Mimesis Law
4 June 2020

Federal Sentencing Guidelines: Good Enough For Government Work

September 9, 2016 (Fault Lines) – There is a seemingly never-ending supply of issues to discuss in criminal law. Maybe admiralty law has as many niches and fertile research areas, but who besides 107 lawyers doing that work would actually know? Part of what makes criminal law interesting is that it illustrates human drama, tragedy and comedy in many of its forms. Thus, it is entertaining in a way that contract law just cannot be.

Dick Wolf isn’t going to make a television show about your average transactional lawyer reviewing contracts and issuing opinion letters. The high stakes mid-season cliffhanger would be a dispute over the wording of a force majeure clause. Commercial lawyers are born, not made.

Within criminal law there are broad fields like constitutional law, procedure, investigations, substantive law, ethics, statutory interpretation, and sentencing. Then, on top of all that, there are policy areas such as race, socio-economic status, social justice, police regulation, and so on. And there have been philosophical inquiries into the nature of crime and punishment going back to ancient Greece. In sum, the field of criminal law is like a delicious version of Golden Corral—if it were possible for that to truly exist anyway.

Through the passage of time, some principles have been essentially resolved, for good or ill. Inchoate crimes are punishable and defendants are held accountable for the natural and probable outcomes of their actions. But punishment has proven to be a vexing problem in criminal law.

The fact that philosophers cannot agree on the purpose of punishment is not, in itself, an issue. Philosophers can’t decide whether Schrödinger’s cat is alive or dead, or even if Schrödinger was being serious. Germans are famously known for their humor. In the real world, explanations tend to follow discovery and practice; so, the lack of a coherent philosophy is usually no matter.

But there are two conflicting ideas about imprisonment, rehabilitation and retribution. Conceptually, they are not mutually exclusive, meaning the sentence could represent just deserts and help the defendant get out of the criminal life. But in reality one will often operate at the expense of another. And which idea has primacy waxes and wanes, leading to some inconsistences over time. Plus, there’s a little deterrence sprinkled into this mix. So, it can be argued that we have a system that tries to do everything and fails completely at doing so.

Judge Kopf, who has argued for more empiricism in sentencing, recently discussed at Fault Lines the U.S. Sentencing Commission’s study regarding criminal history scores and recidivism. In sum, the criminal history and age of offenders correlated with re-offense rates, but the criminal history score did not. This is the same study finding that about half the offenders were not re-arrested in an eight-year period and discussed here.

Judge Kopf makes the following suggestions to criminal defense lawyers:

Since Fault Lines is devoted in large part to criminal defense practitioners, I will highlight two points from this study that a CDL can use to help his or her client.** I turn to those two points next.

First, except for firearms offenses, ask the judge to ignore the total offense level if the judge is (like me) most interested in protecting the public from future crimes of the defendant.  This is very, very important.

Second, age matters, and it matters a lot when a judge is concerned with recidivism. That is, this study confirms that offenders do “age out of crime.”  As the Commission concludes, “Studies have repeatedly shown that older offenders at sentencing are at lower risk for reoffending, and the Commission’s research confirms these findings.” Study at p. 23 (footnote omitted).

One of the big problems with sentencing is that the sort of data discussed in the report is probabilistic, thus not determinative. The defendant in front of you may be part of cohort that in the aggregate has a higher re-offense rate, but you cannot say with certainty that the particular defendant will re-offend. Yet, the defendant is often sentenced as if the probability was a certainty.

We can infer from this data using criminal history to lengthen sentences does not serve rehabilitative, deterrence, or incapacitation goals. But as the Commission quickly points out, “that the offense levels in the federal sentencing guidelines were intended to reflect multiple purposes of punishment, including just punishment and general deterrence (which are unrelated to offender recidivism).” There it is—the desire for just deserts.

At the lowest criminal history level, the difference between the maximum sentence at levels 14 and 15 is three months. But between levels 38 and 39 it’s almost three years. If we are being honest with ourselves, we cannot truly represent the range of human behavior with gradations on a line. Nor can we truly distinguish between those small of gradations. It’s the kind of faux certainty regarding human affairs that later followers of Fredrick Taylor would love.

Yet, it came about because it was seen as an improvement over what was before. And in states that have no such tool, it’s mostly judicial guesswork that hopefully looks Gaussian when aggregated. Even still, the mean and average sentence doesn’t matter a whole lot to the guy who got maxed by the hanging judge.

