"Cost as a sentencing factor": a response.
Professor Chad Flanders offers a normative theoretical critique of including costs of punishment in Sentence Advisory Reports (SARs) that the Missouri Sentencing Advisory Commission (MOSAC) produces. (1) This approach provides a useful lens for understanding divergent opinions on the practice of including cost information in SARs and provides a consistent logical framework for understanding whether this practice squares with more fundamental principles of criminal punishment. In this Response, I complement the normative analysis in the main Article with several observations from a different analytical perspective.
As an empirical social scientist, my analytical approach is based on positive analysis. I am less concerned with how people should be motivated and more concerned about how people actually behave. So without diminishing or contradicting Flanders's analysis, I wish to recast it as a positive analysis. Part II summarizes the main Article. (2) Then, Part III sketches a gametheoretic perspective on MOSAC and SARs. (3) Next, Part IV describes MOSAC's repeated tendency to overstate the social costs of longer prison sentences and thereby conflict with voter and legislator preferences. (4) Finally, Part V discusses the theoretical and practical effect of including cost information in SARs. (5)
II. A Brief Recap of Flanders' Thesis
The fundamental tension between the retributivist and consequentialist motives in sentencing policy identified by Flanders is a useful lens for understanding the controversy over including sentencing costs in SARs. (6) The notion that highlighting the costs of prison versus alternative (and less punitive) sentences offends those commentators who focus on retribution as the purpose of sentencing seems right on target. (7) Flanders then employs these dueling perspectives to motivate what I agree is a conventional and overly simplistic "ideal" division of labor in sentencing policy: that the legislature should be consequentialist in setting sentencing parameters for judges while judges should be retributivist in administering uniform punishments within those parameters. (8) Flanders concludes by offering a more sophisticated view by way of analogy to overlapping circles, arguing that judges should focus more on the core concerns of retributive justice while legislators stand back and take in the larger picture. (9) Given this conclusion, costs of prison versus probation are not of primary relevance to judges and, in Flanders's view, should not be part of SARs. (10)
III. A Game Theoretic Perspective on MOSAC and SARs
In general, sentencing policy has multiple goals, including retribution, cost-effectiveness, rehabilitation, and deterrence. The latter three goals fall under consequentialist motives in Flanders's analysis. (11) However, rather than focusing on which goals should motivate different actors, (12) I posit that voters, legislators, judges, and MOSAC all seek to maximize their own utility, which is in turn a function of the consequentialist goals. However, each of these actors also derives utility from other sources; for example, legislators face frequent reelection so they also must place weight on the opinions of their constituents in order to retain office (or at least more weight than judges and MOSAC). Further, each of these actors has different access to information; for example, voters probably are least informed about the costs and benefits of alternative sentencing policies. Thus voters, legislators, judges, and MOSAC likely may have differing preferences over sentencing policy.
Given this framework, sentencing policy can be seen as the product of a complex game among voters, legislators, judges, and MOSAC. This game has rules (the state constitution) and within the rules each actor seeks to maximize its own utility. And while a full treatment of such a game is beyond the scope of this Response, some insights may emerge from this outline of the sentencing game. For example, one way to interpret institutions such as legislated sentencing guidelines or truth-in-sentencing laws is that they are attempts by the legislature to control judges in a manner that enhances the policy goals of legislators.
Given recent history in Missouri and elsewhere, it is fair to assume that legislators prefer policies to be more punitive than judges might choose on their own. I conjecture that this situation is in part because the general public pushes legislators in this direction. Thus, the least informed actors in this game are attempting indirectly to constrain judges to mete out sentences which are "too punitive" from the perspective of the more informed actors (judges and perhaps even many legislators).
One response by self-interested legislators might be to give judges a wider range of sentencing options and more discretion in reducing actual time served in practice. Judges, in turn, exploit this freedom to enact their preferred sentences, which may be closer to what legislators want. Of course, once voters catch on to these policy responses, they can pressure the legislature to adopt truth-in-sentencing laws or to reform judicial selection and retention methods in a manner that would augment judicial accountability to voters.
Additionally, self-interested legislators could create an expert panel, such as MOSAC, to recommend sentencing policy that is more in line with the preferences of legislators and judges. As an independent and non-partisan body, MOSAC may be more credible to voters and thus be able to sway voters' opinions regarding sentencing policy through the production of expert reports and recommended sentences. These activities also provide cover to legislators and judges, making it easier for them to choose less punitive sentencing policies that may run counter to voters' preferences.
