FEDERAL PUBLIC DEFENDER OFFICE
DISTRICT OF MASSACHUSETTS
51 SLEEPER STREET, 5TH FLOOR
BOSTON, MASSACHUSETTS 02210
TELEPHONE: 617-223-8061
FAX: 617-223-8080
December 3, 2012

Honorable Patti B. Saris
Chair
United States Sentencing Commission
One Columbus Circle, N.E.
Suite 2-500, South Lobby
Washington, D.C. 2002-8002
Dear Judge Saris:
As we discussed at the National Sentencing Policy Institute in October, we have some
suggestions regarding how the Commission’s report on sentencing after Booker v. United States,
543 U.S. 220 (2005) might present the issues in a complete and balanced way. This letter is not
meant to reiterate every issue that we have raised in our previous letters and testimony, but to
provide specific suggestions.
We urge the Commission to include the following information in its report for the reasons
that follow:









Complete statistics and information regarding regional differences, including whether
variation in sentence lengths has grown; the kinds of cases, rates of governmentsponsored departures and variances, extents of variances and departures, average
guideline minimums, sentences imposed; and the interaction between prosecutorial and
judicial sentencing practices
A complete account of current data and research regarding racial disparity in sentencing,
including appropriate cautions regarding the Commission’s findings, the evidence that
the primary cause of racial disparity is not judges, but prosecutorial charging and plea
bargaining practices, and the evidence that judges have reduced racial disparity
Current data through FY 2012 regarding judicial departures and variances
Current data through FY 2012 regarding departures and variances sought or agreed to by
prosecutors
The role of unwarranted severity in causing disparities in the sentencing practices of both
judges and prosecutors
A fair account of the views of appellate judges and the appellate caselaw

1

 
I.

The Report Should Include Complete Statistics and Information Regarding
Regional Differences.

Rather than focusing solely on rates of non-government sponsored below-guideline
sentences,1 the Commission should examine whether variation in sentence length has grown,
should include the kinds of cases, rates of government-sponsored departures and variances,
extents of variances and departures, average guideline minimums, and sentences imposed in
different districts, and should attempt to shed light on the interaction between prosecutorial and
judicial sentencing practices.2
A.

Sentence Length as a Neutral and Relevant Measure of Whether Differences
Among Districts Have Grown

One way to present a more complete picture is to look at the bottom line: whether
variation among districts in sentence length has grown over time. Sentence length reflects the
influence of both judges and prosecutors on guideline ranges, and rates and extents of
variances/departures, as well as the mix of cases, changes in guidelines and statutes, and other
factors that affect the rates and extents of below-range sentences. Moreover, the Commission
cannot accurately compare non-government sponsored rates (or government-sponsored rates)
before and after 2003 because until 2003, the Commission reported a large number of
government-sponsored departures in the same category as non-government sponsored
departures.3 Comparing variation in sentence length over time avoids these problems and
presents a more accurate picture of whether differences among districts really have increased.
                                                       
1

In its October 2011 testimony, the Commission said that there had been “growing disparities among
circuits and districts.” In support of this claim, it listed, from highest to lowest, rates of non-government
sponsored below-range sentences for each district in fiscal year 2010, and set forth the difference between
the highest and lowest rates of non-government sponsored below-range sentences by district for certain
types of offenses during a 33-month period after Gall. See Statement of Judge Patti B. Saris, Chair, U.S.
Sent’g Comm’n, Uncertain Justice: The Status of Federal Sentencing and the U.S. Sentencing
Commission Six Years After Booker: Hearing Before the Subcomm. on Crime, Terrorism, & Homeland
Sec. of the H. Comm. on the Judiciary, 112th Cong. 55-60 (2011) at 25, 28, 31, 33, 36, 38, 41, 43, 46, 48,
50, 53; Appendix D [hereinafter “Commission Testimony].

2

Analyzing sources of regional disparity “cannot be resolved through simple examination of the reported
rates” because “the potential sources are so many, varied, and interacting.” U.S. Sentencing Comm’n,
Fifteen Years of Guideline Sentencing: An Assessment of How Well the Federal Criminal Justice
System is Achieving the Goals of Sentencing Reform 93, 111 (2004).

3

See U.S. Sent’g Comm’n, Downward Departures from the Federal Sentencing Guidelines 60 (2003)
(reporting that after subtracting government-initiated downward departures, the 2001 rate of “other”
downward departures was reduced from 18.1% to 10.9%); U.S. Sent’g Comm’n, 2002 Sourcebook of
Federal Sentencing Statistics tbl.26 (reporting only two categories of downward departures: “substantial
assistance departure[s]” and “other downward departure[s]”); U.S. Sentencing Comm’n, 2004
Sourcebook of Federal Sentencing Statistics tbl.26A (2004) (reporting three categories: “substantial
assistance departure[s],” “gov’t initiated downward departures,” and “other downward departures”).

2

 
A straightforward comparison of inter-district variation in sentence lengths in 2003 with
inter-district variation in sentence lengths in 2011 shows that it has grown very little. The chart
below plots mean sentences by district in 2003 and 2011 using SENTOT0, with mean sentence
in months on the vertical axis and district number on the horizontal axis.
Variation by District in Average Sentence Length (Probation = 0)

Researchers using multivariate methods have also concluded that variation in sentence
length among districts has not grown since Booker, and has even decreased. Ulmer, Light and
Kramer found that the percentage of sentence length variation explained by differences among
districts was 6.6% before the PROTECT Act, 5.8% after the PROTECT Act, 5.2% after Booker,
and 6.3% after Gall through 2009. 4 Lynch and Omori, analyzing drug cases from 1993
through 2009, found that the proportion of variation in sentence length due to differences among
districts was 14.1% before Koon, 12% after Koon, 13.6% after the PROTECT Act, 13.9% after
Booker, and 13.1% after Kimbrough.5
B.

Guideline Severity, Rates, Sentences Imposed

More complete information would help shed light on the reasons for different rates of
non-government sponsored below-range sentences among districts and their ultimate impact.
For example, the relatively low rates of non-government sponsored below-range sentences in the
Middle District of Georgia (5.3%), the Western District of Texas (12.3%), and the Southern
District of Texas (15.7%) have been held up for comparison to districts with higher rates of nongovernment sponsored below-range sentences, in particular the Southern District of New York
(49%), the District of Massachusetts (30.8%), and the District of Minnesota (44.5%). The reason
for these differences is illuminated by the relative severity of guideline ranges. In 2011, the
                                                       
4

See Jeffery Ulmer et al., The “Liberation” of Federal Judges’ Discretion in the Wake of the
Booker/Fanfan Decision: Is There Increased Disparity and Divergence Between Courts?, 28 JUSTICE Q.
799, 816 (2011).

5

See Lynch & Omari, Legal Change and Sentencing Norms in the Wake of Booker: The Impact of Time
and Place on Drug Trafficking Cases in Federal Court at 46, tbl.3, Criminology, Law and Society,
University of California, Irvine, Paper presented at Conference on Empirical Legal Studies, Stanford Law
School (Nov. 2012).

3

 
average guideline minimums in the Middle District of Georgia, the Western District of Texas,
and the Southern District of Texas were only 44 months, 35 months, and 34 months respectively.
The average guideline minimums in the Southern District of New York, the District of
Massachusetts, and the District of Minnesota were twice as high, at 72 months, 71 months, and
83 months respectively. And despite the higher rates of below-range sentences in the latter
districts, the average sentence lengths – 55 months, 59 months, and 72 months respectively –
well exceeded the national average of 49 months. It seems that in some districts, judges are
correcting for overly-harsh guideline ranges, and in others there is less need to do so.
C.

Rates and Kinds of Government-Sponsored Below-Range Sentences

The Commission has cited the growth in the difference between high and low rates of
non-government sponsored below-guideline sentences by circuit from 2006 to 2010 as evidence
of “increased inconsistencies” as a result of Booker.6 Yet the growth in the difference between
the high and low rates of government sponsored below-guideline sentences by circuit has nearly
tripled that of the non-government sponsored rate: the difference between the high and low rates
of government-sponsored below-guideline sentences by circuit grew by 20.7 percentage points
from 2006 to the third quarter of 2012, while the difference in non-government sponsored rates
by circuit grew by only 8.1 percentage points.7 And the difference between the highest and
lowest rates of government-sponsored below guideline sentences by district has been
consistently greater than the difference between the highest and lowest non-government
sponsored rates.8 In short, government-sponsored below-range sentences have consistently
contributed more to regional variation and to the growth in that variation than judicial belowrange sentences.
It would also help to clarify that some districts have low non-government sponsored rates
of below-guideline sentences and others have high rates in certain kinds of cases because of the
interaction with government-sponsored below-guideline sentences. For example, in 2011,
Arizona had a 3.9% rate of non-government sponsored below-range sentences in immigration
cases, while the Southern District of New York had a 59.7% rate. This is explained by the fact
that Arizona had a very high rate of government-sponsored sponsored below-guideline sentences
                                                       
6

U.S. Sent’g Comm’n, Mandatory Minimum Penalties in the Criminal Justice System 347 (2011).

7

See U.S. Sent’g Comm’n, 2006 Sourcebook of Federal Sentencing Statistics, tbl. 26 (government
sponsored rate of 16.4% for the First Circuit and 31.4% for the Ninth Circuit; non-government-sponsored
rate of 7.3% for the Fifth Circuit and 24.1% for the Second Circuit); U.S. Sent’g Comm’n, 2012
Preliminary Quarterly Data Report, Third Quarter, tbl. 2 (government sponsored rate of 14.3% for the
Fifth Circuit and 50% for the Ninth Circuit; non-government-sponsored rate of 10.8% for the Ninth
Circuit and 35.7% for the Second Circuit).
8

For example, in 2011, the difference between the highest and lowest rates of government-sponsored
below guideline sentences by district was 12.5 percentage points higher than the difference between the
highest and lowest non-government sponsored rates. See U.S. Sent’g Comm’n, 2011 Sourcebook of
Federal Sentencing Statistics, tbl.26.

4

 
in immigration cases (60.4%) because it has a fast track program, while the Southern District of
New York had a very low rate of government-sponsored sponsored below-guideline sentences in
immigration cases (1.5%) because it has no fast track program. The end result is that average
sentence length was 17 months in Arizona and 22 months in the Southern District of New York.
Providing information like this would put the statistics in better perspective.9
D.

Prosecutorial Practices and Problematic Guideline Applications in Certain
Districts

A full and accurate assessment of differences among districts requires an in-depth
examination of charging and plea bargaining practices and anomalies in guideline application in
each district. A few examples would help shed light on the issues. For example, the career
offender, firearms, and immigration guidelines apply differently in different districts depending
on peculiarities of state and circuit law with respect to prior convictions. And in some districts
more than others, prosecutors bring minor drug cases that would otherwise be prosecuted in state
court when the career offender guideline applies. This is especially problematic – and creates
unwarranted geographical disparity – when the career offender guideline would not apply in
most other districts. See, e.g., United States v. Curet, 670 F.3d 296 (1st Cir. 2012) (defendant
who pled guilty to selling 12 grams of crack was a career offender as a result of a “guilty filed,”
which is not a conviction under Massachusetts law, for resisting arrest, which is not a felony
under Massachusetts law or the law of most states but is punishable by up to 2 1/2 years).
As a result of peculiarities of Massachusetts law, First Circuit law, and prosecutors’
practices in the District of Massachusetts, career offender cases comprised 11.2% of all cases in
the district in 2011, the third highest career offender caseload in the country. The withinguideline rate in career offender cases was only 11.7%, with 45% non-government sponsored
departures/variances and 41.7% government sponsored departures/variances.10 The two districts
with higher career offender caseloads were the Western District of Pennsylvania and the District
of Maryland, with career offender cases comprising 11.8% and 11.4% of their caseloads
respectively. Like Massachusetts and unlike most other states, Pennsylvania and Maryland have
misdemeanors that are punishable by more than one year. In the Western District of
Pennsylvania, the within-guideline rate was 13%, with 63% non-government sponsored
departures/variances and 0% government sponsored departures/variances. In the District of
Maryland, the within-guideline rate was 23.6%, with 49.4% non-government sponsored
departures/variances and 0% government sponsored departures/variances.11 It appears that
judges in these districts are alleviating unwarranted geographic disparity, as well as the
inappropriate severity of the career offender guideline in many cases.

                                                       
9

Cf. Commission Testimony at 26 (stating that there was a difference of 65.6 percentage points between
two unnamed districts in non-government sponsored rates in immigration cases).

10

USSC 2011 Monitoring Dataset.

11

Id.

5

 
II.

The Report Should Include a Complete Account of Current Data and Research
Regarding Racial Disparity in Sentencing.

The Commission’s research on racial disparity has recently focused on only one potential
source of disparity: judicial discretion. The Commission’s multivariate analyses of judicial
sentencing decisions attempts to measure average differences among racial groups after
controlling for some legally relevant factors, such as the presumptive sentence,
departure/variance status, and other factors. This method fails to capture disparity arising from
prosecutorial discretion before the case ever gets to a judge for sentencing, and fails to measure
structural disparity built into the rules that determine the presumptive sentence. This method
also conceals the benefits to minority defendants of expanded judicial discretion to sentence
outside the range.
The Commission should include the substantial evidence, including its own research, that
unjustified racial disparities continue to result from prosecutors’ decisions, as well as many of
the statutory and guideline rules themselves. It also should acknowledge that judges reduce
racial disparity when they use their discretion to offset disparities stemming from prosecutorial
discretion, unsound rules, or a combination of the two. Without a full assessment of all sources
of and remedies for disparity, the consequences of legislation that would constrain judicial
discretion and give prosecutors greater control over sentencing outcomes will not be understood.
A.

