An offender’s record of prior convictions is a major sentencing factor in all sentencing systems, including all U.S. jurisdictions with sentencing guidelines -- offenders in the highest criminal history category often have recommended prison sentences many times longer than recommended sentences for offenders in the lowest category. For offenders convicted of crimes of low or medium seriousness, a higher criminal history score can also result in a recommended executed-prison sentence instead of the community-based sentence provided at lower history.
These substantial prior record sentencing enhancements have major human and economic costs that are often not justified by the sentencing purposes sought to be achieved. The two most common purposes are retribution and crime control. The first of these asserts that repeat offenders are more blameworthy, and deserve penalty enhancements in proportion to the number and/or seriousness of their prior convictions. But why? Scholars who study and write about retributive theory are in complete disagreement on this point.1 While some theorists argue that each additional prior conviction makes the offender more blameworthy for his new offense, a second group claims the opposite: that offenders with no prior record have mitigated culpability and should not be fully blamed and fully punished (some adherents to this view would extend a lesser degree of mitigation to offenders with only a minor prior record). A third group of theorists insists that prior record is not relevant to desert at all -- that blameworthiness depends entirely on the new crime for which the offender is being punished. Given this total lack of consensus, it is doubtful that retribution can provide a sound basis for criminal history enhancements, especially the high-magnitude enhancements found in many guidelines systems.
The second common rationale for prior-record enhancements is based on punishment goals related to public safety and, in particular, on the assumption that an offender’s record of prior convictions is a good predictor of recidivism risk. There is much more consensus on this rationale, and there is also empirical evidence to support it: offenders with more prior convictions generally have a higher rate of repeat offending. On the other hand, there is very little research validating the predictive value of particular criminal history formulas, or the components of those formulas. Thus it is possible, and indeed it seems quite likely, that most criminal history formulas contain elements that increase sentencing severity without increasing risk prediction accuracy. It also appears that most criminal history formulas are not taking into account other risk factors, such as offender age, that would help to identify lower- and higher-risk offenders. Thus, even though prior record can serve as a useful proxy for offender risk, it’s a crude proxy—both over-inclusive and under-inclusive.
Moreover, even if a jurisdiction’s criminal history score and/or other validated risk factors can accurately predict elevated recidivism risk, a further very important set of policy questions relates to how offender risk should affect the form and severity of the sentence imposed. In particular: (1) Does imposing a prison sentence or a longer prison term on higher-risk offenders reduce future crime enough to justify the substantial added public and private costs? (2) Even if it does, do more severe punishments achieve crime control better than alternative preventive measures within and outside of the criminal justice system? Positive answers to these important questions can no longer simply be assumed, given the substantial amount of research showing the meager crime control benefits of increased sentence severity.2
In short: there is good reason to believe that many guidelines criminal history scores are not achieving public safety goals in the most cost-effective manner.
But while there is uncertainty about the extent to which prior record enhancements can be justified on retributive and/or crime-control grounds, there is little doubt that these penalty enhancements have major, undesirable impacts. In the remainder of this essay, I will briefly examine five of these impacts, using illustrative data from Minnesota. Since many guidelines systems have more substantial criminal history enhancements than Minnesota does, these five problematic impacts are likely to be even greater in those systems.3
1. Increasing the size and expense of prison populations
More than half of executed Minnesota prison sentences are a direct result of the offender’s prior record: these offenders are convicted of medium- or low-severity crimes, and are only recommended for prison because of their prior convictions. Some of these repeat offenders are subject to mandatory-prison statutes triggered by a prior conviction, but the vast majority of them are recommended for prison solely because of their elevated guidelines criminal history score.
In terms of actual bed impact, factoring in the length of the prison terms imposed and assuming that all offenders earn the maximum one-third reduction for good conduct and do not have their post-prison release revoked, the prison terms given to these medium- and low-offense-severity offenders because of their prior convictions require at least 4,000 additional prison beds. That’s about 40 percent of the prison beds required to house all offenders sentenced in a given year.
For high-offense-severity offenders, the prison bed impact of criminal history enhancements is harder to estimate since the guidelines recommend prison even at zero criminal history. But the bed impact here must be substantial since criminal history enhancements greatly lengthen these offenders’ prison terms (which are already very long even at zero criminal history).
Again, it should be kept in mind that prison admission and bed impacts in many other guidelines systems are probably even greater, since their criminal history formulas call for more substantial sentencing enhancements than Minnesota’s formula does.
2. Decreasing the proportionality of sentence severity relative to the seriousness of the offender’s conviction offense
One of the core goals of sentencing guidelines is to make sentences more proportional to the severity of the offender’s conviction offense. Unless there are case-specific aggravating or mitigating circumstances, more serious crimes should receive more severe penalties. But the substantial criminal history enhancements found in many guidelines systems undercut sentencing proportionality. These enhancements result in major overlaps in the recommended sentencing ranges for adjoining severity levels: high-criminal-history offenders at one severity level are recommended for, and usually receive, sentences more severe than lower-history offenders convicted of more serious crimes -- sometimes much more serious crimes.
