Policy, Uniformity, Discretion, and Congress’s
Sentencing Acid Trip

By Mark Osler
Copyright 2008

I.

The Pointless Forest and the Pointless Man
Around 1970, singer Harry Nilsson went on an acid trip. He later

reported that during this experience, he “looked at the trees and… realized
that they all came to points, and the little branches came to points and the
houses came to a point. I thought ‘Oh! Everything has a point and if it
doesn’t, then there’s a point to it.’” 1
Nilsson put these insights to good use, later producing an album and
an animated film, both entitled “The Point.” 2 In either format, “The Point”
told the story of Oblio, who along with his dog, Arrow, is thrown out of the
land of Point because Oblio does not have a point on top of his head like
everyone else. They are banished to the scary Pointless Forest, where they
encounter the Pointless Man, another banished soul who welcomes them as
they begin their journey. The “Pointless Man,” as drawn, has several pointy
1

The Point, Harry Nillson Web Pages, http://www.harrynilsson.com/page-the-point.html (last visited
November 20, 2007).
2
The animated version was televised by ABC on February 2, 1971, and was narrated by Dustin Hoffman.
Later narrators on subsequent versions included Alan Thicke and Ringo Starr. Id.

1
Electronic copy available at: http://ssrn.com/abstract=1130212

faces and actual arrows emanating from his torso, all pointing in different
directions.
Once Oblio enters the Pointless Forest and actually meets the socalled Pointless Man, he has a radical change in perspective:
You see the Pointless Man did have a point. In
fact, he had hundreds of them, all pointing in
different directions. But as he so quickly pointed
out, a point in every direction is the same as no
point at all. 3
Which (of course) brings us to Congress and federal sentencing.
Congress has issued at least 31 separate directives setting general policy
goals in criminal sentencing. 4 Section II describes some of these policy

3

Harry Nilsson, The Pointless Man, The Point Soundtrack (BMG Entertainment 1970).
Those 31 directives would mandate that each of the following be considered in creating guidelines,
sentencing individuals, or both:

4

1)
2)
3)
4)
5)
6)
7)

The nature of the offense (18 U.S.C. § 3553(a)(1)
The circumstances of the offense (18 U.S.C. § 3553(a)(1)
The history of the defendant (18 U.S.C. § 3553(a)(1)
The characteristics of the defendant (18 U.S.C. § 3553(a)(1)
The seriousness of the offense (18 U.S.C. § 3553(a)(2)(A)
Promotion of respect for the law (18 U.S.C. § 3553(a)(2)(A)
Just punishment for the offense (18 U.S.C. § 3553(a)(2)(A) & USSG §1A.1, intro to comment, pt. A,
¶ 2)
8) Deterrence to criminal conduct (18 U.S.C. § 3553(a)(2)(B) & USSG §1A.1, intro to comment, pt. A,
¶ 2)
9) Protection of the public from further crimes by the defendant (18 U.S.C. § 3553(a)(2)(C)
10) To provide defendants with needed education or vocational training (18 U.S.C. § 3553(a)(2)(D)
11) To provide defendants with needed medical care or other correctional treatment (18 U.S.C. §
3553(a)(2)(D)
12) The kinds of sentences available (18 U.S.C. § 3553(a)(3)
13) Policy statements by the sentencing commission (18 U.S.C. § 3553(a)(5)
14) The need to avoid unwarranted sentence disparities among defendants with similar records
found guilty of similar conduct (18 U.S.C. § 3553(a)(6) & USSG §1A.1, intro to comment, pt.
A, ¶ 2)
15) Provision of restitution (18 U.S.C. § 3553(a)(7)
16) Incapacitating the offender (USSG §1A.1, intro to comment, pt. A, ¶ 2)
17) Rehabilitating the offender (USSG §1A.1, intro to comment, pt. A, ¶ 2)

2
Electronic copy available at: http://ssrn.com/abstract=1130212

directives, which all point in different directions, with different degrees of
specificity, clarity, and import. In setting out what some of these policy
directives seek, it becomes clear that federal sentencing policy resembles
nothing so much as it does Nilsson’s Pointless Man.
Section III, in turn, describes some of the underlying conflicts
between these principles, and then describes the effect of combining a 31point policy directive together with a strong mandate for uniformity. This
project-- putting a pointless mish-mash of policy directives together with a
demand for uniform punishments-- doesn’t make much sense. Without a
clear policy goal, after all, uniformity is as likely to be uniformly wrong as it
is to be uniformly right relative to any understandable principle or set of
principles. What is the sense in having consistent and uniform sentencing if

18)
19)
20)
21)
22)
23)
24)
25)
26)
27)
28)
29)
30)
31)

Proportionality in sentencing for conduct of differing severity
(USSG §1A.1, intro to comment, pt. A, ¶ 2)
Input from the Probation system, Judicial conference, DOJ, and Federal Defenders (28 U.S.C. §
994(o)
Directions from Congress (note following 28 U.S.C. § 994)
Maintaining sufficient flexibility to permit individualized sentences when warranted (18
U.S.C. § 991(b)(1)(B))
Advancements in knowledge of human behavior (18 U.S.C. § 991(b)(1)(C))
Neutrality as to race, sex, national origin, creed, and socioeconomic
status of offenders (28 U.S.C. § 994(d))
Fairness in sentencing (28 U.S.C. §§ 991(b)(1)(B) & 994(f))
Sentences need to be near the statutory maximum for crimes of violence or certain drug offenses
(28 U.S.C. § 994(h))
Sentences need to allow for probation for certain first offenders (28 U.S.C. § 994(j))
Average sentences prior to imposition of the Guidelines (28 U.S.C. § 994(m))
Effect on prison populations (28 U.S.C. § 994(q))
Certainty (28 U.S.C. §§ 991(b)(1)(B) & 994(f))
The community view of the gravity of an offense (28 U.S.C. § 994(c)(4))
The current incidence of an offense in the community and nation
as a whole (28 U.S.C. § 994(c)(7))

3

it is consistently and uniformly wrong? To insist on uniformity without
principled directives to create those uniform results does nothing less than
rob sentencing of any sense of real authority, by making it morally
indeterminate. 5
Finally, Section IV suggests a do-over for federal sentencing, in which
a new Sentencing Commission would start with a small number of
reasonable policy goals and then re-make the guidelines in a way which
would allow those goals to be met. Opponents to re-making the guidelines
would no doubt (correctly) fear the specter of greater discretion for judges
being a feature of any new system. This “fear of judging,” as Cabranes and
Stith called it, 6 is our modern equivalent of the Pointless Forest— we
(through our legislators) are scared to enter a world where judges exercise
independent discretion because we don’t know everything that may lie in
wait for us there. Individuals, even individual judges chosen expressly for
their superior discretion and judgment, can be unpredictable, after all. Given
the pointlessness of current “policy,” however, the prospect of reformed
guidelines with individual judges more actively evaluating cases becomes

5

How odd this combination is may benefit from an analogy to denominational religion. Those faiths
most judgmental of moral behavior tend to be those with a defined set of core beliefs that are maintained by
a magisterium. On the other hand, those faiths which embrace a wide variety of beliefs (such as the Bahia
or Unitarian/Universalists), tend to be more tolerant of a wide variety of behavior and belief. The current
federal sentencing system turns this on its head—it is as if the Unitarian/Universalists were suddenly
harshly judgmental of those who violated the tenets of any faith.
6
Kate Stith and Jose A. Cabranes, Fear of Judging in the Federal Courts (1998).

4

more appealing. With a new system and more-empowered judges a greater
degree of judicial discretion, there would be a much better chance that policy
and outcome would match, something that Congress has currently made
impossible at the macro level with too many policy directives and a welter of
statutes mandating uniform sentencing. 7

II.

A Point in Every Direction is the Same as No Point At All
A.

It used to be simple

Once upon a time, the policy goals of federal sentencing were simple.
As continues to be true in many other nations, 8 four simple goals structured
sentencing. 9 These four shifted in importance over time relative to one
another, but as a whole remained constant. One advantage to this
framework was that the goals of sentencing were easily understood— they
simply sought to (1) punish offenders (“retribution”), (2) deter both that
individual and others from committing further crimes (“deterrence”), (3) to
incapacitate dangerous individuals so that they could not cause more harm

7

Even with the elimination of mandatory guidelines, not only the advisory guidelines but several
mandatory minimum sentences serve the purpose of mandating uniformity; e.g. the mandatory minimum
sentences for sexual exploitation of children, 18 U.S.C. § 2251(d).
8

Andrew Dubinsky, An Examination of International Sentencing Guidelines and a Proposal for
Amendments to the International Criminal Court’s Sentencing Structure, 33 New England J. on Crim. &
Civ. Confinement 609, 618 (2007).
9
Patricia M. Wald, Why Focus on Female Offenders?. 16 Crim. Just. 10,11 (Spring, 2001)(listing
traditional goals of sentencing).