In the above scenario, the expert analyses and recommendations of MOSAC provide a counterbalance to the influence of an ignorant public on sentencing policy. This situation is consistent with the role that MOSAC plays in making non-binding sentencing recommendations, as well as in issuing expert reports that support the use of alternative (less punitive) sentencing. (13) Of course, voters may have different preferences from judges and legislators for reasons other than ignorance. For example, ordinary citizens may place greater weight on public safety or may have a greater demand for retribution than MOSAC. The fact that MOSAC is an unelected body and not directly accountable to the public makes such differences in policy goals and preferences all the more likely to exist. (14) In this latter scenario, MOSAC may play a less salutary role by facilitating policy drift away from what the general public would prefer. In either case, MOSAC serves to make legislators and judges less accountable to the public, but the normative implications are potentially different across these scenarios.
IV. SENTENCING COSTS AND SARS: A SECOND LOOK
But there must be more to the story in order to understand the newly adopted practice of including cost information in SARs. After all, if the goal is to inform voters (and perhaps legislators) about the costs of different sentencing dispositions, this objective could be done in an annual report. Why does MOSAC repeatedly send this message to judges via SARs?
As Flanders noted, the inclusion of sentencing costs in SARs makes these costs more salient when judges are deciding among different sentencing dispositions. (15) The intent is fairly obvious: to "nudge" judges. Experimental research in psychology and economics finds that such constant and subtle reminders can frame issues in a way that influences decision-making in a predictable fashion. (16) By including sentencing costs in SARs, MOSAC is trying to nudge judges to issue less punitive sentences than they would if cost information was issued in an annual report. These circumstances suggest that either judges are not competent to make appropriate use of cost information and so must be manipulated into doing so or that MOSAC is trying to push sentencing policy to be more in line with its own preferences. I maintain the latter is a more plausible explanation since there are other and more effective means to deal with incompetent judges. Further, as I argue below, SARs systematically exaggerate the benefits of alternative sentences over prison, which is consistent with the notion that MOSAC prefers less punitive policies than would otherwise be chosen in a game that involved only voters, legislators, and judges.
SARs include information on the relative budgetary costs of prison versus alternative sentencing. (17) But these costs are not necessarily representative of the social costs of alternative sentencing policies. Most importantly, SARs do not tally the social costs of future crimes caused by more lenient sentencing policies, which means that the sentencing costs presented in SARs exaggerate the true costs of prison versus alternative sentences.
SARs also include a "risk assessment" for each offender. (18) The risk assessments estimate the predicted recidivism rate for an offender using data from the experiences of past offenders with similar characteristics. (19) But MOSAC does not report these predicted rates to judges. Instead, the risk assessments are presented as simple qualitative rankings of "good", "aver age", etc. (where good indicates a lower risk of recidivism). (20) But if the goal of MOSAC were to provide objective information to judges about offender risks, then reporting the predicted recidivism rate would be much more objective, informative, and transparent. This practice therefore suggests that MOSAC is again acting in a manner to nudge judges toward less punitive punishments.
In addition to risk assessments, SARs include a recidivism rate for similar offenders assigned to different sentencing dispositions. (21) These recidivism rates are not estimates but actual averages for offenders sentenced to prison, probation, etc. (22) MOSAC presents these average recidivism rates as though they were useful information for predicting whether a particular offender is more likely to recidivate if sentenced to prison versus some alternative sentence. However, it is well known that average recidivism rates exaggerate the causal effect of prison on recidivism. (23) This phenomenon occurs because judges do not randomly assign offenders to prison versus probation; rather, high-risk offenders are more likely to go to prison than receive probation. Thus, average recidivism rates reflect the fact that offenders that typically receive prison sentences are more likely to recidivate in the first place. So, once again, MOSAC presents information to judges that has the effect of exaggerating the social costs of prison versus less punitive alternative sentences.
Finally, MOSAC describes the recommended sentences in SARS as average sentences derived from actual practice in Missouri; (24) however, it is apparent that this conclusion is not quite correct since recommended sentences appear to be lower than average sentences in many instances. It is my understanding that recommended sentences are based on average sentences in some fashion, but despite repeated efforts, I have been unable to learn anything more about how recommended sentences come into being. I have made formal and informal inquiries as to how recommended sentences are generated, but these requests have been denied or ignored. This lack of transparency regarding how recommended sentences are generated and why these deviate from actual practice again suggests that MOSAC is not interested strictly in providing objective and useful information to judges but may be trying to influence sentences in a particular direction.
