The Financial Gains from Criminal Appeals: An Investigation into the Administrative
Costs and Incarceration Cost Savings of Post-Conviction Review
Andrew Chongseh Kim*
Draft : Do not cite or quote without permission of the author.

I. Introduction
Judges, legislators, and voters are strongly sensitive to anything that increases the
financial costs of adjudicating criminal cases. 1 For example, in 1994, Michigan voters amended
their state constitution to provide that appeals by defendants who pled guilty would be by leave
of the court, rather than by right. 2 Defendants who pled guilty could still petition for the right to
appeal, but, after new legislation in 1999, could not receive publicly financed counsel to help
prepare these petitions. 3 These measures were taken in order to reduce significant backlogs in the
Michigan Court of Appeals 4 and, implicitly, to reduce the costs of handling such appeals. 5 The
procedural rights of presumptively guilty defendants did not justify the financial and
administrative burdens on the state.
Judges are also wary of the financial burden that defendants' procedural rights can impose
on the states. When limiting defendants' access to appellate review or new procedural rights, the
Supreme Court has often cited the "costs imposed upon the States" 6 or the "burden on judicial
and prosecutorial resources" that would be necessary to enforce defendant's rights. 7 More
At the same time the Court has asserted that broader access to postconviction review
reduces the deterrent effect of criminal law by adding an element of uncertainty to sentences,
"undermining the principle of finality." 8 9 Though these financial effects very defendants'
postconviction procedural rights as financial burdens and as factors that reduce the deterrent
effect of incarceration by making it less certain that criminal behavior will be met with
punishment.
* Law Clerk for Justice Richard Palmer, Connecticut Supreme Court, University of Chicago B.A. 2000, Harvard
Law School; J.D. 2010.
I thank William Stuntz for comments and inspiration.
1
Cf. Halbert v. Michigan, 545 U.S. 605, 609 (2005) (Thomas, J., dissenting).
2
See Halbert v. Michigan, 545 U.S. 605, 609 (2005) (Thomas, J., dissenting).
3
Shortly after the constitutional amendment, some judges began denying requests for appointed counsel to assist in
preparing a petition for leave to appeal. In 1999, the Michigan Legislature passed a law making such denials
mandatory in most cases. Mich. Comp. Laws § 770.3a. This practice was later found unconstitutional by the United States
Supreme Court. See id. at 609, 625-26.
4
Id. at 625.
5
Cf. id. at 629.
6
Teague v. Lane, 489 U.S. 288, 310 (1989) (internal quotations removed).
7
Stumes, 654, quoted in Teague; see e.g. District Attorney's Office for the Third Judicial District, v. Osborne, 129 S.
Ct. 2308, 2326-29 (2009) (Alito concurring) cataloging a number of costs associated with post-conviction DNA
testing).
8
Teague, 309.
9
See e.g., Griffin v. Illinois, 351 U.S. 12, 24 (1956); Teague v. Lane, 489 U.S. 288, 309-10 (1989) ("Without
finality, the criminal law is deprived of much of its deterrent effect"); Solem v. Stumes, 465 U.S. 638, 654 (1984);
Halbert (630-31,623, Thomas dissenting); Herrera v. Collins, 506 U.S. 390, 417 (1993); cf. Scott and Stuntz
(discussing the reduced deterrent effects if certain punishment through plea bargains were replaced with uncertain
trials).

Electronic copy available at: http://ssrn.com/abstract=1680626

Though the American system of plea bargaining, which resolves 97% of criminal cases in
the United States has been probably criticized, has been heavily criticized long been used to
validate the American plea bargaining system that dominates American criminal justice. 10
Judges and legislatures often conceptualize the post-conviction rights of criminal
defendants as imposing burdens on society that must be balanced against the legal and moral
benefits of providing those rights. 11 However, the assumption that criminal appeals impose a
financial burden on the state ignores the fact that successful appeals produce substantial savings
through reduced incarceration costs. 12 The assumption that the possibility of post-conviction
relief reduces the deterrent effect of criminal law ignores the importance of legitimacy in
obtaining compliance with the law and the negative deterrent effects of wrongful convictions. 13
This paper argues that even though the "finality" that is "disrupt[ed]" by post-conviction claims
of innocence or sentencing errors may serve important social goals, the actual financial costs and
reduced deterrent effects of entertaining these claims are often much lower than generally
believed. In fact, in some cases, increasing access to counseled post-conviction review may
created net financial savings to the state and increase the deterrent value of the law.
This paper proceeds in four parts. In part II, I compile recent research on the costs and
benefits of post-conviction procedures and use back of the envelope calculations to demonstrate
that for cases with lengthy sentences, increased access to post-conviction remedies may yield net
cost savings to the state.
The basic economic model of criminal behavior demonstrates that wrongful convictions
reduce deterrence by decreasing the benefits of not committing crimes. However some scholars
have criticized this model and argued that the reduced deterrent effect of wrongful convictions is
generally minimal. In part III A, I argue that the reduced deterrent effects of wrongful
convictions can be much larger among those populations most at risk for criminal behavior. I
demonstrate that the criticisms of the reduced deterrence of wrongful convictions theory carry far
less weight when describing the incentives of at-risk people and communities.
Under economic theories of criminal law, the deterrent effect of the law is highly
influenced by the expected severity of punishment. Any proposal that would reduce net
incarceration must address the possibility that the net deterrent effect of criminal law would also
be reduced. In part III B, I discuss the deterrent effects of legitimacy and argue that denying
judicial consideration of claims perceived by defendants as potentially valid reduces their
10
11

Cf. Stumes at 654
See generally, Dawn Van Hoek, Penny-Wise and Pound-Foolish: Waste in Michigan Public Defense Spending,
written testimony of Dawn Van Hoek, Chief Deupty Director of the Michigan State Appellate Defender Office, to
the United States House Of Representatives Committee on the Judiciary, House Subcommittee on Crime, Terrorism
and Homeland Security, March 26, 2009. (Hereinafter DVH, SADO). Available at
http://www.sado.org/sado_news/DVH_testimony_3-26-09.pdf, last visited April 18, 2010.
13
See generally, Matteo Rizzolli and Luca Stanca, Judicial Errors and Crime Deterrence: Theory and Experimental
Evidence, Working Paper, (2009) (theorizing that "in the presence of risk aversion, loss aversion, or differential
sensitivity to procedural fairness, [wrongful punishment errors] can have a larger effect on deterrence than [failures
to punish]," and offering experimental evidence to support the claim) available at
http://ideas.repec.org/p/mib/wpaper/170.html; Tom R.Tyler, Why People Obey the Law, (2006).
12

