Cost containment, sentencing and perverse incentives

Yesterday, the Judicial Conference of the United States issued a press release about cost containment as our national government lurches toward yet another October 1 shut-down date. Among other things, that press release contained the following little nugget:

Acting on the recommendation of its Criminal Law Committee, the Conference agreed to seek legislation, such as the Justice Safety Valve Act of 2013 (S. 619), which is designed to restore judges’ sentencing discretion and avoid the costs associated with mandatory minimum sentences. According to the U.S. Sentencing Commission, the average term of supervised release of an offender subject to a mandatory minimum sentence is 52 months, compared to 35 months for an offender who was not subject to a mandatory minimum. The cost of supervising an offender for one month is approximately $279.

I find this statement utterly remarkable. Let’s be clear on what the statement really means. It means that judges should explicitly consider preserving the federal judiciary’s budget by reducing the time that offenders spend on supervision after they get out of prison.

If a sentencing judge is supposed to consider what’s best for the federal judiciary’s bottom-line when sentencing people, haven’t we created a truly twisted and perverse incentive for judges to look out for their own institutional best interests rather than looking out for the best interests of the offender and the well-being of the public?


PS Thanks to Elaine Mittleman for her thoughts on the subject of costs. To be clear, however, Elaine has no responsibility for the thrust of this post.

14 responses

  1. Good morning Judge,

    It would be much better if the judicial discretion were to occur on the front end instead of the back end. Not long ago there was a New York Times article about a convicted felon who had since turned his life around and became a good husband and father. Through a variety of cluster Fs, police executed a search warrant and found some unfired bullets that a neighbor gave him to dispose of. It turns out it’s a 15-year minimum for a felon to possess ammo. The U.S. Attorney thought 15 years is a fair sentence for father a with young children who possesses ammo, but no gun. Even the judge apologized to the defendant for the harsh sentence. I emailed a dissenting opinion to the USA. But what do I know about mandatory minimums? I’m just a citizen who believes that ABOUT HALF THE TIME ON THESE TYPES OF CASES that the real criminals work for the an unjust system. But that’s just me.

  2. There is another way of reading the proposal. Rather than asserting that judges ought to consider what’s best for the federal judiciary’s bottom-line in sentencing, its premise is that, if judges are allowed to exercise their discretion in sentencing, they do (already, without consideration of costs) overall issue shorter periods of supervision than they do when they are not allowed to exercise their discretion. Thus, if allowed to exercise their discretion using only the standard criteria for sentencing (without consideration of costs,) the average time ordered for supervision will decrease, thus reducing the federal judiciary’s bottom-line without ever considering the cost of supervision.

    However, I question the implication of the proposal that simple reduction of time on supervision will result in cost savings of $279 per individual per month. While the average cost of a month’s supervisions is likely $279, reducing the sentence of an individual from 52 months to 35 months will not reduce costs by $1953 (17 x 279,) since only the variable costs associated with the individual will be saved, e.g., the extra paper needed for that person, the extra urine testing materials, etc. Fixed costs will remained fixed. If the overall reduction in supervision is large enough there may be a reduction in fixed costs, e.g, elimination of the salary and override for one out of every ten supervisors, but not nearly $279 per month per individual.

  3. Isn’t this just the latest iteration of an evaluative system where money is the foundational value – thus all decisions should be grounded in terms of an economic analysis? I.e., that by locking criminals up for a period of time we reduce the negative economic effect their crimes have, while shouldering the negative economic effect of housing them. Now the calculus has shifted and the cost (via debt assumpton, polical compromise on other points of economic policy/effects, etc.) is being re-evaluated.

    Of course, I say this out of fatigue with the reductive line of reasoning that everything comes down to the dollar. We see such logic being bandied about all the time in terms of disaster relief, emergency aid, environmental policy, rehabilitation issues, health care, etc. ad nauseum, and while it is effective in it’s simplicity, any system (or rule) that basic is going to lack the dynamism required to be broadly and fairly applied. At least in my opinion, there are too many values that don’t convert well to currency, at least not on a reliable scale.

  4. My math above is clearly wrong – the result of trying to do too many things at once. My calculator assures me $279 x 17 equals $4,743, not $1,953. Fortunately my students ceased believing I was infallible a long time ago so any of them reading the post will not be traumatized by the error.

  5. Dear R.E.,

    As you note in a later comment, your math is wrong. But you’re point is right; that is, since fixed costs will remain the same, the savings are likely to be much less than anticipated. This proposal looks like a Trojan horse to me. It purports to be about cost reduction, but may be more about the hunger of many judges for more sentencing discretion. I don’t mind debating the advisability of greater sentencing discretion, but not in the guise of cost containment.

    All the best.


  6. Mike D.,

    You make a very important point about cost.