While Judge Kopf and the Sentencing Commission cast a critical eye, Scott Greenfield criticized the recidivism, which mostly captured by the criminal history axis:

The problem isn’t a matter of correlation. Sure they correlate, but that does nothing to inform as to causation. Therein lies the problem, and penalizing someone in advance for what they’re likely to do in the future may make a judge feel protective of the public’s future safety, but it’s a lousy way to run a legal system.

Let’s consider how one gets from here to there. A defendant gets convicted of one crime, his first offense. He’s now saddled with a felony, which will preclude a host of future possibilities, from jobs to housing to education. (Sob stories omitted here.) * * *

As Judge Kopf notes, there is a correlation between this guy and a guy who goes on to commit additional crimes after his release.  Go figure. I neither doubt it nor disagree with it. In fact, I can’t imagine how anyone manages to survive post-prison release and not return to a life of crime.  Aside from those fortunate enough to have loving families to support them on release, or fall into one of the handful of employers who will give an ex-con a break, there is little else for them to do to survive than commit crimes.

So it correlates? You bet it does. What did you expect out of this system we’ve created?

It’s a fair question. On one hand, there is research supporting Scott’s overall point. And there is research that suggests that prison has an impact on crime rates and recidivism. Again, the probabilistic nature of this information can convey what the average inmate may do, but it cannot tell us that prison either caused or thwarted future crime in a particular inmate. While there seems to be a point where prison may indeed cause more crime than it stops, it’s difficult to move this down from the policy clouds to the individual sentencing hearing.

The judge can be confident that while in prison the particular inmate will be incapacitated from committing most types of crime against the public. But what level of certainty are we comfortable with during sentencing? If the defendant is part of a cohort where 75% of them re-offend with serious and violent crimes, then from a systematic perspective the risk probably is too great. In another context, we wouldn’t leave our infant with someone we perceive as 75% likely to kill our child. Or leave our property with someone that we perceive as that likely to steal it.

Yet, as Scott illustrates, at the individual level it could be an injustice. The Guidelines avoid addressing this problem by making it one factor and not explicitly using criminal history for any one purpose. Through the magic of making criminal history serve multiple purposes, no one argument or study can fell it. After all, giving a life-long felon a long sentence can arguably be just deserts for a life of crime.

At the end of the day, the criminal history score probably communicates some useful information to the judge but the offense level very little. But the sentences imposed by the guidelines appear to do a fair job at accomplishing both punishment and rehabilitation. And it does all this while not being particularly philosophically coherent. Perhaps Churchill would say the federal guidelines are the worst form of sentencing except all the others.

9 Comments on this post.

Leave a Reply



Comments for Fault Lines posts are closed here. You can leave comments for this post at the new site, faultlines.us

  • Richard G. Kopf
    9 September 2016 at 12:07 pm - Reply


    Great post. However, two things in response are worth mentioning:

    Point One:

    Germans are funny but only in a way the rest of you layabouts are incapable of understanding because you (1) don’t work hard enough and (2) are insufficiently literal. Example:

    Adolf Hitler is speeding through Germany with his chauffeur. While driving through a rural area, the chauffeur accidentally runs over a chicken that walked out onto the road. Hitler demands that the chauffeur should go apologize to the farmer. The chauffeur does so, but returns badly beaten up. A bit later, Hitler sees a pig on the road and orders the chauffeur to hit it, then go apologize. This time, the chauffeur comes back with a broad smile on his face, clutching some money and ham. Hitler asks him “What happened? Didn’t you apologize for killing his pig?”. The chauffeur answers “Yes, I did. I told him ‘Heil Hitler, I just ran over that pig.’”

    What a knee slapper!

    Point Two:

    The following sentence in the seventh paragraph of your post is, I think, incorrect: “In sum, the criminal history and age of offenders correlated with re-offense rates, but the criminal history score did not.” If you substitute “total offense level” for “criminal history score” in the sentence then I understand and agree with your sentence that I question.

    If you disagree, that I remind you that the Commission found that a “federal offender’s criminal history was closely correlated with recidivism rates. Rearrest rates range from 30.2 percent for offenders with zero total criminal history points to 80.1 percent of offenders in the highest Criminal History Category, VI. Each additional criminal history point was generally associated with a greater likelihood of recidivism.” Page 5 of the Sentencing Commission’s study.

    In short,the study did show that the “criminal history score” was correlated with reoffense rates. If you think I am in error, please explain how so.

    Alles Gute.

    Rich Kopf

    • Andrew King
      11 September 2016 at 6:31 pm - Reply


      You are correct; it was a misstatement.

      Thanks for the catch.


  • Jeff Gamso
    9 September 2016 at 12:51 pm - Reply

    “But the sentences imposed by the guidelines appear to do a fair job at accomplishing both punishment and rehabilitation.”

    I suppose that’s theoretically possible. It’s equally possible, of course, that they do an appalling job of accomplishing those goals.

    Consider the punishment part of the equation. Yes, if you agree that the guidelines properly recognize various degrees of seriousness of offense and then more-or-less appropriately impose punishment, then they do a fair job of accomplishing punishment. Ditto with rehabilitation.

    That’s not an argument, it’s a tautology. to one (you for instance) who thinks the guidelines pretty much get it right, it will follow that they pretty much get it right.

    As they say on the internet, your mileage – and certainly mine – may vary.

    • Richard G. Kopf
      9 September 2016 at 12:55 pm - Reply


      If you could redraft 3553(a) (the ink blot for sentencing), what sentencing philosophy(s) would you implement? All the best.


      • Jeff Gamso
        9 September 2016 at 1:16 pm - Reply

        I rather like 3553(a)(1) and (2). I’d scrap the rest.

        But it doesn’t (and regardless of whether you include subsections 3 through 7) actually provide any real guidance as to what is a “sentence sufficient, but not greater than necessary.”

        What’s the calculus that says 38 months is enough but not too much for this guy and this offense while for that gal and that offense 141 months fits the bill? Ultimately, those numbers are plucked from someone’s . . . ok, maybe from the air. We can all (except perhaps Bill Otis) agree that a couple of decades in the Big House is excessive for jaywalking. But that doesn’t tell us much.

        Statutes may provide sentencing floors or ceilings, but there’s no particular empirical basis for them. The USSG doesn’t really know what’s appropriate for an offense level 14 with a criminal history category 1, let alone whether some particular offense belongs in a moral gradient at offense level 14. All those numbers are made up.

        Ultimately, we rely on the judge’s (or in state cases in states like Texas the jury’s) sense of what’s appropriate. I’m not sure there’s a better way, but it’s silly to pretend that the results in any particular case satisfy or even approximate some platonic ideal of the appropriate sentence.

        • Richard G. Kopf
          9 September 2016 at 3:34 pm - Reply


          Thank you.

          For what it is worth, I can’t stand the ambiguity and contradictions of 3553(a) and particularly(a)(2). If I could rewrite the statute I would do as I suggested in WHY DO FEDERAL JUDGES SEND CRIMINALS TO PRISON? Fault Lines (September 9, 2015), available at

          Anyway, I appreciate learning your views. I suppose I look for clarity, consistency, and external constraints ’cause I’m a chicken shit who subconsciously wants to avoid responsibility for the destruction of another human being.

          All the best.


          • Jeff Gamso
            9 September 2016 at 3:46 pm -

            I get it. My problem with the guideline rules (aside from the fact that it’s more interesting – and tends to be better for my clients – to argue around them) is that they’re inherently arbitrary and despite all the qualifications are inherently based on a one-size-fits all model.

            Sentencing ought – and as in the pre-Mistretta and now post-Booker days somewhat is – individualized.

            The arbitrariness is unavoidable. But case-by-case arbitrariness from those on the ground may be fairer (or maybe not, of course) than generic arbitrariness assigned from afar.

    • Andrew King
      11 September 2016 at 6:38 pm - Reply


      I made that conclusion through the inference of this and other studies. There is a very real question whether the guidelines, BoP, or something else is responsible for low recidivism rates. I gave myself a little breathing room by referring to the sentences imposed rather than the guidelines themselves. Also, how you define rehabilitation matters. In the broadest sense, no future recidivism after release can count as rehabilitation.


      • Jeff Gamso
        11 September 2016 at 9:32 pm - Reply

        The problem remains. Unless you assume that if there’s any recidivism at all the guidelines (or imposed sentences if you prefer) are worthless, in which case you can call the entire exercise of imposing sentences to achieve that end hopeless, the evaluation is still inherently arbitrary.

        The principle of lex talonis may have some visceral appeal, but we don’t do that in western society, and the 8th Amendment would put significant limits on it if we tried. (And see State v. Gears, 135 Ohio App.3d 297 (1999).) But beyond that? How do we determine, in any but an arbitrary way, that X is the appropriate punishment for person A who did thing R? And then to weigh in how successful (in the scheme of things) that punishment will be at rehabilitation? Ain’t no way.

        I’m not saying we shouldn’t impose punishments. I’m just saying that pretending the choices are rational – or that we can meaningfully measure whether they’re appropriate – is a fool’s game.