5

(“incapacitation”), and (4) to rehabilitate some offenders for both their
benefit and that of the larger society (“rehabilition”).
Certainly, these four traditional goals were often in tension. 10 For
example, in a given case some might insist that incapacitation of the
defendant through imprisonment was necessary to protect the community,
while others might insist that rehabilitation is possible. These goals would
be served by different means; prison for the former, treatment for the latter.
Despite such tensions, these limited goals allowed judges to weigh them
relative to one another and evaluate each defendant by the same standards.
On a macro level, the simplicity of these traditional goals also allowed
for a national debate over which should predominate, and they shifted in
importance over time. For example, beginning in the late nineteenth century
the goal of rehabilitation was ascendant. 11 As Doug Berman has noted, this
rehabilitative ideal was framed in medical terms, with the criminal viewed as
“sick” and in need of “cure.” 12 The traditional goals, then, provided not
only reasonable guideposts for the use of discretion, but framed the national
debate on sentencing in an understandable way, with a known and limited
number of trade-offs available.
10

Kate Stith and Jose A. Cabranes, Fear of Judging in the Federal Courts 9-22 (1998) (describing the
conflicts between and changing roles of these sentencing goals in American history).
11
12

Douglas A. Berman, Reconceptualizing Sentencing, 2005 U. Chicago Legal Forum 1, 3 (2005).
Id. at 4.

6

These goals of retribution, deterrence, incapacitation, and
rehabilitation still remain in the federal scheme, at least in the sense they are
listed in the statute book. Specifically, 18 U.S.C. § 3553(a)(2)(A)-(D)
directs a judge to consider the need for a sentence to reflect “just
punishment,” 13 to provide “adequate deterrence to criminal conduct,” 14 “to
protect the public from further crimes by the defendant,” 15 and to “provide
the defendant with needed educational or vocational training, medical care,
or other correctional treatment in the most effective manner.” 16
Unfortunately, Congress did not stop there. While the policy goals
encompassed within retribution, deterrence, incapacitation, and
rehabilitation remain in the books, they have been buried beneath an
avalanche of other goals, including (perhaps most importantly) uniformity.
While I have previously addressed the sad fact that the federal Sentencing
Guidelines wholly ignore these goals in the machinery it establishes to
calculate a sentence, 17 here I address a related but different question: Has
the wide variety of diverse policy goals packed into the federal sentencing
system made that system morally indeterminate?
13

18 U.S.C. § 3553(a)(2)(A).
18 U.S.C. § 3553(a)(2)(B)
15
18 U.S.C. § 3553(a)(2)(C)
16
18 U.S.C. § 3553(a)(2)(D)
17
Mark Osler, Must Have Got Lost: Traditional Sentencing Goals, The False Trail of Uniformity and
Process, and the Way Back Home, 54 S.Car. L. Rev. 649 (2003); Mark Osler, Uniformity and Traditional
Sentencing Goals in the Age of Feeney, 16 Fed. Sentencing Rptr. 253 (2004), quoted in United States v.
Booker, 543 U.S. 220, 297 n. 15 (dissent of Stevens, J.).
14

7

B.

The Non-Traditional Policy Goals

I have already described the traditional sentencing policy goals and
how they are included in the sentencing scheme at 18 U.S.C. §
3553(a)(2)(A)-(D). Now, let’s explore the remainder of Congress’s policy
directives to the Sentencing Commission and judges, both of which have
roles in turning policy into action—the Sentencing Commission through
creation and revision of the sentencing guidelines, and judges through the
act of sentencing itself.
The policy directives for sentencing, sadly, are not grouped together
in the federal code despite having been largely enacted together through the
Sentencing Reform Act of 1984. 18 Rather, they are lumped together in three
separate places, leading to frequent redundancies. One of those places is 18
U.S.C. § 3553, which has been at the center of nearly every federal
sentencing controversy since United States v. Booker, 19 which declared in
2005 that the sentencing guidelines were no longer strictly mandatory and
that § 3553 was to be the guiding statute of sentencing judges and courts of
appeal. 20 Though § 3553 itself seems clearly directed to sentencing judges

18

For a worthwhile history of the Sentencing Reform Act of 1984, see Kristin J. Balding, It Is a ‘War On
Drugs’ and It Is Time To Reload Our Weapons: An Interpretation of 21 U.S.C. § 841, 43 St. Louis U. L.J.
1449,
19
543 U.S. 220 (2005).
20
543 U.S. at 259-260.

8

and not to the Sentencing Commission, 21 28 U.S.C. § 991(b) in turn directs
the Sentencing Commission to consider some of those same objectives. 22
As Justice Breyer somewhat famously put it in Rita v. United States, 23 “the
sentencing statutes envision both the sentencing judge and the Commission
as carrying out the same basic § 3553(a) objectives, the one, at retail, the
other at wholesale.” 24
Though 18 U.S.C. § 3553 contains the best-known set of sentencing
policy goals, it by no means contains the only set. 28 U.S.C. § 991, the
statute which established the Sentencing Commission itself, contains not
only specific sentencing policy goals 25 but a sweeping description of what
the guidelines as a whole should look like:
The purposes of the United States Sentencing
Commission are to … provide certainty and
fairness in meeting the purposes of sentencing,
avoiding unwarranted sentencing disparities
among defendants with similar records who have
been found guilty of similar criminal conduct
while maintaining sufficient flexibility to permit
individualized sentences when warranted…. 26

21

For example, the key provisions included at 18 U.S.C. § 3553(a) are prefaced with “The court, in
determining the particular sentence to be imposed, shall consider-- ….”
22
Specifically, 28 U.S.C. § 991(b) directs the Sentencing Commission to “assure the meeting of the
purposes of sentencing as set forth in section 3553(a)(2)….”
23
127 S. Ct. 2456 (2007).
24
Id. at 2463.
25
E.g., that statute’s requirement that sentencing guidelines “reflect, to the extent practicable,
advancement in knowledge of human behavior as it relates to the criminal justice process….” 28 U.S.C. §
991(b)(1)(C).
26
28 U.S.C. § 991(b)1)(B).

9

The real mother lode of policy goals, though, is found at 28 U.S.C. §
994, which is directed at the Sentencing Commission and sets out with both
great specificity 27 and stunning breadth 28 what should be contained in the
guidelines themselves.
Split up as they are into three distinct statutes, the policy goals taken
as a whole suffer from redundancy and overlap in several places. To avoid
replicating those problems and to provide a clearer analysis, I have grouped
some of the sentencing policy goals of 18 U.S.C. § 3553 and 28 U.S.C. §§
991 & 994 into three categories. First, there are the broad dictates, which
focus sentencing in a general way on interests other than the traditional
goals, and which on their face should be considered in all federal
sentencings. The second group contains specific provisions, which require
as a matter of policy the consideration of certain discrete factors in
sentencing defined types of cases. Third, federal law contains at least two
statutes which might be called “Trap-door” provisions, directing that judges
and the Commission obey unnamed existing and yet-uncreated policy
dictates.

27

E.g., that consecutive sentences for both an offense and conspiracy to commit that offense should be
avoided. 28 U.S.C. § 994(l)(1)(B)(2).
28
E.g., these directives echo the others in seeking “fairness.” 28 U.S.C. § 994(f).

10

1.

Broad Dictates.

The broad-dictate provisions of federal law direct sentencing towards
general goals rather than specific objectives. This group includes the four
traditional goals already discussed, of course. It also includes several other
broad goals which in many cases will undercut those traditional goals,
including the following, which reflect only a fraction of the total number of
policy goals: 29
a.

Uniformity

Perhaps the most commonly-recognized non-traditional goal of
sentencing is uniformity, which is codified as part of the long list found at 18
U.S.C. § 3553(a). 30

What Congress sought in mandating sentencing

guidelines was, above all else, uniformity between judges and within a
judge’s own docket in sentencing cases which are at least somewhat
similar. 31 As discussed in the next section at some length, the Department
of Justice and some in Congress have straightforwardly declared uniformity
to be the paramount goal of the sentencing system, and it is fair to say that it
is the pursuit of this goal which has driven the restructuring of the federal

29

For a more complete list, see note 4, supra. This list omits for the sake of efficiency some important
codified principles which also create conflict, such as the principle of proportionality.
30
18 U.S.C. § 3553(a)(6).
31
As Stith and Cabranes described it, “Congress’s concern with reducing perceived or assumed disparities
in federal sentencing is reflected in the debates leading up to the Acts’s passage, in the Senate report
accompanying it, and in the text of the Act itself.” Stith and Cabranes, Fear of Judging: Sentencing
Guidelines in the Federal Courts 104 (1998).