    Half-assed “cost” reduction proposals (like the one discussed in this post) obscure important policy considerations. We were all bent out of shape about the “cost” of crime when we decided to get tough with sentences. Now, we are all bent out of shape about the “cost” of imprisoning people who are subject to tough sentences.

    Serious cost analysis would be one thing. But, the former and present “cost” arguments for criminal justice policies were not and are not serious–they were and are camouflage for normative choices.

    Thanks for taking the time to write. All the best.


  7. Thank you, Judge.

    It’s a pet peeve of mine when people conflate short term savings with long term efficiency. There’s a reason that accounts use a balance sheet, every credit comes with some sort of debit – when we’re talking about public safety, justice, and other core values I think we should keep a critical eye on the sacrifices we make to achieve a better quarterly report.

  8. I’m not sure that reducing terms of supervised release is a good idea, but it should not be rejected out of hand. A lot of my federal clients face terms of supervised release of three to five years after they finish their sentences. I assume the goal of supervised release to reduce recidivism by watching over an offender as he transitions to life outside prison.
    In my experience, those on supervised release tend to violate either by committing a new offense (fairly common) or by “technical” violations of the terms of supervised release such as not following a curfew, failing to seek employment, not completing community service, etc., (not as common because probation officers often will tolerate technical violations if the person is otherwise doing OK).
    My hypothesis is: most of those who will fail badly in supervised release will probably do so within 12 months. Those who make it past that point are probably lower risk, so supervision after that point may not contribute much to reducing recidivism. Those who commit a new offense after the 12-month point can still be punished for the new offense and their criminal history is something that could increase their sentence for the new offense. I think that some districts do terminate supervised release early for those offenders who have done well. Certainly that saves resources. I’ve heard of offenders who have supervised release terms of ten years or life. It is worth looking at whether continuing to supervise an offender for that long provides any real benefit.
    For those who face supervised release revocation for technical violations, we should consider whether expensive prison space should be used to house someone who committed no crime but did decide he would rather not seek a job, keep appointments with his probation officer, pay restitution, etc. Non-compliance with supervised release conditions probably correlates with a higher risk of recidivism, but I do not know how strong the correlation is. I’m not saying that someone who won’t follow the supervised release rules don’t deserve punishment. However, punishment is not free and with finite resources it is possible those resources could better spent elsewhere.
    Whether the proposal would save money should be closely looked at. I suspect that the staffing at probation offices is determined in part by how many persons that office is expected to supervise. If that is not true, I’m sure the judge can set us straight. But theoretically fewer persons under supervision should translate to fewer officers needed to supervise them.
    Finally everyone who pays attention to the federal judiciary knows that Congress is not providing enough money for it to carry out its mission. Federal public defenders have been laid off. Payment to attorneys who do court-appointed work have been reduced 12 percent. Criminal cases (including providing defense counsel for the indigent) are a core Constitutional function of the judicial branch. Supervised release, which did not even exist before 1987, is not.
    Cutting supervised release terms may be a terrible idea, but we should not decide that without looking at some facts and cost/benefit analysis.

  9. There are two other aspects to this, in my mind. One is the validity of the underlying premise that, left to their own devices, judges would sentence below the existing minima more than they do now, when exigent circumstances warrant. Is there any evidence that this actually would occur? Am I conflating statutory minima with sentencing guideline minima?

    The other aspect is the nature of the populations involved. It seems to me that the folks under supervision with whom we can “get away with” shorter supervisory periods already are cheaper to supervise–less prone to calling in over a questionable behavior or what-have-you–than those who need the longer supervisory periods. This would tend to lessen the amount of savings.

    Eric Hines

  10. Economic analysis techniques have a broad reach of applicability beyond just dollars and cents. Nearly any question is accessible to cost-benefit analysis; the analysts just need to keep in mind that there are other units for costs and benefits besides mere dollars.

    Part of the problem is that many of those other costs and benefits are hard to quantify. But they can be categorized and ordered and judgments made based on them.

    Eric Hines

  11. There is no doubt that CBA and associated game theory systems have value – my gripe (and it is little more than that), is when short term CBA thinking is the only tool used.

    This is especially problematic when dealing with values that tend to erupt around people – where there may be a lot of values not addressed by economic thinking. There is no reason to willfully ignore efficiency and tools that can provide valuable data and recommend solutions, but any survey of applied economic theory (or game theory experiments) should caution against total reliance upon such tools.

  12. Bryan,

    I am with you. But proposing cuts in supervised release terms as a cost containment measure without any real analysis of the costs and benefits seems like a sham to me.

    All the best.


  13. Eric,

    As to your first question (point), I don’t know the answer. Additionally, I don’t see that anyone looked at that question.

    As to your second point, I agree with you.

    All the best.


  14. Pingback: Do the costs of incarceration influence sentencing length? « Hercules and the umpire.

%d bloggers like this: