Justice Delayed Report - Death Penalty Appeals

• Topics: Death Penalty/Death Row, Death Penalty

Download original document:
Brief thumbnail

This text is machine-read, and may contain errors. Check the original document to verify accuracy.

The author(s) shown below used Federal funds provided by the U.S.
Department of Justice and prepared the following final report:

Document Title:

Justice Delayed? Time Consumption in Capital
Appeals: A Multistate Study

Author(s):

Barry Latzer, J.D., Ph.D. ; James N.G. Cauthen,
J.D., Ph.D.

Document No.:

217555

Date Received:

March 2007

Award Number:

2004-IJ-CX-0005

This report has not been published by the U.S. Department of Justice.
To provide better customer service, NCJRS has made this Federallyfunded grant final report available electronically in addition to
traditional paper copies.

Opinions or points of view expressed are those
of the author(s) and do not necessarily reflect
the official position or policies of the U.S.
Department of Justice.

justice delayed?

time consumption

in capital appeals:

a multistate study
BY

barry latzer, JD, Phd

james n.g. cauthen, JD, Phd

This document is a research report submitted to the U.S. Department of Justice. This report has not
been published by the Department. Opinions or points of view expressed are those of the author(s)
and do not necessarily reflect the official position or policies of the U.S. Department of Justice.

This project was supported by Award No. 2004-IJ-CX-0005, awarded by the
National Institute of Justice, Office of Justice Programs, United States Department
of Justice. The opinions, findings, and conclusions or recommendations expressed in
this publication are those of the authors and do not necessarily reflect the views of
the Department of Justice.

Abstract
This study focuses on the time taken to process direct appeals of capital cases in
fourteen states. The fourteen states are: Arizona, Florida, Georgia, Kentucky, Missouri,
Nevada, New Jersey, North Carolina, Ohio, South Carolina, Tennessee, Texas, Virginia,
and Washington. For each state, we examined every capital case resolved on direct
appeal by the court of last resort (“COLR”) between January 1, 1992 and December 31,
2002. This generated a database of 1,676 cases.
Measuring from date of death sentence, it took a median 966 days to complete
direct appeal. Petitioning the U.S. Supreme Court added 188 days where certiorari was
denied, and a median 250 days where certiorari was granted and the issues were decided
on the merits. Virginia is the most efficient of all states in the study, with a median
processing time from sentence to COLR ruling of 295 days. Measuring from notice of
appeal to COLR decision, Georgia, at 297 days, is the fastest court of last resort. Ohio,
Tennessee, and Kentucky were the least efficient COLRs, consuming respectively, 1,388,
1,350 and 1,309 days. However, Ohio subsequently reduced its time consumption by 25
percent by eliminating intermediate appeals court review. Median time consumption of
capital appeals from notice of appeal to COLR decision was 921 days, far in excess of
American Bar Association guidelines, which call for 50 percent of all appeals to be
completed in 290 days.

Executive Summary
Despite public controversy over the length of death penalty appeals, little
empirical work has been done on the time allocated to the capital appeals process. The
instant study is one of the few multistate empirical analyses of that aspect of the process.
To ascertain the time consumed by capital appeals and provide explanations for the
differences across states, we selected fourteen representative states with enforceable
death penalty laws and then determined the time allotted to direct appeal in the capital
cases of each state. The fourteen states are: Arizona, Florida, Georgia, Kentucky,
Missouri, Nevada, New Jersey, North Carolina, Ohio, South Carolina, Tennessee, Texas,
Virginia, and Washington. For each state, we examined every capital case resolved on
direct appeal by the court of last resort (“COLR”) between January 1, 1992 and
December 31, 2002. This generated a database of 1,676 cases.
Descriptive Results
Here are the highlights of our results:
• The frequency of COLR decisions by year reflects national homicide trends. The
number of capital appeals resolved annually rose steadily from 1992 to 1997 and declined
thereafter.
• Three-quarters (.737) of the appeals upheld the capital conviction and sentence.
The reversal rate was 26.3 percent. Six out of ten reversals overturned the sentence alone.
In eleven percent of the cases the conviction was overturned.
• Measuring from date of death sentence, it took a median 966 days to complete
direct appeal. Petitioning the U.S. Supreme Court added 188 days where certiorari was
denied, and a median 250 days where certiorari was granted and the issues were decided
on the merits.
• Virginia is the most efficient of all states in the study, with a median processing
time from sentence to COLR ruling of 295 days. Measuring from notice of appeal to
COLR decision, Georgia, at 297 days, is the fastest court of last resort.
• Ohio, Tennessee, and Kentucky were the least efficient COLRs, consuming
respectively, 1,388, 1,350 and 1,309 days. Ohio reduced its time consumption by 25
percent by eliminating intermediate appeals court review.
Explanatory Results
We hypothesized that three broad categories of variables affect the processing
time of capital appeals: 1. case characteristics, 2. institutional arrangements and
resources, and 3. court attributes. We developed a multivariate model to test this
hypothesis. Overall, model results indicate that variation in processing times is driven by

factors in all three categories, but is most significantly influenced by case attributes and
institutional factors. Specifically, we found that:
• For each additional page of the majority opinion (our measure of case
complexity) processing time increased by 1.3 percent.
• For each dissenting opinion filed in an appeal, processing time increased 7.1
percent.
• A reversal of the lower court decision increased processing time by about 7
percent.
• Prior review of an appeal by an intermediate appellate court had a major impact
on processing time, adding about 43 percent to the time necessary to process an appeal
through COLR review.
• The existence of a state law or rule specifically directed to expediting capital
appeals decreased processing time by about 14 percent.
Policy Recommendations
We recommend two general reforms to expedite the direct appeals process: 1.
eliminate intermediate court review, and 2. adopt rules or statutes that impose deadlines
on actors in the capital appeals process. In addition, our research flagged the following
state-specific inefficiencies:
• Georgia defense attorneys take over eleven times the aggregate median to file
notices of appeal in capital cases. Georgia needs to reexamine a statute that provides that
a motion for a new trial tolls the filing deadline for the notice of appeal.
• Kentucky, Washington, and Tennessee take the longest to prepare the record and
appellate briefs. Enforced and meaningful deadlines for trial court personnel, appellate
counsel and prosecutors could help move the process along.
• Ohio, Arizona and Kentucky allocate the most time to the scheduling of oral
arguments before the COLR. The high courts of these states should consider establishing
deadlines through court rules. Alternatively, the state legislatures might adopt overall
deadlines for resolving capital appeals.
• The Texas Court of Criminal Appeals is the slowest court to reach a decision
once oral argument is completed and the briefs have been filed. Court rules or statutes
imposing deadlines could help change this.

Table of Contents
Executive Summary ........................................................................................................3

Acknowledgements .........................................................................................................7

Interviews .......................................................................................................................7

Introduction.....................................................................................................................9

Causes of Protracted Time Consumption in the Capital Appeals Process ..................12

Impact of Protracted Time Consumption in the Capital Appeals Process...................13

Design of the Study .......................................................................................................18

Scope .......................................................................................................................21

Findings .......................................................................................................................23

Outcome of Cases.....................................................................................................23

Distribution of Cases ................................................................................................25

Aggregate Time Consumption ..................................................................................28

The Postsentence Stage.............................................................................................35

The Preparation Stage...............................................................................................37

The Argument Stage.................................................................................................38

The Decision Stage...................................................................................................39

The Supreme Court Stage .........................................................................................41

Explaining Variation in Processing Time.......................................................................41

Independent Variables Included in the Model ...........................................................42

Methods to Investigate Variable Influence ................................................................45

Regression Results: Determinants of Capital Case Processing Times ........................46

Policy Recommendations ..............................................................................................48

Conclusion ....................................................................................................................51

Bibliography .................................................................................................................52

Figures
Fig. 1. Time from Sentence to Execution in the United States........................................10

Fig. 2. Typology of Death Penalty States.......................................................................19

Fig. 3. Capital Reversal Rates on Direct Appeal in Fourteen States, By Year.................24

Fig. 4. Trend in Capital Reversal Rates on Direct Appeal, 1973-2002............................25

Fig. 5. Frequency of Decisions in Direct Appeals of Capital Cases, By State,

1992-2002 ....................................................................................................................26

Fig. 6. Frequency of Death Penalty Decisions by State Supreme Courts, By Year,

1992-2002 .....................................................................................................................27

Fig. 7. U.S. Murder Rates By Year ................................................................................27

Fig. 8. Frequency of Death Penalty Decisions by State Supreme Courts, Related to

Homicide Rates, By Year, 1992-2002............................................................................28

Fig. 9. Median Time (In Days) From Sentence to State Supreme Court Decision, By

State, 1992-2002 ...........................................................................................................30

Fig. 10. Median Time (In Days) From Notice of Appeal To State Supreme Court

Decision, By State, 1992-2002 ......................................................................................31

Fig. 11. Time Consumed (In Days) by Discretionary and Capital Appeals in Five

States in 1993 ................................................................................................................32

Fig. 12. Time Consumed (In Days) by Mandatory, Discretionary and Capital

Appeals in Five States ...................................................................................................33

Fig. 13. Median Time (In Days) From Sentence to Notice of Appeal, By State,

1992-2002 .....................................................................................................................36

Fig. 14. Median Time (In Days) From Notice of Appeal to Last Brief Filed, By State,

1992-2002 .....................................................................................................................38

Fig. 15. Median Time (In Days) From Last Brief to Oral Argument, By State,

1992-2002 .....................................................................................................................39

Fig. 16. Median Time (In Days) From Oral Argument to State Supreme Court

Decision, By State, 1992-2002 ......................................................................................40

Fig. 17. Processing Time of Direct Appeals in Capital Cases, 1992-2002 ......................46

Acknowledgements
We appreciate the guidance and help of Janice Munsterman and Linda Truitt, National
Institute of Justice, Grant Managers for this project.
For their many hours of conscientious labor, we are indebted to Matthew D’Anna and
Julia Wood, our Research Assistants, MA students in Criminal Justice at John Jay
College of Criminal Justice.
We also thank Laura Langer, Professor of Political Science, University of Arizona, for
the use of her dataset on state judges.
Finally, we are grateful to the many state judges, court clerks, prosecutors, defense
attorneys, and other state personnel who graciously provided time, information and
expertise. Without their support, our research efforts would have been unsuccessful.
We especially acknowledge the kind assistance of the following:
Abel Acosta, Chief Deputy Clerk, Texas Court of Criminal Appeals
Lisa Bausch, Capital Case Manager, Washington Supreme Court
Janette Bloom, Clerk, Nevada Supreme Court
Michael Catalano, Appellate Court Clerk, Tennessee Supreme Court
Brenda Deckard, Deputy Clerk, Arizona Supreme Court
Ebby Edwards, Deputy Clerk, Virginia Supreme Court
Karen Gehres, Deputy Clerk, Arizona Supreme Court
David Gormley, Director, Legal Resources Division, Ohio Supreme Court
Boris Moczula, Assistant Attorney General, New Jersey
Bob Northrup, Information Technology Director, North Carolina Supreme Court
Daniel Shearouse, Clerk, South Carolina Supreme Court
Jennifer L. Smith, Associate Deputy Attorney General, Tennessee
Lynn M. Stinchcomb, Chief Deputy Clerk, Georgia Supreme Court
Bill Thompson, Assistant to the Missouri Supreme Court for the Death Penalty
Craig Waters, Director of Public Information, Florida Supreme Court
Paula Yeast, Chief Deputy Clerk, Kentucky Supreme Court

Interviews
Georgia
Susan Boleyn, Senior Assistant Attorney General, March 22, 2005
Beth Burton, Assistant Attorney General, March 22, 2005
The Hon. Norman S. Fletcher, Chief Justice, Georgia Supreme Court, March 23, 2005
The Hon. Leah J. Sears, Justice, Georgia Supreme Court, March 23, 2005

Missouri
Deborah Daniels, Chief Counsel, Criminal Division, Office of the Attorney General,
April 29, 2005
Dan Gralike, Deputy Director for Litigation Services, Office of the State Public
Defender, April 21, 2005
The Hon. William Ray Price, Jr., Justice, Missouri Supreme Court, March 8, 2005
Bill L. Thompson, Assistant to the Supreme Court for the Death Penalty, March 9, 2005
The Hon. Ronnie L. White, Justice, Missouri Supreme Court, March 8, 2005
New Jersey
Robert E. Bonpietro, Deputy Attorney General, Appellate Bureau, Division of Criminal
Justice, May 4, 2005
Mark H. Friedman, Assistant Deputy Public Defender, April 27, 2005
Ohio
David H. Bodiker, Public Defender, April 5, 2005
The Hon. Thomas J. Moyer, Chief Justice, Ohio Supreme Court, April 5, 2005
Timothy Pritchard, Senior Deputy Attorney General, April 5, 2005
The Hon. Evelyn Lundberg Stratton, Justice, Ohio Supreme Court, April 4, 2005
Tennessee
The Hon. Adolpho A. Birch, Jr., Justice, Tennessee Supreme Court, March 3, 2005
Donald E. Dawson, Post-Conviction Public Defender, March 3, 2005
The Hon. Frank Drowata, Chief Justice, Tennessee Supreme Court, March 2, 2005
Michael E. Moore, Solicitor General, March 2, 2005
Jennifer L. Smith, Associate Deputy Attorney General, March 2, 2005
Amy Tarkington, Deputy Attorney General, Criminal Division, March 2, 2005
Texas
Bill Delmore, Chief, Legal Services Bureau, Harris County District Attorney, April 21,
2005
Gary Hart, Esq., Austin, Texas, April 28, 2005
The Hon. Sharon Keller, Presiding Judge, Court of Criminal Appeals, March 16, 2005
The Hon. Lawrence Meyers, Justice, Court of Criminal Appeals, March 16, 2005
Washington
The Hon. Gerry L. Alexander, Chief Justice, Washington Supreme Court, April 27, 2005
J. Scott Blonien, Senior Assistant Attorney General, Chief, Criminal Justice Division,
April 26, 2005
The Hon. Richard P. Guy, Chief Justice (Retired), Washington Supreme Court, April 28,
2005
Joanne Moore, Director, Office of Public Defense, April 27, 2005
Paul D. Weisser, Assistant Attorney General, Criminal Justice Division, April 26, 2005

Justice Delayed? Time Consumption in Capital Appeals:

A Multistate Study
by
Barry Latzer, J.D., Ph.D.
James N.G. Cauthen, J.D., Ph.D.