The Commission should present more complete data bearing on the effect of
judicial discretion at the sentencing stage.

One straightforward way to identify the effect of increased judicial discretion is to
compare the average sentences recommended or required for different groups by the applicable
guidelines and statutes—i.e., the presumptive sentence—with the average sentences actually
imposed after the exercise of judicial discretion. In fiscal year 2011, the gap between black and
white males in the mean guideline minimum was 31.4 months, but the gap in the mean sentence
imposed was 29.3 months.12 That is, the racial gap is smaller after the exercise of judicial
discretion than before it – the opposite effect than suggested by the Commission’s multivariate
analysis. We urge the Commission to include this data.
Further, while the Commission previously reported “growing” demographic disparities,
the Commission’s measure of a difference in sentence length between black and white males
after Gall and Kimbrough has decreased as the period lengthens, from 23.3% as reported in its
2010 report, to 20% as reported in its October 2011 testimony, to 19.5% as stated by Mr.
Blackwell in October 2012. This decrease should be noted in the report. As we explain below,
the Commission should also report the fluctuations that have occurred in shorter time periods not
defined by dates of Supreme Court decisions, in order to provide context for evaluating whether
any changes after those decisions are greater than might be expected from other sources.
                                                       
12

For this analysis, SENTOT0 was used for average sentence length, and GLMIN was used for guideline
minimum. Both the guideline minimum and average sentence length were capped at 470 months.

6

 
B.

The Commission should give clear and appropriate cautions.

Many factors that legitimately affect sentences and that might explain the differences in
sentences imposed on different groups are not included among the control variables in the
Commission’s multivariate studies. We were happy to hear Mr. Blackwell acknowledge in
Memphis that the Commission’s study does not take account of all relevant factors. The
Commission acknowledged that unmeasured factors likely affected its results in U.S. Sent’g
Comm’n, Demographic Differences in Federal Sentencing Practices: An Update of the
Booker Report’s Multivariate Regression Analysis 4, 9-10 (2010). In two submissions to
Congress, however, the Commission not only omitted this information but strongly suggested
that its study controlled for all relevant factors. See Commission Testimony at 53 (“The
principal benefit of this tool is that it accounts, or controls, for the effect of each factor in the
analysis. Each factor is separately assessed and the extent to which each factor influences the
outcome is measured.”); U.S. Sent’g Comm’n, Mandatory Minimum Penalties in the Criminal
Justice System 347 (2011) (“[A] recent Commission analysis found that, after controlling for
relevant factors, Black male offenders received longer sentences than White male offenders, and
that those differences in sentence length have increased steadily since Booker.”). We hope that
the Commission will correct this impression in its Booker report.
The Commission should also clearly state appropriate cautions about causation. While
Commissioners and staff have denied that the Commission is suggesting that the “race effects” it
finds are unfair or are caused by judges, this point has been undercut by the fact that the
Commission is calling for legislation that would constrain judicial discretion. The fact that data
is missing on many factors that may differ among groups that judges legitimately consider in
sentencing is one reason to question any attribution of unfairness to judicial decision making, as
the Commission cautioned in its 2010 report.
In addition, previous Commission reports discussed year-to-year fluctuations in the
regression coefficient for race as evidence that these coefficients vary due to a variety of factors
having nothing to do with discrimination by judges. But the Commission’s recent work ignores
these fluctuations. Moreover, by creating time periods defined by legislation (the PROTECT
Act) and Supreme Court decisions, the Commission invites readers to assume that differences
among these time periods are due to the legal changes defining the periods. Commission
research examining neutrally-defined time periods found race effects in some years but not in
other years, and only for drug offenses in some years and non-drug offenses in other years. See
U.S. Sentencing Comm’n, Fifteen Years of Guideline Sentencing: An Assessment of How
Well the Federal Criminal Justice System is Achieving the Goals of Sentencing Reform 121127 (2004). These fluctuations, observed during periods of relative legal stability, led the
Commission to caution that “[o]ffense-to-offense and year to year fluctuations in racial and
ethnic effects are difficult to reconcile with theories of enduring stereotypes . . . or overt
discrimination” on the part of judges. Id. at 125. The Commission gave the same warning in
its first Booker report. See U.S. Sent’g Comm’n, Final Report on the Impact of United States v.
Booker on Federal Sentencing 108 & n.320 (2006).
Cautions such as this remain valid, and are even more important today, because the
Commission has suggested that changes in the coefficient associated with race can be attributed
7

 
to the legal changes it has used to define its time periods. For example, the Commission finds
race effects for males but not females. Further, there have been fluctuations in shorter time
periods since Booker, thus suggesting that many factors other than legal changes affect the
coefficient for race. Hispanic males reportedly received sentences 6.8% longer than white males
during the twenty-one-month period after Gall through 2009, but this dropped to statistical
insignificance during the thirty-three-month period through 2010 reported in the Commission’s
2011 testimony. Likewise, the reported aggregate black male/white male difference decreases as
the study period grows longer, from 23.3% as reported in 2010, to 20% as reported in October
2011, to 19.5% as reported by Mr. Blackwell in October 2012. The Commission should use
shorter, neutrally-defined time periods, and more sophisticated statistical models that measure
long term trends and random fluctuations, to help prevent the false inference that the differences
among the time periods were caused by the legal changes used to define the time periods. Short
of that, the Commission should caution that studying shorter time periods and different offense
types would likely show different results.
Most important, the Commission should acknowledge the growing body of research that
questions its conclusions and applies different methods to study disparity arising from a wider
range of sources. The Penn State study, released shortly after the Commission’s study, reached
different results using the same data and the same presumptive sentence model but a
methodology that differed from the Commission’s in other respects. Assuming that plausible
arguments can be made on either side as to who made the right choices,13 this only proves the
point that “findings that are sensitive to minor changes in model specifications such as these
must be interpreted with caution.” Douglas C. McDonald & Kenneth E. Carlson, U.S. Dep’t of
Justice, Sentencing in the Federal Courts: Does Race Matter? The Transition to Sentencing
Guidelines, 1986–90, AT 106 (1993); see also Fifteen Year Review at 127 (noting the same
regarding the Commission’s own fluctuating results); U.S. Sent’g Comm’n, Final Report on the
Impact of United States v. Booker on Federal Sentencing 108 (2006) (same).
Other researchers, using entirely different models, have questioned the Commission’s
apparent inference of causation and provided strong evidence that racial disparity in sentencing
outcomes is not caused by increased judicial discretion at all. See Part C.2 and 3, infra. This is
further reason to caution against inferring causation.