11

sentencing system 32 through guidelines, mandatory minimum sentences, and
Court of Appeals opinions on what is “reasonable” in the period after United
States v. Booker. 33
b.

Parsimony 34

What is commonly called the “parsimony clause” of 18 U.S.C. §
3553(a) sets forth that “The court shall impose a sentence sufficient, but not
greater than necessary, to comply with the purposes set forth in paragraph
(2) of this subsection….” This directive sets out a clear principle: That a
sentence should not exceed that necessary to fulfill the factors set out in the
following paragraph, which include the traditional goals of retribution, 35
deterrence, 36 incapacitation, 37 and rehabilitation. 38
Unfortunately, post-guideline courts have rarely tried to give the
parsimony clause much meaning. 39 It’s no wonder, either, as trial courts
generally follow the sentencing guidelines, which serve a multitude of other
goals (and sometimes, seemingly, no goal at all). The guidelines are simply
32

Michael M. O’Hear, The Original Intent of Uniformity in Federal Sentencing, 74 U.Cin. L. Rev. 749,
756-791 (2006).
33
534 U.S. 220 (2005).
34
18 U.S.C. § 3553(a).
35
18 U.S.C. § 3553(a)(2)(A)
36
18 U.S.C. § 3553(a)(2)(B)
37
18 U.S.C. § 3553(a)(2)(C)
38
18 U.S.C. § 3553(a)(2)(D)
39
It is important to distinguish the principle of parsimony from the rule of lenity, which historically has
little to do with sentencing. Rather, the rule of lenity is a rule of statutory construction which insists that
application of a criminal statute be construed in favor of a defendant when it is unclear whether or not that
law applies to the defendant’s actions at all. See Phillip M. Spector, The Sentencing Rule of Lenity, 33 U.
Tol. L. Rev. 511 (2002).

12

not calibrated to follow the simple parsimony directive, as Justice Breyer
seemed to acknowledge in Rita v. United States. 40 In that opinion, Justice
Breyer noted that in the course of trying to use the four traditional goals and
the parsimony provision to arrive at a foundational set of guidelines, a
conflict arose among those drafting the guidelines “when the Commission
attempted to reconcile the different perceptions of the purposes of criminal
punishment.” 41 Rather than resolving this conflict, the Commission simply
punted on the issue, instead choosing to codify past practices by averaging
out the results from thousands of prior cases. 42 Thus, the Commission
tossed out any real consideration of the parsimony provision becoming one
of the structuring mechanisms for the guidelines, while at the same time 18
U.S.C. § 3553(a) affirmatively directed that it be considered by judges in
sentencing individual defendants.
This, predictably, set up constant conflicts between the guidelines and
a sentencing judge’s attempt to use the parsimony provision as it applied to
any specific case, since the guidelines were developed without active
reference to that guiding principle, yet judges were both bound to the

40
41
42

127 S. Ct. 2456 (2007)
127 S. Ct. at 2464, quoting United States Sentencing Guidelines § 1A.1, intro to comment, pt. A, ¶ 3.
Id.

13

guidelines and charged with employing parsimony. 43 These conflicts have
extended beyond the ruling in Booker that the guidelines were no longer
mandatory, and will likely be litigated into the future as courts resolve what
type of application of the parsimony clause is “reasonable.” 44 While at some
point in the past it might have been said that the parsimony clause was of no
significance, the more recent Supreme Court decisions point in the other
direction, meaning that this clause will step up among the many others
competing for the attention of the Sentencing Commission and judges.
c.

Following advancements in knowledge of
human behavior 45

One of Congress’s principle directives to the Sentencing Commission
was to establish policies which “reflect, to the extent practicable,
advancements in knowledge of human behavior as it relates to the criminal
justice process….” 46 On its face, this directive tells the Commission that it
must monitor scientific progress that would relate to things such as possible
rehabilitation through new therapies.
Frank Bowman has argued that the guidelines themselves are not only
unscientific, but constitute “a reaction against the notion that science has
43

It should be no surprise, then, that both Kimbrough v. United States, 128 S. Ct. 558 (2007) and Gall v.
United States, 128 S. Ct. 586 (2007) arose out of this foundational conflict between, on the one hand, the
traditional goals and the parsimony provision as applied by a judge, and, on the other, the sentencing
guidelines.
44
Id.
45
28 U.S.C. § 991(b)(1)(C)
46
Id.

14

very much to say about criminal punishment.” 47 In a broad sense this is
absolutely correct. The available histories of the development of the
guidelines reflect no substantive reference by the framers of the guidelines to
the social or biological sciences. However, in some instances the
Sentencing Commission itself has developed policies while relying on
specific scientific findings. One such instance involves crack cocaine
sentencing, in which two decades after adopting an unscientific approach the
Commission reversed course based on scientific data.
In developing the sentencing guidelines for crack cocaine, the
Sentencing Commission adopted the 100-to-1 powder-to-crack cocaine ratio
contained in 21 U.S.C. § 841(b), even though that ratio had no scientific
foundation. 48 In 2002 49 and in 2007, 50 the Commission issued lengthy
reports which relied on current scientific studies to refute its own 100-to-1
ratio. 51 Ultimately, in 2007, the Commission adjusted that ratio in the
guidelines based in part on the findings in its own report, 52 despite the fact

47

Frank O. Bowman, Fear of Law: Thoughts on Fear of Judging and the State of the Federal Sentencing
Guidelines, 44 St. Louis U. L.J. 299, 316 (2000).
48
CITE
49
CITE
50
CITE
51
For example, the 2007 Report refutes the idea that crack cocaine affects fetal development
disproportionately relative to powder cocaine. United States Sentencing Commission, Cocaine and Federal
Sentencing Policy, 68-71 (2007) (available at http://ussc.gov).
52
Amendments to the Sentencing Guidelines for United States Courts, 72 Fed. Reg. 28571-28572.

15

that Congress had taken no action to change that ratio as contained in the
corresponding set of mandatory minimum sentences. 53
While some may dismiss the imperative of considering new science as
toothless, in at least one high-profile sentencing realm it has played a role in
a major change.

d.

Neutrality as to race, sex, national origin,
creed, and socioeconomic status 54

In one of the few absolutes among the directives by Congress, 28
U.S.C. § 994(d) mandates that “The Commission shall assure that the
guidelines and policy statements are entirely neutral as to the race, sex,
national origin, creed, and socioeconomic status of offenders.” In turn, the
Commission placed equally strong language in the guidelines themselves,
providing flatly that race, sex, national origin, creed, religion, and
socioeconomic status of a defendant are all factors which “are not relevant in
the determination of a sentence.” 55
The breadth of this prohibition is striking. Though it is contained
within the chapter of the guidelines which describes departures from a
guideline range, 56 the language there is distinct from surrounding guidelines,

53
54
55
56

21 U.S.C. § 841(b)
28 U.S.C. § 994(d)
United States Sentencing Guideline § 5H1.10.
United States Sentencing Guidelines § 5.

16

all of which limit their affects to departure considerations. 57 In contrast, the
prohibition against consideration of race, sex, national origin, creed,
religion, and socioeconomic status apply to all aspects of determining a
sentence—including the establishment of a sentence within a guideline
range. In other words, by the plain language of the guideline, it is improper
for a sentencing judge to even consider the fact that the defendant is female
when sentencing within a guideline range.
This absolutist nature of this rule of neutrality has brought the
race/sex/national origin/creed/ religion/socioeconomic status ban into
conflict with other of the policy directives. For example, the mandate to
follow current science (discussed above) runs into a wall when it conflicts
with the bar on consideration of these factors. The idea of whether such
science mandates consideration of a defendant’s sex is an actively debated
question. In relation to female offenders, for example, some have employed
reams of statistical analysis to oppose the ban on taking gender into account
when sentencing, arguing that this masks important and relevant gender
effects which pervade society as a whole. 58

57

E.g. United States Sentencing Guideline § 5H1.5, which states that “Employment record is not
ordinarily relevant in determining whether a departure is warranted.”
58
E.g., Myrna S. Raeder, Gender and Sentencing: Single Moms, Battered Women, and Other Sex-Based
Anomalies in the Gender-Free World of the Federal Sentencing Guidelines, 20 Pepp. L. Rev. 905 (1993).
Using criminological data to back up her point, Raeder argues that “Treating men and women fungibly for
sentencing purposes overlooks the role played by gender in criminality.” Id., at 908. See also Nekima

17

The strict bar on considering protected class status continues to have a
strong impact on sentencing, even as it comes under harsher attack from
those who would bend this rule to allow for certain factors (such as gender)
into account. 59
e.

