Smart Sentencing

Sentencing for Public Safety and Harm Reduction

last updated April 17, 2008

     I am an Oregon trial judge determined to aim the sentencing process at crime reduction through evidence-based sentencing. Our existing archaic approach produces outrageously high recidivism rates, enormous waste and avoidable victimizations. Most offenders sentenced for most crimes offend again; most offenders sentenced for serious, violent crime have been sentenced before with no responsible attempt to prevent future crimes by that offender.
     Judges have a wide range of discretion in many sentencing and probation violation hearings, but - until we developed sentencing support tools - no information about which choices are most likely to prevent further criminal behavior by the offender. A good part of academia, and much of the corrections community, continue to research best practices, but the vast bulk of sentencing culture ignores them while they, in turn, ignore sentencing. Yet every sentence we impose has a public safety outcome whether or not we pay attention.
      There are probably three essential flaws in existing sentencing culture and mythology.

The role of "just deserts" remains paramount, so that participants devote their energies to what is an appropriate severity in terms of moral equivalency, courts think of "aggravation" and "mitigation," and all can claim success as long as a sentence is "proportional." Instead,

we should acknowledge that proportionality sets a limit on severity, and focus on how sentencing choices rationally further some social purpose -- typically but not exclusively, crime reduction.
"just deserts" should support a deviation from crime reduction only when and to the extent that deviation is actually necessary to serve one or more of these functions (which capture all social purposes of punishment per se): to serve a legitimate need of a victim, to prevent vigilantism or private retribution, to maintain respect for legitimate authority, or to enhance respect for the persons, property, or rights of others.
Just deserts is not enough - within the limits of proportionality, we fail if we do not exercise best efforts at public safety (modified only as necessary to serve some other legitimate public purpose).

Our sentencing culture thinks of "public safety" as the function of prison and as in competition with "rehabilitation" and programs; therefore prison beds and programs are largely (and dysfunctionally) allocated on the basis of blameworthiness and symmetry (theives go to theft counseling, bullies to anger management, and so on) rather than best efforts at crime reduction. Instead,

prison, programs, and all sentencing devices should be allocated by evidence-based choices based on risk, results, and resources (within the limits of proportionality);
we should not ask merely "what works"; we need to focus on " what works on which offenders" -- different things work or not for different offenders

We allow plea bargaining to drive most sentencing outcomes with no attention to public safety or any other social function. Instead,

courts should direct attention to public purposes when reviewing plea agreements;
Prosecutors, as advocates for the "people," ought to bargain for the best public safety outcome -- exclusively, in most cases
Defenders should be ready to address "what works" for their clients when and to the extent that that role furthers a client's interests.

      Modern decision support technology holds great promise for improving our attention to and accomplishment of crime reduction. The "sentencing support tools" described here are not intended to generate the "best" sentence based on the input related to an offender and an offense -- but merely to inform the process and encourage attention to the issue of public safety outcomes by all involved. This includes encouraging other sources of information - including variables unique to the offender and the community, as well as other support for best practices -- including legitimate research from fields such as criminology, corrections, and the other applicable social and medical sciences.
      All sentencing hearings (as well as pretrial release, post conviction supervision, and custodial curriculum decisions) should be based on readily available evidence about which sanctions and programs seem to work on which offenders. "What works?" should be the question routinely addressed by advocates and judges who have ready access to operational data to help answer that question.
      We have developed sentencing suport tools in Mulnomah County, and have made them available to all judges and counsel in the County. These tools allow all involved to run queries to determine how offenders who are like the subject have fared after being sentenced to any of the sanctions (custodial and otherwise) available for that offender, with success measured by various standards - but all keyed to reduced criminal behavior.
      The use of these tools is slowly growing, as is their impact in making attention to public safety outcomes a routine part of sentencing arguments and hearings. The state Judicial Department has adopted sentencing support technology as a goal. They are a part of the "decision support" initiative which is in turn part of a major overhaul of our court technology -- which is on a five-year plan subject to legislative support.
       Sentencing support tools are but one strategy for pursuing evidence based sentencing in service of public safety. We have made major progress, but we have a very long way to go. Sentencing is not going to abandon its ancient and brutally dysfunctional liturgy without the difficult struggle for cultural change that we have encountered at least for the last decade.

Michael Marcus, Judge, Circuit Court, Multnomah County, Oregon