                                                       
13

For example, according to Rehavi and Starr, either the Commission’s approach of combining the
incarceration decision with the sentence length decision or Ulmer et al’s approach of separating them is
appropriate, but there were problems with both in the way they were implemented. See Sonja B. Starr &
M. Marit Rehavi, Racial Disparity in the Criminal Justice Process: Prosecutors, Judges, and the Effects
of United States v. Booker 30 n.109 & 46 n.145 (U. Mich. L. & Econ. Research Paper No. 12-021, 2012),
available at http://ssrn.com/abstract=2170148.

8

 
C.

The Commission Should Present Data and Research That Shed Light on
Whether Constraining Judicial Discretion and Enlarging Prosecutorial
Power Would Increase Racial Disparity.

More recent studies using different methodologies than the Commission or Penn State
conclude that racial disparity after Booker is not caused by judicial discretion, but primarily by
prosecutors’ use of mandatory minimums, and that increased judicial discretion appears to have
lessened racial disparity. In the interest of a fully informed understanding of the tradeoff
between judicial discretion and prosecutorial power resulting from any legislative fix, we urge
the Commission to present the findings of Professors Fischman and Schanzenbach and those of
Professors Starr and Rehavi. We also urge the Commission to include its own research bearing
on whether it would be wise to transfer more sentencing power to prosecutors.14
1.

The Commission’s data show that racial disparity results from the
exercise of prosecutorial discretion when it is unchecked by the
balancing influence of judges.

The Commission’s data show that black offenders are more likely to be charged with
more severe mandatory minimums than similarly situated white offenders. Among drug
trafficking defendants in 2010 who received either a § 924(c) conviction or a weapon
enhancement under § 2D1.1, 36% of black defendants, but only 26% of white defendants,
received the harsher § 924(c) enhancement. See Paul J. Hofer, Review of U.S. Sentencing
Commission’s Report to Congress: Mandatory Minimum Penalties in the Criminal Justice
System, 24 Fed Sent Rep 193, 198 (2012). The Commission previously reported that among
offenders who possessed or used a gun during a drug offense in 1995, black offenders were 48%
of those who qualified for a mandatory minimum of five or more years under 18 U.S.C. § 924(c),
but 56% of those charged with a mandatory minimum and 64% of those who actually received it.
Data from 2000 showed the same pattern. See U.S. Sent’g Comm’n, Fifteen Years of Guideline
Sentencing: An Assessment of How Well the Federal Criminal Justice System is Achieving
the Goals of Sentencing Reform 90-91, 131 (2004).
The recent mandatory minimum report found the same kind of disparity under 21 U.S.C.
§ 851, which requires the government to file an information: 29.9% of eligible black drug
offenders received an increased mandatory minimum under § 851, while only 25% of eligible
white offenders, 19.9% of eligible Hispanic offenders, and 24.8% of offenders of “other” races
received the increase. U.S. Sent’g Comm’n, Mandatory Minimum Penalties in the Criminal
Justice System 257-58 (2011). The Commission also reported that “stacked” § 924(c) counts,
which result in sentences that are “excessively severe and disproportionate to the offense
committed,” are charged against black defendants at a higher rate than defendants of other races
and that this creates “perceptions of unfairness and unwarranted disparity.” Id. at 359-60, 363                                                       
14

As the Commission and many others have recognized, discretion is hydraulic: reducing judicial
discretion has the effect of giving prosecutors greater control over sentencing outcomes. See, e.g., U.S.
Sent’g Comm’n, Fifteen Years of Guideline Sentencing: An Assessment of How Well the Federal
Criminal Justice System is Achieving the Goals of Sentencing Reform 92 (2004); Rehavi & Starr, at 6-8
(reviewing the literature).

9

 
64. Further, black offenders receive government-sponsored substantial assistance departures less
often than defendants of other races. Id. at 159-60, 179, 214-15, 221, 291.
This data regarding racial disparity in prosecutors’ use of mandatory minimums and
substantial assistance motions has direct implications for the likelihood that a legislative “fix”
that would constrain judicial discretion at the cost of enlarging prosecutorial control over
sentencing outcomes would increase racial disparity. The Commission should include these
findings in its report on sentencing after Booker, just as it included discussion of its multivariate
study of judicial discretion in its report on mandatory minimums. Id. at 347.
2.

Fischman and Schanzenbach’s research shows that racial disparity
narrows in periods of deferential review, that the primary cause of
racial disparity after Rita, Gall and Kimbrough is mandatory
minimums that impede judicial discretion, and that judicial discretion
does not contribute to, and may in fact mitigate, racial disparities.

Innovative research by econometricians has recently employed sophisticated models that
are able to examine a wider scope of decision making than the models employed by the
Commission or Penn State. Professors Fischmann and Schanzenbach focus on the effect of the
standard of review and the effect of prosecutors’ use of mandatory minimums on racial disparity.
Like the Commission and Penn State, they examined data only at the sentencing stage, but unlike
the Commission and Penn State, they did not use the presumptive sentence, departures/variances,
or the presence of mandatory minimums as control variables. Instead, they examined these
factors as outcome variables that could themselves be affected by the response of prosecutors
and/or judges to changes in the law, i.e., Koon, the PROTECT Act, Booker,
Rita/Gall/Kimbrough. Joshua B. Fischman & Max M. Schanzenbach, Racial Disparities Under
the Federal Sentencing Guidelines: The Role of Judicial Discretion and Mandatory Minimums, 9
J. Empirical Legal Stud. 730-31, 737-43 (2012).
The results shed a very different light on the effects of judicial and prosecutorial decision
making. They found that the “disparity in departure rates and prison sentences [for black
offenders] relative to whites narrows in periods of deferential review, but black offenders are
sentenced more often at the statutory minimum even as their offense levels do not change. In
other words, when judges are freer to depart, they do so proportionally more often for blacks
than whites, resulting in lower prison sentences. However, judges appear to be constrained more
frequently by mandatory minimums when sentencing black defendants.” Id. at 746; see also id.
at 743-44.
After Rita, Gall and Kimbrough, departures substantially increased for both white and
black offenders. Id. at 757. However, the percentage of black offenders sentenced at the
mandatory minimum also increased. Id. at 752. Prison sentences declined for white offenders,
but for black offenders, prison sentences declined only in cases unlikely to involve a binding
mandatory minimum and therefore stayed relatively flat overall. Id. at 757, 761. In cases not
likely to involve a binding mandatory minimum, sentences fell by the same proportion for whites
and blacks. Id. at 761.

