Moneyballing Criminal Justice?

I recently learned about TED Talks.  In short, these are presentations by people who have “ideas worth spreading”.  If interested, you can learn more about TED Talks at TED.com.

Anne Milgram, the former New Jersey Attorney General, gave a TED Talk where she discusses what she calls “moneyballing criminal justice”.  In brief, Ms. Milgram wants to focus more efforts upon collecting statistical data surrounding arrests, convictions, and other aspects of the criminal justice process, and use it to make better decisions concerning basic issues, including whether to release or incarcerate someone after they have been arrested for a particular offense.  Ms. Milgram notes correctly that many judges rely upon instinct and experience when making this important decision, and that such decisions are therefore based upon subjective reasoning and are frequently incorrect.  She wants to develop a universal risk assessment tool that State and Federal judges can use to guide their decision making as to the “release or incarcerate” decision, which is one of the first major issues addressed in any criminal matter.  She likens her efforts to those used by baseball teams that selected good players using statistical analysis, as opposed to the purely subjective observations and instincts of scouts.

Many of Ms. Milgram’s observations are absolutely correct.  First, the only surprise relating to the fact that we spend $75 Billion annually on criminal justice is that the number seems small.  I always thought it was much larger.  Prisons and jails in this country are an industry unto themselves.  Next, our recidivism rate is obscenely high, leading to the obvious conclusion that a lot of this money is wasted.  Indeed, it often seems that the only thing we accomplish by the end of the case is to temporarily disable a defendant from having another encounter with the system.  We seem to focus more on “processing” defendants through the system, with no eye on long-term solutions to the problems associated with criminal conduct.  Additionally, Ms. Milgram notes correctly that there is astonishingly little helpful statistical data describing the operations of our system.  Interestingly, she failed to note that there is a federal agency called the Bureau of Justice Statistics that collects statistical data concerning various criminal justice issues.  However, the manner in which BJS identifies the subjects concerning which it will collect and analyze data has always been a mystery to me.  Further, many of the reports available on the BJS website are dated to the point where they are probably no longer valid.  Moreover, even if the material is helpful, I don’t know too many judges who make a habit of reading statistical reports from the BJS website.

Finally, Ms. Milgram correctly notes that the vast majority of criminal defendants are in the system because of relatively low-level (drug) offenses and other non-violent crimes, that we currently devote huge resources to address “small” cases which stem from non-violent conduct, and that such cases could probably be addressed more efficiently.  As to this group of offenders, the problem manifests itself almost immediately when the defendant is incarcerated in a local facility at taxpayer expense, and must remain there while their case is pending because they cannot make bail.  Releasing these defendants would save taxpayer dollars, as well as foster a more efficient criminal justice system that could more effectively utilize scarce resources.

Ms. Milgram’s observation that defendants with small drug cases and non-violent petty crimes make up the vast majority of our offender population is absolutely true.  The overwhelming majority of criminal defendants are, in fact, in the system because of relatively small offenses.  These offenders consume significant resources that could be more effectively used to address the issues created by more violent offenders.  Clearly, the system as a whole would undoubtedly benefit if non-violent offenders with petty offenses were released either on their own recognizance or with a nominal bail, as opposed to being detained in a county facility at taxpayer expense.  However, her notion that it is possible to create a universal risk assessment tool to guide the “release or incarcerate” decision seems far fetched.  Ms. Milgram states that this risk assessment tool is being used in Kentucky, apparently with positive results.  However, the fact that such a tool works in Kentucky does not mean that it will work equally well in New York City, Newark (New Jersey), Chicago, or Los Angeles.  A universal risk assessment tool simply fails to recognize that different environments have different needs.  Ms. Milgram indicates that judges would still rely upon instinct and experience along with the risk assessment tool, but numbers with their apparent certainty can be highly seductive, with the result that the facts and circumstances of the individual case could easily get lost in the scoring process.  Ms. Milgram’s proposal amounts to a “one-size-fits-all” approach to criminal justice, which clearly will not work in a country as diverse as ours.

Further, efforts to develop risk assessment tools in other contexts have failed.  New Jersey (and probably other jurisdictions) use a risk assessment tool to score convicted sex offenders for purposes of tier assignment and level of parole/community supervision.  New Jersey rates sex offenders subject to supervision in 13 categories.  The score determines whether they are placed in tier one (low risk/low level supervision), tier two (medium risk/medium level supervision), or tier three (high risk/most intense supervision).  There is absolutely no data demonstrating that this system provides an effective mechanism for making decision concerning the supervision of sex offenders living in the community.  In fact, it is my understanding that the available research demonstrates that this framework accomplishes nothing, and is a complete waste of resources.  At least in this analogous context, the “score sheet” approach has accomplished little, if anything.

I would suggest that a more effective means of relieving the criminal justice system of the immense burden created by low-level offenders would be the implementation of drug court and mental health court programs.  We have at least some data indicating that the vast majority of low-level non-violent offenders suffer from drug and/or alcohol addiction, as well as mental health issues.  Diverting these cases away from the mainstream criminal justice system to specialized courts designed to provide these defendants with needed treatment relieves the system of overcrowding, reduces recidivism, and allows valuable resources to be used to address the issues created by more violent offenders.  Moneyballing works for corporations trying to increase profits, or for baseball teams trying to win games.  But a criminal case is not a game, and should not be viewed in the same light.