Electronic copy available at: http://ssrn.com/abstract=1680626

perception of the legitimacy of the law. While post-conviction acquittals and sentence reductions
based on "technicalities" may impair the legitimacy of the law in the eyes of many law-abiding
citizens, I suggest that the delegitimization effects in the eyes of criminals and their communities
may be more directly related to actual deterrence of crime.
Part IV concludes.
II. Calculating the Financial Costs and Benefits of Post-Conviction Procedure
Many appellate courts are overworked, and appellate judges are likely well aware that
only one out of every five criminal appeals that they process yields any relief to the defendant. 14
Crafting rules that reduce the availability of post-conviction review reduces the number of the,
usually meritless, criminal claims they have to deal with. Similarly, disposing of an appeal on
procedural grounds, without having to weigh the actual merits of the claim may often reduce the
amount of time a judge must spend on a particular case. At the same time, the costs of continued
incarceration are incurred by, and have already been budgeted for, the Department of
Corrections. Judges are directly affected by decisions that increase judicial workloads, but do not
internalize the cost savings to the state of reduced incarceration. Because of this, the
administrative burdens of post-conviction review is likely much more salient for judges than the
costs of continued incarceration. This bias is visible in the large number of judicial opinions that
balance the state's interest in finality and the financial costs of post-conviction review against the
rights of defendants without any mention of the costs to the state of continuing to house
defendants longer than legally appropriate. 15 The failure of judges to consider the costs of legally
wrongful or excessive incarceration while consistently giving explicit weight to the
administrative burdens on the judiciary is essentially an agency problem and may cause judges to
inefficiently prefer policies that favor judicial efficiency over the rights of defendants and the
financial interests of the states. 16 This is not to say that every restriction of defendants' rights in
the interests of judicial efficiency or "finality" is undesirable. The nonfinancial interests of
finality can be substantial. For instance, contemporaneous objection rules, wherein the failure to
promptly object to certain errors forfeits the right to have the error corrected, give incentives to
defense counsel to avoid "sandbagging," 17 wherein defense counsel consciously decides not to
bring an error to the court's attention in hopes of later securing a retrial if the first trial ended
unfavorably. Theoretically, if the defense had no incentives to object to errors as they occurred
and could always obtain a new trial in the event of an error, trials might balloon exponentially, at
14

See Joy A. Chapper and Roger A. Hanson, Understanding Reversible Error in Criminal Appeals, final report
submitted by the National Center for State Courts (1989) (hereinafter NCSC) (empirical analysis of criminal
appeals); cf e.g. Michigan, in Halbert.
15
See e.g. Teague v. Lane, (discussing the financial burdens on states of having to relitigate finalized cases when a
new constitutional rule is made retroactive without mentioning the costs of continued incarceration); Justice
Thomas's dissent in Halbert; Harlan's dissent in Chapman v. California, 386 U.S. 18, 57 (1967) ("It cuts sharply into
the finality of state criminal processes; it bids fair to place an unnecessary substantial burden of work on the federal
courts; and it opens the door to further excursions by the federal judiciary into state judicial domains"); United
States v. John, 597 F.3d 263, (5th Cir. 2010) (dissent) (not every sentencing error that increases a sentence merits
relief under plain error doctrine).
16
Cf. Stuntz, Unequal Justice (discussing the agency problem where county prosecutors are not forced to internalize
the costs of lengthy sentences, which are often paid for out of state, not county budgets).
17
See Puckett v. United States, 129 S. Ct. 1423, 1431-32 (2009).

Electronic copy available at: http://ssrn.com/abstract=1680626

substantial financial cost. 18 In this section, I attempt to estimate the magnitude of the
incarceration savings from direct appeals to gain a better sense of how significant the state's
financial interest in avoiding the administrative burdens of post-conviction procedure really are. I
conclude that states do not incur substantial net costs by providing counseled direct appeals and
may well realize net financial savings through reduced unnecessary incarceration costs. 19
Much of the research comparing the costs of post conviction proceedings with the costs
of punishment comes from death penalty literature. Studies have consistently found that the full
costs of criminal procedural protections for capital defendants, including trials, appeals, and
habeas proceedings, outstrips the costs of life incarceration and the less thorough protections
offered life sentence defendants. 20 When the outcomes of capital post-conviction review favor of
the defendant, the costs to the state often increase substantially, through civil settlements in cases
of wrongful conviction and through incarceration costs when only the death penalty itself is
overturned. For capital cases, these post-conviction protections cost states significant amounts of
money that could be saved by reducing post-conviction review, seeking non-death sentences, or
eliminating appellate review and proceeding straight to execution. 21 The publicity around these
studies may contribute to the perception that criminal procedural rights are offered to generally
guilty defendants at great financial cost to the state.
Post-conviction proceedings outside of capital cases are much cheaper and outcomes that
favor the defendant produce substantial savings to the state through reduced incarceration costs.
Incarcerating a prisoner for one year costs around $30,000, though this number varies
significantly by jurisdiction and type of facility. 22 When a conviction is overturned, or remanded
and not prosecuted further, the state no longer has to pay to imprison the defendant for the
remainder of his term, which can be a considerable savings. 23 When, as occurs frequently, postconviction proceedings lead not to acquittals, but to reduced sentences, the state accrues
substantial savings on the future incarceration that will no longer occur. 24 Recent research
suggests that even though success rates for post-conviction procedures tend to be relatively low,
the net "costs" to the state of providing these procedures is negative in many cases.

18

Cf. id.
Cf. DVH, SADO (implying the same for appeals and public defense in Michigan); Richard C. Goemann, First
You Cripple Public Defense: Musings on How Policymakers Dismantle the Adversarial System In Criminal Cases,
9 Loy. J. Pub. Int. L. 239 (suggesting that because defendants represented by public defenders are incarcerated more
often than those represented by private counsel, the state might realize cost savings from better funding for public
defenders).
20
See generally http://www.deathpenaltyinfo.org/costs-death-penalty#financialfacts (tracking studies).
21
See e.g. INDIANA, $298,585, 2002, $363,797 in 2010 dollars: cost of execution is around $25,000; Maryland,
$661,000, $805,000 $2010 NC p. 3.
22
See Indiana, 2001 death penalty study, translated to 2010 dollars.
23
When a defendant is released, he may impose direct costs on the state through public assistance benefits, or
indirectly if he renews a life of crime. However, he may also contribute directly to state coffers through the payment
of taxes or indirectly by contributing productive labor to society. Though relevant, a full economic analysis of the
costs or contributions of recidivists to society is beyond the scope of this paper.
24
See DVH, SADO.
19

In Michigan, the State Appellate Defender Office (SADO) has compiled records of its
success rates for a Special Unit of three attorneys that handle appeals from guilty pleas. 25
Because guilty plea defendants have already conceded their guilt in open court, most of the
appeals handled by this unit relate to sentencing errors, in which defendants were sentenced to
more serious punishment based on inaccurate information or inaccurately calculated guidelines
recommended sentences. 26 These are generally not cases in which defendants exploit a legal
technicality to avoid punishment, but those in which defendants have been given lengthier, and
costlier, punishment than legislatively authorized or recommended. SADO has calculated that
the Special Unit obtained sentencing reductions in about a third of the plea cases that they
pursued. For successful cases, Special Unit attorneys obtained an average 6.3 month reduction to
the minimum sentence and an average 25.3 month reduction to the maximum sentence. If
incarceration costs $30,000 a year, this translates to an average minimum savings of $18,900 per
successful case. The average incarceration savings for Special Unit plea appeals is $3654 per
case, when all cases, including those in which defendants, after consulting with their attorney,
voluntarily dismiss, are considered. 27 In 2007, the work of the three attorneys, whose salaries
totaled around $200,000, 28 saved the state of Michigan at least $855,000 in incarceration costs. 29
Dawn Van Hoek, Chief Deputy Director of SADO, estimated that if the experience of the
Special Unit was representative of the success rate for all Michigan represented appeals, the total
incarceration savings from sentencing error correction for 2007 was at least $11,850,000. 30
Because defense costs are some of the largest expenses in post-conviction proceedings, 31 it is
likely that after prosecutor and court costs are considered, rather than imposing a financial
burden on the state, providing these defendants with post-conviction remedies and state-funded
attorneys saves the state money while ensuring that defendants are given only legally appropriate
levels of punishment.
It should be noted that around 40% of Special Unit appeals were concluded internally
when clients, after discussions with their attorney, voluntarily dismissed their appeals, allowing
the cases to be resolved no pleadings filed, incurring no additional expenses on behalf of
prosecutors or the courts. 32 Given the relatively high success rate of appeals that are pursued and
25

The Michigan Constitution was amended to eliminate appeals as of right for defendants who plead guilty and
subsequent legislation denied appointed counsel to assist defendants in petitions for discretionary appeal. However,
the Supreme Court in Halbert restored the right to appointed counsel for petitions for appeal. Halbert upheld the
denial of appeals as of right as constitutional.
26
See DVH, SADO, at 3.
27
Calculated from numbers in the 2007 Annual Report for the State Appellate Defender Office, and Michigan
Appellate Assigned Counsel System, by the Appellate Defender Commission. 5-6 (Hereinafter SADO) (available at
http://www.sado.org/commission/#annualreports, last visited April 18, 2010). The total annual per attorney
reduction was 9 1/2 years for minimum terms and 38 years for maximum terms. Each attorney found some success
in an average of 18 cases.
28
Cf. Performance Audit of the Appellate Defender Commission, 23 (2002) available at
http://audgen.michigan.gov/comprpt/docs/r0515501.pdf.
29
See SADO at 6
30
See DVH SADO at 6
31
Cf. Indiana, post-conviction proceedings for noncapital, life without parole cases cost $3724 in public defender
expenses and $3601 in Attorney General staff time. For death penalty cases, the costs were $123,101 for the defense
and $37,600 (figure partially illegible) for the Attorney General.
32
See SADO at 5, 5 n4. ("Historically, after Unit attorneys review the file, conduct research and fact investigation,
consult with and advise the client, their clients on average voluntarily dismiss between 38% and 42% of their cases.
In 2007, the average dismissal rate for the Unit was 31%.").

the fact that defendants have the right to petition for discretionary appeals, it appears that Justice
Ginsburg's contention in Halbert that providing appointed counsel for petitions for discretionary
appeal may alleviate the workload of the judiciary 33 has some merit. Conversely, Justice
Thomas's implication that the provision of appellate counsel for indigent defendants would force
the judiciary to expend large amounts of resources on frivolous claims 34 appears to have been
mistaken. The Special Unit, perhaps in response to their own budget constraints, appear to be
quite adept at weeding out genuinely frivolous claims. As a third of the claims they do pursue are
granted some relief, it would be difficult to label these claims as a group, "frivolous."
The problem of sentencing errors, and the potential cost savings from error correction, is
not limited to Michigan or jurisdictions that sentence under the guidelines. Sentencing errors
occur in state and federal courts. 35 A 1989 study by the National Center for State Courts (NCSC)
on criminal appellate practices in five different states found that sentencing issues were raised in
one quarter of direct appeals. 36 In these cases, courts found error 25% of the time. Sentencing
errors discovered through appeal include problems with aggravating and mitigating factors that
affect statutory and guidelines ranges, 37 the accuracy of the presentence report, 38 miscalculations
of the guidelines, improper imposition of consecutive as opposed to concurrent terms, and
procedural problems with the sentencing hearing, like the denial of allocution, or the judge's
consideration of legally irrelevant factors in sentencing, like a trial defendant's failure to plead
guilty. 39 Sentencing error is not limited to guideline states. The NCSC study found a 38.5% error
rate for Rhode Island, which has indeterminate sentencing. 40 7.3% of all appeals in the study
were granted a new sentencing hearing. 41
The most obvious incarceration cost savings occur when the convictions of both guilty
and innocent defendants are reversed. Defendants are rarely released through appellate
proceedings. In the NCSC study, only 1.9% of all appeals were straight "acquittals." 42 It is much
more common for appellate courts to vacate the conviction and remand the case to the trial court,
where the prosecutor can attempt a retrial, if she chooses. In the NCSC study, 6.6% of all appeals
were remanded for new trial 43 while in the federal system, 6.3% of appeals disposed on the

33

Halbert at 623, ("While the State has a legitimate interest in reducing the workload of its judiciary, providing
indigents with appellate counsel will yield applications easier to comprehend. . . . when a defendant's case presents
no genuinely arguable issue, appointed counsel may so inform the court.").
34
Halbert at 630-31 ("Today's decision will therefore do no favors for indigent defendants in Michigan--at least,
indigent defendants with nonfrivolous claims. While defendants who admit their guilt will receive more attention,
defendants who maintain their innocence will receive less.").
35
See John, 597 F.3d 263 (5th Cir) (discussing plain error doctrine in the context of miscalculated federal
guidelines).
36
See NCSC at 18-19
37
Id. at 19
38
DVH SADO Congress 3
39
NCSC at 19; insisting on trial and refusing to plead guilty is not a constitutionally valid basis for imposing a
harsher sentence. However, pleading not guilty and then committing perjury by falsely testifying in one's own
defense has been viewed by the Supreme Court as an acceptable basis for sentence enhancement.
40
NCSC at 19.
41
Id. at 5
42
Id. at 5
43
Id. at 5

merits are reversed and vacated. An additional 12.1% of cases were partially affirmed, granted a
new sentencing hearing, or received other relief. 44
How do the cost savings from reduced incarcerations in the relatively few cases in which
some relief is granted compare with the prosecutorial, judicial, and public defender costs of
providing appellate review in the vast majority of cases in which the defendant's conviction and
sentence are upheld? To date, no study appears to have attempted to directly address this
question. However, by making a number of educated guesses based on data from the NCSC
study and other studies on the costs of criminal procedures, including expenditures on
prosecutors and public defenders, 45 it is possible to make back of the envelope calculations, 46 of
the systemic financial impact of direct appeals. See Table 1. The results are somewhat surprising.
I estimate that the total cost to the state of providing appellate representation and adversarial
direct appellate procedure for 1000 cases is approximately $7.5 million, or $7,500 per defendant,
a significant sum. 47 However, the expected incarceration savings from the 20% of cases that
obtain at least some relief is nearly $10 million. 48 Due to the rough nature of these calculations,
we cannot say with any certainty that the total incarceration savings from direct appeals is larger
than the procedural costs of appeals. However, it is safe to say that, contrary to the intuitions of
many, when incarceration cost savings are considered, the net financial costs of providing
counseled access to direct appeals on states is not large, and might actually result in financial
savings. 49

III.

Improving criminal procedure while saving money

The calculations also do not establish that a general expansion of post conviction rights
would yield financial savings to the state. For one thing, these are estimates of the incarceration
savings from direct appeals and subsequent retrials or resentencing hearings only. The
incarceration cost savings that result from collateral review are likely quite small, as very few
44

Id. at 35
Most of my assumptions are detailed in the footnotes to Table 1. One assumption that bears highlighting is the
assumption that the average sentence length for successful appeals is no greater or less than the average sentence
length for appeals filed. Without this assumption, the calculations become much more complicated. This assumption
is somewhat supported by the NCSC study, which found that there was little relationship between sentence length
and the likelihood of success in direct appeals. Defendants with the shortest sentences were more likely to have their
convictions vacated, but defendants facing the longest sentences were most likely to receive a new sentencing
hearing or succeed on only some issues. See NCSC at 39. These relationships were statistically weak. Because there
is no clear, statistically strong relationship between the likelihood of receiving at least some success and sentence
length, this assumption does not obviously bias my estimates.
46
To steal a term from physics, one might call these calculations "Fermi approximations." Fermi approximations are
quick and dirty estimates of quantitative values used when precise values of the relevant variables are not easily
available. So long as the estimates of the relevant variables, like the costs of an appeal, average success rates, and
average sentence reductions per case, are unbiased, the calculated values will usually be in the same ballpark as the
actual values.
47
This figure includes estimated costs of new trials in one third of cases that are vacated and remanded, under the
assumption that most we many cases are not pursued further and estimated costs of resentencing hearings.
48
These figures assume that a year of incarceration costs $30,000.
49
Cf. DVH SADO.
45

defendants, outside of capital cases, obtain any measure of success. More importantly, these
numbers approximate the costs and benefits of direct appeals as they exist today. If an expansion
of post-conviction rights provided appellate review to defendants whose cases were generally
worse on the merits than those that are currently reviewed, states might end up paying more to
adjudicate those cases than they would save in incarceration costs. This might be the case if, for
instance, the additional appeals that would be brought as a result of better funding of appellate
defenders were generally less meritorious than those currently brought. However, these numbers
do suggest that, so long as these marginal cases were not entirely frivolous, the net financial
burden would be significantly less than the increased spending on appellate counsel.
While general expansions of post conviction rights would not necessarily lead to net cost
savings for states, as marginal cases would not necessarily yield the same cost savings as those
currently reviewed, expansions of post conviction procedural rights would clearly come at a
lower financial cost than most realize. Additionally, certain more particularized reforms could be
expected to yield large cost savings to the state. One realm in which the net savings from
incarceration costs might outweigh the costs of increased procedural rights is the probability
standards for new trials based on newly discovered evidence. Under the Berry standard for new
trials based on newly discovered evidence, a new trial will not be ordered if defense counsel
failed to discover the evidence through lack of due diligence, "sandbagged" the new evidence, or
if it is less than probable that the defendant would be acquitted at a new trial. 50 The "probable,"
or "more likely than not" standard prevents a new trial if there is less than a 50% chance that the
defendant would be acquitted. While there may be strong "finality" interests in enforcing this
fairly high standard, the financial burden to the states will rarely be one of them. If, as I assume,
a retrial costs around $20,000 and a year of incarceration costs $30,000, allowing a new trial
when a defendant has even a one in four chance of acquittal would, on average, yield large
savings to the state whenever there are at least 32 months left on the defendant's sentence.
Allowing greater access to new trials based on newly discovered evidence does raise questions
about the effects on deterrence. 51 However, as I discuss in the next two sections, these deterrent
effects may not be as severe as some believe.
In the application of the plain error rule to correct unobjected-to guidelines
miscalculations, the issues of sandbagging and deterrence are much less relevant and the tradeoff between judicial workloads and incarceration costs is stark. There is now some consensus in
the federal circuits that the plain error rule can be used to overcome the forfeiture of an error
when defense counsel fails to object to a judge's clearly erroneous calculation of the guideline
sentence range, so long as correcting the error "would likely significantly reduce the length of
the sentence." 52 However, there is disagreement about how significant the reduction must be in
order to merit resentencing. 53 In John, the Fifth Circuit found that the district court judge, who
had sentenced the defendant to 108 months, had miscalculated the guidelines sentencing range as
97-121 months, rather than the correct 70-87 months. Because defense counsel did not object to
this error at sentencing, the defendant would have lost the right to appeal the error unless the
50

See United States v. Johnson, 327 U.S. 106 (1946).

51
52

See United States v. Meacham, 567 F.3d 1184, 1190 (10th Cir. 2009); FRCP 52(b) ("A plain error that affects
substantial rights may be considered even though it was not brought to the court's attention").
53
See e.g. John, 597 F.3d 263 dissent

miscalculation was a "plain error that affects substantial rights." 54 The Fifth Circuit found that
this plain error "affect[ed] substantial rights" because there was a "reasonable probability" that
defendant would have received a lower sentence if not for the error. 55 However, in an earlier
case, the Fifth Circuit had found that a an incorrectly calculated range of 46 to 57 months did not
affect substantial rights when the defendant had been sentenced to a minimum and the correct
sentencing range was 41 to 51 months. 56 The dissent in John similarly argued that a sentence that
"is only 21 months above the maximum of 87 months in the proper guideline range" does not
"affect substantial rights," and that it would be "unfair[] . . .[to] district judges," who already
have large workloads, to force them to hold a another sentencing hearing when defense counsel
could have caught the mistake. 57 Absent from the dissent, and from the majority opinion, is any
consideration of the burden on the state, the ultimate party of interest, of paying for sentences
that are lengthier than legislatively intended. 58
In the case of post-conviction guidelines miscalculations that increase sentencing ranges,
it is difficult to imagine what strategic advantage defense counsel might gain by "remaining
silent about his objection and belatedly raising the error only if the case does not conclude in his
favor." 59 After all, it is unlikely that an error that raises the recommended sentence would ever
favor the defendant, and any "second bite at the apple" 60 would be before the same judge,
substantially eliminating the risk that defense counsel would use the plain error rule to game the
system. Additionally, while failing to correct sentencing errors that increase sentences might
have a small deterrent effect, by marginally raising the sentence a defendant expects to receive,
these sentences are in excess of what the legislature, through the guidelines, has chosen as
appropriate retribution and deterrence for the criminal acts. Applying the plain error rule in cases
like these would clearly increase judicial workloads, either by making judges hold a resentencing
hearing or by forcing them to take greater care in calculating the sentences in the first place.
However, while it is debatable whether defendants or judges should suffer for mistakes made by
judges and defense counsel, 61 the financial cost of the defendant's excessive incarceration is
borne by the government. In John, even if the district court resentenced the defendant to the
corrected maximum 87 months, the government would still save over $50,000 in incarceration
costs on 21 months of punishment that were not justified by the guidelines. If, as I have
estimated, the total cost of an appeal and resentencing hearing to the government is around
$8,600, allowing the appeal and resentencing hearing in John saved the government over
$40,000. In other words, greater deference to judicial workloads and the interests of finality
would have come at a cost of over $40,000. The fact that neither the majority nor the dissent
54