10

 
The departure results showed that judges prefer more lenient prison sentences for black
offenders, but this did not translate into lower prison sentences overall because lower sentences
were impeded by trumping mandatory minimums more often for black offenders than for white
offenders. Id. at 757, 761. Thus, while the “disparity between white and black offenders in
prison sentences increase[d]” after Rita, Gall and Kimbrough, “this is largely a consequence of
the mandatory minimums.” Id. at 752. They found “no evidence that the increase in racial
disparities after RGK is due to the biased exercise of judicial discretion.” Id. at 730. They
concluded that “judicial discretion does not contribute to, and may in fact mitigate, racial
disparities in Guidelines sentencing.” Id. at 761. Thus, “[p]olicymakers interested in redressing
racial disparity today should pay much closer attention to the effects of mandatory minimums
and their effect on prosecutorial and judicial discretion.” Id.
3.

Rehavi and Starr’s research shows that prosecutors’ charging and
plea bargaining decisions in the use of mandatory minimums is the
primary driver of racial disparity in sentencing, and that there is no
evidence that Booker increased racial disparity in the exercise of
judicial discretion.

Professors Rehavi and Starr used data from multiple sources to study the criminal justice
process from arrest through sentencing. See Sonja B. Starr & M. Marit Rehavi, Racial Disparity
in the Criminal Justice Process: Prosecutors, Judges, and the Effects of United States v. Booker
18-19 (U. Mich. L. & Econ. Research Paper No. 12-021, 2012), available at
http://ssrn.com/abstract=2170148. They rejected use of the presumptive sentence as a control
variable because that approach assumes that the presumptive sentence does not itself reflect
unwarranted disparity in prosecutors’ charging and plea bargaining decisions and their
substantial control over the facts that eventually get to the judge; that approach could lead to
misguided policy changes because disparities found at the sentencing stage cannot safely be
attributed to judicial discretion without examining disparity at the earlier stages that produced the
presumptive sentence. Id. at 10-16. They also questioned whether changes in disparity can be
attributed to changes in the law regarding judicial discretion without testing long term trends and
other changes that have occurred over time. Id. at 3, 20, 30-31.
Rehavi and Starr used the arrest offense rather than the presumptive sentence as the key
case-severity control variable, and assessed disparities introduced throughout the post-arrest
process. Id. at 16-17. They found a 10% black-white disparity in non-drug cases, which rose to
14% when drug cases were added. Id. They also found that black men were twice as likely as
white men to be charged with a mandatory minimum offense. When they controlled for the
charged mandatory minimum, half of the otherwise-unexplained racial disparity in average
sentence disappeared. When they controlled for the mandatory minimum of conviction, all of
the otherwise-unexplained racial disparity disappeared. Id. at 19. In other words, disparity in the
charging of mandatory minimums caused all of the racial disparity in average sentence length—a
finding consistent with Fischman and Schanzenbach, but contrary to studies that examine only
sentencing decisions using the presumptive sentence as the primary control.15
                                                       
15

Rehavi and Starr combined the incarceration decision and the sentence length decision as the
Commission did, but counted non-prison sentences as 0 months, rather than .01 months as the

 
11

 

Rehavi and Starr used a regression discontinuity design to isolate the effect of Booker on
racial disparity attributable to judges. Id. at 32-40. Again, like Fischman and Schanzenbach,
they found “no evidence that Booker increased racial disparity in the exercise of judicial
discretion; if anything it may have reduced it.” Id. at 40. And notably, in light of the
Commission’s proposal to restrict judicial discretion to consider certain offender characteristics,
they found that there was no socioeconomic disparity in sentencing after Booker. Id. at 22, 50.
Rehavi and Starr concluded that it is “myopic” to focus on judicial sentencing alone
because “racial disparities in recent years have been mostly driven by the cases in which judges
have the least sentencing discretion: those with mandatory minimums.” Id. at 48. They warned
that using the Commission’s contrary results as the basis to constrain judicial discretion would be
counterproductive, first, because it would not reduce disparities but increase them, and, second,
because increased rigidity would also increase severity, primarily for black men. Id. at 49.
4.

Evidence showing that judges have reduced racial disparity that is
built into the rules

None of these studies takes account of racial disparity that is built into the statues and
guidelines themselves, including not just crack, but the career offender guideline, the criminal
history score, and various mandatory minimums, particularly 18 U.S.C. § 924(c) and 21 U.S.C.
§§ 841 and 851. Judges can and do exercise their discretion to reduce unwarranted racial
disparity caused by guidelines that have a disproportionate impact on black offenders without
clearly serving any purpose of sentencing. For example, in fiscal year 2010, by imposing
below-guideline sentences they would not (and could not) have imposed under the mandatory
guidelines, judges spared more than 860 African American defendants sentenced under the crack
or career offender guidelines over 3300 years of unnecessary incarceration.16 We hope that the
Commission will include some analysis of how judges have reduced racial disparity built into
certain rules.
It has not been our experience that judges exercise discretion in a biased manner against
black men. It has been our experience that prosecutors’ use of harsh rules appears to have a
racially disparate impact, and that judges, when given the opportunity, mitigate that harshness in
an evenhanded way. There is mounting empirical evidence that this is in fact the case. Thus,
                                                                                                                                                                               
Commission did, and also excluded immigration offenses as Ulmer et al did. Id. at 30 n.109, 45. They
also tested the incarceration decision separately, and found “no significant disparity in the incarceration
decision after controlling for arrest offense.” Id. at 30 n.109.
16

This estimate was made using the Monitoring Datasets for fiscal years 2003 and 2010. It is based on
the increase in the rate of non-government sponsored below-guideline sentences for crack and career
offenders in fiscal year 2010 as compared to the rate in 2003 and the average extent of these reductions.
Fiscal year 2003 was used as the comparison year because it preceded the Supreme Court’s decision in
Blakely v. Washington, 542 U.S. 296 (2004), which affected how cases were handled in anticipation of
Booker.

12

 
giving prosecutors more control over sentencing outcomes – through a stricter standard of review
and various other devices the Commission has proposed to make the guidelines more binding –
is no solution to racial disparity and in fact would be harmful to racial minorities. While we
understand that the Commission has committed itself to proposing this legislation, we believe it
should present all of the relevant evidence.
III.

The Report Should Present Current Data Regarding Both Judicial and
Prosecutorial Departures and Variances.