Introduction
There is a widespread sense that death penalty appeals in the United States take
too long. Nevertheless, the disquiet stemming from this perception has not frequently led
to empirical study. This is one of the few multistate empirical analyses of the capital
appeals process.
This Introduction first sets forth some fundamentals about how the capital appeals
process works. We then will address the causes and effects of the protracted time
consumption. This will segue into the study proper: how it was constructed, its findings,
some causal explanations, and our policy recommendations.
Here is what we already know about overall time consumption in death penalty
appeals. Measuring from sentence to execution, it takes more than twelve years to carry
out fully reviewed and implemented death sentences in the United States.1 A substantial
number of cases, however, have not completed the review process and the sentence has
yet to be carried out. Forty-three percent of those sentenced to death between 1973 and
2005 are still in prison, awaiting the application of the sentence or the resolution of their
appeals.2 At year-end 2005, 339 inmates were under sentence of death for 20 years or
more.3 This is roughly ten percent of the death row population (339/3254 = .104).

Justice Department data for 2005, the most recent available, indicate a 147 month average. 147/12 =
12.25. Bureau of Justice Statistics 2006:11, reprinted here as Fig. 1. This twelve year figure does not
include the average time of the review process in cases in which death sentences were imposed but
ultimately overturned. 39.7 percent of the death sentences imposed between 1973 and 2005 were
overturned through judicial reversals or executive commutations. 2006:16, Appendix table 4. Nor does it
include the numerous cases in which a death sentence was imposed and not (yet) carried out.
2
From 1973 to 2005, 7,662 were sentenced to death, 3,254 of which were still in prison at the end of 2005
(3,254/7,662 = .425). 2006:14, Appendix table 2.
3
Bureau of Justice Statistics 2006:14, Appendix table 2. This was calculated by adding the number of death
row inmates at the end of 2005 who had been sentenced between 1973 and 1985.

Fig. 1. Time from Sentence to Execution in the United States.4

Most of this lengthy post-conviction time allocation is, of course, devoted to the
capital appeals process. That process has three stages: 1. Direct Appeal, which is the
focus of this study, 2. State Postconviction Review, and 3. Federal Habeas Corpus. Each
of these stages is now described in turn.
1. Direct Appeal
Thirty-eight United States jurisdictions – thirty-seven states plus the federal
government – have enforceable death penalty laws.5 Thirty-five of the death penalty
states provide for direct appeal of a capital conviction to the state court of last resort
(“COLR”). By contrast, in noncapital cases, appeals ordinarily must be filed with the
intermediate appeals courts (“IAC”).6 Two death penalty states, Alabama and Tennessee,
provide for direct appeal in capital cases to the IAC, called, in those states, courts of

Table reproduced from Bureau of Justice Statistics 2006:11, table 11.

New York currently has an unenforceable death penalty statute, as its highest court struck down a key

provision. See People v. LaValle, 817 N.E.2d 341 (N.Y. 2004) (holding that a statutory jury deadlock

instruction violates the state constitution and that the defect could only be cured by a revised statute).

6
Ten states have no intermediate appeals courts. In these states, all appeals are lodged with the state court

of last resort.

criminal appeals.7 Ohio, which is, along with Tennessee, included in this study, directed
death penalty appeals to its intermediate appeals court until 1995, at which time it
changed to the more common direct COLR review process.8
The direct appeal process also encompasses United States Supreme Court review
of the federal constitutional issues resolved by the state appellate courts. The party that
was unsuccessful in the highest state court to review the case (either the defendant or the
prosecution) may petition the Supreme Court for a writ of certiorari. If the petition is
granted, which occurs infrequently, the ensuing review by the Supreme Court is
considered part of the direct appeal process. Two-thirds of the cases in our study (.667)
sought Supreme Court review, but only 1.2 percent (n = 20) actually were reviewed by
the Court.
2. State Postconviction Review
Following direct appeal, defendants may file a motion in the trial court for
postconviction review (“PCR”). (In some states, PCR is called “habeas corpus,” but such
state level proceedings should not be confused with federal habeas corpus.) The purpose
of PCR is to permit review of claims that could not be raised on direct appeal, such as
errors that are not manifest from the trial record.9
If the PCR motion is denied by the trial court, the defendant may appeal the denial
to the IAC and then, if necessary, to the COLR, or, in some states, directly to the COLR.
Likewise, if the trial court grants the PCR motion, the prosecution may appeal.
Where the PCR claim involves federal constitutional issues, an adverse ruling by
the highest state court to review the case may be appealed by a petition for a writ of
certiorari to the U.S. Supreme Court. Review is discretionary with the Court. However,
the typical case does not seek Supreme Court review at this point, moving instead to the
third stage of the process.
3. Federal Habeas Corpus
Federal law provides for the writ of habeas corpus for state prisoners, capital or
noncapital, who claim that they are “in custody in violation of the Constitution or laws or
treaties of the United States.”10 The prisoner must first “exhaust” his state remedies, i.e.,
he must first raise (unsuccessfully) the federal claims in the direct appeal or state PCR
process. The petition for the writ of habeas corpus is filed with the U.S. district court in
7

See Ala. Code 1975 § 13A-5-53 and Tenn. Code Ann. § 39-13-206. The Tennessee statute provides the
right of direct appeal from the trial court to the court of criminal appeals with (as of 1992) automatic review
by the Tennessee Supreme Court if the conviction and sentence of death are affirmed by the IAC.
Tennessee, but not Alabama, is included in this study.
8
Art. IV, § 2 of the Ohio Constitution, by amendment adopted in 1994, provides for direct appeal to the
state supreme court in cases in which the death penalty was imposed for crimes committed on or after
January 1, 1995.
9
For example, it might be considered ineffective assistance of counsel if a defense attorney fails to put an
exculpatory witness on the stand, but the existence of such a witness is not likely to be apparent from the
trial transcript.
10
28 U.S.C. § 2241(c)(3). The writ of habeas corpus also is available for federal prisoners.

the federal judicial district in which the inmate is incarcerated. If the writ is granted, the
case may be remanded to the state courts for possible retrial. However, the state may
appeal the issuance of the writ. If the petition is dismissed or denied, the inmate may
appeal to the U.S. court of appeals provided he receives a “certificate of appealability”
from a federal court. Petitioners who are unsuccessful in the federal courts of appeals
may seek discretionary review by the U.S. Supreme Court.
The aggregate number of judicial reviews for all three stages of the capital
appeals process outlined above ranges from six to ten, depending on the state, the
aggressiveness of appellate counsel, and the willingness of the courts to grant review,
where it is discretionary. (This count does not include remands to the state courts for
additional consideration following a reversal.) As we noted, this entire process consumed
an average 12.25 years where the outcome was execution.11
We next address the causes and effects of this protracted time consumption.
Causes of Protracted Time Consumption in the Capital Appeals Process
Several factors appear to contribute to the length of death penalty appeals. A
significant part of the time overhead in death cases surely is attributable to United States
Supreme Court mandates. The Supreme Court has required, as a matter of constitutional
law, extraordinary procedures in capital cases, characterized as “super” due process
(Radin 1980). Some of the prolongation of the capital appeals process is a necessary byproduct of these procedural demands. For instance, bifurcated (two-part) capital trials,
endorsed by the Supreme Court in Gregg v. Georgia,12 are a potent source of appellate
error claims and therefore increase the number of legal issues that must be reviewed on
appeal. The use of juries in the sentencing phase of death penalty trials greatly increases
the incidence of appellate claims and is comparatively uncommon in noncapital cases.13
As discussed below, this study found that 59 percent of capital reversals were due to
errors in the second or sentencing phase of the capital trial.
In addition to these Supreme Court mandates, the states sometimes impose special
procedural requirements for capital cases that prolong the appellate process. For example,
although the U.S. Supreme Court held that so-called comparative proportionality review
is not a constitutional requisite, several state courts of last resort engage in such review
pursuant to state law. 14

Bureau of Justice Statistics 2006:11.
Gregg v. Georgia, 428 U.S. 153 (1976) (upholding Georgia law providing two-part trials in capital
cases).
13
All death penalty states require actual sentencing or sentence recommendations by juries at the
conclusion of the separate penalty phase of a capital trial, whereas only six states provide for general felony
sentencing by juries (King and Noble 2004).
14
Pulley v. Harris, 465 U.S. 37 (1984) established that comparative proportionality review is not mandated
by the Eighth Amendment. Nevertheless, twenty states require it. (Latzer 2001:1161) Proportionality
review requires appeals courts to determine if a death sentence is disproportionate by comparing the case
under review to cases of similar reprehensibility.
12

Third, more claims are raised and more time and effort are expended in defending
and responding to such claims in capital, as opposed to noncapital, appeals (Kozinski and
Gallagher 1995). This undoubtedly reflects the sensitivity of the actors to the life and
death stakes involved.
However, even if conscientious compliance with Supreme Court and state law
directives requires lengthier review, and notwithstanding the high stakes, the capital
appeals process is not irreducible. All state COLRs must conform to constitutional and
state law mandates, but, as we will demonstrate, not all courts take the same amount of
time to do so.
Impact of Protracted Time Consumption in the Capital Appeals Process
The length of the capital appeals process is believed to have significant negative
consequences for the administration of justice. We now discuss eight different effects of
this extended time consumption. The existence of some of these effects, such as delays in
the implementation of death sentences, is self-evident, whereas the occurrence of other
consequences, e.g., a loss of public confidence in the criminal justice system, may be a
matter of dispute.
1. Decreases finality in the administration of justice.

2. Increases corrections costs for maintaining death row inmates.

3. Decreases public confidence in the criminal justice system.

4. Weakens the deterrent benefits of capital punishment.

5. Increases the emotional trauma of the murder victims’ family and friends.

6. Increases the difficulty of conducting legally permitted retrials.

7. Delays the resolution of meritorious capital appeals.

8. Creates a basis for additional litigation.

Below we discuss in the order listed above, and in detail, each of the preceding
effects.
1. Decreases finality in the administration of justice.
The appeals process does not affect non-capital sentences in the same way as
capital sentences. Whereas a death sentence may not be carried out while appeals are
pending, a sentence of imprisonment ordinarily will run its course notwithstanding
appellate review. Because appeals stay, even if temporarily, the implementation of death
sentences, there is a strong incentive to file numerous and even frivolous postconviction
petitions, further prolonging the appellate process. Since no other penological sanction
takes twelve years to implement, and since numerous capital appeals are perceived to be
frivolous, the process has been the subject of sharp criticism. 15
15

Former President Bill Clinton once commented: “For too long, and in too many cases, endless death row
appeals have stood in the way of justice being served” (U.S. President 1996). An ad hoc committee headed
by former Supreme Court Justice Lewis F. Powell, Jr., wrote: “[O]ur present system of multi-layered state
and federal appeal and collateral review has led to piecemeal and repetitious litigation, and years of delay
between sentencing and a judicial resolution as to whether the sentence was permissible under the law. The
resulting lack of finality undermines public confidence in our criminal justice system” (Powell 1989).

Furthermore, although some defense attorneys employ tactics to delay carrying
out death sentences (Alfieri 1996), other actors also may contribute to protracted time
consumption. It has been reported, for example, that in California, “years may pass
before the court reporter, court clerk, trial attorneys, and trial judge review the transcript
and certify it for appellate review” (Aarons 1998:14). Moreover, prosecutors also may
contribute to delay, perhaps because of high caseloads or the use of different
prosecutorial offices for death penalty trials and appeals (Aarons 1998:20). It is also
possible that prosecutors may be discouraged by the lengthy process. The instant study
sheds light on the time consumption by prosecutors, court personnel and other such actors
in the capital appeals process.
2. Increases corrections costs for maintaining death row inmates.
Each year that elapses between a capital sentence and its implementation or
reversal adds thousands of dollars per inmate to the imprisonment costs of the death
penalty. In California, for example, it cost $22,400 to house a prisoner for one year on
death row at San Quentin, compared with $20,760 for non-death row prisoners.16 Since
the average national stay on death row (for those executed) is 12.25 years, the
correctional cost (using the California death row figure) is $274,400 per executed
prisoner (12 x $22,400 = $274,400), $20,090 more per prisoner than the non-death row
cost of imprisonment over a comparable period. Clearly, the lengthier the capital appeals
process, the more expensive the imprisonment. It should be noted, however, that this does
not address the imprisonment costs of the 3,254 death row inmates who have not yet been
executed (Bureau of Justice Statistics 2006:14, App. table 2). Nor does our calculation
include the relative correctional costs of life imprisonment and capital punishment.17
3. Decreases public confidence in the criminal justice system.
Multiple posttrial reviews of death sentences over a period of many years create a
sense of lack of finality in death penalty cases. They contribute to an impression among
the general public that justice is not being served. This is especially likely with highly
publicized crimes and trials that have captured popular attention. (Of course, media
reports of miscarriages of justice also erode public confidence in the criminal justice
system.)
Where it takes many years to carry out a death sentence, there is a risk that the
public will come to believe that the sentence will never be applied. As noted above, fortythree percent of those sentenced to death are still in prison. In some states, death
16

