keenly aware of how much she would have to give up to join us. Not only did she take leave from the University of Pennsylvania, but she also had to give up her Chair of Vera’s Board of Trustees. I know that the polite thing to do would be to apologize for taking her from you – but the truth is, your loss is our gain. We hope that Laurie stays at the Department for a long time.

It is a privilege to join you this evening as your keynote speaker. Your past speakers have been Nicholas Katzenbach and James Comey, who reflected on the law after their government service. Perhaps one day I might have that kind of conversation with you as a former Attorney General. For now, I stand before you in a different posture, to share some ideas about how I think the American people can best be served by the Department of Justice going forward.

The Vera Institute of Justice has been an extraordinary partner to government in the administration of justice. I thank you in particular for your work with the federal government across a range of issues – from your contributions to the national commission to eliminate prison rape to the administration of Legal Orientation Programs for non-citizens in immigration proceedings. Your practical, rational, data-driven, results-oriented approach can best be described as post-partisan. In the five months that I have served as Attorney General, I have tried to take that same approach, and that is what I would like to talk about this evening: how we can move past politics and ideology in order to get smart on crime.

Getting smart on crime requires talking honestly about which policies have worked and which have not, without fear of being labeled as too hard or, more likely, as too soft on crime. Getting smart on crime means moving beyond useless labels and instead embracing science and data, and relying on them to shape policy. And it means thinking about crime in context – not just reacting to the criminal act, but developing the government’s ability to enhance public safety before the crime is committed and after the former offender is returned to society.

It is imperative that we get smart on crime now, for much has changed since some of our basic, governing assumptions about criminal law enforcement were developed. In the middle years of the twentieth century, America went through an historic increase in crime and illegal drug use. In the 1960s and 70s, the overall crime rate increased more than five-fold. Violent crime nearly quadrupled. The murder rate doubled. And heroin, cocaine and other illegal drug use surged.

Many lawmakers in the 1980s responded by declaring, in rhetoric and in legislation, that we needed to get tough on crime. States passed truth-in-sentencing and three strikes and you’re out laws. Some state parole boards became more cautious, while other states eliminated discretionary parole altogether. The federal government adopted severe mandatory minimum sentencing laws, eliminated parole, and developed the federal sentencing guidelines.

The federal government and states spent billions of dollars in new prison construction. The result was dramatic: the number of inmates in American prisons has increased seven-fold since 1970. Today, one out of every 100 adults in America is incarcerated – the highest incarceration rate in the world.

Few would dispute that public safety requires incarceration, and that imprisonment is at least partially responsible for the dramatic drop in crime rates nationwide in recent decades. By 2007, the nation’s violent crime rate had dropped by almost 40% from its peak in 1991. But just as everyone should concede that incarceration is part of the answer, everyone should also concede that it is not the whole answer. Simply stated, imprisonment is not a complete strategy for criminal law enforcement.

To begin with, high rates of incarceration have tremendous social costs. And, of course, there also is the matter of simple dollars and cents, and the principle of diminishing marginal returns. Every state in the union is trying to trim budgets. States and localities are laying off teachers and canceling sanitation department shifts, but in almost all cases, spending on prisons continues to increase. Not only is this unsustainable economically, but it is also not proving to be effective at fighting crime. For while prison building and prison spending continue to increase, public safety is not improving. Since 2003, spending on incarceration has continued to rise, but crime rates have flattened. Indeed, crime rates appear to have reached a plateau, and no longer respond to increases in incarceration.

So what can we do to lower the crime rate further, to make American communities safer, to get smarter on crime? We need new tools – and one way to develop new tools is to look several steps past getting people into prison, and to consider what happens to people after they leave prison and reenter society.

We know that offenders who have participated in the federal Bureau of Prisons’ residential drug abuse treatment program are 16% less likely to be re-arrested, have their supervision revoked, and be returned to prison, than similar inmates who did not receive such treatment before their reentry into society. They are also less likely to use drugs once released. We also know that inmates who work in prison industries are 24% less likely to commit crimes again, compared to inmates who have not participated in such programs – which, incidentally, operate at no cost to the taxpayer. The Bureau of Prisons’ educational programs designed to address educational deficiencies – ranging from Adult Basic Education to high school level classes – are also effective in reducing recidivism: inmates who participate in these programs are 16% less likely to commit crime again as compared to their non-participating peers. And inmates who are released through halfway houses are more likely to be gainfully employed, and therefore less likely to commit crime again, as compared to inmates who are released from prison directly to the community.

That recitation of statistics might not sound exciting, but what we do with it is. We rely upon evidence-based methods to innovate in agriculture, transportation, environmental safety, and public health – and it is my belief, that the Department of Justice likewise should embrace modern, evidence-based methods for developing policy.

In particular, it is critical that we work to develop policies – rooted in data – to address what happens after incarceration. For the statistics I cited are even more compelling when coupled with another fact: most crimes in America are committed by persons who have committed crime before. About 67% of former state prisoners and 40% of former federal prisoners are rearrested within three years. Logically, if we reduce the recidivism rate, we will directly lower the crime rate. Even a modest reduction in recidivism rates would prevent thousands of crimes and save hundreds of millions of taxpayer dollars. In other words, being smart on crime means understanding that our work does not end when prison time begins.

Smart risk assessments can identify which offenders can safely remain in their communities and which require continued detention and more intensive supervision. Data analysis can determine which offenders pose a higher recidivism risk based on the type of crime the offender was charged with and the offender’s prior record. For example, risk assessments might determine that removing a 16-year-old, non-violent, first-time offender from his family and school and placing him in a juvenile detention facility is a bad idea because it would actually increase the risk of recidivism, and waste taxpayer dollars besides.

