chevron-down Created with Sketch Beta.
November 01, 2012

Proportionality, Disparity, and Recidivism

By Judge Patti Saris

Last February, the U.S. Sentencing Commission held a hearing on the state of federal sentencing seven years after the landmark Supreme Court opinion in United States v. Booker (2005).1 At that hearing, Bureau of Prisons (BOP) Director Charles Samuels informed the Commission about the overcrowding of federal prisons: federal prisons are 38 percent over rated capacity and the rate of overcrowding is 53 percent in high-security facilities. Further, Assistant Attorney General Matthew Axelrod stated that, in 2011, the federal prison population grew by over 7,500 prisoners. The number of federal convictions has continued to increase, with 86,201 documented individual federal offender cases in fiscal year 2011, up from 83,946 the year before. Approximately 87 percent received prison sentences. This is not to say that crime rates are up. In a recent letter to the Commission, the Department of Justice (DOJ) reported that violent crime is at the lowest levels in generations. Further, the DOJ referred to a Bureau of Justice Statistics’ National Crime Victimization Survey finding that fewer than 4 million Americans were victims of violent crime in 2010, down from over 10 million in 1991. In light of these statistics, sentencing policy has never been more important.

The Commission is an independent, bipartisan agency within the judicial branch. It was established by the Sentencing Reform Act of 1984, and Congress charged it with writing guidelines that judges use to determine sentences and thereby increase the proportionality, uniformity, and transparency of federal sentences.2 It also has the mission of assessing sentencing, penal, and correctional practices.3 The Commission is currently addressing three predominant issues in the area of sentencing: proportionality, disparity, and recidivism. This article describes the Commission’s work on these important issues.


The Commission has undertaken several initiatives to ensure proportionality. The Commission took the lead in examining crack cocaine sentencing and concluded that the sentences were disproportionate to the seriousness of the offense. It issued four reports on crack sentencing and ultimately recommended that the powder/crack ratio be no greater than 20:1. In 2007, it reduced the sentencing guidelines by two levels of severity for crack defendants. Consistent with the Commission’s report, the Fair Sentencing Act of 2010 reduced the crack to powder cocaine ratio from 100:1 to 18:1, resulting in an average sentence reduction of 37 months. The Commission made those amendments retroactive, thereby potentially reducing the guidelines and prison terms for about 12,000 prisoners who were sentenced under the previous crack guidelines, to the extent public safety was not jeopardized.

On October 31, 2011, the Commission also issued a 645-page report that contains comprehensive data on and legal analyses of mandatory minimums.4 The Commission’s findings were striking. More than one-quarter of the 73,239 fiscal year 2010 offenders sentenced in the federal courts and included in this analysis were convicted of an offense carrying a mandatory minimum penalty, but almost half of the offenders convicted of an offense carrying a mandatory minimum penalty were relieved from the application of such a penalty at sentencing because they provided substantial assistance to the government or qualified for the “safety valve” provision, or both. While 14.5 percent of all federal offenders were subject to a mandatory minimum penalty at sentencing, such offenders made up 39.4 percent of the 191,757 offenders in BOP custody as of September 30, 2010. Over three-quarters of convictions of an offense carrying a mandatory minimum penalty were for drug-trafficking offenses. Hispanic offenders accounted for the largest group (38.3 percent) of offenders convicted of an offense carrying a mandatory minimum penalty, followed by black offenders (31.5 percent) and white offenders (27.4 percent).

Black offenders received relief from a mandatory minimum penalty least often (in 34.9 percent of their cases), compared to white (46.5 percent) and Hispanic (55.7 percent) offenders. Black offenders also qualified for relief under the safety valve at the lowest rate of any other racial group (11.1 percent), compared to white (26.7 percent) and Hispanic (42.8 percent), either because of their criminal history or the involvement of a dangerous weapon in connection with the offense. Prior to the enactment of the safety valve (in fiscal year 1994), black and Hispanic offenders received relief from the mandatory minimum penalty at comparable rates (34.3 percent and 34.2 percent, respectively), although at a lower rate than white offenders (44.2 percent). After enactment of the safety valve provision, the rate at which Hispanic and white offenders obtained relief from a mandatory minimum penalty increased appreciably, while the rate for black offenders did not. Offenders subject to a mandatory minimum at sentencing received an average sentence of 139 months; offenders who received relief from a mandatory minimum received, on average, a 63-month sentence.