It should be apparent that MOSAC has different preferences than judges and exploits its role as a provider of information to try to manipulate judges into selecting less punitive sentences than they would otherwise. These same tendencies to exaggerate the social costs of prison are evident in annual re ports MOSAC issues, which suggest that MOSAC likewise has preferences over sentencing policy that are out-of-synch with voters and legislators.
Flanders presents a normative argument against the inclusion of cost information in SARs. (25) I disagree that including cost information is necessarily deleterious to good public policy. But I do agree that there is good reason to be concerned about the information contained in SARs, albeit for different reasons.
In theory, objective expert reports on the costs of sentencing, recidivism, and the like can improve sentencing policy by counteracting the influence of voters who ignorantly prefer excessively punitive sentences. However, it is also plausible that expert information of this sort can distort policy when voters are well informed. Finally, we must recognize that MOSAC has its own preferences that may deviate from the preferences of voters, legislators, and judges. I argue that the manner in which costs of sentencing, risk assessments, and recidivism rates are presented has the effect of systematically overstating the social costs of prison. This discrepancy suggests that in practice MOSAC has preferences for less punitive sentencing and exploits its role as a provider of information to manipulate other actors in the sentencing game. From a positive theoretical perspective, one would expect nothing less of a rational actor in the sentencing game.
Flanders also raises concerns about how cost information might undermine the goal of uniformity in sentencing. (26) Again, I concur, although the concern is not specific to cost information. If some judges are more susceptible to nudging than other judges, then the activities of MOSAC may increase sentencing disparities across otherwise comparable offenders.
As I have emphasized, the role of policy experts like MOSAC can be beneficial to the formation of better sentencing policy. However, other actors in the sentencing game need to exercise sufficient oversight to ensure that MOSAC provides objective and useful information.
(1.) See generally Chad Flanders, Cost as a Sentencing Factor: Missouri's Experiment, 77 Mo. L. Rev. 391 (2012).
(2.) See infra Part II.
(3.) See infra Part III.
(4.) See infra Part IV.
(5.) See infra Part V.
(6.) See Flanders, supra note 1, at 397-98.
(7.) See id.
(8.) Id. at 398.
(10.) See id. at 399, 410.
(11.) Id. at 400-03.
(12.) See id. at 398-400.
(13.) See Mo. Sent'g Advisory Commission, http://www.mosac.mo.gov/page. jsp?id=45392 (last visited Feb. 13, 2012) (discussing the functions of MOSAC).
(14.) The current composition of MOSAC members is available at Missouri Sentencing Advisory Commission Members and Staff, Mo. Sent'g ADVISORY Commission, http://www.mosac.mo.gov/page.jsp?id=45351 (last visited Mar. 15, 2012).
(15.) Flanders, supra note 1, at 392.
(16.) See, e.g., Richard H. Thaler & Cass R. Sunstein, Nudge: Improving Decisions About Health, Wealth, and Happiness (2008).
(17.) See Mo. Sentencing Advisory Comm'n, Sentencing Information on www.mosac.mo.gov Now Includes Costs of Recommended Sentences and Risks of Reincarceration, SMART SENTENCING, Aug. 17, 2010, at 2-3, available at http://www.mosac.mo.gov/file.jsp?id=45502.
(18.) Id. at 1-2.
(19.) See id. at 2.
(20.) See id. at 3-6.
(21.) See id. at 2-3.
(22.) See id.
(23.) See Donald P. Green & Daniel Wink, Using Random Judge Assignments to Estimate the Effects of Incarceration and Probation on Recidivism Among Drug Offenders, 48 Criminology 357, 358-59 (2010).
(24.) See Mo. Sentencing Advisory Comm'n, supra note 17, at 2.
(25.) See Flanders, supra note 1, at 395.
(26.) Id. at 404-10.
Jeff Milyo *
* Middlebush Professor of Social Sciences, Professor of Economics, University of Missouri. Ph.D., Stanford University 1994.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||response to article by Chad Flanders in this issue, p. 391|
|Publication:||Missouri Law Review|
|Date:||Mar 22, 2012|
|Previous Article:||Cost as a sentencing factor: Missouri's experiment.|
|Next Article:||Canines on campus: companion animals at postsecondary educational institutions.|