Kozinski and Gallagher (1995:20 n. 64) (1994 data).
The cost of a life sentence is difficult to calculate because of the paucity of data on time actually served
by parolable lifers. As for nonparolable life, the total corrections cost may be estimated as follows. The
median age of murderers at sentencing is 26 (Bureau of Justice Statistics 2001:6), and the typical life
expectancy of newborn American males is 74 years (National Center for Health Statistics 2001), which
suggests an average term of imprisonment of 48 years. (Life expectancy may, however, be shorter for
prisoners because of the high incidence of AIDS and other communicable diseases in prison, and
susceptibility to prison assaults.) Using the California annual non-death-row prison cost of $20,760
(Kozinski and Gallagher 1995:20 n. 64), the cost of imprisoning a single murderer for life is roughly
$996,480 (48 x 20,760 = 996,480), 3.6 times the cost of incarceration in death cases (996,480/274,400 =
3.631).

sentences are imposed but rarely or never effectuated. In 1994, the Ohio Supreme Court
observed that it had affirmed eighty-seven death penalties and not one had been carried
out.18 “That fact,” the justices warned, “creates doubt about the ability of the justice
system to carry out the death penalty and, perhaps even more importantly, a perception
that the entire criminal justice system is not working.”19
It is reasonable to infer that public confidence is undermined by a lengthy and
often ineffectual death penalty system. Approximately two-thirds of the American public
supports the death penalty. 20 An equal number believe that courts in their area do not deal
harshly enough with criminals. 21 Nearly half the American public believes that death row
inmates have too many opportunities to appeal.22 And more recently, a Gallup Poll
revealed that a majority of Americans think that capital punishment is imposed too
infrequently. When asked whether the death penalty was imposed “too often, about the
right amount, or not often enough?,” 21 percent of the respondents said “too often,”
whereas 51 percent said “not often enough.”23 In California, in the 1980s, state supreme
court disapproval of death sentences was a significant factor in a public referendum
leading to the removal of the chief justice (Culver and Wold 1986). There also have been
several initiatives to unseat state supreme court judges because of their votes in capital
cases (Blume and Eisenberg 1999:470-475).
4. Weakens the deterrent benefits of capital punishment.
The extent to which the death penalty deters crime is a matter of great controversy
and is not addressed by this study (Bailey and Peterson 1994; Donohue and Wolfers
2005). We simply assert here that if capital punishment has a deterrent effect over and
above life imprisonment, a lengthy implementation process weakens that effect.
Deterrence is based on an association of the crime with its attendant punishment, and the
passage of time has long been thought to diminish that association.24 Delays in carrying
out death sentences would undermine the deterrent effect of the death penalty by creating
in the minds of potential offenders uncertainty about its imposition. A recent empirical
examination of the relationship between deterrence and the length of time on death row
before execution found that shorter waits on death row increase deterrence (Shepherd
2004). Specifically, the Shepherd study found that one extra murder is deterred for every
18

State v. Steffen, 70 Ohio St.3d 399, 639 N.E.2d 67 (1994).

Ibid. 70 Ohio St.3d at 406, 639 N.E.2d at 73.

20
In 2001, 67% of the American public answered “in favor” when asked if they were in favor of the death

penalty for a person convicted of murder (Bureau of Justice Statistics 2000, table 2.64).

21
In 2000, 68% of the American public chose “not harshly enough” when asked if they thought the courts

in their area deal too harshly, about right, or not harshly enough with criminals. Ibid. table 2.54.

22
The Fox News/Opinion Dynamics Poll of June 28-29, 2000, asked the following: “Admittedly the system

varies from state to state, but in general, do you think that death row inmates are given too much help and

opportunity to appeal their cases and try to prove their innocence, or are inmates not given sufficient help

and opportunity to appeal their cases and prove their innocence?” The response was: 48%-Too much; 19%

Not enough; 23%-About the right amount; 10%-Not sure. N=900 registered voters nationwide. MoE ± 3.

Reported at http://www.pollingreport.com/crime.htm (visited 10/11/05).

23
Gallup Poll, May 8-11, 2006. N=510 adults nationwide. MoE ± 5. Reported at

http://www.pollingreport.com/crime.htm (visited 6/19/06).

24
Beccaria noted that “an immediate punishment is more useful” because the shorter the time between the

crime and the punishment, the stronger and more lasting will be the association (Beccaria 1764 (1963)).

2.75-year reduction in the death row wait before execution. However, a reanalysis of
Shepherd’s data found no evidence that the death penalty provides additional deterrence
(Berk 2005).
5. Increases the emotional trauma of the murder victims’ families and friends.
In all probability, a lengthy appeals process contributes to the frustration, anger
and disappointment of the families and friends of murder victims. The California
Attorney General’s Office warns that delays in the capital appeals process “can make an
already traumatizing experience even more difficult for the victims of these heinous
crimes and their loved ones” (California Attorney General’s Office 2000). Although there
is a dearth of social science evidence on psychological trauma among the families of
murder victims, what little evidence there is indicates that the distress is great and that it
does not lessen over time (Thompson, Norris and Ruback 1998). We also would note that
death penalty cases have a unique characteristic that makes them different from all other
criminal cases: the sentence takes, on average, 12 years to apply, or may, after years of
appeals, never be imposed at all. Although we are not aware of any studies that
specifically address the effect of length of capital appeals on families of victims, we
believe that this unique characteristic is likely to add to their anger, frustration and
distress.
6. Increases the difficulty of conducting legally permitted retrials.
Large numbers of capital convictions or sentences are reversed on appeal and
remanded for retrial or resentencing.25 Generally, the longer the passage of time between
the crime and the reversal and remand, the more difficult it is to reassemble witnesses and
evidence. Consequently, retrial may be difficult or futile. It is not uncommon for
prosecutions to be abandoned when a capital reversal occurs many years after the crime.26
The interests of justice may be thwarted under such circumstances.
7. Delays the resolution of meritorious capital appeals.
While appellate delays are assumed to benefit some capital inmates by staving off
the death sentence, such slowdowns may be harmful to other capital defendants. For
those death-sentenced prisoners with meritorious appeals a lengthy appellate process
extends their time on death row with uncertainty about the ultimate resolution of the case.
This needless anxiety affects at least one-fourth of death row prisoners, as we found that
26 percent of direct appeals end in reversals.
It is sometimes claimed that the lengthy capital postconviction process serves the
interests of justice by providing more time for the discovery of exculpatory evidence.
Data presented by an anti-death penalty organization indicates that since 1973, 123 death
25

A study of death penalty appeals found that from 1973 to 1995, “2349 death verdicts were reversed and
sent back for retrial” (Liebman, et al. 2002:36). Our research found 441 reversals in 1,676 direct appeals
during the years 1992 to 2002 (= .263). We did not tally the number of remands for retrial or track the
results of retrials.
26
The Death Penalty Information Center (2002) found 93 cases from 1976 to the end of 2001 in which
capital prosecutions were terminated following a reversal on appeal. These terminations included many
abandoned reprosecutions.

row prisoners have been “released with evidence of their innocence.”27 We calculate the
median time from trial to release in these cases to be eight years. Consequently, it seems
possible that a radically abbreviated postconviction process would obviate the
presentation of bona fide evidence of nonculpability.
However, evidence of innocence that was not presented at trial usually may not be
considered until the state post-conviction review or federal habeas corpus stages, which
take place after the initial appeal. Consequently, speeding up direct appeals would benefit
wrongfully convicted death row inmates by enabling them to present evidence of
innocence sooner.
8. Creates a basis for additional litigation.
The length of the capital appeals process itself has become a ground for litigation,
potentially adding more time to postconviction review. In four different, though
unsuccessful, certiorari petitions to the United States Supreme Court, one or two of the
justices suggested that prolonged imprisonment under sentence of death could constitute
“cruel and unusual punishment” in violation of the Eighth Amendment.28 However,
another justice opined that “lengthy delay” was an unavoidable concomitant of capital
punishment and should not provide grounds for judicial review.29
Two different theories underlie Eighth Amendment claims based on the length of
imprisonment prior to execution.30 One is that confinement under such circumstances
(decades of great psychological duress combined with the physical and social restrictions
that inhere in life on death row) constitutes cruel and unusual punishment. The other
theory is that execution following such lengthy imprisonment is an inherently excessive
punishment because it no longer serves any legitimate purpose due to the attenuation of
retribution and deterrence over time. So far, these claims have been unsuccessful in state
courts.31 Some of the courts noted that the inmate himself was responsible for extending
27

The Death Penalty Information Center, http://www.deathpenaltyinfo.org (visited Aug. 9, 2006), asserts
that in each of these 123 cases, the conviction was overturned and the death row inmates were “acquitted at
a re-trial or all charges were dismissed”; or, that the inmates were “given an absolute pardon by the
governor based on new evidence of innocence.” We have not independently verified these assertions. An
unsuccessful retrial or a failure to reprosecute may not be indicative of factual innocence as retrials are
often problematic due to the difficulty of finding witnesses after a lapse of many years.
28
See Foster v. Florida, 537 U.S. 990 (2002) (Breyer, J., dissenting from denial of certiorari; petitioner on
death row for 27 years); Knight v. Florida, 528 U.S. 990 (1999) (Breyer, J., dissenting from denial of
certiorari; petitioner Knight on death row for 24½ years; petitioner Moore on death row for 19⅓ years);
Elledge v. Florida, 525 U.S. 944 (1998) (Breyer, J., dissenting from denial of certiorari; petitioner on death
row for over 23 years); Lackey v. Texas, 514 U.S. 1045 (1995) (mem. opin. of Stevens, J. respecting denial
of certiorari; petitioner on death row for 17 years).
29
“Consistency would seem to demand that those who accept our death penalty jurisprudence as a given
also accept the lengthy delay between sentencing and execution as a necessary consequence.” Knight v.
Florida, 528 U.S. at 992 (Thomas, J., concurring in denial of certiorari).
30
These theories are spelled out in the dissenting opinion in Ceja v. Stewart, 134 F.3d 1368, 1376 (9th Cir.
(Ariz.) 1998) (refusing to grant a stay of execution or address a claim that execution after 23 years of
incarceration would violate the Eighth Amendment).
31
E.g., Ex parte Bush, 695 So.2d 138 (Ala. 1997) (defendant's incarceration for 16 years awaiting the
execution of his death sentence was not cruel and unusual punishment); People v. Frye, 18 Cal.4th 894, 77
Cal.Rptr.2d 25 (Cal. 1998) (continued incarceration for seven years following imposition of death penalty

the appeals process and therefore would not be heard to complain that the process was
too long.
Summary
Overall, the protracted capital appeals process presents several serious and costly
problems for the administration of justice in the United States. Undoubtedly, some will
defend the process, claiming that the additional time is needed to ensure that the sentence
is legally justified. Others will say that the negative consequences of lengthy appeals can
only be obviated by the abolition of capital punishment. In assessing whether or not
efforts to shorten the process are advisable, the numerous disadvantages of current
practice should be weighed against the need to ensure a full and fair review of claims of
error appropriate to the seriousness of the sentence. By this standard, we believe that
those sentenced to death can be afforded the extraordinary review to which they are
entitled in a much more efficient manner. As this study shows, some states have
established more efficient appellate procedures while still maintaining guarantees of
fairness and “super due process.”
Design of the Study
This study focuses on the time taken to process direct appeals of capital cases in
fourteen states. These direct appeals are the first stage in the capital appeals process. In
all but two death penalty states, only the state court of last resort hears these appeals.32
Selection of States for Inclusion in the Study
We include in the study decisions from fourteen states that we believe to be
representative of the thirty-seven states that have enforceable death penalty laws. For
state selection we relied initially on the work of Lofquist (2002). Lofquist classified states
in terms of their application of the death penalty as measured by three criteria: the
number of death sentences, the number of reversals, and the number of executions. He
then established six categories of states, which he called Abolitionist, Inactive, Active,
Symbolic, Inefficient, and Aggressive. The figure below depicts Lofquist’s typology
without the Abolitionist states.

was result of need to assure careful review of his conviction and sentence and did not constitute cruel and
unusual punishment); Lucas v. State, 841 So.2d 380 (Fla. 2003) (25 years on death row is not cruel and
unusual punishment as defendant's exercise of his constitutional rights prevented his sentence from being
carried out and he may not now claim that his punishment has been cruel and unusual as a result of his own
actions); State v. Lafferty, 20 P.3d 342 (Utah 2001) (executing defendant after he has spent fourteen years
on death row would not violate his federal constitutional rights against cruel and unusual punishment).
32
As noted previously, if upheld on direct review, these appeals may then proceed to other stages of review
– i.e., state postconviction review and federal habeas corpus review. The time to complete these subsequent
stages is outside the scope of this study.

Fig. 2. Typology of Death Penalty States

Classification

Death Sent.
Rate

Reversal
Rate

Execution
Rate

Inactive

very low

CO, CT, KS, MT,
NH, NM, SD, WY

Active

avg or below

average

avg or below

DE, ID, IN, KY,
MD, NE, NJ, OR,
UT, WA

Symbolic

above avg or
high

avg or below

very low

CA, IL, NV, NY,
OH, PA, TN

Inefficient

high or very
high

high

moderate or
low

AL, AZ, AR, FL,
GA, MS, NC, OK

Aggressive

high or very
high

limited

high

LA, MO, SC, TX,
VA

States

In an effort to allow for generalization of our results, we included in the study
states from each of Lofquist’s categories except the “Abolitionist” and the “Inactive.”
The twelve Abolitionist states (thirteen, if one includes New York) were excluded for the
obvious reason that they do not provide for capital punishment. We also rejected the six
Inactive states because we concluded that they do not impose enough death sentences to
permit fruitful study.
The following states were chosen:

Three Active states – Kentucky, New Jersey and Washington;

Three Symbolic states – Nevada, Ohio and Tennessee;

Four Inefficient states – Arizona, Georgia, Florida and North Carolina; and

Four Aggressive states – Missouri, South Carolina, Texas and Virginia.