The Commission made a number of recommendations to Congress about statutory mandatory minimum penalties.


  • If Congress decides to exercise its power to direct sentencing policy by enacting mandatory minimum penalties, the Commission believes that such penalties should (1) not be excessively severe, (2) be narrowly tailored to apply only to those offenders who warrant such punishment, and (3) be applied consistently. Sentencing data and interviews with prosecutors and defense attorneys indicate that mandatory minimum penalties that are considered excessively severe tend to be applied inconsistently.
  • Congress should consider whether the statutory “safety valve” mechanism should be expanded.
  • Congress should request prison impact analyses from the Commission as early as possible in its legislative process whenever it considers enacting or amending mandatory minimum penalties.
  • Congress should eliminate the “stacking” requirement of 18 U.S.C. § 924(c), which requires that the penalties for multiple firearms offenses charged in the same indictment be served consecutively, to give the sentencing court discretion to impose sentences for multiple violations of section 924(c) concurrently with each other.5

The Commission has also undertaken a comprehensive report on child pornography offenses and sentencing, which will be released by year end. The child pornography report will include a thorough analysis of child pornography possession, receipt, distribution, and production offenses. These offenses make up a small but growing part of the federal docket.6 In October 2009, the Commission released the first report in a series providing a comprehensive history of the development of the child pornography guidelines. Over the last decade or so Congress has repeatedly expressed its concern in this area by creating new offenses, increasing penalties, and issuing directives to the Commission regarding child pornography offenses. In recent years, the Commission has received feedback from a number of groups, district courts prominently included, indicating that a review of the guidelines for child pornography offenses is appropriate at this time because of the evolving technology and changes in how these offenses are committed. In February 2012, we held a hearing about this issue, and the Commission heard testimony from leading experts in the fields of forensic computer technology and psychology.

One main purpose of the Sentencing Reform Act was to ensure white-collar offenses are treated as seriously as street crimes. A Senate Judiciary Committee report explained that, in the estimation of some in Congress, “some major offenders, particularly white collar offenders . . . frequently do not receive sentences that reflect the seriousness of their offenses.”7 The recent Dodd-Frank Act directed the Commission to ensure that economic fraud crimes receive adequate sentences. The Commission held hearings on the issue and increased penalties in certain areas such as insider trading. However, the Commission also heard that in high-loss securities fraud cases the loss table and victim table sometimes combine to produce an offense level that substantially overstates the seriousness of the offense. This can occur because in fraud on the market cases, for example, the high loss flows necessarily from the large pool of victims, and so increasing penalties for both may be impermissible double-counting. Thus judges are less likely to follow the guidelines, and the government is more likely to sponsor a below-range sentence; in FY2011, less than one-half of offenders sentenced under USSG § 2B1.1 loss amounts above $70,000 were sentenced within their guideline range and the within-range rate for loss amounts above $7 million dropped below one-third. In the 2012–13 amendment cycle, the Commission will continue its work on economic crimes, including a comprehensive, multiyear study of USSG § 2B1.1.

The Commission is required by statute to consider the impact of changes to the Federal Sentencing Guidelines on the federal prison population.8 A prison and sentencing impact assessment is the Commission’s appraisal of the impact that changes to the sentencing guidelines or criminal statutes will have on the federal prison population. To make these estimates, the Commission employs a predictive model that can, for example, assess a proposed change to the drug amounts that correspond to base offense levels (BOLs) in the Guidelines Manual. The Commission’s prison and sentencing assessment Web page was launched on July 3, 2012, and it contains information starting with analyses completed during the FY2012 amendment cycle.