Fairness 60

Fairness, mandated as a part of the guideline scheme at both 28 U.S.C.
§§ 991 & 994, seems so vague a concept that we might imagine it has not
been a substantive policy issue in the grand debates over sentencing.
However, as the guidelines were formed the idea of “fairness” was given
two contradictory but precise meanings, and each played a role in how those
guidelines were constructed.
In recounting the creation of the sentencing guidelines’ structure,
now-Justice Stephen Breyer has spoken quite clearly about some of the
compromises which were made by the first Sentencing Commission. 61 In
the course of that discussion, he describes fairness in two clear but opposing

Levy-Pounds, From the Frying Pan Into the Fire: How Poor Women of Color and Children Are Affected
by Sentencing Guidelines and Mandatory Minimums, 47 Santa Clara L. Rev. 285 (2007),
59
Despite the ban on considering gender, there do seem to be gender effects in sentencing, with women
getting lighter sentences than similarly situated men. Anne Martin Stacey & Cassia Spohn, Gender and the
Social Costs of Sentencing: An Analysis of Sentences Imposed on Male and Female Offenders in Three U.S.
District Courts, 11 Berkeley J. of Crim. L. 43, 48-49 (2006).
60
28 U.S.C. §§ 991(b)(1)(B) & 994(f)
61
Stephen Breyer, The Federal Sentencing Guidelines and the Key Compromises Upon Which They Rest,
17 Hofstra L.Rev. 1 (1988).

18

ways—first, in terms of procedural fairness, and second as substantive
fairness. 62
To Breyer, procedural unfairness results when a judge determines
facts which enhance a sentence in an informal way, without jury
determinations, the rules of evidence, or the requirement of proof beyond a
reasonable doubt. As Breyer puts it, “the more facts the court must find in
this informal way, the more unwieldy the process becomes, and the less fair
that process appears to be.” 63 Intriguingly, each of these shortcomings
became a feature of the sentencing system, and they are at the core of the
issues raised in Booker and its progeny.
On the other hand, Breyer seems to think that allowing the judge to
adjust the range to account for “real” offense conduct (that is, acts beyond
those charged in an indictment or information) almost always with a higher
sentence, provides “substantive” fairness by allowing the punishment to fit
the real crime. 64 The inclusion of this principle in the sentencing scheme
has brought us such controversial features as sentencing a given defendant
under the guidelines for which the defendant has been acquitted. 65

62

Id. at 8-12.
Id. at 11.
64
Id. at 11-12.
65
For a thorough explanation of how courts have justified this, see John Lanny Lynch v. United States,
437 F. 3d 902 (9th Cir. 2006)(en banc).
63

19

In the end, according to Breyer, the Commission compromised
between these two types of fairness, essentially by giving up on both. 66 That
compromise resulted in a number of troubling features of the guidelines,
including the lack of jury findings which brought us to Booker and the
crucial effect of relevant conduct in calculating a sentence. 67 Far from
being meaningless, the principle of “fairness” played a major role in shaping
federal sentencing. In fact, it can be safely said that the failure to give the
idea of “fairness” a single and unique meaning within the mandated
sentencing goals at the time the guidelines were framed (and thus requiring a
“compromise”) was a major factor in the disruption within federal
sentencing we have experienced in Booker and beyond.
f.

Consistency with prior practices 68

In 28 U.S.C. § 994(m), Congress gave a very specific mandate to the
Sentencing Commission:
The Commission shall ensure that the guidelines
reflect the fact that, in many cases, current
sentences do not accurately reflect the seriousness
of the offense. This will require that, as a starting
66

In a broad sense, it is difficult to reconcile these two ideas of fairness. Breyer defines “procedural
fairness” as the opposite of complexity, and “substantive fairness” as the opposite of efficiency. Assuming
that simple things are more efficient than complex things, it is hard to reconcile these two definitions, yet
apparently each played a role in the framing of the guideline scheme in the form of what Breyer describes
as a “compromise.” Id.
67
The guidelines expressly direct a court to consider uncharged “relevant conduct” in calculating a
guideline sentence, which means that acts never charged will be the basis of sentencing if they are found to
have been committed by the judge under a mere preponderance standard. U.S.S.G. § 1B1.3.
68
28 U.S.C. § 994(m).

20

point for its development of the initials sets of
guidelines for particular categories of cases, the
Commission ascertain the average sentences
imposed in such categories of cases prior to the
creation of the Commission….” 69
In other words, the Commission was directed to first survey the thencurrent sentences being given, then create guidelines that reject those
averages and instead create guideline ranges which “accurately reflect the
seriousness of the offense.” 70 In other words, the Sentencing Commission
was directly told to ascertain the collective judgment of hundreds of
experienced judges, and then substitute it with their own.
It is beyond dispute that they did exactly that, in an imprecise,
contentious, and hurried way.
First, the commission gathered data on past practices. Those with a
better knowledge of statistical analysis than I possess have been critical of
this process, in that the Commission’s methods did not meet social science
standards, focused on a small number of relevant variables, and were
shrouded in mystery even to those with an expertise in such analysis. 71
With the data they did gather and analyze, however inadequately, the
Commission then followed the mandate of Congress to reject that

69

Id.
Id.
71
Bernard E. Harcourt, From the Neer-Do-Well to the Criminal History Category: The Refinement of the
Actuarial Model in Criminal Law, 66 SUM Law & Contemp. Probs. 99, 123 (2003).
70

21

accumulated wisdom. Among other adjustments, they significantly raised
sentences under the guidelines for violent crimes, white-collar crimes, and
narcotics crimes, relative to prior practice. 72
It is hard to underestimate the profound impact or bizarre nature of
this task. The first step, a comprehensive study of existing practices,
certainly makes sense. What is odd is the second step—not to consider
those prior practices, or hold them up against an objective standard, but
rather to reject them. A single obscure command from Congress told the
Commission both to gather data for the first time, draw a specific conclusion
about that data (that it represents under-punishment of some crimes), and to
take action on that fore-drawn conclusion (raise sentences for those crimes).
This process is bizarre not only in that it is devoid of respect for social
science, but negates any consideration of the many other sentencing goals
Congress was mandating at the same time.

The dictate to assess current

practice and crank it up a notch does not take into account parsimony, has
nothing to do with science, probably incorporates sexist and racist
assumptions, and reflects neither procedural nor substantive fairness. It was,
however, relatively easy to do. The command was followed, and largely
accounts for the sentencing structure we struggle with today.

72

Id. at 125-126.

22

g.

Certainty 73

The very act of creating guidelines in part fulfilled the mandate of 28
U.S.C. § 991(b)(1)(B) that the Sentencing Commission establish policies
and practices that provide “certainty” in meeting the purposes of sentencing,
in the sense that it made sentencing more predictable and “certain” from the
perspective of the defendant being sentenced. Justice Breyer called this
factor “honesty,” and described it as one in which “the sentence the judge
gives is the sentence the offender will serve….” 74 More significant to the
achievement of certainty than the guidelines, though, may have been the
elimination of parole and drastic reduction in ‘good time’ credit allowed
those who had already been sentenced.
This project as a whole is perhaps best understood through the
guidelines’ own description of “The Basic Approach.” 75 There, the
Commission explained its understanding of such certainty, relative to the
regime it was replacing:
… Congress first sought honesty in sentencing. It
sought to avoid the confusion and implicit
deception that arises out of the present sentencing
system which requires a judge to impose an
indeterminate sentence that is automatically
reduced in most cases by ‘good time’ credits. In
73

28 U.S.C. §§ 991(b)(1)(B) & 994(f).
Stephen Breyer, The Federal Sentencing Guidelines and the Key Compromises Upon Which They Rest,
17 Hofstra L. Rev. 1, 4 (1988).
75
U.S.S.G §1A1.1(A)(3).
74

23

addition, the parole commission is permitted to
determine how much of the remainder of any
prison sentence an offender actually will serve. 76
Getting rid of parole 77 and limiting ‘good time’ credit to 15% of the
total sentence, 78 though not directly related to the guidelines, nonetheless
had an impact on the operation of the guidelines and the problems that
resulted. Because the term of imprisonment was determined in whole at the
time of sentencing (rather than being subject to later revision by parole
boards and the Bureau of Prisons), that made the sentence issued in the
judgment more important, and amplified the distortions and anomalies
contained therein, including those caused by the confusion cloud of policy
goals discussed here.
Importantly for this discussion, the elimination of parole and good
time meant that where there previously had been multiple chances to achieve
the goals of sentencing, now there was only one—the sentencing itself.
Without the possible mitigating effects of parole and good time, in any
individual case only the courtroom judge (and the Sentencing Commission
directing her via the guidelines) was left to make real the many and
conflicting goals of sentencing articulated by Congress.

76
77
78

Id.
18 U.S.C. § 3624(a).
18 U.S.C. § 3624(b).

24

h.