Id. at 27-28, quoting FRCP 52(b).
Id. at 29-31 (applying the plain error rule described in Puckett, 129 S. Ct. at 1429).
56
See United States v. Jasso, 587 F.3d 706, 713-14 (5th Cir. 2009).
57
See generally John, at 41-42, dissent ("Judges are not like pigs, hunting for truffles . . . . Faced with hundreds of
sentencings, raising thousands of issues, a district judge should be able to rely on counsel, as officers of the court
and zealous advocates, to call arguable error to the court’s attention." Internal quotes removed).
58
Though the guidelines merely set sentencing recommendations that judges can depart from based on the details of
a particular case, it is fair to assume that Congress, by creating the guidelines, intended that the sentence for a
defendant should be anchored around a sentencing range that accurately describes a defendant's crime, and not
anchored around a sentencing range that describes a crime the defendant did not commit.
59
Puckett at 1428
60
John, dissent at 42.
61
Cf. Puckett.
55

made any reference to this financial cost of finality suggests that the court may have under
weighted this factor in their ruling. 62
Incarceration savings from correcting sentencing errors and improperly obtained
convictions are a significant factor that appears to be under appreciated by judges and
lawmakers. This analysis and the Michigan study by SADO demonstrate that, contrary to the
assumption of some, in its current form, providing defendants with represented access to direct
appeals does not impose substantial financial burdens on the state. In fact, some post-conviction
procedures that would improve the accuracy of the legal process would likely produce net cost
savings to the state. More lenient application of the plain error rule for miscalculated guidelines
would impose slightly heavier burdens on the judiciary, but would yield much larger savings to
the state by reducing unnecessary incarceration. The financial expenses of continued erroneous
incarcerations are substantial. However, because judges do not benefit from reduced
incarceration costs but do bear the burden of increased workloads, they tend to undervalue the
importance of the cost savings to the state. In fact, they rarely mention it at all. Deference to
"finality" interests reduces the administrative burdens on the judiciary but increases the financial
burden on the state of continued erroneous incarceration. While courts easily recognize the
"state's interest in finality" as an important policy concern, they fail to acknowledge that finality
imposes a substantial financial cost to the state.

62

It bears noting that a significant number of defendants in Michigan file motions to reconsider sentences in the trial
court, allowing the trial judge to correct sentencing errors without the expense of full proceedings in the higher
appellate courts. See SADO at 6 (the Special Unit's success rate for cases initiated in the trial court is around 85%).
Encouraging a similar procedure in federal courts in instances of truly "plain," non-debatable, error would reduce the
administrative costs of obtaining the incarceration savings while allowing district court judges to correct their
mistakes without having to suffer a reversal of their decisions by an appellate court.

IV. Deterrence
Many have argued that allowing convicted defendants greater opportunities to challenge
their sentences and convictions “undermines the principle of finality” and “deprive[s] [criminal
law] of much of its deterrent effect.” 63 The deterrence argument is that by marginally decreasing
the certainty that criminal behavior will be met with punishment and marginally decreasing the
sentences criminals expect to receive, post-conviction procedure reduces the incentives on
potential criminals to avoid crime. 64 This rational actor analysis has been critiqued and criticized
by studies that demonstrate that these rational incentives often do not have a strong effect on
people's decisions to break the law, 65 that certainty of punishment is generally more important
than severity, 66 and that people's subjective sense of the fairness of the law has a strong influence
on their willingness to obey the law. 67
These deterrence concerns are significantly less problematic in the context of post
conviction review. First, the vast majority of defendants who succeed in appellate review are in
fact incarcerated for a substantial amount of time. Second, appellate review can contribute to the
deterrent effect of criminal law by reducing the number of wrongful convictions, increasing the
benefits of not committing crimes. Third, appellate review gives defendants a voice with which
to air their concerns about the fairness of their procedures and helps ensure that defendants'
sentences and convictions themselves comply with the law, increasing the appearance of
legitimacy in the eyes of defendants. Because an individual's beliefs about the fairness of legal
systems has a substantial effect on their willingness to comply with the law, increased access to
post-conviction remedies, particularly to correct mistakes by criminal "insiders," like
prosecutors, judges, and public defenders, may actually improve the deterrent effect of criminal
law. 68
A.Sentencing Reductions through Appellate Review
Several important factors distinguish the deterrence question for sentence reductions through
post-conviction review from sentence reductions generally. The first is that almost all defendants
who receive some success in post conviction review still receive substantial punishment. Many
studies show that increased certainty of receiving some substantial amount of punishment has a
greater deterrent effect than increasing the severity of the punishment. Post conviction review
takes time. In federal courts, the median time between conviction in trial court and appellate
63

Teague v. Lane, 309.
Cf. Scott & Stuntz, at 1934, (increasing trial rates, as opposed to guilty plea rates, would creat a “loss of certainty
of punishment [that] would lead to a loss of deterrence”); Stephanos Bibas, Transparency and Participation in
Criminal Procedure, 81 N.Y.U. L. Rev. 911, 947 (2006) (describing "[t]he traditional Benthamite" that deterring
would-be criminals requires "making the expected punishment for the crime exceed the expected benefit").
65
See e.g. Robert J. MacCoun, Drugs and the Law: A Psychological Analysis of Drug Prohibition, Psychological
Bulletin 1993. Vol. 113. No.3, 497-512 at 501 (Certainty and severity effects are quite modest in size, generally
accounting for less than 5% of the variance in marijuana use reported in perceptual deterrence surveys).
66
See George Antunes and A. Lee Hunt, The Impact of Certainty and Severity of Punishment on Levels of Crime in
American States: An Extended Analysis, 64 J. Crim. L. & Criminology 486, 489-90 (1973); Scott & Stuntz, Plea
Bargaining as Contract, at 1939 n. 105 (citing studies).
67
See Tom Tyler, Why People Obey the Law, 161.
68
See generally Bibas, Transparency and Participation, (discussing how the opacity of the criminal justice system
damages the legitimacy of the law in the eyes of "outsiders," the general public).
64

ruling for first-time appeals is two years and fewer than 9% of appeals are resolved less than a
year after conviction. 69 This means that for the relatively few defendants are released through
direct appeals, "success" comes only after serving at least a year or two in prison. 70 For the
majority of appellate defendants, success means that instead of serving a sentence that was
arbitrarily lengthy due to a procedural error, they will serve sentences that conform with the law.
These corrected sentences are not arbitrarily short, but are instead in line with the levels of
retribution and deterrence chosen by the legislature. Additionally, less than 1% of all federal
convictions are reversed through appellate procedures and fewer than 3% of convictions receive
appellate relief of any kind. 71 Though the incarceration cost savings from appellate procedure are
quite significant compared to the cost of post-conviction review, successful appeals have a very
small effect on the punishment would be criminals should expect to receive if convicted of their
crime. Additionally, because post-conviction review, when it is successful, reduces, but does not
eliminate punishment, offering more generous post-conviction review would likely have only a
small effect on deterrence.