To ensure that each selected state provided a sufficient number of death sentences
and executions for meaningful analysis, we examined the Justice Department list of states
that sentenced defendants to death from 1973 to 2000 (Bureau of Justice Statistics
2002:15). Among the states selected, those with the lowest number of death sentences
were Kentucky (76), New Jersey (50) and Washington (37). Each of these offered a
sufficient number of cases to contribute to the study.
Our selections were subjected to two other confirmatory checks. Liebman, et al.,
(2002:367) found a significant relationship between high numbers of capital verdicts
awaiting appeal and low rates of progress in moving capital verdicts through the system,
19

either to approval and execution or to reversal. As processing time is the focus of our
study, we made use of this analysis as well in selecting states. According to Liebman, et
al. (2002:369, n. 788), the five states with the highest capital “backlogs” are California,
Texas, Florida, Pennsylvania and Ohio. Three of these states – Florida, Ohio and Texas –
were chosen for the instant study. 33
The five states with the lowest capital backlogs, according to Liebman, et al.
(2002:369, n. 788), are Nebraska, Montana, Washington, Connecticut and Wyoming,
each with less than one backlogged capital case on average. To ensure that we study
efficient processors of death penalty appeals as well as those that appear to be inefficient,
we included Washington in our study. The remaining states were excluded because of
their low number of death sentences. 34 In addition, two states with moderately low
backlogs according to Liebman, et al. – New Jersey and Kentucky – are a part of our
study. 35 Therefore, we feel confident that our project includes states with both high and
low numbers of capital cases awaiting review.
Justice Department data shed light on efficiency in executing death sentences by a
different measure: the average time under sentence of death of prisoners, broken down by
state (Bureau of Justice Statistics 2002:14). States with the longest incarceration time on
death row (in descending order; years in parentheses) are: Tennessee (10.8), Idaho (10.6),
Illinois (10.4), Indiana (10.4), Utah (10.3) and Florida (10.2). States with the shortest
incarceration time (in ascending order) are: Virginia (2.6), Delaware (4.4), Oregon (5.2),
North Carolina (6.0), Arkansas (6.2), Louisiana (6.4) and South Carolina (6.4). Three of
the short-incarceration states, Virginia, North Carolina and South Carolina, were selected
for our time allotment study, and among the long-incarceration states, Tennessee and
Florida were chosen for analysis.
Finally, the proposed study was able to determine whether the use of intermediate
appeals courts in capital case review affects efficiency. Two states, Alabama and
Tennessee, direct capital appeals to their intermediate courts, and one state, Ohio, did so
for part of the study period. Tennessee and Ohio are on our list.
The fourteen states ultimately selected constitute thirty-eight percent of all states
with enforceable death penalty laws. Each produces enough death sentences to provide a
sufficient base for analysis. There also is a sufficient mix of states with high and low
numbers of appellate reversals and high and low numbers of capital cases awaiting
33

A paper by Blume, Eisenberg and Wells (2004), published after our research had begun, adds additional
confirmation that our selection of states was representative. Measuring death sentences as a percentage of
murders from 1977 to 1999, Blume et al. found that the five most aggressive states in imposing death
sentences were Nevada, Oklahoma, Delaware, Idaho and Arizona. We had selected Nevada and Arizona
for our study. The five least aggressive states were Colorado, Maryland, New Mexico, Washington and
New Jersey. Washington and New Jersey also were selected for our study.
34
Over a twenty-seven year period, each of these states imposed less than one death sentence per year on
average: Nebraska, 26 sentences; Montana, 15 sentences; Connecticut, 8 sentences; and Wyoming, 11
sentences (Bureau of Justice Statistics 2002:15).
35
New Jersey was ranked ninth among the 34 states studied by Liebman and associates (low rank = low
backlog); Kentucky ranked sixteenth (2002:369, n. 788).

review to make meaningful cross-state comparisons of efficiency. Moreover, our states
represent a good geographical distribution as well, with states from the south (Florida,
Georgia, North Carolina, South Carolina and Virginia), the northeast (New Jersey), the
midwest (Missouri and Ohio), the midsouth (Kentucky and Tennessee), the southwest
(Texas and Arizona), and the west coast (Washington). In short, we believe that the
fourteen states listed below are a good representative sampling of death penalty
jurisdictions throughout the United States.
The Fourteen States Selected for the Time Consumption Study
Arizona
North Carolina
Florida

Ohio

Georgia

South Carolina

Kentucky

Tennessee

Missouri

Texas

New Jersey

Virginia

Nevada

Washington

Scope
For each of the fourteen states included in the study, we examined every capital
case decided on direct appeal by the courts of last resort between the dates January 1,
1992, and December 31, 2002.
This time period was selected for three reasons. First, it is long enough to provide
sufficient cases from which to draw generalizable conclusions. Our eleven-year time
frame generated a database of 1,676 cases. A second advantage is the recency of the time
period, which makes our conclusions and policy recommendations more relevant to
contemporary policymaking. Death penalty laws, policies and practices have changed
sufficiently since the early years following the Supreme Court’s reinstatement of the
penalty that generalizations drawn from cases decided in the 1970s and 1980s will have
to be reappraised. Finally, eleven years is long enough to investigate possible trends in
time consumption.
We developed our case database by examining a variety of sources. First, we
turned to the clerk’s office of each state court of last resort. Practically all of the offices
had electronic docket control systems that identified capital appeals. Second, to confirm
completeness, we searched the online legal databases, Westlaw and Lexis, for all years of
the study. In addition, for cases decided through 1995, we reviewed the data on direct
appeal produced by Liebman, et al. (2002). Once all the cases were identified, the clerks’
offices provided, either electronically or though docket sheets, the dates for the
completion of various steps in each appeal. Trial court sentencing dates for each case

were obtained from the state supreme court clerks’ offices, the trial court clerks’ offices,
appellate briefs, or through public information available over the World Wide Web.36
To better understand the time allocated to direct appeal, we modified the approach
used in our study of capital case processing by the New Jersey Supreme Court (Cauthen
and Latzer 2000). There, we had identified three stages in the review process: 1) the
“preparation stage,” covering the period from the filing of the notice of appeal with the
state court of last resort to the filing of the last brief; 2) the “argument stage,” covering
the period from the filing of the last brief to the date of oral argument; and 3) the
“decision stage,” covering the period from the date of oral argument to the date of the
COLR’s decision.
Here, we also collected trial court sentencing dates for each of the cases included
in the study. Because one of the states in our study, Virginia, does not require notice of
appeal in death penalty cases – perhaps because appeals in capital cases are automatic –
we began our measurements with a “postsentence stage,” which runs from the imposition
of sentence by the trial court to the filing of the notice of appeal.
In addition, because we also are interested in the role of the United States
Supreme Court in the direct review process, we determined time allotments for the
“Supreme Court stage,” which runs from the COLR’s ultimate disposition to the ruling of
the U.S. Supreme Court on the petition for the writ of certiorari. We also examined the
time to the Supreme Court’s decision on the merits in those few cases (n = 20) in which
the certiorari petition was granted.
Thus, for each of the 1,676 cases in our multistate database, we collected time
consumption data for each of the following five phases of the direct appeal process:
1. The Postsentence Stage
2. The Preparation Stage
3. The Argument Stage
4. The Decision Stage
5. The Supreme Court Stage
a. Certiorari Denied
b. Review Granted

For each case included in the data we have both a starting date and an ending date. Because we selected
cases for inclusion based on the ending date falling within the study period, there are no cases included in
the data that remained undecided at the end of the study period. Also, as a result of this case selection
approach, the starting date in some cases preceded January 1, 1992, especially in those cases completed
during the early years of the study period. This case selection method excludes those direct appeals that
started but did not end during the study period. We do not believe that this biases the results, however,
given the eleven-year scope of the study, the large number of decisions contained in the database, and the
fact that our data are complete with regard to all of the cases included.

Findings
This study examined 1,676 capital cases from fourteen states. In each case, a
death sentence was imposed and the state court of last resort (“COLR”) subsequently
made a decision on direct appeal, including the resolution of any motions for
reconsideration or rehearing, between January 1, 1992 and December 31, 2002. In twelve
states the appeal went directly to the COLR; in two states, Ohio and Tennessee, the
appeal first went to the intermediate appeals court (“IAC”). (During the study period,
Ohio changed its practice to direct COLR review.) The principal focus of the study was
the time allocated to the resolution of these appeals.
88 percent of the cases in the database were appeals of both the capital conviction
and the sentence. Twelve percent were appeals of sentence only. These latter cases
included cases on direct appeal after remand and resentence, pleas of guilty followed by a
death sentence, and appeals of sentence alone.
Outcome of Cases
Given that the capital appeals process has three stages – direct appeal, state
postconviction review and federal habeas corpus – and that we examined only the first of
these stages, we do not know the ultimate outcomes in these cases, or whether the
defendants actually were executed. (We are examining this in a follow-up study now
underway.)
We do, however, know the results of the direct appeal process. The state courts of
last resort affirmed the lower court’s decision in 73.7 percent of the cases and reversed in
26.3 percent. 59.2 percent of the reversals overturned the sentence alone, leaving the
conviction intact. This tracks our earlier finding of a 61 percent sentence-reversal rate,
although the latter applied to direct appeal as well as postconviction review for a different
time period 1990-1999 (Latzer and Cauthen 2000). In raw numbers: of the 1,676 cases in
our database, 1,235 affirmed the conviction and sentence, 261 reversed the sentence only,
and 180 overturned the conviction.
Our 26 percent reversal rate also is consistent with federal Bureau of Justice
Statistics findings of a 28 percent rate for all death penalty cases (Bureau of Justice
Statistics 2005:14). However, our analysis and the Bureau’s measured different but
overlapping phenomena. Our study included all direct appeals in capital cases for 14
states resolved between 1992 and 2002, whereas BJS figures encompass all death
sentenced cases from 1973 to 2004 and all court reversals, including postconviction
review and federal habeas corpus.
Our reversal rates are considerably at odds with the findings of Liebman, et al. in
A Broken System (2000).37 That study found a 41.2 percent reversal rate on direct appeal,
fifteen percentage points higher than our results. Liebman and associates asserted that the
37

We acknowledge that Liebman, et al. studied a larger sample of states (28 versus our 14), and that the
discrepancy in reversal rates may be attributable to the “missing” states. However, this seems doubtful
since our sample is large and we took pains to ensure that it is representative of all death penalty states.

41.2 percent reversal rate, coupled with the reversals in the two subsequent stages of the
capital appeals process, were indicative of a “broken system.” Our results suggest that
that conclusion was premature.
First of all, the 41.2 percent figure – the reversal rate found by Liebman, et al. – is
not extraordinarily high by historical standards. A now classic study of a century of state
supreme court cases (1870-1970) found an overall reversal rate of 38.5 percent and a rate
of 35.6 percent in criminal cases (Mason 1978:1215, 1217).
Second, and more importantly, Liebman and associates adopted a time frame –
1973-1995 – dominated by early uncertainties in death penalty law. As is well known, in
1976, the U.S. Supreme Court dramatically changed the procedural requirements for
administering the death penalty, and for the next fourteen years or so there was a stream
of Supreme Court cases clarifying or modifying the law (Latzer 2002). It is not
surprising, therefore, that in this period of legal turmoil there would be a high number of
procedural errors in capital trials, and concomitantly, an increased number of reversals on
appeal.
Consistent with the preceding, we also would expect the error rate to decline over
time as death penalty law became more stable and certain. Liebman, et al. offers some
evidence that this occurred. After 1990, reversal rates at the direct appeal stage declined
fairly steadily from 40 percent or more to approximately 30 percent in 1994 and 1995, the
last years of Liebman’s study. This reduction in reversals over time is borne out by our
results. We found reversal rates approximately ten percent higher in the earliest years of
our study as compared with 2001-2002, the last two years.
Fig. 3. Capital Reversal Rates on Direct Appeal in Fourteen States, By Year
Year
1992
1993
1994
1995
1998
1999
2001
2002

Reversal Rate
37.4%
38.3%
28%
21.7%
24.6%
21%
27.9%
28.8%

In short, from 1991 to 2002, there has been a fairly steady drop-off in the number
of capital reversals at the direct appeal stage. The rate declined steadily in the last years
of Liebman’s study and even more thereafter. Figure 4 below vividly illustrates this, with
the red line representing the reversal rates found by Liebman, et al. for 1973-1995, and
the green line representing the reversal rates we found for 1992-2002.

Fig. 4. Trend in Capital Reversal Rates on Direct Appeal, 1973-2002
1.0

REVERSAL RATE

0.0
1973
1975

1977
1979

1981

1985

1983

1989

1987

1991

1993

1997

1995

2001

1999

YEAR
We believe that the most likely explanation for the downward trend in reversal
rate is a reduction in the number of reversible errors over time due to:
1. a reduction in the number and scope of changes in the procedural law
governing the administration of the death penalty, resulting in greater clarity and
certainty in the law;
2. improved handling of capital cases as trial judges and prosecutors gained
experience with them; and
3. greater selectivity in the cases chosen for capital prosecution.
However, these explanations are provisional, as they have not been tested.
Distribution of Cases
As expected, there were marked differences among the states studied in the
number of cases resolved. The top courts of Florida, North Carolina and Texas were far
more burdened with capital appeals than their sister states. More than half the cases in the
study (.540) came from these three states, and Florida alone resolved one-quarter (.252)

of the total. These results are consistent with Bureau of Justice Statistics findings on the
frequency of death sentences by the various states.38
When we examined the frequency of decisions by year we found a fairly steady
upward trend peaking in 1997, followed by a downward movement. This, too, was not
unexpected as it correlates with national homicide trends. After rising in the early 1990s,
the homicide rate in the United States steadily declined from 1995 to 2000. Since there is
(as detailed below) roughly a three year gap between the imposition of a death sentence
and the resolution of the direct appeal, one would expect a reduction in the number of
death penalty appeals starting in 1998, three years after the 1995 dip in the homicide rate,
and a steady decline thereafter.
Fig. 5. Frequency of Decisions in Direct Appeals of Capital Cases, By State, 1992
2002 (N=1676)
500
423

CASES

400

300

260
222

200
137

114

100

0
AZ

STATE

Bureau of Justice Statistics data indicate that the four states that imposed the most death sentences
between 1973 and 2004 were Texas (979), Florida (890), California (828) and North Carolina (511)
(Bureau of Justice Statistics 2005:16). All of these states except California were selected for the instant
study. We excluded California on the ground that the states with a high volume of capital appeals are
adequately represented.