In Minnesota, over 40 percent of offenders have a recommended prison duration, enhanced due to high criminal history, that is equal to or greater than the recommended term applicable to some low-history offenders convicted at higher offense severity levels. Again, since many guidelines systems give criminal history even more weight than Minnesota does, the loss of offense proportionality in those systems is even greater.
3. Undercutting the goal of ensuring that limited and expensive prison beds are reserved for serious and violent offenders
Another frequently stated goal of guidelines sentencing is to give priority in the use of prison resources to violent offenders and those convicted of high-severity crimes. But if any combination of prior crime types can produce a high criminal history score and a severe recommended sentence, the result is to increase the number and proportion of nonviolent, lower-severity offenders in prison. A few guidelines systems address this problem by stipulating that an offender cannot be eligible for the highest criminal history categories unless he or she has at least one violent or high-severity prior conviction.
Minnesota, like most guidelines jurisdictions, does not follow this limiting rule; as a result, many nonviolent, lower-severity offenders are recommended for and receive prison sentences. This effect was documented in a recent year for which criminal history data included the offense type of each prior felony conviction. Among medium- and low-offense-severity offenders recommended for and receiving an executed-prison sentence because of their elevated criminal history scores, 60 percent were being sentenced for a property, drug, non-person sex (e.g., failure to register), or other nonviolent crime); of these offenders, 57 percent also had no violent crimes in their prior criminal record. These 638 never-violent4 offenders required over 1,000 additional prison beds.
4. Shifting the age composition and risk level of prison inmates
Many offenders with high criminal history scores are older, and thus are likely to be past their peak crime years. One of the best-documented findings in criminological research is called the “age-crime curve”: for most offenders and most types of crime, offending and re-offending rates begin to fall rapidly as offenders enter their 40s and 50s. Yet due in part to the impact of criminal history sentencing enhancements, these aging, declining-risk offenders are taking up a lot of prison beds; they also consume a lot of resources for medical care.
In Minnesota, again looking only at medium- and low-severity offenders recommended for and receiving prison sentences because of their prior convictions, about one-third are age 40 or older at the time of sentencing. Those aging, mostly nonviolent offenders require over 1,300 additional prison beds.
A related issue is the criminal history “look-back” formula: how long ago were the offender’s prior crimes? The Robina Institute’s survey of criminal history formulas around the country5 found that almost half of the jurisdictions with a criminal history score had no limit on how far back in time adult felonies are counted as part of the criminal history score, and almost three-quarters of the jurisdictions had unlimited look-back for at least some felonies. Such rules raise important sentencing policy issues, for example: if an offender is now in his mid-40s or older, does it make sense – on either retributive or crime-control grounds -- to greatly enhance his sentence and consume expensive prison resources based on things he did in his late teens?
5. Increasing the degree of racial disproportionality in prison populations
Data from Minnesota and other jurisdictions show that black and Native American offenders tend to have significantly higher criminal history scores, and thus are subject to more severe recommended and imposed sentences. For example, in Minnesota for a recent 10-year period, an examination of the higher recommended-prison rate for black defendants compared to white defendants found that almost two-thirds of the higher black rate was due to higher black criminal history scores.
No matter how it is formulated, any criminal history enhancement is likely to have some disparate racial impact. But such impacts should be kept as low as possible, and should only be imposed for very sound reasons. Any component of the criminal history score that has a strong racially disparate impact should be dropped or given less weight unless it can be shown to be closely related to expected risk of recidivism for violent or other serious crime. And, as noted above, there should also be substantial reason to believe that increased sentence severity based on predicted higher risk will actually prevent more crime, and do so in a cost-effective manner.
- 1These conflicting retributive accounts are reviewed and critiqued in Rhys Hester, Richard S. Frase, Julian V. Roberts, & Kelly Lyn Mitchell, Prior Record Enhancements at Sentencing: Exploring the Unsettled Justifications and Unsettling Consequences, Crime & Justice: A Review of Research (forthcoming).
- 2This research is summarized in Hester et al., supra.
- 3For a survey of criminal history scoring and enhancement effects in 18 American guidelines systems, conducted by the Robina Institute of Criminal Law and Criminal Justice, see Richard S. Frase, Julian V. Roberts, Rhys Hester, & Kelly Lyn Mitchell, Criminal History Enhancements Sourcebook (2015), http://robinainstitute.umn.edu/publications/criminal-history-enhancements-sourcebook/.
- 4The figures in text here treat all “person” crimes in the Minnesota data as potentially violent, but this overstates the proportion of truly “violent” crimes committed by these offenders; many “person” crimes are ranked at low severity, and involve no serious injury or weapon use (only threats, minor injury, or potential injury).
- 5Frase et al., supra note 3, chapter 3.