Reflecting community beliefs 79

Among the many sentencing goals which promote national standards
and uniformity, there is a striking anomaly: 28 U.S.C. § 994(c)(4) directs
that the guidelines shall take into account to the extent it is relevant “the
community view of the gravity of the offense.” 80 While this goal is not
explicitly reflected in the guidelines themselves, they are built into the
structure they are a part of, through the great deal of discretion the
guidelines and related statutes give to localized federal prosecutors.
Stephanos Bibas has chronicled the significance of these local variations
through prosecutorial discretion, 81 which the Sentencing Commission has
continued to allow. Specifically, Bibas describes significant variations in
the way, for example, that different federal prosecutors employ substantial
assistance departures. 82
Such localized variations, of course, undermine many of the other
goals described here. Most obviously, it cuts against uniformity, the goal
some see as first among many. 83

79

28 U.S.C. § 994(c)(4).
Id.
81
Stephanos Bibas, Regulating Local Variations In Federal Sentencing, 58 Stanford L. Rev. 137 (2005)
82
Id. at 148-153.
83
Stith and Cabranes, Fear of Judging, 104 (1998)(“Reduction of ‘unwarranted sentencing disparities’
was a—probably the—goal of the Sentencing Reform act of 1984.”)(emphasis in original)
80

25

i.

Commonality of the offense 84

The Commission is charged in 28 U.S.C. § 994(c)(7) with crafting
guidelines that take into account (where relevant) “the current incidence of
the offense in the community and in the Nation as a whole.” This, it must
be acknowledged, the Commission has certainly done, at least when it is
reacting to the emergence of a new narcotic. Often, this has been in
response to a direct Congressional directive. One good example of this
dynamic is the recent treatment of anabolic steroids.
In February, 2004, Attorney General John Ashcroft personally
announced the indictment of several men connected with the BALCO lab in
San Francisco, who were charged with making and selling steroids. 85
President Bush even denounced steroid use in his State of the Union address
that year, 86 and subsequently Senator John McCain and others promoted
bills in Congress which would require mandatory uniform testing of
professional athletes for the use of anabolic steroids. 87 Clearly, the nation’s
politicians perceived a growing epidemic of steroid use.

84

28 U.S.C. § 994(c)(7)
Richard D. Collins, Of Ballparks and Jail Yards: Pumping Up the War on Steroids, 30 Nov. Champion
22 (2006).
86
Id.
87
Lindsay J. Taylor, Congressional Attempts to “Strike Out” Steroids: Constitutional Concerns About
the Clean Sports Act, 49 Ariz. L. Rev. 961, 961-962 (2007).
85

26

Consistent with 28 U.S.C. § 994(c)(7), Congress then directed the
United States Sentencing Commission to consider an increase to the steroid
guidelines. 88 The Sentencing Commission took this advice and acted,
bumping up the guidelines sentence for a given amount of steroids. 89
Whether or not there was an upsurge of steroid use around 2004, the
guidelines reacted to the perception that there was such a relevant change in
the commonality of that particular type of drug abuse, reflecting yet another
Congressional directive to sentencing.
j.

Effect on prison populations 90

At the time the guidelines were created, Congress ordered that the
Sentencing Commission, in conjunction with the Bureau of Prisons, report to
Congress on the “maximum utilization of resources to deal effectively with
the federal prison population.”91 This implies, at least, that the guidelines
are to be created with an eye to the effect that the guidelines would have on
prison populations. Presumably, this would be to limit the effect of the
guidelines on prison populations—that is, to avoid the need for a prisonbuilding binge due to the impact of sentencing guidelines.

88

Anabolic Steroid Control Act of 2004, Pub. L. No. 108-358, 118 Stat. 1661, § 3 (2004)(codified at 21
U.S.C. § 801, 802, & 811 (2004).
89
The commission achieved this increase by adjusting the dosage amount for steroids. Richard D.
Collins, Of Ballparks and Jail Yards: Pumping Up the War on Steroids, 30 Nov. Champion 23 (2006).
90
91

28 U.S.C. § 994(q).
Id.

27

It is unclear that this implication had much effect. Prior to the
guidelines, about four in ten federal offenders went to prison. 92 By 2006,
that number was 9.5 out of ten: 93 Twenty-some years down the guideline
path, nearly all defendants are going to prison. Predictably, the federal
prison population shot up. In 1984, there were 32,317 people in federal
prisons. 94 By 1992, that figure had doubled, and in 2007 the federal prison
population stood at a shocking 198,656 95 -- six times the population at the
time the guidelines were created.
Unlike the other broad directives discussed above, it seems as if the
mandate to consider prison populations had little direct effect on the
guidelines themselves. However, it could be that the Commission’s reports
may have deterred Congress from passing some laws which might have had
a drastic effect on prison populations. For example, in analyzing the
proposed Gang Deterrence and Community Protection Act of 2007, the
Commission reported that that act would create the need for $9,000,000,000
to construct about 23,600 additional prison beds.96 It’s easy to imagine that

92

Timothy P. Cadigan, Pretrial Services in the Federal System: Impact of the Pretrial Services Act of
1982, 71 Federal Probation 10, 13 (September 2007).
93
Id.
94
Id.
95
Id.
96
Tiffany Sykes, Much Ado About Something: Reconciling Roper v. Simmons With the Gang
Deterrence and Community Protection Act of 2007 and the Possibility of Inconsistent Jurisprudence, 34
New Eng. J. on Crim. and Civ. Confinement 163, 171 (Note)(2008), citing Jeffery A. Kidder, Gang
Deterrence and the Community Protection Act of 2005: Why the Federal Response to MS-13 is Flawed

28

in tough economic times, this may have dampened chance for the passage of
that bill.
2.

Specific Provisions

Some of Congress’s policy goals apply only to certain types of cases.
Listed below are only a fraction of the total number of these specific
sentencing goals.
a.

Restitution

In concert with the creation of the guidelines, Congress directed that
the court in a given case consider the need for the defendant to “provide
restitution to any victims of the offense.” 97 The guidelines, consistent with
statute, 98 direct that the payment of restitution be required for the full
amount of the victim’s loss, 99 even if that would reduce the amount of a
fine. 100
Conceivably, this commitment to restitution would auger against a
sentence of imprisonment, so that the defendant would be free to work and
earn money in order to pay off the restitution amount. This tradeoff—
disfavoring prison for probation where restitution is possible—seems to be

and How it Will Have An Adverse Impact On Your State, 33 New Eng. J. on Crim. and Civ. Confinement
639, 648 n. 85 (2007).
97
18 U.S.C. § 3553(a)(7).
98
18 U.S.C. 3572(b).
99
U.S.S.C. § 5E1.1(a)
100
U.S.S.C. § 5E1.1(c).

29

in tension with some of the other principles articulated above, including the
call to uniformity 101 and the ban on considering socioeconomic status. 102
b.

Harsh punishment for certain crimes 103

18 U.S.C. § 994(h) created what we now know as the career offender
provisions of the guidelines, 104 which advise harsh punishments for those
charged with drug crimes or crimes of violence and who have at least two
such prior convictions. 105 Such offenders receive a stiff upward adjustment
not only in their offense level score, 106 but in the other axis of the sentencing
grid, the criminal history category, regardless of their actual criminal
history. 107 Thus, an offender with three minor marijuana trafficking
offenses spread over two decades may end up with a more severe sentence
than some drug kingpins. As set out in the next section, this creates a direct
conflict with the competing principle of parsimony. 108
c.

Rewarding Co-operators109

Congress further mandated that the guidelines encourage cooperation
with the government through the promise of lower sentences and the waiver
101

There would be conflict because in otherwise similar cases, defendants with the ability to pay would
get probation, while those with lesser job skills might end up in prison.
102
Presumably, socioeconomic status relates to earning power, which will equate to a greater ability to
pay restitution if the defendant receives a sentence of probation.
103
28 U.S.C. § 994(h).
104
U.S.S.C. §4B1.1.
105
U.S.S.C. §4B1.1(a).
106
U.S.S.C. §4B1.1(b).
107
Id.
108
Section III(A)(1), supra
109
28 U.S.C. § 994(n).

30

of mandatory minimum sentence provisions for those who provide the
government with what the government decides is “substantial assistance.” 110
This mandate was fulfilled in the guidelines through the provisions at §
5K1.1, which allow downward departures for cooperators. These authorized
departures are now particularly important in federal criminal law, as they not
only are an essential tool for prosecutors, but hold out for many defendants
the only hope to escape harsh mandatory minimum sentencing provisions. 111
By making the breaks given to cooperators so important within federal
sentencing, Congress and the Commission, of course, sacrifice the hope of
fulfilling certain of the other principles they have set out. Uniformity, of
course, loses out, as does neutrality as to race and socioeconomic class. 112
3.