69

For guidelines cases that were not reopened or remanded. Calculated from Federal Appellate Court Data of all
criminal cases for 2006 and 2007:

70
71

Estimated by comparing the total number of federal convictions for 2004 and 2005 (152,824) with the total
number of cases for 2006 and 2007 that were reversed and vacated (1,438), or reversed and vacated, reversed in part,
or remanded (3,759). From BJS tables.

B. Wrongful Convictions
Under the economic model of deterrence, a rational person will commit a crime when the
expected costs and benefits of committing the crime outweighs the expected costs and benefits of
not committing the crime. 72 The more likely it is that people who commit crimes will be caught
and the harsher the punishment, the less likely they are to commit crimes. At the same time,
when wrongful convictions occur, there is a chance that a person who chooses not to reap the
benefits of committing a crime will be punished for it nonetheless. Higher wrongful conviction
rates decrease deterrence by decreasing the benefits of remaining factually innocent, giving
people less to lose by committing crimes. Some scholars believe that changes in the odds of
being wrongfully convicted can actually have a greater impact on deterrence than changes in the
odds of apprehension. 73 RIZZOLLLI Post-conviction review can help reduce wrongful convictions
by identifying and overturning factually incorrect convictions and exposing practices that
contribute to wrongful convictions in the first place. 74
One important critique of the reduced deterrence effects of wrongful convictions is that
wrongful convictions will only affect the incentives for potential criminals if the criminals are
aware that the risk of punishment without committing a crime exists. 75 If potential criminals
believed that all convictions were accurate, wrongfully convicting an innocent defendant of a
crime that would otherwise go unpunished would appear to raise the likelihood that a person who
commits a crime would in fact be punished. In this case, reversals of convictions as a result of
DNA testing or other newly discovered evidence would in themselves reduce deterrence by
exposing the risk of wrongful conviction. However, potential criminals, like most of America,
are already aware that wrongful convictions occur with some frequency.

72

See generally Png, Ivan P. L., _Optimal Subsidies and Damages in the Presence of Judicial Error,_ International
Review of Law and Economics, 1986, 6 (1), 101_05.; Richard Posner (1999, p. 1484)
73
See e.g. Rizzolli and Stanca, (theorizing that "in the presence of risk aversion, loss aversion, or differential
sensitivity to procedural fairness, [wrongful punishment errors] can have a larger effect on deterrence than [failures
to punish]," and offering experimental evidence to support the claim).
74
IMPORTANT NOTE: Henrik Lando wrote an article in 2006 arguing that under the basic economic model for
wrongful convictions, wrongful convictions of mistaken identity for crimes that factually occurred do not reduce
incentives to refrain from crimes. See Henrik Lando, Conflict or Credibility: Does Wrongful Conviction Lower
Deterrence?, 35 J. Legal Stud. 327 (2006). His basic math checks out, but I haven't been able to look more closely at
it. I have seen a reference to an unpublished article that critiques Lando.
75
See Isaac Ehrlich, The Optimum Enforcement of Laws and the Concept of Justice: a Positive Analysis
International Review of Law and Economics (1982), 2 (3-27) at 17.

Professor Michael Risinger has estimated that in the 1980s, around 3-5% of capital rapemurder convictions were of factually innocent people. 76 If the wrongful conviction rate for
crimes in general is in this ballpark, the rational deterrent effects of wrongful convictions are
most likely small for the vast majority of Americans. Roughly 7,000,000 felony and
misdemeanor convictions occur each year in America, or about one for every 40 Americans aged
15 and older. 77 If even 5% of these convictions are of innocent people, then the average
American would have a one in 800 chance of being wrongfully convicted each year and would
be unlikely to personally know someone who had been wrongfully convicted. Thus, for the
majority of Americans, the effect of wrongful convictions on deterrence are likely small.

76

See D. Michael Risinger, An Empirically Justified Factual Wrongful Conviction Rate, 97 J. Crim. L. &
Criminology 761, 780 (2006). Justice Scalia famously opined in his concurrence to Kansas v. Marsh that the
wrongful conviction rate for felony convictions in America is around .027%. 126 S. Ct. 2516, 2538 (2006)(quoting
an op-ed piece by Joshua Marquis). As Risinger explains in his article, this number has no empirical basis
77
In 2004, were 1,078,900 felony conviction in state courts,
(http://bjs.ojp.usdoj.gov/content/glance/tables/felcovtab.cfm), and 67,464 felony convictions in federal courts. (BJS
Table 4.2, 2004: Federal). In 1998-99 in North Carolina, there were 24,168 felonies and 157,611 misdemeanors
sentenced under North Carolina's structured sentencing, or around six times as many misdemeanors as felonies.
(http://www.ncjrs.gov/App/Publications/abstract.aspx?ID=184300). The current American population is around 3
million people, 80% of whom are 15 or older. Wikipedia.

However, wrongful convictions may have a significant impact on deterrence for people who
have spent time in prison. For one thing, people who already have criminal records are more
likely to be suspected of crimes that occur, which may make the risk of wrongful conviction
more salient for such people. 78 Additionally, people who have spent time in prison are more
likely to have first or second hand reports about wrongful convictions. If the wrongful conviction
rate is even one percent, then any sizable prison facility will contain a number of inmates who
are factually innocent of the crimes for which they were convicted. There will also be many
more who falsely profess their innocence. A prisoner is unlikely to believe every protestation of
innocence may hear in prison. Indeed, studies have shown that prisoners are significantly better
at detecting lies than detectives, customs agents, prison guards, and laypeople. 79 Additionally,
other studies have demonstrated that the more feedback people receive from individual liars, the
better they become at detecting lies from those individuals. 80 Because of this, the perceptions
that former inmates have about their own likelihood of being wrongfully convicted will depend
in large part on the number of factually innocent defendants in prison. Additionally, the
willingness of prisoners to believe stories of innocence will undoubtedly depend on the
plausibility of the innocence stories other inmates present. An "innocent" inmate who claims to
have an ironclad new eyewitness but was denied a new trial will be more credible than the same
inmate who still protests his innocence after a second trial. The innocence claim of an inmate
who offers to pay for the DNA testing he claims will prove his innocence but is refused the
opportunity will be much more credible than the innocence claim of the same inmate after the
DNA test has confirmed his guilt. 81 Post-conviction review helps alleviate the reduced deterrent
effects of wrongful convictions both by reducing the number of actual wrongful convictions and
by reducing the credibility of wrongful conviction claims.
Though wrongful convictions may have little effect on deterrence for most people, they may
significantly reduce deterrence for former inmates. Because former inmates are much more
likely to commit additional crimes than most people, more generous post-conviction review may
actually increase the deterrent value of the law. 82
C. Legitimacy and Compliance
Much of criminal law attempts to reduce crime rates through deterrence: by increasing
the likelihood that people who break the law will be caught and punished. However, studies fear
of punishment is generally not the primary motivator for compliance with the law. For instance,
one study by Robert MacCoun found that people's perceptions of their likelihood of being caught
using marijuana and the severity of punishments they expect explained only 5% of the variance