Fig. 6. Frequency of Death Penalty Decisions by State Supreme Courts, By Year,
1992-2002 (N=1676)
250
204
200

FREQUENCY

175

166

154
150

171

164

138

131

133

129
111

100

0
1992

1993

1994

1995

1996

1997

1998

1999

2000

YEAR

Fig. 7. U.S. Murder Rates By Year
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002

9.4
9.8
9.3
9.5
9.0
8.2
7.4
6.8
6.3
5.7
5.5
5.6
5.6

Source: FBI, Uniform Crime Reports, 1990-2002

2001

2002

Fig. 8. Frequency of Death Penalty Decisions by State Supreme Courts, Related to
Homicide Rates, By Year, 1992-2002 (N=1676)

250

10
9
8

FREQUENCY

7
150

6
5

100

4
3

HOMICIDE RATES

200

2
1
2002

2001

2000

1999

1998

1997

1996

1995

1994

1993

0
1992

Aggregate Time Consumption
The main focus of our analysis is on the time allocated to the first stage of the
capital appeals process, direct appeal to the state court of last resort. Measuring from the
date of the death sentence to the decision of the COLR for all fourteen states in the study,
the entire direct appeal process consumed a median 966 days, or 2.65 years (n = 1,676).39
In 68 percent of these cases litigants sought relief in the United States Supreme
Court following the state COLR decision (n = 1,141). This Supreme Court review is
considered part of the direct appeal process. Petitioning the Supreme Court for a writ of
certiorari (the legal mechanism to obtain Court review) added 188 days, or 6.13 months,
to the process where certiorari was denied (n = 1,121). Where certiorari was granted and
the Supreme Court decided the issues on the merits – which occurred only in twenty
cases (1.8 percent) – a median 250 days (8.15 months) were consumed (from state COLR
decision to Supreme Court decision). The total time allocated to direct appeal, measuring
from death sentence to U.S. Supreme Court decision (whether that decision was to deny
certiorari or to resolve the case on the merits) (n = 1,141), was a median of 1,180 days
(3.23 years).
To summarize, capital cases take a median 2.7 years (966 days) to move from the
state trial courts to a COLR decision on direct appeal, and 3.2 years (1,180 days) if we
include U.S. Supreme Court review. This does not include the two subsequent stages of
the capital appeals process, state postconviction review and federal habeas corpus.

One year was defined as 365.25 days; one month as 30.67 days.

Two Time Periods
To measure the efficiency of state courts of last resort it is preferable to start with
the notice of appeal (“NOA”) rather than the sentence as the NOA signals the appellate
court that a case is being sent up. Prior to the filing of the NOA, the state high court is not
directly responsible for the efficient handling of the case. On the other hand, the state
supreme courts usually are not without influence over the Postsentence Stage, i.e., the
period between sentence and NOA. They can, on their own authority, create rules for the
expeditious filing of appeals documents and can impose informal pressures on trial court
personnel to act with greater speed. For instance, the Missouri Supreme Court used to
require monthly reports from trial judges presiding over death penalty cases.40
From a research perspective it appeared to be problematic that one of the study
states, Virginia, does not require a NOA in capital cases. 41 To enable state-to-state
comparisons it was necessary to have uniform measuring posts for the appeals process.
To accommodate Virginia, therefore, we decided to measure two different time periods
for all states: from sentence to COLR ruling, and from NOA to COLR ruling. This dual
measure turned out to be revealing as the sentence-to-COLR analysis provided outcomes
that did not appear in the NOA-to-COLR data. Most noticeably, measuring from the date
of sentence disclosed that Virginia (ironically) is the most efficient of all study states in
the processing of capital appeals. The median time for sentence-to-COLR-decision for all
fourteen states was 966 days; for Virginia it was a mere 295 days, less than one-third of
the aggregate figure. Nevada, with a median of 661 days, was the next most efficient.
In addition to revealing that Virginia was by far the most expeditious in
processing capital appeals, the sentence-to-COLR data also indicate that Ohio,
Tennessee, and Kentucky were the least efficient. It is noteworthy that during our study
time frame, 1992-2002, Ohio and Tennessee were the only states that directed death
penalty appeals to their intermediate appeals courts prior to ultimate COLR review. 42
(Ohio ended this practice for murders occurring in 1995 or thereafter.) Our results
suggest that intermediate review is less efficient than direct appeal to the state supreme
court, and we will examine this hypothesis more closely at a later point in this study.
When we measure time consumption in the more typical way – from NOA to
COLR decision – we get another new insight. Ohio (1,331 days), Kentucky (1,296 days)
and Tennessee (1,182 days) are still among the slowest states, this time joined by
Washington (1,193 days). At the other end of the spectrum, however, Georgia, with a
mere 297 day median, is the fastest. The runner-up is Nevada (639 days). The differences
between the fastest and slowest states were glaring: Georgia took less than one-fourth
(.223) of the time consumed by Ohio – a difference of 2.8 years. The median for all
thirteen states was 921 days or over 2½ years. Using this as a benchmark, the four least
40

Interview with Bill Thompson, Assistant to Missouri Supreme Court for the Death Penalty, March 9,

2005.

41
North Carolina has a “First Acquired” date in its electronic docket information that reflects the date the

court first received any information on the appeal. We used this for the North Carolina notice of appeal

date.

42
Alabama, not included in this study, is the only other state that provides for direct appeal to the

intermediate appeals court in capital cases.

efficient states take an average of 330 additional days, or nearly one extra year, to resolve
capital appeals.
Fig. 9. Median Time (In Days) From Sentence to State Supreme Court Decision, By
State, 1992-2002 (N=1676)
1600

1200

1388

1309

1400
1062

1207

1126
991

1000
DAYS

1350

798

761

750

800

883

825

661

600
400

295

200
0
AZ

NJ NC
STATE

State

114

423

260

137

222

A comparison of the NOA-to-COLR data with the sentence-to-COLR results tells
us something about Georgia’s capital appeals process. Measuring from the imposition of
sentence, Georgia is an above-average performer, ranking fifth among all fourteen states.
But when we measure from the notice of appeal point, Georgia moves up to Number 1,
with 297 days compared to the 921 day total state average. Clearly, Georgia’s
Postsentence Stage, the time between the sentence and the NOA, is relatively inefficient.
Indeed, at 441 days, it is by far the longest such period.
Ohio’s situation is complicated by a change in its capital appeals process during
the time period of our study. Appeals involving pre-1995 murders were first directed to
the intermediate appeals court and then to the Ohio Supreme Court. For murders
occurring in 1995 or thereafter, all capital appeals went directly to the COLR. This gave
us an opportunity to gauge the effect of intermediate court review. Measuring from notice
of appeal to COLR decision, all Ohio cases took 1,331 days, whereas the post-1994 cases
that skipped the Court of Appeals took 1,006 days to complete, a difference of 325
days.43 In other words, Ohio streamlined its capital appeals process by 25 percent (10.6
months) when it switched to direct COLR review. Had we counted the latter cases alone,

For cases subjected to intermediate appellate court review, the notice of appeal date is the date of filing in
the intermediate court.

Ohio’s standing among its sister states would have improved: Ohio would have moved
from 13th place, the slowest state, to eighth place.

DAYS

Fig. 10. Median Time (In Days) From Notice of Appeal To State Supreme Court
Decision, By State, 1992-2002 (N=1607)

1600
1400
1200
1000
800
600
400
200
0

1331

1296
1050

1193

1182
1034

945

859
707

639

718

725

297

STATE

State

114

423

260

137

222

Bar Association Standards
The median time consumption from the notice of appeal to the high court decision
for all thirteen states (excluding Virginia) is 921 days. Is this long? While there are no
benchmarks for efficient review of death penalty cases per se, the American Bar
Association developed time standards to serve as an administrative goal for timely
disposition of appellate cases in general (ABA 1994). The ABA targets do not
differentiate between civil and criminal cases, mandatory and discretionary appeals, or
capital and noncapital cases. The ABA’s markers simply provide that fifty percent of a
state supreme court’s caseload should be resolved within 290 days of the filing of the
notice of appeal, ninety percent within one year of filing, and the remaining ten percent
should be decided “as expeditiously as possible” (1994:100).
We found that the resolution of death penalty cases in our thirteen state database
far exceeded ABA guidelines. At the 50th percentile the capital cases took 921 days,
which is more than three times the ABA recommendation (921/290 = 3.176) for cases at
that percentile. At the 90th percentile, death penalty cases consumed 1555.2 days, or over
four years (1555.2/365.25 = 4.26), instead of the one year suggested by the ABA.
These are, of course, aggregate data that do not indicate the efficiency of
individual states. The work of Roger Hanson and the National Center for State Courts
allowed us to examine performance by selected individual states. In Time on Appeal,

Hanson presented data for one year (1993) for twenty-two state COLRs (Hanson 1996).
This study distinguished between mandatory and discretionary appeals. Death penalty
appeals are not discretionary. However, since the mandatory appeals in most of Hanson’s
states included capital cases, we reproduce his data for the discretionary cases in which
full review was granted. Such cases may be expected to consume more time than
mandatory cases since time is needed to determine whether or not to grant review. We
report only on the states that are also included in our study, provided there were at least
50 cases per state. Bar Association benchmarks are in parentheses.
Fig. 11. Time Consumed (In Days) by Discretionary and Capital Appeals in Five
States in 1993
State

(ABA)

50th Percentile
Discretionary
Capital

90th Percentile
Discretionary Capital

(290)

(365)

517

1,044

721

1,428

290

725

598

1,098

284

1,182

501

1,792

295

295*

345

424*

435

1,193

888

1,430

* Measured from date of sentence.

At the 50th percentile, the differences between capital and noncapital appeals were
pronounced: 2:1 for Arizona, 2.5:1 for North Carolina, 4:1 for Tennessee, and 2.7:1 for
Washington. The differences were nearly as significant at the 90th percentile. Virginia
alone proved equally expeditious with death sentence appeals and appeals not involving a
death sentence. Moreover, Virginia virtually met ABA guidelines at the 50th percentile,
while falling short by only two months at the 90th. Virginia’s feat is all the more
noteworthy since we measured its time allocation from an earlier starting point – the
sentencing date – because, as we noted before, Virginia waives notice of appeal in capital
cases.
After the publication of Time on Appeal, Hanson and the National Center for State
Courts conducted a follow-up study covering the supreme courts of Florida, Georgia,
Minnesota, Ohio and Virginia over a two-year period, 1996-97 (Hanson 2001). This
study measured, among other things, the length of mandatory and discretionary appeals,
including death penalty cases. Once again, we can gain insights into individual state
performance by comparing Hanson’s results for each of these states with the outcomes of
our study. (Minnesota, which has no death penalty, was excluded.)

The table below presents only Hanson’s data, except for the column on the far
right, labeled “Capital 1992-2002,” which provides our findings.
Fig. 12. Time Consumed (In Days) by Mandatory, Discretionary and Capital
Appeals in Five States
50th Percentile
State

Mandatory

Discretionary

Capital
1996-1997

Capital
1992-2002

(ABA)

(290)

141

311

955

946

227

150

255

296

162

409

433*
569†

1443*
1006†

–

270

183**

295††

90th Percentile
State

Mandatory

Discretionary

Capital
1996-1997

Capital
1992-2002

(ABA)

(365)

323

528

1492

1467

384

298

551

703

624

574

558*

2469*

569†

1361†

281**

424††

–

355

* Cases first reviewed by the intermediate appeals court.
†
Cases only reviewed by the court of last resort.
** Measured from filing of trial record.
††
Measured from date of sentence.

First, as to the discrepancies between Hanson’s capital case findings and our
results, it must be kept in mind that Hanson’s study covered a much shorter time period
than ours – two years, as opposed to eleven – and, concomitantly, included far fewer
cases. With Ohio, for example, where the discrepancies are especially glaring, Hanson
examined 27 cases to our 137. Some disparity in outcomes also may be attributable to
methodological differences. For instance, for Virginia cases, Hanson measured from the

date when the record was filed with the state supreme court, whereas we measured from
the date of death sentence.
Nevertheless, it is clear from both Hanson’s work and ours that, with the
exception of Virginia and, to a lesser extent, Georgia, death penalty appeals take a far
longer time to process than all other appeals.
In Florida, at both the 50th and 90th percentiles, Hanson’s results were strikingly
similar to ours. For the former, the Florida Supreme Court took nearly seven times longer
to resolve capital appeals than other mandatory appeals, and three times longer than
discretionary appeals. At the 90th percentile, the Florida court took 4½ times as long with
death penalty cases than other mandatory appeals, and nearly three times longer than with
discretionary noncapital cases.
For Ohio, Hanson’s results and ours are divergent. A complicating factor for Ohio
was its midstream change to direct COLR review (for murders on 1/1/1995 or thereafter)
instead of the previous two-level review process. To eliminate this complication we now
compare only cases that went directly to the Ohio Supreme Court. Hanson’s results (569
days at the 50th percentile) and ours (1,006 days) still diverge considerably, an artifact,
we believe, of the number of cases involved. Using our figures, it takes Ohio’s COLR
more than six times as long to resolve death cases than other mandatory appeals (50th
percentile), two and one-half times as long as the time devoted to discretionary review.
At the 90th percentile the ratios are not as great: 2.18 times as long as the mandatory
cases; 2.4 for the discretionary.
The Georgia Supreme Court’s performance is quite respectable at the 50th
percentile level, but is much less impressive at the 90th, where we found that it took
nearly twice as long to complete a capital as any other mandatory appeal.
And Virginia is, once again, the efficiency leader, requiring, at the 50th percentile,
less than an additional month to complete a capital case. At the 90th percentile Virginia
requires another 69 days.
As for the Bar Association standards, note that nearly all of the noncapital appeals
are resolved within ABA time recommendations, Ohio being the most obvious exception.
For death penalty appeals, however, only Virginia and Georgia, at the 50th percentile,
essentially meet the 290 day benchmark. (Georgia, however, falls woefully short at the
90th percentile.) This demonstrates that even for capital cases it is possible to have full
appellate review and reasonable efficiency.
Summary of Findings
• Three state courts of last resort – Florida, North Carolina and Texas – provided
more than half of the capital appeals in this study, reflecting the number of death
sentences imposed in those states.