One of the Commission’s core functions under the Sentencing Reform Act is to promulgate sentencing guidelines, which became advisory after Booker. The guidelines continue to exercise a gravitational pull on sentences and, in FY2011, over 80 percent of the over 86,000 cases processed by the Commission were sentenced either within the applicable guideline range or pursuant to a government-sponsored departure/variance. Non-government-sponsored below-range sentences accounted for approximately 12.5 percent of sentencings in the year after Booker. That number rose steadily until FY2010, when the rate of non-government-sponsored below-range sentences hit 17.8 percent.9 While the non-government-sponsored below-range rate dropped to 17.4 percent in FY2011,10 the fact remains that the variance rate is higher than it was.11 The Commission has observed increasing geographical differences in sentencing, with rates of non-government-sponsored below-range sentences by district ranging from 5.7 percent in the Middle District of Georgia to almost half (49.5 percent) in the Southern District of New York.

Particularly troubling, the Commission’s multivariate data analysis indicates that demographic differences in sentencing have increased in the wake of Booker and its progeny. Since the Court’s decision in Gall v. United States,12 black male defendants have received sentences that are, on average, 20 percent longer than those received by white male offenders. This is not to suggest that racial animus is driving sentencing; there may be variables, from the offender’s employment history and the violence of their criminal history, that may explain some of these differences. These explanatory variables notwithstanding, though, the trend is troubling. This finding has been replicated by other studies, albeit to differing magnitudes.

To combat the rise of unwarranted sentencing disparity, the Commission made a set of proposals to Congress that the Commission believes will strengthen the sentencing guidelines system. The proposals reflect existing Supreme Court case law. Briefly put, the Commission made the following proposals. Congress should require appellate courts to presume within-guideline sentences to be substantively reasonable.13 It also should require that the greater a sentencing judge’s variance from a guideline, the more compelling should be that judge’s justification for the variance.14 A heightened standard of review for sentences outside the guideline range that are based on sentencing judges’ policy disagreements with the guidelines should be established. The Supreme Court allowed for policy-based variances in Kimbrough15 and Spears.16 The Court explained, however, that “closer review may be in order when the sentencing judge varies from the guidelines based solely on the judge’s view that the guidelines range ‘fails properly to reflect § 3553(a) considerations’ even in a mine-run case.”17

The Commission recommends that Congress make clear the weight to be given the guidelines by district courts when imposing sentences. In Booker, the Supreme Court said that courts must “consider” the properly calculated guideline range but did not express exactly how much weight the guidelines should be given. For example, the Commission suggested in October 2011 testimony that Congress statutorily requires district courts to give “substantial weight” to the guidelines. Other possibilities include “due regard” or “respectful consideration.” Regardless of precise language, the Commission believes it is important for Congress to impose one uniform standard that conveys that the guidelines must be taken seriously.

The Commission also recommends to Congress that it codify the three-part sentencing protocol set out in Booker, in which sentencing courts begin by calculating the guidelines range, proceed to calculate departures within the guidelines, and then consider the sentencing factors listed at section 3553(a) taken as a whole.18

The Commission has also requested that Congress address the tension between 28 U.S.C. § 994 and 18 U.S.C. § 3553(a). Section 994 of Title 28 instructs the Commission not to incorporate certain offender characteristics (for example, family ties) into the guidelines, but some judges view 18 U.S.C. § 3553(a) as an instruction to courts to consider those same characteristics.


In the past, the Commission has published reports on recidivism, most recently with respect to crack offenders. For example, the Commission found that offenders who benefitted from the retroactive application of the Commission’s 2007 crack cocaine guidelines amendment were no more likely to recidivate than similarly situated offenders who served their (longer) original sentence. The Commission, however, has recently undertaken a more ambitious project with respect to recidivism. A Commission priority for the 2012–13 amendment cycle is to begin a comprehensive, multiyear study of recidivism, including (a) examination of circumstances that correlate with increased or reduced recidivism, (b) possible development of recommendations for using information obtained from such study to reduce costs of incarceration and overcapacity of prisons, and (c) consideration of any amendments to the Guidelines Manual that may be appropriate in light of the information obtained from such study. The Commission also hopes to begin a new effort in collecting supervised release and modification data. Recidivism statistics can help enable the assessment and improvement of much in the federal criminal justice system—from the guidelines themselves to drug and reentry courts.