The Commission’s first Booker report, released in March 2006, was based on data
extracted as of February 22, 2006. It has now been two months since fiscal year 2012 ended.
We urge the Commission to provide the most current picture of the sentencing data in this
Booker report, including data through 2012.
That data demonstrates that the “increasing” trend previously cited by the Commission
has reversed itself. The non-government sponsored rate of below-guideline sentences has
decreased in a sustained manner, from 17.8% in 2010 to 17.4% in 2011 to 16.7% during the first
three quarters of 2012.17 And the extent of non-government sponsored below-range sentences
has decreased significantly or remained the same in every category from 2011 to 2012.18
At the same time, contrary to the Commission’s recent statement that the rate and extent
of government-sponsored below-range sentences have remained “about the same” since Booker,
the government-sponsored rate has grown more than the non-government sponsored rate, despite
the fact that it was so much higher to begin with. The government-sponsored rate has grown by
4.1% (from 23.8% in 2005 post-Booker to 27.9% in the first three quarters of 2012), while the
non-government sponsored rate has grown by 3.8% (from 12.9% to 16.7%).19 Further, while the
extent of substantial assistance departures has decreased by .1%, the extents of fast track
departures and government-sponsored variances have increased by 3.8% and 1.7%
respectively.20
                                                       
17

U.S. Sent’g Comm’n, 2012 Preliminary Quarterly Data Report, Third Quarter, tbl. 1; U.S. Sent’g
Comm’n, 2011 Sourcebook of Federal Sentencing Statistics, tbl.N; U.S. Sent’g Comm’n, 2010
Sourcebook of Federal Sentencing Statistics, tbl.N.

18

Compare U.S. Sent’g Comm’n, 2011 Sourcebook of Federal Sentencing Statistics, tbls. 31A-31D
(degree of decrease for non-government sponsored below-range sentences: 33.3%, 46.4%, 35.1%, 58.5%)
with U.S. Sent’g Comm’n, 2012 Preliminary Quarterly Data Report, Third Quarter, tbls. 10-13 (degree of
decrease for non-government sponsored below-range sentences: 32.7%, 39.3%, 35.1%, 33.5%).
19

U.S. Sent’g Comm’n, 2005 Sourcebook of Federal Sentencing Statistics, Section III, tbl. 26; U.S.
Sent’g Comm’n, 2012 Preliminary Quarterly Data Report, Third Quarter, tbl. 1.

20

Compare U.S. Sent’g Comm’n, 2011 Sourcebook of Federal Sentencing Statistics, tbls. 30-31 (degree
of decrease for government-sponsored below-range sentences: 50%, 30%, 36.4%) with U.S. Sent’g
Comm’n, 2012 Preliminary Quarterly Data Report, Third Quarter, tbls. 7-9 (degree of decrease for
government-sponsored below-range sentences: 48.9%, 33.8%, 38.1%).

13

 

In addition, the nature of government-sponsored below-guideline sentences has shifted.
The rate of substantial assistance departures has dropped from 14.7% to 11.5%, while the rates of
fast track departures and government-sponsored variances have each nearly doubled (fast track
departures from 6.2% to 11.5% and variances from 2.9% to 4.9%).21 In child pornography cases,
the rate of government-sponsored variances is now 17.9%.22 Further, the government agrees to
or does not oppose more than half of sentences classified as non-government sponsored below
range.23 Taken together, the government moves for or does not object to a below-guideline
sentence in at least 36.3% of all cases, while judges impose below-range sentences without a
government motion and over the government’s objection in at most 8.4% of cases. Put another
way, the government moves for 62.6% of all below-guideline sentences and does not object to at
least another 18.7% of such sentences.
It is important that the Commission include information regarding prosecutorial practices
in its Booker report for several reasons. Where more than a quarter of defendants are found or
plead guilty to a charge carrying a mandatory minimum sentence and nearly 97% of defendants
plead guilty, it is obvious that prosecutors have an enormous influence over sentencing
outcomes. After Booker, prosecutors still control the applicable guideline, as well as the
minimum and maximum sentence, through their charging and plea bargaining decisions, and
they are largely in control of the severity of the guideline range by deciding which aggravating
facts to convey to the judge. In addition, prosecutors seek below-guideline sentences – and at a
significantly increasing rate – including variances under § 3553(a).
While Booker brought needed balance between judges and prosecutors in their influence
on sentencing outcomes,24 the Commission proposes a series of legislative changes that would
constrain judicial discretion and consequently increase prosecutorial power over sentencing
outcomes. Some members of Congress, apparently based on the Commission’s presentation last
October, voiced support for a mandatory guidelines system which would place the mandatory
range, and thus the final sentence, solely in the hands of prosecutors. These proposals cannot be
fairly evaluated without knowing the consequences of greater prosecutorial control over
sentencing outcomes. If prosecutors introduce unwarranted disparity – and there is substantial
                                                       
21

U.S. Sent’g Comm’n, 2005 Sourcebook of Federal Sentencing Statistics, Section III, tbl. 26; U.S.
Sent’g Comm’n, 2012 Preliminary Quarterly Data Report, Third Quarter, tbl. 1.

22

Id., tbl. 5.

23

U.S. Sent’g Comm’n, 2012 Preliminary Quarterly Data Report, Third Quarter, tbls. 1, 6. See Paul J.
Hofer, How Well Do Sentencing Commission Statistics Help in Understanding the Post-Booker System?,
22 Fed. Sent. Rep. 89, 92 (Dec. 2009) (explaining the 2009 data); Amy Baron-Evans & Kate Stith,
Booker Rules, 160 1160 U. Pa. L. Rev. 1631, 1679 n.262 (2012) (explaining the 2011 data).

24

As the Commission has observed, “in a tightly structured sentencing system like the [mandatory]
federal sentencing guidelines,” the “[d]isparate effects of charging and plea bargaining [were] a special
concern . . . because the ability of judges to compensate for disparities in presentence decisions [was]
reduced.” U.S. Sentencing Comm’n, Fifteen Years of Guideline Sentencing: An Assessment of How
Well the Federal Criminal Justice System is Achieving the Goals of Sentencing Reform 92 (2004).

14

 
evidence that they do – then undoing the current balance between prosecutors and judges would
be counterproductive. Without this information, Congress could be misled to adopt a solution to
the wrong problem that would make matters much worse.
Further, the chief problem facing the federal criminal justice system is that sentences are
too severe in many cases to serve any legitimate purpose of sentencing. The Bureau of Prisons
population is consuming the Department of Justice’s budget. As the Commission saw firsthand
at FCI Memphis, too many people who no longer need to be there – primarily African American
men – will be languishing in prison for another five, ten or twenty years. The fact that
prosecutors – whose interest is in retribution for the seriousness of the offense and protecting the
public from further crimes of defendants – frequently seek or agree to sentences below the
guideline range strongly supports the need for lower sentences.
Finally, we believe that sentences outside the guideline range are often necessary to
comply with the purposes of sentencing. But if they are a problem, one would expect the
Commission to be at least as concerned about those initiated by prosecutors as those initiated by
judges. Not only are there far fewer judge-initiated variances and departures, but judges are
neutral, are charged with imposing sentences that meet the purposes of sentencing, state their
reasons in open court, and are subject to appellate review. Prosecutors are adversarial parties
whose decisions are not made in public, are not required to take into account the purposes of
sentencing, and are not subject to judicial review.
For these reasons and as shown above in Parts I and II, the post-Booker sentencing
system cannot be fairly assessed without assessing prosecutorial practices. In its mandatory
minimum report, the Commission counterbalanced its criticism of inconsistent use of mandatory
minimums by prosecutors with criticism of judicial discretion under the guidelines. See U.S.
Sent’g Comm’n, Mandatory Minimum Penalties in the Criminal Justice System 346-47
(2011). Likewise, a balanced picture of the advisory guidelines system must include an
assessment of prosecutorial practices.
IV.