• The frequency of state supreme court decisions by year reflects national
homicide trends plus three years, the time it takes for appellate review. The number of
capital appeals resolved annually rose steadily from 1992 to 1997 and declined thereafter.
• Nearly three-quarters of the appeals (.737) upheld the capital conviction and
sentence. The reversal rate was 26.3 percent, fifteen percent lower than in the aftermath
of the Supreme Court’s 1976 death penalty cases, when the law was in flux. Six out of
ten reversals overturned the sentence alone, leaving the conviction intact. In eleven
percent of the cases the conviction was overturned. Ordinarily, reversals were followed
by retrial or resentencing hearings.
• Measuring from date of death sentence, it took a median 966 days (2.65 years)
to complete the direct appeal process in the state courts. Petitioning the U.S. Supreme
Court added 188 days (6.13 months) to the process where certiorari was denied, and a
median 250 days (8.15 months) where certiorari was granted and the Court decided the
issues on the merits.
• Virginia is the most efficient of all states in the study, with a median processing
time from sentence to COLR ruling of 295 days (less than ten months). Measuring from
notice of appeal to COLR decision, Georgia, at 297 days, is the fastest court of last resort.
However, Georgia is the slowest state in the postsentence stage, and this reduces its
overall efficiency when measured from sentence to COLR ruling.
• Ohio, Tennessee, and Kentucky were the least efficient state supreme courts,
consuming respectively, 1,388, 1,350 and 1,309 days (= over 3.6 years). However, Ohio
reduced its time consumption by 25 percent by eliminating intermediate appeals court
review.
• Median time consumption of capital appeals from NOA to COLR decision was
921 days, far in excess of American Bar Association guidelines, which call for 50 percent
of all appeals to be completed in 290 days. The ABA benchmark for cases at the 90th
percentile is one year; capital appeals at the 90th percentile require 4.26 years.
The next section of this study will consider the length of each of the stages of the
appeals process. We divided the direct appeal process into five stages: 1. the Postsentence
Stage; 2. the Preparation Stage; 3. the Argument Stage; 4. the Decision Stage; and 5. the
Supreme Court Stage. Close analysis of each stage will provide additional insights into
the factors contributing most to inefficiencies.
The Postsentence Stage
We refer to the period between the imposition of the death sentence and the filing
of the Notice of Appeal (NOA) as the Postsentence Stage.44 Ordinarily, the only activity

For Ohio and Tennessee appeals to the intermediate appellate court, the NOA date is the date of filing in
that court. For all other Ohio and Tennessee appeals, the NOA date is the date of filing in the COLR. In ten
cases in the dataset, excluded for analysis of this stage, the notice of appeal was filed before the imposition

involved in this phase of the process is the preparation and filing of the NOA by appellate
counsel, a fairly simple exercise. Most states impose a short filing deadline for notices of
appeal. For instance, Ohio requires a filing within 45 days of the entry of the trial
judgment.45 However, some states do not use the term “notice of appeal.” North Carolina
has a “First Acquired” date in its docket system, reflecting the initial receipt by the state
supreme court of information on a case, and which could pre-date the filing of any notice
of appeal. And Virginia has no notice of appeal or functional equivalent for death penalty
cases.
For thirteen of the states studied (not including Virginia), the median length of the
Postsentence Stage is 39 days. However, a breakdown by state reveals some interesting
variations.
Fig. 13. Median Time (In Days) From Sentence to Notice of Appeal, By State, 1992
2002 (N=1597)
500
441
400

DAYS

300

200
145
85

100
32
7

47
24

0
AZ

STATE

State

114

423

260

137

222

Georgia takes a median 441 days (= 1.2 years) from sentence to NOA, over

eleven times the aggregate median 39 days. The explanation is that Georgia law provides

that a motion for a new trial, which is entered in the trial court, suspends the 30 day filing

of the sentence by the trial court. The exclusion of these ten cases had no significant impact on the findings

for this stage.

45
Rule XIX, Rules of Practice of the Ohio Supreme Court.

deadline for the NOA.46 Such motions may not be disposed of for a considerable period
of time. Representatives from the Georgia Attorney General’s Office confirmed that this
has created a “bottleneck” in the appeals process.47 Here is an especially egregious
example. On March 30, 1991, William C. Sallie was sentenced to death for murder. His
lawyer filed a motion for new trial on April 26, 1991. Two and one-half years later, on
November 1, 1993, the motion was amended. Nearly four years after the amended motion
was filed, on July 28, 1997, following a hearing, the motion for new trial was denied. A
NOA was then filed less than 30 days afterward, on Aug 20, 1997. The total time from
sentence to notice was almost six and one-half years. We strongly recommend that the
Georgia Supreme Court develop rules to expedite the resolution of motions for a new
trial, especially as such motions are rarely granted and serve little purpose other than to
delay the appellate process.48
It is difficult to explain the length of the Postsentence Stage in Tennessee.
Tennessee continues to route death penalty cases to its intermediate appeals court, and we
used as the notice of appeal date the date of filing in the intermediate court. We know of
no reason, however, why IAC review should cause delay at this stage.
The Preparation Stage
This phase of the appeals process runs from the NOA to the last brief filed with
the COLR. After the notice of appeal is filed the preparation of appellate documents
begins in earnest. This includes the creation of the trial transcript by trial court personnel
and the preparation of the briefs by appellate counsel. Because death penalty trials are
usually lengthier than noncapital proceedings, the trial transcript may be expected to take
longer to prepare. Likewise, because the legal claims asserted in death cases often are
more numerous and complex, the briefs are likely to be longer and require more research
and writing time.
The chart below shows the time allocated by each state to the Preparation Stage.

Ga. Code Ann. § 5-6-38(a).
Interview with Senior Assistant Attorney General Susan Boleyn and Assistant Attorney General Beth
Burton, March 22, 2005. Boleyn observed that “sometimes the cases languish at the motion for new trial
stage. You just can’t get that transcript filed or you can’t get the judges to rule because once that thing’s
over [the capital trial] nobody has incentive to get them out.”
48
We recognize that Georgia could simply shift the transcript preparation to a later point in the process.
However, filing the NOA informs the state supreme court that the appeals process has been set into motion
and therefore may be expected to generate on all actors pressure from “above” to perform their duties. By
contrast, failing to file a NOA allows delays at the trial level to fly below the radar. We therefore think it
advisable to amend Ga. Code Ann. § 5-6-38(a) to eliminate suspension of the NOA filing.
47

Fig. 14. Median Time (In Days) From Notice of Appeal to Last Brief Filed, By State,
1992-2002 (N=1607)
1000
824

DAYS

600

771

765

800

504

573

529

560

514

501

415
400

289
204

167

200
0
AZ

STATE

State

114

423

260

137

222

Kentucky takes the longest to prepare: 824 days, or 2.3 years. We believe a
significant factor causing this delay may be multiple extensions of time granted to the
parties to complete steps in the appeal. Washington (771 days) and Tennessee (765 days)
are not far behind Kentucky; each takes over two years.
Tennessee’s high score is puzzling since death penalty appeals in that state first
go to the intermediate appeals court. Consequently, when the case advances to the state
supreme court, trial documents and much of the substance of the appellate briefs will
have already been produced, and one would not expect a need for much additional
preparation.
Georgia takes the least time to prepare appellate documents, just 167 days (= 5.5
months), but this may be due to a shifting of the transcript preparation to the earlier
Postsentence Stage, which, as we saw, takes an inordinately long time in Georgia.
Virginia is, once again, not included in the list because of the lack of a notice of
appeal. However, when we measure sentence to last brief, Virginia takes the laurels for
speed, consuming just 177 days (under six months).
The Argument Stage
Once the briefs have been filed, the prosecuting and defense attorneys will appear
for oral argument before the state’s highest court. This should occur rather quickly, the

principal considerations being the court’s schedule and the availability and preparation of
the lawyers.
Occasionally, oral argument will not be held (n = 9), or briefs will be
supplemented after oral argument (n = 120). These cases were not included in our time
calculations.49
The Argument Stage takes a median of 64 days, and Georgia accomplishes it in
about two weeks. There are, however, some perplexingly slow jurisdictions. Ohio
requires 256 days (over eight months); Arizona, 182 days (six months); and Kentucky,
125 days (almost four months). We could find no good reason for such lengthy time
consumption.
Fig. 15. Median Time (In Days) From Last Brief to Oral Argument, By State, 1992
2002 (N=1547)
350
300
256

DAYS

250
200

182

150

125
83

100
50

81
50

54
29

0
AZ

STATE

State

114

423

260

137

222

The 120 cases in which supplemental briefs were filed after oral argument come from nine states, with
the most cases coming from Georgia (47), Texas (27) and Florida (18). The median number of days after
oral argument when these 120 briefs were filed was 55. If all of these 120 cases were included in the data,
the median number of days between last brief and oral argument would be 59 days instead of 64, not a
significant change in findings.
The Virginia Supreme Court docket system did not include the exact date of oral argument; it only
included the term during which oral argument was held. The court meets for five-day terms at regular
intervals throughout the year. We used as the oral argument date the midpoint (i.e., the third day) of each
term.

The Decision Stage
Once oral argument is complete (and the supplemental briefs, if any, have been
filed), the resolution of the case becomes the sole responsibility of the state supreme
court justices. They must debate and decide the issues among themselves, and prepare
and present written opinions explaining their positions. The median time consumption per
case for this activity is 206 days, or just under seven months.
Fig. 16. Median Time (In Days) From Oral Argument to State Supreme Court
Decision, By State, 1992-2002 (N=1667)
600
500

459

DAYS

400
300

258

200

310

279

274

150

261

260

247
143

123

155

186
96

100
0
AZ

STATE

State

114

423

260

137

222

The Virginia Supreme Court works the fastest, producing opinions in little more
than three months. Missouri’s highest tribunal takes four. Somewhat surprisingly, the
Texas Court of Criminal Appeals is the slowest court, requiring fifteen months to
complete a capital case. Washington State’s court is the second slowest at over 10
months. In a subsequent section of this study we will offer explanations for some of these
results.
The time consumption of the New Jersey Supreme Court would have been greater
had we included its decisions on what is called “proportionality review.” Proportionality
review, which purports to determine evenhandedness in the application of the death
penalty, is a statutorily mandated proceeding in New Jersey, not a part of the direct
appeal process.50 Therefore, we did not include it in our calculations.

In re Proportionality Review Project, 161 N.J. 71, 735 A.2d 528 (1999), directed that proportionality
review be conducted along with direct appeal. However, proportionality review remains a legally separate
process.

The Supreme Court Stage
As was already stated, United States Supreme Court review of the direct appeals
decisions of the state courts is itself a part of the direct appeal process. In 1,141 of the
1,676 cases in our database (.681), the litigants sought relief from the U.S. Supreme
Court. Of the 1,141 cases seeking Supreme Court relief, the Court declined review in
1,121, or 98 percent (.982). In twenty cases, 1.8 percent, certiorari was granted and the
Supreme Court decided the case on the merits. This is slightly higher than the ratio of
certiorari grants for Supreme Court cases generally.51 Nevertheless, it is clear that a
miniscule number of death penalty cases are fully reviewed by the Supreme Court.
Cases in which Supreme Court review was denied took a median 188 days, just
over six months, from COLR decision to denial of certiorari. The twenty cases in which
certiorari was granted added 250 days (= eight months) to the process.
In short, U.S. Supreme Court review adds one-half year to the direct appeal
timeline, eight months in those exceptional cases in which there is a decision on the
merits.
Explaining Variation in Processing Time
The findings above demonstrate that there is significant variation in processing
time across states and cases. However, these results do not identify factors that may
explain this variation. Understanding why one case takes longer to process than another
may aid legislatures and courts in improving the efficiency of capital appeals processing.
To investigate possible explanations, therefore, we developed a multivariate model,
allowing us to determine the influence of individual factors on processing time.
Our unit of analysis is the individual case that proceeds through direct appeal. We
seek to explain processing time from the filing of the notice of appeal (start of direct
appeal) to the final decision of the state court of last resort, including any rulings by that
court on reconsideration or rehearing motions (end of direct appeal). 52 Many of these
appeals proceed to other levels of review, viz., state and federal habeas corpus, but our
focus here is to explain variation in processing times during this first stage of review. For
the dependent variable in our model, i.e., the variable whose change across cases we are
seeking to explain, we used the log transformation of the total number of days from
notice of appeal to decision.53

In the Supreme Court’s 2004 Term, for example (October 2004 to July 2005), the Court considered 7,542
petitions, granting review in 80, or 1.1 percent. Harvard Law Review 119:426 (2005).
52
As noted earlier, there is not a formal notice of appeal filed in Virginia; therefore, we used the trial court
sentencing date as the start date for Virginia cases.
53
We used a log transformation instead of the actual number of days for two reasons. First, using the log of
total days aids in model interpretation, as the individual effect of the independent variables on the variation
in the dependent variable can be addressed in percentage terms (Bonneau 2005). Second, using the log
transformation may avoid problems with heteroscedasticity, or non-constant variance of the residuals from
the predicted values of the model, an effect that causes model estimates to be inefficient (Fox 1991).