One specific area where I think we can do a much better job by looking beyond incarceration is in the way we deal with non-violent drug offenses. We know that people convicted of drug possession or the sales of small amounts of drugs comprise a significant portion of the prison population. Indeed, in my thirty years in law enforcement, I have seen far too many young people lose their claim to a future by committing non-violent drug crimes.

One promising, viable solution to the devastating effect of drugs on the criminal justice system and on American communities is the implementation of more drug treatment courts. Drug court programs provide an alternative to incarceration for non-violent offenders by focusing on treatment of their underlying addiction. Program participants are placed in treatment and routinely tested for drug use – with the imposition of immediate sanctions for positive tests balanced with suitable incentives to encourage abstinence from drug use. These programs give no one a free pass. They are strict and can be extraordinarily difficult to get through. But for those who succeed, there is the real prospect of a productive future.

New York has been a leader in this area, diverting some non-violent offenders into drug court programs and away from prison, and extending early release to other non-violent offenders who participate in treatment programs. And while national prison populations have consistently increased, in New York the state prison population has dropped steadily and has 12,000 fewer inmates now than it did in 1999. And since 1999, the overall crime rate in New York has dropped 27%. Other states have followed New York’s example. And most importantly, studies show significant reductions in re-arrests, from about 15 to 30 percentage points, for drug-court participants as compared to criminals simply incarcerated.

Furthermore, smart criminal justice policies are not, of course, exclusively reactive – we can also use data and evidence-based methods to prevent crime before it occurs. We have models, for example, in New York’s CompStat program: it uses data to map where crime is most likely to occur, deploy police to those areas to disrupt criminal activity, and evaluate the effectiveness of the enforcement strategies. We can also extrapolate from available data to identify youth that are highly at-risk to commit crimes in the future. For example, it seems that children who are exposed to domestic violence at home are more at-risk. Once we have identified at-risk youth, we can intervene with targeted programs, and I have asked the Department to make a priority of focusing on the issue of children exposed to violence. There is much work to be done in this area, but the underlying premise is already clear: we need to understand crime in context in order to prevent it – and with better understanding and more information, we can develop new approaches to old and seemingly intractable problems.

Although this Administration is still young, we have already started to put into practice what I believe is a data-driven, non-ideological, post-partisan approach to crime. For example, I have asked attorneys throughout the Department to conduct a comprehensive, evidence-based review of federal sentencing and corrections policy. Specifically, the group is examining the federal sentencing guidelines, the Department’s charging and sentencing advocacy practices, mandatory minimums, crack/powder cocaine sentencing disparities, and other racial and ethnic disparities in sentencing. The group is also studying alternatives to incarceration and strategies that help reduce recidivism when former offenders reenter society. We intend to use the group’s findings as a springboard for recommending new legislation that will reform the structure of federal sentencing.

I have also called upon the Department to focus on another part of the criminal justice system: the very difficult issue of indigent defense. Putting politics aside, we must address the fact that, simply put, there is a crisis in indigent defense in this country. Resources for public defender programs lag far behind other justice system programs, constituting only about 3 percent of all criminal justice expenditures in our nation’s largest counties. In many cases, contract attorneys and assigned lawyers receive compensation that does not even cover their overhead. We know that defenders in many jurisdictions carry huge caseloads that make it difficult for them to fulfill their legal and ethical responsibilities to their clients. We hear of lawyers who cannot interview their clients properly, file appropriate motions, conduct fact investigations, or do many of the other things an attorney should be able to do as a matter of course.

This growing crisis is troubling not just because of the government’s constitutional duty to ensure the right to counsel. When defendants fail to receive competent legal representation, their cases are vulnerable to costly mistakes that can take a long time to correct. Lawyers on both sides can spend years dealing with appeals arising from technical infractions and procedural errors. When that happens, no one wins. Addressing the American Council of Chief Defenders last month, I committed to several steps to help improve the indigent defense system, including hosting a national conference with the goal of developing a set of best practices and practical solutions.

I have also made it clear that this Department of Justice will use the available data to improve our handling of the forensics sciences – such as fingerprints, trace evidence, firearms matching. We are studying a recent report from the National Academies of Science that diagnosed problems in the use of forensics sciences and suggested ways forward, and we are working with our partners in the Executive Branch and Congress to act on the report’s insights and recommendations. Our goal is to ensure that forensic science is practiced at the highest level possible, and always in the pursuit of truth. Because we put a premium on truth-seeking – because, indeed, this Administration is committed to using the best science possible whenever possible, including in criminal justice – I also believe that defendants should have access to DNA evidence in a range of circumstances. DNA testing has an unparalleled ability to exonerate the wrongfully convicted as well as to identify the guilty. Federal law already guarantees access to DNA evidence held by the federal government under specific conditions, and I hope that all states will follow the federal government’s lead on this issue.

Many of the things I have mentioned in these remarks are still in early stages or under review. There are numerous areas I have not even mentioned – for example, the prevention and detection of economic crimes and on-line crimes – where we can similarly get smarter with a research-driven approach. I am already certain, however, that change is both necessary and it is possible – if we are willing to make it. Challenges have changed with time: Prison populations are at an all-time high and still climbing, yet the crime rate is no longer declining. States are in serious financial distress. But opportunities have changed too. We are able to compare the cost and suitability of different criminal justice strategies. We no longer must choose between more crime and more prisons: we can reduce crime rates and reduce our dependence on incarceration, and at the same time increase the integrity of our criminal justice system. We can harness science and data to tackle emerging problems and also to preserve our foundational principles. The more we know, the better we can do, the more sophisticated we can be. With the help of the scholars and experts in this room, state and local law enforcement, corrections officials across the country, judges, victims of crime, and always with the fine work of attorneys at the Department of Justice, there is no question that a smarter and better criminal justice system is within our grasp.

Thank you very much.


Updated August 20, 2015