Trap-door provisions

A final category of Congressional requirement would be trap-door
provisions, which allow for an unlimited number of additional policy
provisions to enter into the calculations that supposedly are determining a
federal sentence. The legislative equivalent of using a magical wish to ask
for more wishes, federal law contains two primary trap doors, allowing
unforeseen new principles to be introduced to the scheme through further

110

Id.
E.g. 21 U.S.C. § 841.
112
Those defendants with access to other defendants, often through bonds of race and class, have the best
chance of successfully getting a break for cooperating with federal investigators.
111

31

directions from Congress, 113 and directions from Sentencing Commission. 114
This allows both bodies to develop even more guiding principles on the fly,
as if the welter of provisions we already have is not enough. Periodically we
see this happen, such as when Congress decided to protect the integrity of
professional sports against the threat of steroids, as discussed above. 115
In a sense, the trap-door provisions may tell us more about the
problems with the federal sentencing project than anything else. Not content
with having created tens of distinct and competing policy principles,
Congress reserved the right to create even more and inject them into an
already confused and effectively random system of sentencing.
III.

The Problems of Pointlessness and Uniformity
The listings above only describe less than half of the policies

embedded in federal sentencing statutes, but should serve to illustrate the
dynamic at work—one where so many policy strands are knit together that
the resulting fabric resembles none of them. Below, I will first explore just
a few of the resulting conflicts within this mess, and then describe the effect
of combining this project with a consistent and unyielding desire (on the part
of Congress) for uniformity.

113
114
115

28 U.S.C. § 994 & note following.
18 U.S.C. § 3553(a)(5).
Section II(B)(1)(i), supra.

32

A.

The conflicts within the policy swamp

While describing 116 some of the policies involved in federal
sentencing and the guidelines, I have mentioned a few of the conflicts
created when these policy goals conflict. I would like now to evaluate a few
more in order to exemplify the workings of these conflicts at ground level. I
am able to describe only a small fraction of the total conflicts; at some level,
of course, each of the policy goals opposes all the others for primacy in
affecting the sentence of any given defendant.
1.

Parsimony v. Harsh Punishment for Certain
Offenders

The principle of parsimony, as described above, 117 requires that a
sentence should be “sufficient, but not greater than necessary,” to comply
with the traditional sentencing goals. 118 Congress chose to place this
principle at the heart of its description of sentencing process. 119
At the same time, however, Congress created the career offender
provisions, 120 which direct especially harsh sentences for those convicted of
narcotics or violent crimes who have two prior convictions for drug

116
117
118
119
120

Section II, supra.
Section II(B)(1)(b), supra.
18 U.S.C. § 3553(a).
Id.
Section II(B)(2)(b), supra.

33

trafficking or violent crimes. 121 This is a particularly blunt instrument, as it
covers both drug kingpins and those who have three relatively minor
convictions for selling small amounts of marijuana. 122
It’s not hard to see how these two policies conflict. 123 If the career
offender provision is followed, in many cases it will run contrary to the
parsimony provision which requires more individualized consideration. A
good example of this conflict was described in United States v.
Fernandez, 124 in which the district judge considered a defendant who
qualified as a career offender, based on two relatively minor prior
convictions. 125 The judge there noted that the career offender provisions
applied but would double the sentence, 126 and rejected the application of
those provisions in the case, as “the advisory guideline range was greater
than necessary to satisfy the purposes of sentencing.” 127 Other judges seem

121

U.S.S.C. § 4B1.1(a).
Id.
123
The parsimony provision, of course, conflicts with a number of other policies as well. For example, it
conflicts with the directive to base the guidelines on prior experience, which may not reflect parsimony.
Parsimony also conflicts with the desire for guidelines to reflect community beliefs, which may exaggerate
the threat posed by a given category of crime based on sensationalistic media reports.
124
436 F. Supp. 2d 983 (E.D. Wisc. 2006).
125
436 F. Supp. 2d at 987.
126
436 F. Supp. 2d at 990.
127
Id.
122

34

to be reaching the same conclusion, 128 a result which over time will erode
the uniformity sought by the career offender provision. 129
2.

Neutrality v. Consistency with prior sentencing
practices

If the guidelines are clear in stating anything, it is that both the
guidelines themselves and sentencing judges are to be strictly neutral as to
race, sex, national origin, creed, and socioeconomic status. 130 Nonetheless,
Congress decreed that the starting point for creating the new guidelines be a
survey of existing practices. 131
What was not done in basing the guidelines on prior practices was to
filter those practices for racial, gender, and other disparities. In other words,
the guidelines, while expressing strict neutrality, began with the simple step
by the Commission of building into the guidelines structure, root and branch,
any bias and prejudice that may have existed in the sentencing practices of
the judges who were surveyed. 132

128

E.g., United States v. Ortiz, 502 F. Supp. 2d 712 (N.D. Ohio 2007); United States v. Vigorito, 2007
WL 4125914 (N.D. Ohio 2007).
129
One could argue that the erosion of one Congressional mandate by judicial action is not an inherent
conflict, but rather a dialogue. Remember, however, that the provisions of 18 U.S.C. § 3553(a) (including
the parsimony provision) are to direct the Sentencing Commission as well in formulating the guidelines. 28
U.S.C. § 991(b).
130
Section II(B)(1)(d), supra.
131
Section II(B)(1)(f), supra.
132
One exception to this general observation would be that the directive to adjust under-punished crimes
(18 U.S.C. § 994(m)) was employed to raise sentences for white-collar criminals, a result that may have
countered a pre-existing bias in favor of the wealthy as a socioeconomic group. See Stephen Breyer, The

35

3.

Certainty v. Rewarding Co-operators

The federal sentencing scheme changed drastically at the time the
guidelines were first employed, in part because of the initial emphasis on
certainty of sentencing at the time sentence is announced, achieved in large
part by eliminating parole and diminishing the effect of ‘good time’
credit. 133 At the same time, Congress insisted that defendants who assist the
prosecution be rewarded for their efforts with a break in their sentences. 134
The tension between these two directives comes from the fact that
many cooperators are rewarded with a break on their term of incarceration
after they are sentenced, as is authorized by Federal Rule of Criminal
Procedure 35(b). Thus, certainty at the time of sentencing is undone. In
fact, the amount of uncertainty created by the re-sentencing of cooperators is
exacerbated by the fact that the guidelines do not restrict the size of a
departure once the court has found that “substantial assistance” has been
given. 135 Thus, the judge is free to change the sentence as much as she
wants, giving her powers similar to that of the parole board before certainty
became a central aspect of federal sentencing. 136

Federal Sentencing Guidelines and the Key Compromises Upon Which They Rest, 17 Hofstra L. Rev. 1
(1988).
133
Section II(B)(1)(g), supra.
134
Section II(B)(2)(c), supra.
135
U.S.S.G. § 5K1.1.
136
The analogy to a parole board is not a tight fit, of course, since the judge reviewing a defendant’s
assistance after sentencing will often be the same person who issued the initial sentence, and looks to

36

Needless to say, this is just a thimbleful from the swamp of conflicts
created by the very fact that at least 31 policy goals fight for attention in the
realm of federal sentencing. The biggest problem with this is simple: The
failure to articulate a reasonable set of goals robs the guideline system of any
hope of moral authority. There is no way to measure success when so many
factors are in play and in tension. By its nature, a system with so many
goals has the moral trajectory of a toy boat in a baby pool being splashed by
a group of toddlers. At the heart of what should be the most transparently
moralistic of our governmental functions, we have nothing less than moral
relativism, where a virtually limitless set of principles are at play with no
sorting mechanism at hand.

B.

The Dangerous Combination: Uniformity and Pointlessness

Federal sentencing policy isn’t really a policy; it’s a grab-bag of too
many ideas and priorities. On its own, this could be seen as typical of
Congressional action in many areas where it is creates an administrative
agency and then hands off power to that agency. Congress is free in those
circumstances to decree what is important by laying out guiding principles,
and then leave the messy work of implementation to others—that is the
different criteria that a parole board would. The analogy is apposite only in that certainty at the point in
time of judgment is undermined.