The Report Should Acknowledge the Role of Unnecessary Severity in Causing
Disparity in the Advisory Guideline System.

As the Commission noted in its mandatory minimum report, mandatory minimums that
apply too broadly, are set too high, or both, lead to inconsistencies in application of mandatory
minimums by prosecutors. U.S. Sent’g Comm’n, Mandatory Minimum Penalties in the
Criminal Justice System 345 (2011). At the same time, the Commission asserted that the
Supreme Court’s decisions have “increased inconsistencies” in sentencing under the guidelines.
Id. at 346. But the root of inconsistencies under the advisory guidelines system is the same as
under mandatory minimums: guidelines that apply too broadly, are overly severe, or both, lead to
differences within and among districts in the sentencing practices of both judges and prosecutors.
The data noted in Part III, showing a decreasing trend in judicial below-guideline
sentences, suggests that when the Commission fixes unsound guidelines, judges follow them
more often. For instance, from 2010 to 2011, the within-guideline rate in crack cases increased

15

 
by 5.9 percentage points and the non-government sponsored rate of departures/variances dropped
by 6.5 percentage points,25 and we suspect this trend has continued in 2012.
V.

The Report Should Give a Fair Account of the Views of Appellate Judges and the
Appellate Caselaw.

In its October 2011 testimony, as well as in New Orleans and Memphis, the Commission
suggested or stated that appellate judges are generally frustrated by the standard of review. We
do not believe that this is an accurate summary of what appellate judges have said or of the
appellate caselaw. We therefore urge the Commission to provide a fair account of the views
expressed by appellate judges and of the appellate caselaw.26
Appellate judges testifying at the Commission’s regional hearings recognized that the
current standard is necessary if the guidelines are to remain constitutional, declined to support
statutory change when asked, recognized that sentencing judges are best situated to decide
sentences and most often get it right, and urged the Commission to facilitate appellate review by
justifying its guidelines. See Statement of Raymond Moore, Federal Public Defender for the
District of Colorado, Before the U.S. Sent’g Comm’n at 53-58 (Feb. 16, 2012) (discussing
testimony of appellate judges at regional hearings). In its October 2011 testimony, however, the
Commission relied on snippets of testimony taken out of their full context and isolated
complaints in two dissenting opinions. See Commission Testimony at 15-17.
Appellate judges at the hearing in February 2012 advised the Commission not to seek a
stricter standard of review. See Transcript of Public Hearing Before the U.S. Sent’g Comm’n at
170-71 (Feb 16, 2012) (Judge Lynch) (“[I]n going from the district court to the court of appeals,
my sense of the desirability of more appellate review of sentences has drooped . . . because now I
see it also from that perspective, and I see that we don’t have the same degree of information, the
same feel for the case. I think appellate judges are very reluctant to get pushed into this.”); id. at
171(Judge Davis) (“I really agree with Judge Lynch, and . . . we really have settled into a
comfort level I think in the Fourth Circuit. It ain’t broke. . . . And I think the court is really quite
comfortable with where we are.”); Letter from Hon. Myron H. Bright, U.S. Circuit Judge, to
Hon. Patti B. Saris, Chair, U.S. Sentencing Comm’n (Jan. 10, 2012) (criticizing the
Commission’s proposals for stricter appellate review as likely unconstitutional, and recognizing
sentencing judges’ greater competence in sentencing and ability to compensate for
unreasonableness of some guidelines).
Appellate judges speaking in New Orleans expressed no frustration with the appellate
standard of review. They did not say that they think sentences are unreasonable, but are unable
to reverse them. In fact, Judge Riley and Judge Owen encouraged “bullet proofing” sentences in
                                                       
25

U.S. Sent’g. Comm’n, Interactive Sourcebook.

26

We have previously explained why the Commission’s proposed changes to the standards of review
would likely be held unconstitutional. See Statement of Henry Bemporad, Federal Public Defender for
the Western District of Texas, Before the U.S. Sent’g Comm’n (Feb. 16, 2012).

16

 
the district court to cut down on appellate review. When pressed, Judge Riley and Judge Owen
said that Congress could codify what Rita said regarding a rebuttable presumption of
reasonableness, but Judge Riley warned that going further may not pass muster. Judge Motz said
that codifying a presumption of reasonableness would be unwise and saw no need for or benefit
from congressional intervention. Judge Chin declined to comment.
Of course, the proper and constitutional standard of review is not up to a vote of appellate
judges. If it were, we believe, based on testimony and comments over the last four years, that the
majority of appellate judges would oppose a stricter standard. That belief also is supported by
appellate decisions.
Numerous courts of appeals have recognized that the district courts are best situated to
determine sentences.27 At the same time, the courts of appeals are fully equipped to reverse
sentences that are procedurally or substantively unreasonable.28 In some cases, dissenting judges
                                                       