Independent Variables Included in the Model
Based on existing case processing scholarship, we hypothesized that three broad
categories of variables affect the processing time of capital appeals: 1. case
characteristics, 2. institutional arrangements and resources, and 3. court attributes. We
identified specific variables to include in our model for each of these categories, and now
discuss each of the three categories in turn.
Case Characteristics
Previous research identified case characteristics as a possible factor affecting case
processing time before state appellate courts (Hanson 1996; Chapper and Hanson 1990).
We expected the same to be true with capital appeals before state high courts. Not all
capital appeals brought to state courts of last resort are the same, of course, and
differences across these cases may affect the time needed to resolve the appeals. For
example, the issues that come to the courts may vary based on whether the appeal seeks
review of both the conviction and sentence or review of the sentence only. Capital trials
are bifurcated into guilt determination proceedings and sentence determination
proceedings. An appeal of the conviction and sentence – in essence, a review of alleged
errors in two separate proceedings – may present the appellate court with more issues
than an appeal of sentence only, necessitating more time to address and resolve issues.
Accordingly, we expected that a state court of last resort would take longer to process a
case when the appeal is of conviction and sentence than of the sentence alone. We
therefore included in the model a Scope of Review variable, coded “0” for review of
sentence only and “1” for review of conviction and sentence.
Regardless of the scope of review, there may be significant variation among cases
in the number and complexity of issues presented to the COLR. These differences also
may affect processing time. One might expect that it takes longer for courts to resolve
cases presenting more issues, particularly when these issues are of first impression before
the court (Martin and Prescott 1981). In our previous study of capital cases before the
New Jersey Supreme Court, we found that opinion length, which we used as a measure of
case complexity, helped explain the time needed to resolve capital and non-capital
criminal appeals (Cauthen and Latzer 2000). Similarly, in his study of civil appeals
before a single state high court, Hanson (1996) found that opinion length influenced time
on appeal. Thus, we included in the model a variable for Case Complexity, measuring this
variable for each case by the number of pages in the court’s majority opinion as reported
in the West Publishing Company’s regional reporter for that state.54
The outcome of a capital appeal also may affect processing time. On one hand, we
might expect that a court, when upholding rather than reversing a conviction and/or death
sentence, would present a more detailed and reasoned opinion, adding to processing time.
On the other hand, a decision overturning a jury’s determination might take longer to
process, as the court may feel the need to address the legal issues in more detail,
particularly given the high level of public support for capital punishment. Here, we

In calculating the number of pages for each opinion, we excluded the pages devoted to the syllabus of the
decision or the headnotes. Each part of a page was counted as one full page.

hypothesized that cases in which the sentence and/or conviction is overturned (coded “1”)
will take longer to resolve than affirmances (coded “0”) (Outcome).
There is one final case characteristic that we investigated in the model.
Oftentimes, decisions in capital cases are nonunanimous. Indeed, in our data, 27 percent
of the decisions included one or more dissenting opinions. If there were disagreement on
the court, one would expect case processing time to increase, as the dissenting justices
devote time to the drafting of dissenting opinions, and as such opinions usually are
circulated for review among the justices. As a result, we hypothesized that the more
disagreement there is in a case, the more time it will take the court to reach a resolution
of the appeal. We measured disagreement by the total number of dissenting opinions filed
in the case (Disagreement).
Institutional Arrangements and Resources
Beyond case characteristics, we hypothesized that institutional arrangements and
resources contribute to the amount of time it takes for a state court of last resort to
process a capital case. All states have specific rules for addressing appeals to the state
COLR and previous research has found that variation in these rules may affect case
processing time (e.g., Martin and Prescott 1981). Some states have enacted laws or court
rules intended to speed up the decision-making process, and given the history of long,
drawn-out appeals, some of these are designed specifically to increase efficiency in
capital appeals processing. These include rules requiring that decisions in capital appeals
be reached within a stated time period,55 rules giving docket preference to capital
appeals,56 and rules ensuring that the trial court record in capital cases is transcribed in a
timely fashion. 57 As a result, we included in the model a dichotomous variable measuring
whether, at the time the appeal was filed, there was any law or rule in the state
specifically directed to efficient processing of capital appeals (not appeals generally). We
expected that the presence of such laws or rules (Law/Rule) would reduce the time
devoted by the state high court to processing capital appeals. 58
In addition to laws and procedural rules, court resources may impact time on
appeal. Much of the existing research on case processing before state appellate courts
focuses on court resources, including caseload pressures, as a major factor explaining
variation in time to decision (Hanson 1996, 1998; Hoffman and Mahoney 2002). Large
capital caseloads may have a negative impact on the ability of the court to process those
cases efficiently given the extraordinary time needed for those appeals. Consequently,
we expected that courts with larger capital caseloads would take longer to process capital
appeals than courts with fewer such cases on the docket. We measured caseload by the
number of capital appeals decided by that court during the same calendar year as the unit
of analysis (Capital Caseload).
55

E.g., Nev. Rev. Stat. § 177.267 (decision in appeal from sentence of death must be reached within 150

days from receipt of the record).

56
E.g., Va. Supreme Court Rule 5:23 (giving docket preference to reviews of death sentences).

57
E.g., Ky. Rule of Civ. Proc. 75.01(4) (sets guidelines for preparation of record in capital appeals with

penalties imposed on court reporter for failure to comply).

58
Our research uncovered such laws or rules for some or all of the years studied in the following states:

New Jersey, North Carolina, Tennessee, Virginia, Florida, Kentucky, Nevada, and Arizona.

Given the complexity of legal issues arising in capital cases, legal resources
available to the judges also could affect court efficiency. These resources can come in
many different forms, including staff available to perform legal research and assist in the
resolution of motions, and clerks to help with the writing of opinions (Hanson 1996).
Thus, we anticipated that the larger the number of staff attorneys on the court, the shorter
the time needed to process capital appeals. However, because these staff attorneys may
have to handle both capital and non-capital cases, their availability to work on death
penalty appeals may be affected by overall caseload on the court. Consequently, we
measured these legal resources by the number of staff attorneys per 100 cases on the
court’s docket (Staff Attorneys). Data on the number of staff attorneys on each court as
well as overall caseload were obtained from the annual editions of State Court Statistics,
published by the National Center for State Courts (1985-2001; 2002).59
The final institutional variable controls for the effect of intermediate court review
of death penalty cases prior to final review by the state court of last resort. For all of the
years covered by our data, direct appeals in Tennessee death penalty cases proceeded first
to the state’s intermediate appellate court and then, if affirmed, to the state court of last
resort. For a portion of our study period a similar process existed in Ohio.60 For the cases
from these states that were subjected to intermediate appellate court review, the notice of
appeal date used to compute the value of the dependent variable was the date of filing in
the intermediate court. Thus, we included a variable in the model (IAC Review) expecting
that appeals subjected to review by both the intermediate and the state high court would
take longer to process than cases reviewed by the state court of last resort only.
Court Attributes
As addressed above, the filing of dissenting opinions in a case may evidence
conflict among the state high court justices in death penalty appeals. However, conflict
may exist even without the filing of dissenting opinions. If justices hold differing
positions on issues raised by the cases, it may be more difficult to form a majority. In
fashioning a majority opinion, the opinion writer must ensure that a sufficient number of
differing views are satisfied in order to maintain majority support for the opinion.
Many studies addressing individual decision-making by state high court judges in
capital appeals have found that the ideology of a justice influences the decision (e.g.,
Brace and Hall 1997; Traut and Emmert 1998). We hypothesized that these ideological
views also may impact case processing time in that a court with more uniform ideological
views will take less time to prepare a majority opinion than a court with conflicting
ideological views. Previous research developed party-adjusted ideological scores for
individual justices on state courts of last resort, taking into account partisan affiliation
and ideology at the time of the justice’s ascension to the bench (Brace, Langer and Hall
59

This measure does not differentiate between lawyers who are on the court’s central staff and lawyers who
are assigned to specific justices. Our measure determined whether the increase in lawyers on staff in any
capacity affects processing time in capital appeals.
60
As explained earlier, during our study period Ohio abandoned intermediate review. In capital crimes
occurring on 1/1/1995 or thereafter, defendant may appeal as of right from the trial court directly to the
Ohio Supreme Court. For capital crimes occurring earlier, defendant could appeal to the Ohio Court of
Appeals, and if the death penalty was affirmed, could further appeal as of right to the state supreme court.

2000). One can identify ideological conflict on a tribunal by using the standard deviation
of these ideology scores for justices on the natural court per court year; the higher the
standard deviation, the more ideological conflict on the court.61 Therefore, we included in
the model a variable representing ideological conflict on the court at the time of the
decision (Ideological Conflict), expecting that the more conflict, the higher the
processing time.
Methods to Investigate Independent Variable Influence
The dependent variable in the model is the log of the number of days to reach a
decision in each direct appeal (a continuous variable), so we used ordinary least squares
(OLS) regression to investigate the influence of the independent variables on case
processing time. 62 As our data cover an eleven year period, there is a risk that the errors
associated with one observation are correlated with errors with another observation, i.e.,
the observations are not independent, a violation of an assumption for using OLS
estimation techniques. For example, errors in observations in and around the same year
will be more highly correlated than observations more separated in time. To control for
temporal effects, if any, we included dummy variables for each year covered in the data.
In addition, given the cross-sectional nature of the data, we included dummy variables for
each state in order to control for state-specific effects not otherwise captured in the model
(Stimson 1985; Harrington and Ward 1995).63
After initial runs of the model, we found through regression diagnostics that
certain cases were having an unusually influential effect on the regression coefficients.
Some cases were so influential that they were distorting the results. This was not
unexpected, as we noticed in the data collection that for various idiosyncratic reasons
some cases were taking years and years or, conversely, very little time to process. Using
standard regression diagnostic techniques, we found that 4.3 percent of the cases had this
impact, thereby distorting the results for the remaining 95.7 percent of the cases in the
dataset. As a result, we constructed the model without these cases, leaving us with a
dataset of 1,602 cases from the fourteen states.64
61

A “natural court” is one on which, over time, the same set of judges sit. The ideology scores and standard
deviations for the years 1970-2004 for each natural court per court year were obtained from a dataset
compiled by Professor Laura Langer, University of Arizona. Unfortunately, Langer’s data do not include
ideological scores for justices on the Texas Court of Criminal Appeals, so we used as surrogates scores for
justices on the Texas Supreme Court, assuming that the ideological division on the two courts was roughly
the same.
62
With models addressing time to an event, problems may arise with censored data, particularly rightcensored data. By the end of the study period, the event under study may not yet have occurred with regard
to some observations in the data, requiring the use of statistical approaches to address the censoring. We do
not believe this is an issue in our study because we have complete data for all observations, i.e., a starting
date and ending date. We did not include in our data appeals that were filed during the study period but had
not been resolved by the end of the study period. However, given the length of the study time frame and the
completeness of the data we have for all appeals that were decided (i.e., ended) during the study period, we
do not believe that the exclusion of the unresolved cases resulted in selection bias.
63
Due to problems of collinearity, the state dummy variable for Florida was excluded from the final model.
64
We identified unusually influential cases using a standard regression diagnostic tool, Cook’s Distance
values. We excluded from the data cases with Cook’s Distance values over 4/n-k-1, with n signifying the
number of cases in the data and k signifying the number of variables included in the model. We also

Regression Results: Determinants of Capital Case Processing Times
The model results are set out below. The overall relationship between the log of
processing time and the set of independent variables was significant (F = 102.681, p <
.001), demonstrating that we can reject the null hypothesis that there is no relationship
between processing time and these independent variables. The adjusted R2 value for the
model is .663, meaning that together, the independent variables in the model explain 66.3
percent of the variance in processing time.65
Fig. 17. Processing Time of Direct Appeals in Capital Cases, 1992-2002
Variable
Coefficient
Std. Error
Scope of Review

.025

.023

Case Complexity

.013**

.001

Outcome

.069**

.018

Disagreement

.071**

.014

Law/Rule

-.135**

.045

Capital Caseload

.003**

.001

Staff Attorneys

.029

.015

IAC Review

.427**

.060

Ideological Conflict

.008**

.002

Constant

6.578**

.067

Note: Year and state dummy variables are included in the estimation but not
reported in the table.

F = 102.681**
R2 = .670
Adj. R2 = .663
N = 1602
*p < .05; **p < .01

Given that the dependent variable is the log of the days between the filing of the
notice of appeal and the decision of the court of last resort, the effect of each significant
variable on processing time can be presented in percentage terms. Specifically, for
statistically significant variables, all other factors being equal, a one unit increase in the
independent variable will increase (decrease, if the coefficient is negative) processing
excluded two cases from the model because they were not reported in the West Regional Reporters (though

available through Westlaw), which prevented us from computing the page length variable.

65
Because the R2 value increases with the inclusion of additional independent variables, the adjusted R2

(.663) takes into account the inclusion of additional variables.

time in a percentage equal to 100 times the value of the coefficient. For example, the
results confirm our expectation that case complexity increases time to decide appeals. For
each additional page of the majority opinion (our measure of case complexity) processing
time increased by 1.3 percent, holding other factors constant.
Among other case characteristic variables Outcome also was significant. A
reversal of the lower court decision increased processing time by about 7 percent. As
noted above, the cases ultimately reversed may be the more problematic appeals for the
court to resolve, and the court may need to spend more time working out the legal issues
and justifying its decision in an opinion. This may especially be true given the high level
of public support for the death penalty and crime reduction in general; the court may need
additional time to prepare the justification of a decision to overturn the conviction or
death sentence of a notorious murderer.
The results also supported our hypothesis that the extent of disagreement among
justices increased the time necessary to resolve capital appeals. For each dissenting
opinion filed in an appeal, processing time increased 7.1 percent. The final case
characteristic variable, Scope of Review, was not significant, suggesting that courts spent
just as much time resolving appeals of sentences alone as they did resolving appeals of
convictions and sentences.
The overall results for the institutional arrangements and resources variables also
were impressive, with large contributions of a number of variables in this group found
significant in explaining variation in processing time. Initially, and as expected, prior
review of an appeal by an intermediate appellate court had a significant impact on
processing time, holding all other variables constant. Specifically, it added about 43
percent to the time necessary to process an appeal through to the state court of last resort.
Recall that the study does not include reversals by the intermediate appellate court that
were not subsequently reviewed by the COLR. (COLR review in Tennessee and Ohio
was mandatory when the death penalty was affirmed by the IAC and discretionary when
the sentence or underlying conviction was reversed.) Such cases do reduce state high
court time devoted to capital cases. However, a sampling indicates that unreviewed IAC
reversals are so few in number that their impact on COLR efficiency is marginal. 66 In
addition, the overall process gains no time advantage from duplicative review by two
appeals courts (the IAC and COLR).