78
79

Vrij, A. (1996). ‘Lie expert’s beliefs about nonverbal indicators of deception’, Journal of Nonverbal Behavior,
20(1): 65–80.
80
Zuckerman, M., Koestner, R., & Alton, A. Learning to detect deception, Journal of Personality and Social
Psychology, 46, 519-528 (1984).
81
This example is based on Osborne, in which the Supreme Court upheld the right of Alaska to refuse to allow the
defendant to test DNA samples that were already in the state's possession at his own expense.
82
Within three years of release, 67% of former inmates are rearrested and 52% are reincarcerated. See John J.
Gibbons and Nicholas de B. Katzenbach, the Vera Institute, Confronting Confinement (2006).

in their actual use. 83 Studies show that compliance with the law in everyday life is motivated
primarily by an internalized sense that one should obey the law, even when they think the law is
wrong. 84 When people believe they have been treated fairly by legal authorities, they are more
likely to accept the law as legitimate and obey the law even without the fear of reprisal. 85
Studies by Tom Tyler reveal the central importance of procedural fairness in people's
assessments of system legitimacy. 86 Favorable outcomes play a role in people's perceptions of
the fairness of the procedures they receive, but are often not the dominant factor. 87 People are
more likely to feel that they received fair procedure when they "hav[e] an opportunity to present
their arguments . . . and hav[e] their views considered by authorities," are judged by authorities
they believe are motivated to be fair, are treated with respect, and are able to a access some
further procedure to correct mistakes. 88
The vast majority of criminal defendants who appeal do not receive favorable
outcomes. 89 However, even for these defendants, providing meaningful access to post-conviction
review may help legitimize the criminal justice system and increase the likelihood that they will
obey the law in the future. If a defendant, through his lawyer, feels he may have a legitimate
claim, like if the jury was given a biased instruction, or if the defendant has discovered a new
witness he feels is compelling, allowing these claims to be heard on the merits in an appellate
court offers an opportunity to be heard. Conversely, if a defendant is not provided with counsel
to help him prepare an appeal or his claims are dismissed on technicalities without reaching the
merits, he may feel that, even though he was factually guilty, the procedures used to convict and
sentence him were not fair in the first place. 90
The delegitimizing effects of overly strict post-conviction procedures may be especially
acute when the defendant feels that he has been prejudiced by the mistakes or misbehavior of
criminal justice "insiders." 91 Consider the different standards for using the plain error rule to
correct unobjected to sentencing miscalculations discussed in section II. In these cases, due to
technical errors by a judge that were not objected to by defense counsel, who is usually
employed by the government, defendants are sentenced under recommended sentence ranges
longer than the law prescribes. Appellate courts will usually apply the plain error rule to allow a
resentencing hearing when the miscalculation is very large. Doing so gives defendants an
opportunity to voice their concerns, demonstrates a judicial commitment to fairness, and shows
respect for the rights of defendants, helping to legitimize the system in the eyes of defendants.
However, some circuits decline to do so when the miscalculation is not "substantial." In these
83

See MacCoun at 501.
See Tom Tyler, Obeying the Law in America: Procedural Justice and the Sense of Fairness, Issues of Democracy,
July 2001, at 17 (2001), http://usinfo.state.gov/journals/itdhr/0701/ijde/tyler.htm.
85
Cf. Tyler, Why People Obey the Law
86
See e.g. Why People Obey the Law, at 163-65, 106-107.
87
See Why People Obey the Law, at 107; Obeying the Law 19
88
See Why People Obey the Law, at 163-64.
89
See supra
90
See Why People Obey the Law, at 109-112.
91
Bibas uses the terms "insiders" and "outsiders" in discussing how the informational gap between institutional
members like judges, prosecutors, and public defenders and the public makes it difficult for the public, and
especially victims, to understand the criminal justice process or accepted as legitimate. Bibas, Transparency and
Participation
84

cases, defendants end up serving sentences imposed out of compliance with law as a result of
clear mistakes by criminal justice insiders which the justice system refuses to take responsibility
for. It is difficult to imagine that defendants would accept these procedures as fair.
The Supreme Court recently denied application of the plain error rule in a case that might
have even worse delegitimizing effects. In Puckett, the defendant reached a plea agreement with
prosecutors and pled guilty. 92 However, at sentencing, the prosecutors reneged and the defendant
received a significantly longer sentence than agreed upon. Defense counsel made objections at
the sentencing hearing, but failed to explicitly object to the prosecution's violation of the plea
agreement. The Supreme Court ruled that in spite of the fact that the prosecutor's behavior was
most likely unlawful, because defense counsel failed to raise the appropriate objection, the
defendant would have to serve a sentence that was years longer than agreed upon. The defendant
was given the opportunity to voice his concerns but may have been left with the impression,
shared by Justice Souter, that the Court was not motivated primarily to reach fair results. 93 The
interests of judicial efficiency that the Court cited in this case 94 are unlikely to alleviate the
delegitimizing effects of this decision. As a result of the intentional misconduct by the prosecutor
the defendant received a much longer sentence than he agreed to in good faith. When the
defendant raised this abuse by a judicial insider in appellate review, the Court did not allow the
defendant the opportunity for relief, but instead laid the blame on the defendant, for the failure of
his defense counsel, another judicial insider, to prevent the abuses by the Government. Justice
Souter observed on moral grounds that "[r]edressing such fundamentally unfair behavior by the
Government . . . is worth the undoubted risk of allowing a defendant to game the system and the
additional administrative burdens." 95 He might have also added that failing to redress these
abuses may have delegitimized the justice system in the eyes of convicted criminals, making
them less willing to obey the law.
While the failure to provide fair post-conviction processes delegitimizes the justice
system in the eyes of convicted criminals, releasing criminals on "technicalities" and providing
costly appeals to criminals may reduce the legitimacy of the system in the eyes of the public, and
particularly of victims. 96 The relative magnitude of these effects is an empirical question that
merits further research. Citizens who have never been to prison vastly outnumber convicted
criminals in the United States, so that even small changes in the public's sense the legitimacy of
the justice system could conceivably have a greater impact on crime rates than larger changes for
convicts. However, only about 3% of all American adults have ever been in prison 97 while 52%
of former inmates will return to prison within three years of release. 98 It is conceivable that
convicted criminals, whose senses of system legitimacy are clearly very low and would be more
directly affected by reforms, would be more responsive to these changes than most. On the other
hand, many victims who live in high crime areas may also be at high risk of offending and may
be less willing to obey laws that are more lenient to their victimizers. Because the costs of
incarceration have gained increased public salience in recent years and the national mood has
92