37

structure of our government. What is perhaps unusual about sentencing,
though, is that in creating the guidelines and the Sentencing Commission,
there was also an overarching goal at work, which was the perceived
primary virtue of the guidelines—the desire to achieve uniform sentences
from case to case and judge to judge. To anyone paying attention, it is
perfectly clear that Congress’s primary sentencing goal has been the same
for the past 20-some years: Eliminate disparities and create uniform
sentences across the nation. 137
Congress has attempted to achieve this goal through the imposition of
sentencing guidelines, the passage of mandatory minimum sentences, 138 and
the reporting of individual judges’ sentences to the United States Sentencing
Commission for evaluation. 139 Intriguingly, within the debate over
sentencing, uniformity is consistently discussed as an end in itself rather
than as a tool to best fulfill other policies. 140 The result, even after Booker,
has been the most restrictive sentencing system in the nation—one that
imposes more uniformity and restricts judicial discretion more severely than

137

Stith & Cabranes, Fear of Judging: Sentencing Guidelines in the Federal Courts 104 (1998).
E.g., 21 U.S.C. § 841(b).
139
These reports are made within 30 days of the entry of judgment in every federal case, and are
mandated by 28 U.S.C. § 994(w)(1).
140
For example, in the Kimbrough opinion on crack cocaine sentencing, even while acknowledging that
Booker marked a departure from strict uniformity, the Supreme Court simply stated without elaboration
that “it is unquestioned that uniformity remains an important goal of sentencing.” 128 S. Ct. 558, 573
(2007).
138

38

any of the 50 state systems that overlap with federal courts in their common
project of regulating crime. 141
Certainly, the argument can be made, and Albert Altschuler has made
it quite convincingly, 142 that despite these efforts, uniformity has not been
achieved. Altschuler points out that regional disparities tripled after the
guidelines went into effect, rather than decreasing as intended, and further
argues that disparities as a whole increased in the first fifteen years of the
guidelines’ existence. 143 Credibly, in explaining this unexpected result, he
points to prosecutorial discretion being employed in far different ways in
different parts of the country. 144
The desire to create uniformity, whether it has been fulfilled or not,
never made sense in the first place unless those uniform results were
consistent with understandable, limited, and discrete goals. The failure to
articulate a simple set of goals before imposing the machinery of
uniformity 145 had two major effects. First, this robbed the guidelines of the
chance of cleanly measuring success in achieving any one of the too-many
goals. Second, it amplified the problems associated with the lack of a moral
141

Kevin R. Reitz, The Enforceability of Sentencing Guidelines, 58 Stanford L.Rev. 155 (2005).
Albert W. Altschuler, Disparity: The Normative and Empirical Failure of the Federal Guidelines, 58
Stanford L. Rev. 85 (2005).
143
Id. at 101.
144
Id. at 102. Stephanos Bibas describes in some detail the way in which this works as to two components
of judicial discretion—the use of fast-track programs and employment of substantial assistance departures.
Stephanos Bibas, Regulating Local Variations in Federal Sentencing, 58 Stanford L. Rev. 137 (2005).
145
Mandatory minimums and the guidelines.
142

39

compass described in the preceding section, 146 by prescribing bright
normative lines that are unmoored from a simple, understandable moral
anchor.
1.

The Unmeasurable Goals

Like a sports league for six-year-olds where the score is not kept
because it might make one of the teams feel bad, Congress has created a
system in which cause and effect cannot be measured, and which as a result
is without accountability to anything. Having a huge number of conflicting
policy goals makes it almost impossible to measure the success of progress
towards any one of those goals. 147 Even the goal which should be simple to
measure, uniformity, seems not to have worked out the way people hoped, or
easy to measure. 148 As Amy Baron-Evans points out, fifteen years after the
guidelines were imposed it was still unclear whether the increased severity
of the guidelines 149 had accomplished any sentencing purpose. 150

146

Section III(A), supra.
William Stuntz has compellingly described the natural tendency to constantly increase the number of
crimes on the books, to the point where the criminal code is so broad that it covers an astonishing array of
activities, to the point where the penal code becomes almost indeterminate. William Stuntz, The
Pathological Politics of Criminal Law, 100 Mich. L. Rev. 505, 555547-558 (2001). What I critique here is
distinct but parallel to that analysis—I am making the same argument as to Congress’s policy goals as
articulated in statute. The two intersect, of course—those policy goals at times will serve as justification
for expanding the penal code as well as jacking up guideline ranges.
148
Albert W. Altschuler, Disparity: The Normative and Empirical Failure of the Federal Guidelines, 58
Stanford L. Rev. 85, 100-103 (2005).
149
Discussed supra, § II(B)(1)(f).
150
Amy Baron-Evans, The Continuing Struggle for Just, Effective and Constitutional Sentencing After
United States v. Booker: Why and How the Guidelines Do Not Comply With § 3553(a), 30 Oct Champion
32, 34-35 (September/October 2006).
147

40

Thus we are left with a guideline structure created with no clear
relationship to its founding principles (other, perhaps, than uniformity), and
which continues to function without an ability to measure a relationship to
those founding principles. The sentencing guidelines are not a success, in
part because the goals they aim to meet are so numerous that they defy the
effort to collect and analyze data which could support the conclusion that
any one of them were a success or failure. 151 Much like a game of Chinese
checkers that 31 people can play at the same time, it is extremely difficult to
figure out what is going on.
2.

Harshness in Service of Nothing

Because the guidelines are pointless, they cannot be rational in
relationship to any discrete sentencing goal. Yet, in striving for uniformity,
Congress has made them uniformly harsh. On a playground, we know what
the combination of irrationality combined with harshness is—it is a bully,
looking to pick on those with less power. The same combination in the
person of our federal sentencing scheme produces a system, a machine
really, that resembles nothing so much as that playground bully—
unreasoning, uncompassionate, and unprincipled.

151

The problem is not a simple lack of data relative to sentencing—the sentencing commission has
produced thousands of pages of data relative to sentencing in the federal courts. (These reports are
available at http://ussc.gov). Rather, the problem is that it is impossible to analyze that data so that it can
reveal how any one of the tens of policy goals are being fulfilled.

41

This moral relativism born of too many goals is especially sad in an
era where very often criminal law is said to be about “sending signals.” 152
Those signals are, or should be, moral signals about the bounds of socially
acceptable behavior, and the price to be paid for differentiated acts (which is
what the guidelines are about—normative price-setting for specific wrongful
acts). Included in these signals are messages about why an act is especially
reprehensible. Without a set of clear policy goals behind it, sentencing
practice loses the value of this important function. In fact, it seems that if
there is one message conveyed to the public under the contemporary
scheme, it is one of simple retribution for any type of crime, a message
inconsistent with much (though certainly not all) of what Congress has
articulated as the policy goals 153 of the guidelines. 154
In particular, this perceived message of widespread, consistent, and
harsh retribution is at odds with the parsimony principle which Congress has
established as the fulcrum of a trial court’s sentencing mechanism. 155 That
provision specifically demands that a sentencing judge impose a sentence
152

See Dan M. Kahan, The Secret Ambition of Deterrence, 113 Harvard L. Rev. 413 (1999); Eric A.
Psner, Law and Social Norms (2000).
153
Section II, supra.
154
In addition, this message is inconsistent with what highly selective federal prosecutors actually do.
Federal prosecutors handle less than 5% of the nation’s felonies, with the rest going to the state systems.
The great majority of these could have been tried at either the state or federal level. William Stuntz, The
Pathological Politics of Criminal Law, 100 Mich. L. Rev. 505, 542-543 (2001), citing Bureau of Justice
Statistics, U.S. Department of Justice, Sourcebook of Criminal Justice Statistics—1998, at 387-388 table
5.6 (Kathleen Maguire and Ann L. Pastore, eds., 1999).
155
18 U.S.C. § 3553(a).

42

“not greater than necessary” to accomplish the aims of the four traditional
sentencing goals. While retribution is one of those goals, it is both one
among equals and is held in check by the others, in particular the principled
goal of rehabilitation. Beyond that necessary balancing which is obscured
by the call to broad retribution, the claim that retribution is primary ignores
the other goals which can often be seen pulling the other way—including the
findings of social science, 156 the idea of fairness (either procedural or
substantive), 157 the effect of mass incarceration on prison populations,158 the
need to encourage restitution, 159 and the benefits given to often highlyculpable cooperators under guideline section 5K1.1. 160
This cry of retribution, while not supported by Congress’s articulated
sentencing goals as a whole, does accurately reflect the general harshness of
the federal scheme. 161 Were the guidelines to be underpinned by a
reasonable and understandable principle or set of principles, harshness (or
surprising lenience) might be a cost worth paying. However, when we
cannot say that uniformity or any other goal is being achieved, and recognize

156

Section II(B)(1)(c), supra.
Section II(B)(1)(e), supra.
158
Section II(B)(1)(j), supra.
159
Section II(B)(2)(a), supra.
160
Section II(B)(2)(c), supra.
161
For a good description of the mechanisms that lead to harsher sentences in the federal system, see
Frank O. Bowman III, Pour Encourager Les Autres?, 1 Ohio State J. Crim. L. 373, 387 (2004).
157

43

that the goals of sentencing have become convoluted, multifarious, and
unworkable, it may be time to seek change.

IV.