27

See, e.g., United States v. Poynter, 495 F.3d 349, 351 (6th Cir. 2007) (“Unlike the trial court,” the court
of appeals “[does] not see” or hear the evidence, and “most appellate judges have little experience
sentencing individuals.”); United States v. Aleo, 681 F.3d 290, 300 (6th Cir. 2012) (noting that “the
sentencing court observed the defendant and the witnesses firsthand, and . . . has a wide variety of
sentencing cases to compare each case to, unlike an appellate court,” and “[o]ur role is not to usurp the
sentencing judge’s position as the best interpreter of the facts”); United States v. Begin, 696 F.3d 405, 414
(3d Cir. 2012 ) (“[T]he sentencing judge, not the court of appeals, is in a superior position to find the facts
and judge their import under § 3553(a) in the individual case.”); United States v. Abu Ali, 528 F.3d 210,
265-66 (4th Cir. 2008) (“[T]he sentencing judge is in a superior position to find facts and conduct the
individualized assessment that is such an integral part of the sentencing process.”) (internal citation and
quotation marks omitted); United States v. Ressam, 679 F.3d 1069, 1086 (9th Cir. 2012) (“[W]e are to
afford significant deference to a district court’s sentencing decision. . . . [W]e may not reverse just
because we think a different sentence is appropriate. . . . ‘The sentencing judge is in a superior position to
find facts and judge their import under § 3553(a) in the individual case.’”) (quoting Gall , 552 U.S. at 51)
(internal citation omitted); United States v. Friedman, 554 F.3d 1301, 1312 (10th Cir. 2009) (“[D]istrict
courts enjoy a strong institutional advantage in arriving at sentencing decisions, and []those decisions are
therefore entitled to substantial deference.”) (internal citation omitted); United States v. Pugh, 515 F.3d
1179, 1192 (11th Cir. 2008) (“We . . . recognize the wide discretion afforded to district courts in
sentencing, especially since the district court is in a superior position to find facts and judge their
import.”) (internal quotation marks omitted); United States v. Cavera, 550 F.3d 180, 191 (2d Cir. 2008)
(en banc) (“[R]esponsibility for sentencing is placed largely in the precincts of the district courts.”);
United States v. Gardellini, 545 F.3d 1089, 1095 (D.C. Cir. 2008) (“The District Court’s conclusion rests
on precisely the kind of defendant-specific determinations that are within the special competence of
sentencing courts, as the Supreme Court has repeatedly emphasized.”).

28

See, e.g., Poynter, 495 F.3d at 353 (reversing upward variance to statutory maximum as substantively
unreasonable); Aleo, 681 F.3d at 300-02 (reversing upward variance to statutory maximum as
substantively unreasonable); Begin, 696 F.3d 405, 414 (3d Cir. 2012) (reversing upward departure as
procedurally unreasonable); Abu Ali, 528 F.3d at 269 (reversing downward variance as substantively
unreasonable); Ressam, 679 F.3d at 1087 (reversing downward variance as substantively unreasonable);
Friedman, 554 F.3d at 1312 (reversing downward variance as substantively unreasonable); Pugh, 515
F.3d at 1194 (reversing downward variance as substantively unreasonable); United States v. Miller, 601
F.3d 734 (7th Cir. 2010) (reversing upward variance as substantively unreasonable); United States v.

 
17

 
have expressed the view that the court affirmed a sentence using a standard more deferential than
Gall requires, both for sentences viewed as too short,29 and for sentences viewed as too long.30
In other cases, dissenting judges have expressed the view that the court reversed a sentence using
a standard less deferential than Gall requires.31 These are not complaints about the standard of
review itself, but reasonable disagreements about the way it was applied in a particular case.
Such disagreements have existed, and would exist, under any standard of review.
We have compiled a list of all appellate reversals in certain categories since Gall and
Kimbrough through September 27, 2012.32 Seventeen sentences were reversed as substantively
unreasonable on the defendant’s appeal (5 within-guideline sentences, 10 above-guideline
sentences, and 2 below-guideline sentences), and twenty-one below-range sentences were
reversed as substantively unreasonable on the government’s appeal. One hundred twenty-one
sentences were reversed for a failure to adequately explain the sentence or to adequately address
non-frivolous arguments on the defendant’s appeal (81 within-guideline sentences, 29 aboveguideline sentences, and 11 below-guideline sentences), and seventeen below-guideline
sentences were reversed for the same kind of procedural error on the government’s appeal.
This demonstrates that courts of appeals are fully capable of reversing outlier sentences
for substantive unreasonableness, and are fully equipped to review and reverse for procedural
error. And, contrary to the view that reversal for procedural error has no effect on remand – a
misperception stated in the Commission’s October 2011 testimony, see Commission Testimony
at 16, reversals for procedural error have resulted in a different sentence 57% of the time. See
Jennifer Niles Coffin, Where Procedure Meets Substance: Making the Most of the Need for
                                                                                                                                                                               
Cutler, 520 F.3d 136 (2d Cir. 2008) (reversing downward variance as substantively unreasonable); United
States v. Akhigbe, 642 F.3d 1078, 1087-88 (D.C. Cir. 2011) (reversing upward variance as procedurally
unreasonable).
29

See United States v. Tomko, 562 F.3d 558, 578 (3d Cir. 2009) (en banc) (Fisher, J., dissenting); United
States v. Feemster, 572 F.3d 455, 470-71 (8th Cir. 2009) (en banc) (Beam, J., dissenting); United States v.
Edwards, 622 F.3d 1215, 1216 (9th Cir. 2010) (en banc) (Gould, J., dissenting).

30

See United States v. Heath, 559 F.3d 263, 268-70 (4th Cir. 2009) (Gregory, J., dissenting); United
States v. Overmyer, 663 F.3d 862, 867 (6th Cir. 2011) (Merritt, J., dissenting); United States v. Docampo,
573 F.3d 1091, 1104 (11th Cir. 2009) (Barkett, J., concurring in part and dissenting in part).

31

See, e.g., United States v. Abu Ali, 528 F.3d 210, 272-73 (4th Cir. 2008) (Motz, J., dissenting); United
States v. Irey, 612 F.3d 1160, 1271-72 (11th Cir. 2010) (Edmondson, J., dissenting); United States v.
Amezcua-Vasquez, 586 F.3d 1176, 1178-80 (9th Cir. 2009) (O’Scannlain, J., dissenting from the denial of
rehearing en banc); United States v. Hunt, 521 F.3d 636, 653 (6th Cir. 2008) (Martin, J., dissenting).

32

See Appellate Decisions After Gall (Sept. 27, 2012), http://www.fd.org/docs/select-topics--sentencing/app_ct_decisions_list.pdf. It does not include reversals where the district court erred in
calculating the guideline range, or treated the guidelines as mandatory or presumptively reasonable, or did
not address an argument because circuit precedent at the time of sentencing precluded it from doing so.

18

 
Adequate Explanation at 18 (Dec. 2011), http://www.fd.org/docs/select-topics--sentencing/Procedure_Substance.pdf.

Thank you for the opportunity to provide these suggestions.
Very truly yours,
/s/ Miriam Conrad
Miriam Conrad, Federal Public Defender
Vice-Chair, Federal Defender Sentencing
Guidelines Committee
on Behalf of the Federal Defender Sentencing
Guidelines Committee and Legislative Expert Panel

cc:

William B. Carr, Jr., Vice Chair
Ketanji Brown Jackson, Vice Chair
Hon. Ricardo H. Hinojosa, Commissioner
Dabney Friedrich, Commissioner
Hon. Beryl A. Howell, Commissioner
Isaac Fulwood, Jr., Commissioner Ex Officio
Jonathan J. Wroblewski, Commissioner Ex Officio
Judith M. Sheon, Staff Director
Kenneth Cohen, General Counsel

19