129 S. Ct. at 1429.
See Puckett, Souter, dissenting,
94
Id. at 1429
95
Puckett, Souter, dissenting
96
See Why People Obey the Law, at 109-10; see generally Bibas, Transparency and participation.
97
http://www.csmonitor.com/2003/0818/p02s01-usju.html
98
See Confronting Confinement.
93

shifted significantly to a less anti-criminal mentality, 99 it may be that fewer people would be
significantly vexed by post-conviction procedures that are more fair to defendants. Some may
even feel that more generous appellate procedures that improve the legality of sentences at little
net costs to the states increases the legitimacy and fairness of the system. To the extent that
increased legitimacy in the eyes of criminals would be balanced by marginally decreased
willingness to obey laws by the population at large, crime reductions in high crime areas might
be balanced by slight increases in low crime areas, not necessarily a bad tradeoff. However, if
the legitimacy effects for victims who live in high crime areas were large, increased postconviction review might not reduce crime in high crime areas, but would be unlikely to increase
it.
By making the criminal justice system appear more fair to convicted criminals, who
reoffend at alarming rates, providing more meaningful access to post-conviction review and
relief may significantly increase the willingness of criminals to obey the law after release.
V. Conclusion
This paper argues that a significant judicial and legislative bias exists that causes
policymakers to overweight the importance of the administrative burdens of protecting
defendants' post-conviction rights while generally ignoring the financial costs of continued
erroneous incarceration. When the financial savings from correcting erroneously lengthy
sentences and convictions are taken into account, direct appeals do not impose substantial net
costs on states, and may in fact produce cost savings for some states. Offering better
representation and more generous opportunities for review post conviction is unlikely to
significantly reduce the deterrent effect of punishment because releases through post-conviction
procedures generally occur after at least a year of incarceration and sentence corrections have a
small impact on average incarceration length. On the other hand, denial of fair post-conviction
review and relief, particularly with wrongful convictions or when the errors are the result of
mistakes or misbehavior of criminal justice insiders, significantly decreases the legitimacy of the
law in the eyes of convicted criminals. It may be that sentencing reductions or other relief for
criminals based on "technicalities" reduces legitimacy for the public at large, but because post
release criminality rates for convicted criminals are so high, this trade-off may be worth it.
The cognitive bias that causes judges to underweight the financial costs of incarceration
is likely a problem of agency; judges bear the burden of increased judicial review but are not
responsible for and do not see the costs of the punishment they order or refuse to correct.
Increasing the salience of these costs could help correct this bias. States could legislate that
presentence reports that offer sentence length recommendations include estimates of the financial
cost of the recommended punishment. At the same time, appellate counsel arguing about a
mistake in sentencing might mention in their briefs that, in addition to the injustice of punishing
a defendant more severely than legislatively intended, failing to correct the error would likely
cost the state thousands or tens of thousands of dollars in unnecessary punishment. Though such
reforms might be seen as crass attempts by the defense bar to distort the pursuit of justice with
morally irrelevant financial concerns, the fact is that the financial concerns already exist in the
99

See Robert Weisberg, How Sentencing Commissions Turned Out To Be a Good Idea, 12 Berkeley J. Crim. L. 179,
184-85, 199-202.

debates, in the guise of "the state's interest in finality," the costs to the state of preserving
evidence, 100 or the burdens to the state of relitigating convictions that "conformed to thenexisting constitutional standards." 101 The financial costs of ensuring the accurate administration
of justice may or may not be a legitimate factor in determining the procedural rights of
defendants. However, if, as has occurred, the costs to the state are to be considered, it is
important that all the costs be considered, and not just those that are most salient to judges.
Forcing judges to be aware of the full impact to the states of their decisions should lead to
policies that more efficiently utilize the limited resources of the state. To borrow language from
Justice Alito, my point is not to denigrate the importance of the administrative costs of accurate
justice to the state. Instead, my point is that while refusing to correct errors of justice is free to
judges, prosecutors, and the police, it is not free to the state. 102 Though the interests of finality
includes the administrative burdens of reviewing often meritless claims, finality itself comes at a
price.

100

Osborne, Alito concurring 2326-29 (cataloging a number of costs associated with post-conviction DNA testing to
make the point that).
101
Teague v. Lane
102
Cf. Osborne, Alito concurring 2329 ("My point in recounting the burdens that postconviction DNA testing
imposes on the Federal Government and the States is not to denigrate the importance of such testing. Instead, my
point is that requests for postconviction DNA testing are not cost free.").

Table 1
Rough Approximations of Adjudication Costs and Cost Savings for State Direct Appeals
Appeal Outcomes
Outcomes 103 Incarceration Incarceration Incarceration
Cost of
Procedural
Months
Costs Saved
Costs Saved
Procedure per
Costs For
Saved per
per Case 104
For 1000
Case
1000 Cases
Case
Cases
Win nothing
79.4%
0
$0
$0
$6,722 105
$5,337,268
106
107
"Acquittal"
1.9%
48
$120,000
$2,280,000
$7,333
$139,327
Vacated and
6.6%
39 108
$86,250
$5,692,500
$14,333 109
$945,978
Remanded
Resentencing/Other
12.1%
6.3 110
$15,750
$1,905,750
$8,555 111
$1,035,155
Totals
$9,878,250
$7,457,728
103

Total Costs of
Providing
Appeals For
1000 cases
$5,337,268
(-$2,140,673)
(-$4,746,522)
(-$870,595)
(-$2,420,522)

Taken from NCSC 35, Table 3
Assuming one year of incarceration costs $30,000 and ignoring discounting of future costs.
105
A very rough approximation based on numbers from a 2002 study on the costs to the state of death penalty cases compared with murder cases in which the
death penalty was not sought. In Indiana, on direct appeal, defendants are only allowed to raise legal challenges, and not reopen the case or present new evidence.
The costs for direct appeals for felony murder cases in a life sentence was granted and the death penalty was not sought was around $9000, or $11,000 in 2010
dollars. (This figure includes the expenses to the county, which pays for the defense and some other expenses, and expenses to the Attorney General's office). See
Indiana at F-G, T-V.
For this approximation, I assume that the appellate costs for most cases are a third lower than the Indiana costs for felony murder cases in which the death penalty
could have been, but was not sought, or $7,333 per case. In federal appellate cases that "win nothing," a third are dismissed on procedural grounds or deemed
frivolous. 06-07 Federal Data. Because of this, I assume that a third of these cases are disposed of more cheaply, at $5,500 per case.
106
Assuming average sentence of six years and two years are served before acquittal in appellate court.
107
See supra, note 105. I assume that the costs of direct appeals for acquittals is more expensive than those that may have been disposed of on more obvious
grounds.
108
Prosecutors usually decline to retry cases that are vacated and remanded. Assuming that prosecutors do not pursue new trials for 2/3 of these cases and that
half of the cases that are tried again lead to acquittal in six months.
109
For the two thirds of vacated and remanded cases that I assume are not pursued, I assume that the cost of the post-conviction procedure is equal to the costs of
acquittals plus $500 for the added paperwork of declining to pursue a new trial. For the cases that are tried again, the costs would be the costs of the appellate
procedure plus the costs of the new trial. A noncapital murder trial costs around $45,000 in 2010 dollars. See IN, TN, KS, NC studies. Assuming that most trials
cost a third less than this, and assuming that a retrial, (for which most of the investigation, research, and trial preparation have already been done), cost a third
less than a trial in the first instance, the cost of a retrial would be about $20,000.
110
6.3 months was the average sentence reduction for successful plea appeals
111
For cases that are granted a new sentencing hearing, the total procedural costs would be the cost of the appellate procedure plus the cost of the new sentencing
hearing. Assuming, as above, that the direct appeal costs $7,333, I arbitrarily guess that the costs of the new sentencing hearing, for which the main arguments
were already fleshed out in the appellate procedure, is cheap, 1/6 of the cost of the direct appeal, or $1,222.
104