Into the Pointless Forest
If the guideline system is pointless and amoral, shouldn’t it matter

that they are no longer mandatory? Certainly, the effects of Booker may
over time mitigate some of the harms created by the guidelines by allowing
sentencing judges more discretion. However, the sentencing guidelines are
still at the center of the process of sentencing in federal court. As Booker
made clear, the mechanisms of sentencing are the same—including revision
of the guidelines and calculation of a guideline range in every case. 162
Given that, it is fair to say that the guidelines have less authority, but still
play a major role in federal sentencing. The critique above 163 applies
regardless of whether the guidelines are mandatory, advisory, or somewhere
in between. Do we want an irrational and pointless construct at the center of
our sentencing structure, even if it is not strictly mandatory? I would hope
not. So long as the guidelines remain at the center of the mechanism for
sentencing, they should be tethered to a few understandable and easily
articulated principles.
162
163

543 U.S. at 767.
Section II and III, supra.

44

If there is to be change, it seems there are three options. First,
Congress could scrap the guidelines altogether. While this might solve some
of the problems associated with the system as it exists, it seems politically
unlikely, given the sentiment in Congress for retaining some measure of
uniformity in federal sentencing.
Another option would be for Congress to restrict the statutory goals of
sentencing to the traditional goals while leaving the guidelines in place,
hoping that they evolve into something better over time. One could read this
article (at least up to this point) as an argument for doing exactly that-limiting the sentencing policy goals to a few understandable points. While
the need for fewer and simpler goals is certainly a part of my argument,
redefining the goals of sentencing would not make much sense unless the
guidelines themselves were either revised or removed from federal
sentencing. Unfortunately, the guidelines are now filled with the numerical
results of thousands of actions taken on behalf of one or another of the goals
described above. It would be impossible to undo the moral relativism of the
guidelines system without taking them apart and remaking them in a better
and more understandable fashion.
Finally, Congress could start the process over again with fewer goals
and advisory guidelines which are written from scratch. If we are to have

45

principled, understandable sentencing in the federal courts, this might be the
single best politically palatable option.
A.

A Project for Principles: Rewrite the Guidelines

Unless we are comfortable with the amoral strictness of the guidelines
we have, they must be remade.
Is it possible to take apart the guidelines and remake them? To do so
would require a massive effort involving the convening of a new guideline
Commission charged with starting from scratch the project of coming up
with federal sentencing guidelines.
Still, this undertaking might be worthwhile. A second-generation
Sentencing Commission starting with a clean sheet of paper would have
significant advantages over the group which came up with the first edition.
First, a new founding Commission would have the advantage of
learning from the problems with the current guidelines. For example, it
would allow for the thorough rethinking of charge v. real offense conduct as
the basis for sentencing. The first Commission’s compromise on this, which
allowed for relevant conduct including acquitted conduct to be
considered, 164 has been subjected (properly) to withering criticism for
importing into the current system, without context, a single feature of a

164

U.S.S.G. § 1B1.3.

46

bygone era in which rehabilitation was a primary goal of sentencing. 165 In
relation to this, a new commission would have the benefit of the reams of
data gathered by the Sentencing Commission staff over the past two
decades. This could be combined with other social science data from other
sources, in order to broaden the perspective of these second-generation
framers.
Second, the new Commission would be much better positioned than
the first to learn from the examples provided by several states with advisory
guidelines in place. The first Commission looked only the example of
Minnesota and Washington, both of whom had fairly new guideline regimes
at that time, and seemed to summarily reject them as too simplistic. 166 Now,
however there has been two decades of guideline experiments in a number
of states, all of which provides trial-and-error lessons for a new federal
system.
Third, starting from scratch at this point would allow the new
Commission to draw from the body of scholarship which has developed
since the mid-1980’s by writers such as Douglas Berman, Michael O’Hear,

165

Daniel J. Freed, Federal Sentencing in the Wake of Guidelines: Unacceptable Limits on the Discretion
of Sentencers, 101 Yale L.J. 1681, 1712-1713 (1992); Mark Osler, Must Have Got Lost: Traditional
Sentencing Goals, The False Trail of Uniformity and Process, and the Way Back Home, 54 S.C. L. Rev.
649, 669 (2003).
166
Stephen Breyer, The Federal Sentencing Guidelines and the Key Compromises Upon Which They
Rest, 17 Hofstra L. Rev. 1, 3 (1988).

47

Frank Bowman, Steven Chanenson, and Stephanos Bibas, each of whom has
focused on this field and who have had a significant impact on its
development.
Finally, a new Commission would have the advantage of a limited and
understandable group of directive principles. Part of the legislation creating
a new guideline Commission could start by phasing out 18 U.S.C. § 3553(a),
28 U.S.C. § 991 and 28 U.S.C. § 994 in favor of a much briefer articulation
of goals. It might be that the parsimony provision of 18 U.S.C. § 3553(a)
plus the four traditional sentencing goals would serve this purpose, 167 were
they expressly made the basis for guidelines in an active way, freed from the
command to place sharp limits on judges and increase sentences.
2. Change as the Pointless Forest
Predictably, there are those who would oppose any change in the
essential structure of the sentencing guidelines, especially given the strong
likelihood that the revision would result in greater judicial discretion. Those
who gain the most from the current regime would raise the strongest
objections. The present system gives tremendous power to federal

167

Like the parsimony provision, the four traditional goals are currently contained in 18 U.S.C. § 3553(a).

48

prosecutors, 168 and the risk of losing that power would no doubt cause them
to employ their significant lobbying abilities 169 to stop any such change.
One would expect in such an instance that the Department of Justice
would argue, in part, that changing the current system in a way which might
give judges more discretion would bring back disparities and destroy
uniformity. There are many counter-arguments to this, of course, including
some already made here: That the guidelines have increased, not decreased,
disparity,170 and that prosecutors themselves create great disparities under
the current system. 171 The best counter-argument to the Department of
Justice would be, though, that the guidelines must be reformed if they are to
give us hope for principled justice through federal criminal law.
Conclusion
Predicting this fight over uniformity and judicial discretion, of course,
brings us back to the Land of Point and Harry Nilsson’s acid trip.
Remember that Oblio and Arrow were banished to the pointless forest, a
place greatly feared by the pointy-headed types in the Land of Point. Once
there, though, Oblio realized that the land of point was not what he expected:

168

Albert Altschuler has described the employment of this discretion as being in the nature of the “good
cop” to Congress’s mean “bad cop.” Albert Altschuler, Disparity: The Normative and Empirical Failure
of the Federal Guidelines, 58 Stanford L. Rev. 85, 112-113 (2005).
169
William Stuntz, The Pathological Politics of Criminal Law, 100 Mich. L. Rev. 505, 534 (2001).
170
Albert Altschuler, Disparity: The Normative and Empirical Failure of the Federal Guidelines, 58
Stanford L. Rev. 85, 101-102 (2005).
171
Id.; Stephanos Bibas, Regulating Local Variations in Federal Sentencing, 58 Stan. L. Rev. 137 (2005).

49

… one of the first things Oblio and Arrow noticed
about the pointless forest was that all the leaves on
all the trees had points, and all the trees had points.
In fact, even the branches of all the trees pointed in
different directions, which seemed a little strange
for a pointless forest. 172
The critics of increasing judicial discretion are right in saying that it
creates disparities—that the branches (judges) do tend to point in different
directions. Judges differ, and they do so in substantive and occasionally
troubling ways. Nonetheless, they tend to sentence on principles which do
directly bear on the question at hand—that is, they have a concrete reason
for doing what they have chosen with a given defendant. There will be, in
other words, at least one principle at play which directly underlays the
crafting of the sentence, a principle which is going to be articulated,
explained, and connected expressly to the result by that judge. 173 In short,
the sentence that results from the discretion of a judge has a point, even if it
is not the same point another judge might make in similar circumstances.
While this does risk endangering the (perhaps false) perception of
uniformity under the guidelines, it at the least is better than our current
guideline-driven system, which is so awash in conflicting policy goals that a
principled point is not even possible.
172

Harry Nilsson, The Pointless Man, The Point Soundtrack (BMG Entertainment 1970)
As Judge Cabranes put it, when a judge has discretion “Judgment proceeds from principles. These
principles can and should be stated, rationally discussed, attacked, and defended.” Stith and Cabranes, Fear
of Judging: Sentencing Guidelines in the Federal Courts 82 (1998).

173

50

The present guidelines, even in advisory form, are hopelessly amoral
because they are not informed by understandably simple policy goals. If the
legislative will requires we have guidelines, the ones we have now need to
be scrapped and re-made from the ground up. Doing so will likely create
greater judicial discretion, but for most Americans, I suspect that some
disparity is an acceptable cost for the hope of a sentencing system with
actual principles at play.
It may seem harsh to describe Congress’s actions as an “acid trip,” but
the analogy is not entirely inapt. One symptom of LSD 174 use is the “fear of
losing control.” 175 It is indisputable that if a new guideline system was put
into place, it might lessen Congress’s control relative to sentencing judges.
However, this shift is necessary if we are to regain moral credibility in
sentencing.

174

Lysergic acid diethylamide.
National Institute on Drug Abuse, “NIDA Infofacts: LSD,” available at
http://www.nida.nih.gov/infofacts/lsd.html.

175

51