The U.S. Sentencing Commission sits at the crossroads of all three branches of government. It relies on many stakeholders, including the American Bar Association, to ensure sentences are proportionate and uniform. We thank the American Bar Association for its important work in the sentencing area, particularly the thoughtful testimony it provides at our hearings. 


1. 543 U.S. 220 (2005).

2. 18 U.S.C. § 3553(a)(2). These purposes include the need for a sentence imposed “(A) to reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense; (B) to afford adequate deterrence to criminal conduct; (C) to protect the public from further crimes of the defendant; and (D) to provide the defendant with needed educational or vocational training, medical care, or other correctional treatment in the most effective manner.”

3. 28 U.S.C. § 991.

4. This report is available on the Commission’s web site,

5. An offender convicted of an underlying offense and two counts of an offense under section 924(c) will receive consecutive mandatory minimum penalties of at least 5 years and 25 years of imprisonment, in addition to any term of imprisonment imposed for the underlying offense and other counts of conviction. An offender charged with three counts of an offense under section 924(c) will face another consecutive 25-year mandatory minimum penalty. Such a result may occur even if the offender has no prior record.

6. Child pornography offenses made up approximately 2 percent of the federal docket in FY2011. U.S. Sentencing Comm’n, Fiscal Year 2011 Sourcebook of Federal Sentencing Statistics (2012).

7. S. Rep. No. 98-225 at 77 (1983).

8. 28 U.S.C. § 994(g); 18 U.S.C. § 4047.

9. U.S. Sentencing Comm’n, Fiscal Year 2010 Sourcebook of Federal Sentencing Statistics, tab. N (2011). The rate of sentences imposed above the guidelines range remained trivial throughout, and was 1.8 percent in 2010.

10. Fiscal Year 2011 Sourcebook, supra note 6, at tab. N.

11. The within-guidelines rate was 54.5 percent in FY2011, the lowest rate in 16 years.

12. 552 U.S. 38 (2007).

13. Rita et al. v. United States, 551 U.S. 338, 347 (2007) (“[i]f the sentence is within the guidelines range, the appellate court may, but is not required to, apply a presumption of reasonableness.”); id. at 350–51 (no such presumption is available to district courts).

14. In Gall v. United States, 552 U.S. 38, 50 (2007), it is “uncontroversial that a major departure should be supported by a more significant justification than a minor one.” This is not to create an “appellate rule that requires ‘extraordinary’ circumstances to justify a sentence outside the Guidelines range,” a rule rejected by the Supreme Court. Id. at 46, 47.

15. Kimbrough v. United States, 552 U.S. 85 (2007).

16. Spears v. United States, 555 U.S. 261 (2009).

17. Kimbrough, 552 at 89 (“On the other hand, while the Guidelines are no longer binding, closer review may be in order when the sentencing judge varies from the Guidelines based solely on the judge’s view that the Guidelines range ‘fails properly to reflect § 3553(a) considerations’ even in a mine-run case.”) (emphasis added). See also Pepper v. United States, 131 S. Ct. 1229, 1255 (2011) (Breyer, J., concurring) (“The appellate courts should review [sentences ‘that depart or vary from a specific guideline’] more closely when they rest upon disagreement with Guidelines policy.”).

18. The Commission amended USSG § 1B1.1(a)–(c) (Amendment 741, effective Nov. 1, 2010) to set out the three-step process.

The material in all ABA publications is copyrighted and may be reprinted by permission only. Request reprint permission here.

By Judge Patti Saris

Judge Patti Saris has served as a United States district judge for the District of Massachusetts since 1994. President Obama nominated Judge Saris to be a member and chair of the United States Sentencing Commission on April 28, 2010.