To determine the impact of IAC reversals, we examined all cases in which a death penalty was imposed
and a direct appeal was decided by the intermediate courts of Tennessee and Ohio, respectively, between
1992 and 1995. (We chose the 1992 start date because our study began with 1992. The 1995 end date was
selected because Ohio discontinued IAC review of capital murders occurring in 1995 or thereafter.) Of the
53 appeals resolved by IACs, 44 of which were affirmances, 48 COLR appeals were granted. Thus, in 91
percent of the cases (48/53 = .905), IAC review did not obviate COLR reconsideration, and therefore did
not contribute to appellate efficiency. In nine cases the IACs reversed the death sentences or underlying
convictions. Eight of these cases were appealed to the COLR, which granted discretionary review in four.
(In one case we could not determine if there was COLR review.) This suggests that half of the IAC
reversals do not receive additional COLR consideration, a potential time saver for the COLR. The savings
is limited, however, due to the small number of IAC reversals.

The other significant variable in this group with noteworthy impact was the
existence of a law or rule in the state specifically directed to expediting capital appeals.
The presence of such a statutory or procedural rule decreased processing time by about
14 percent, holding all other variables constant. The Capital Caseload variable also was
significant, but its impact was slight. Specifically, for each additional capital appeal on
the court’s docket for that year, processing time in an appeal increased by less than one
percent.
The Staff Attorneys variable was not significant. This finding could be a result of
the measure used in the model, as the variable does not take into account how these staff
attorneys are utilized, just that they are on the court’s staff. We do not know, for instance,
the extent to which the courts are using these staff attorneys on capital cases, if at all.
Turning finally to the court attribute variable, Ideological Conflict, our results
show that ideological differences among a state’s high court justices did increase the
processing of these often-controversial cases. Courts with ideologically diverse justices
took longer to resolve direct appeals in capital cases. The impact was small, but it
suggests that conflict between liberal and conservative justices increases the time
necessary to resolve these appeals as justices work through conflict
Overall, the model results indicate that variation in processing times is driven by
factors in all three categories, but is most significantly influenced by case attributes and
institutional factors. Cases that are complex, that generate dissenting opinions, and that
reverse trial court decisions, require more time to resolve. Courts seeking to increase
efficiency in capital appeals will have little control over these case attribute factors, but
they do have an influence over the institutional factors helping to explain time on appeal.
First, avoiding intermediate appellate court review markedly speeds up time on appeal.
Cases going directly to the COLR for review proceed through the direct appeal stage
significantly more rapidly than those going though the intermediate appellate court.
Second, specialized laws or rules directed to increasing appellate efficiency in capital
cases are having their desired effect.
Policy Recommendations
There are, as we detailed in the Introduction to this study, numerous reasons for
seeking to streamline death penalty appeals. We listed eight detrimental effects of the
current protracted capital appeals process:
1. Decreases finality in the administration of justice.

2. Increases corrections costs for maintaining death row inmates.

3. Decreases public confidence in the criminal justice system.

4. Weakens the deterrent benefits of capital punishment.

5. Increases the emotional trauma of the murder victims’ family and friends.

6. Increases the difficulty of conducting legally permitted retrials.

7. Delays the resolution of meritorious capital appeals.

8. Creates a basis for additional litigation.

These effects all boil down to this: the length of the current process adds to the
expense and diminishes the effectiveness of the criminal justice system. Those sentenced
to death are entitled to a complete and thorough review of their convictions and sentences
to ensure that there is no miscarriage of justice. Will improving the efficiency of the
direct appeals process prevent this from occurring? We do not believe so, as the
efficiencies we propose would hardly reduce the appellate process in a way that
undermines full and fair review. If the entire process were reduced by one-third, for
example, it would still take, on average, over seven years to go from death sentence to
actual execution – scarcely assembly line justice. Moreover, increasing efficiency is not
just in the interest of the state. The death-sentenced inmate with a meritorious appeal
should be equally concerned with efficiency in the appellate process so that he or she
does not remain on death row any longer than necessary.
We recommend two general reforms to expedite the direct appeals process. These
improvements are empirically grounded, i.e., they derive from our comprehensive
fourteen state, eleven year study of actual death penalty appeals. Because our analysis
was limited to direct appeals, the first stage in the process, our recommendations are
likewise confined to that part of the process. This is not to imply that there are no other
worthy reforms to consider. To the contrary, our experience with state postconviction
review and federal habeas corpus suggests that these steps are very much in need of
reform. 67
1. Eliminate intermediate court review.
We found that a two-step direct appeals process – intermediate appellate review
followed by high court review – added significant amounts of time with insufficient
benefit. While some judges on the courts of last resort said they appreciated the
narrowing of the issues and the second opinion provided by the intermediate courts,68 we
doubt that these benefits outweigh the loss of efficiency. We found that a two-step
process takes roughly one-third more time to administer. One of the justices of the Ohio
Supreme Court told us that the change to direct COLR review “definitely speeded the
appellate process up because it cut a whole layer out that took two or three years to
complete.”69 We confirmed this impression, finding that Ohio reduced its time
expenditure by 25 percent when it switched to direct high court review. Tennessee and
Alabama – the only states to maintain the two-level process – should abolish it.
2. Adopt rules or statutes that impose deadlines on actors in the capital appeals
process.
Deadlines work! Those states with statutes or court rules that set reasonable but
enforceable deadlines for such crucial activities as the preparation of the record by trial
court personnel, the perfecting of appellate briefs by defense lawyers and prosecutors,
67

The 109th Congress considered, but did not approve, reforms of federal habeas corpus, styled the
“Streamlined Procedures Act of 2005.”
68
Tennessee, at the urging of its supreme court justices, adopted intermediate court review in 1992. IAC
review may lighten the workload of the state supreme court by enabling it to ignore issues resolved to its
satisfaction by the intermediate court and focus on the issues most important to the COLR. Interview with
Donald E. Dawson, Tennessee Post-Conviction Defender, March 3, 2005.
69
Interview with Ohio Supreme Court Justice Evelyn Lundberg Stratton, April 4, 2005.

and the resolution of a case by the appellate court, complete the direct appeals process
more expeditiously.
For example, Virginia, one of the most efficient states, has a supreme court rule
giving docket preference to death-sentenced cases.70 Likewise, a Nevada statute provides
that decisions in capital appeals must be reached within 150 days from receipt of the
record.71
Enforcement of deadlines can be problematic, however, especially with capital
cases. Bill Thompson, Assistant to Missouri Supreme Court for the Death Penalty, told us
that in 1988, as part of a broader reform of death penalty procedures, Missouri
established numerous filing and proceeding deadlines.72 While some of these were
maintained, others were later abolished because, as Thompson put it, “what are you going
to do” if the deadline isn’t met?
While each state must determine for itself where its process needs streamlining
and how best to achieve efficiencies, our research flagged several areas of state
inefficiency.
2.1 Georgia defense attorneys take over eleven times the aggregate median to file
notices of appeal in capital cases. The explanation is that Georgia law provides that a
motion for a new trial tolls the 30 day filing deadline for the notice of appeal. 73 This law
ought to be reexamined.
2.2 Kentucky, Washington, and Tennessee take the longest to prepare the record
and appellate briefs. Enforced and meaningful deadlines for trial court personnel,
appellate counsel and prosecutors could help move the process along.74
2.3 Ohio, Arizona and Kentucky allocate more time than their sister states to the
scheduling of oral arguments before the state court of last resort. This is a management
problem that the justices can rectify themselves by establishing deadlines through court
rules. Alternatively, the legislature might consider overall deadlines for resolving capital
appeals.
2.4 Although it might surprise those who associate Texas with efficient
administration of the death penalty, the Texas Court of Criminal Appeals is the slowest
court to reach a decision once oral argument is completed and the briefs have been filed.
Court rules or statutes imposing deadlines could help change this.

Va. Supreme Court Rule 5:23.

Nev. Rev. Stat.§ 177.267.

72
Interview with Bill Thompson, Assistant to Missouri Supreme Court for the Death Penalty, March 9,

2005.

73
Ga. Code Ann. § 5-6-38(a).

74
Kentucky adopted Rule 75.01(4) in 1991 setting guidelines for preparation of the record in death cases,

but that rule still allows up to 230 days without penalty.

Conclusion
The death penalty appeals process can be administered in a more efficient manner
and still ensure fairness to the accused. This study demonstrates that the states can meet
U.S. Supreme Court due process mandates and provide a speedier capital appeals
process. Indeed, Virginia and Georgia prove that even for death penalty appeals the states
can meet American Bar Association efficiency standards for undifferentiated appeals.
Our two general recommendations – eliminate intermediate court review and adopt rules
or statutes that impose deadlines on actors in the capital appeals process – point the way
toward reasonable and low-cost reform. In fact, we would expect these reforms to save
money by reducing some of the costly by-products of the current protracted appellate
process (see the eight detrimental effects, listed above).
We do not foresee a time when, in the United States, the death penalty will ever
be resorted to for all or even for a substantial number of first-degree murders. On the
other hand, given the breadth, depth and persistence of public support, plus the approval,
however qualified, of the U.S. Supreme Court, we cannot envision the wholesale
abolition of capital punishment either. Some form of death penalty is probably here to
stay, and it is likely to be a highly selective sanction, producing fewer than 300 cases
nationwide per year. The question is, therefore, can appeals for this relative handful of
cases be administered in a more efficient manner? This study marshals compelling
evidence that they can.

Bibliography
Aarons, D. (1998). Getting out of this mess: Steps toward addressing and avoiding
inordinate delay in capital cases. Journal of Criminal Law & Criminology 89:1-80.
Alfieri, A. V. (1996). Mitigation, mercy, and delay: The moral politics of death penalty
abolitionists. Harvard Civil Rights-Civil Liberties Law Review 31:325-352.
American Bar Association (1994). Standards Relating to Appellate Courts. Chicago, IL:
American Bar Association.
Bailey, W. C. (1980). Deterrence and the celerity of the death penalty: A neglected
question in deterrence research. Social Forces 58:1308-1333.
Bailey, W. C. and R. D. Peterson (1994). Murder, capital punishment, and deterrence: A
review of the evidence and an examination of police killings. Journal of Social Issues
50:53-74.
Beccaria, C. 1963 (1764). On Crimes and Punishments. Trans. Henry Paolucci.
Indianapolis, IN: Library of Liberal Arts/Bobbs-Merrill.
Berk, R. (2005) New claims about executions and general deterrence: deja vu all over
again? http://repositories.cdlib.org/uclastat/papers/2005031101/
Blume, J. and T. Eisenberg (1999). Judicial politics, death penalty appeals, and case
selection: An empirical study. Southern California Law Review 72:465-503.
Blume, J., T. Eisenberg, and M. T. Wells (2004). Explaining death row’s population and
racial composition. Journal of Empirical Legal Studies 1:165-207.
Bonneau, C. W. (2005). What price justice(s)? Understanding campaign spending in state
supreme court elections. State Politics and Policy Quarterly 5:107-125.
Brace, P., L. Langer and M.G. Hall (2000). Measuring the preferences of state supreme
court justices. Journal of Politics 62:387-413.
Brace, P. and M.G. Hall. (1997). The interplay of preferences, case facts, context, and
rules in the politics of judicial choice. Journal of Politics 59:1206-1231.
Bureau of Justice Statistics (2002). Capital Punishment 2001. Washington, DC: U.S.
Department of Justice. http://www.ojp.usdoj.gov/bjs/abstract/cp01.htm.
________. (2005). Capital Punishment 2004. Washington, DC: U.S. Department of
Justice.

________. (2006). Capital Punishment 2005. Washington, DC: U.S. Department of
Justice.
________. (2001). Felony Sentences in State Courts, 1998. Washington, DC: U.S.
Department of Justice.
________. (2000). Sourcebook of Criminal Justice Statistics 2000. Washington, DC: U.S.
Department of Justice. http://www.albany.edu/sourcebook/
California Attorney General’s Office (2000). A victim’s guide to the capital case process.
http://ag.ca.gov/publications/pdf/deathpen.pdf
Cauthen, J. N. G. and B. Latzer (2000). Time on appeal: Empirical evidence on capital
case processing in New Jersey. Judges’ Journal 39:12-17, 48.
Chapper, J. A. and R. A. Hanson (1990). Intermediate appellate courts: Improving case
processing. Williamsburg, VA: National Center for State Courts.
Culver, J. H. and J. T. Wold (1986). Rose Bird and the politics of judicial accountability
in California. Judicature 70:81-89.
Death Penalty Information Center (2002). http://www.deathpenaltyinfo.org.
Donohue, J.J. and J. Wolfers (2005). The ethics and empirics of capital punishment: Uses
and abuses of empirical evidence in the death penalty debate. Stanford Law Review
58:791-845.
Fox, J. (1991). Regression Diagnostics. Newbury Park, CA: Sage Publications.
Hall, M. G. and P. Brace (1994). The vicissitudes of death by decree: Forces influencing
capital punishment decision making in state supreme courts. Social Science Quarterly
75:136-151.
Hanson, R. A. (2001). Jurisdiction, Caseload, and Timeliness of State Supreme Courts.
Williamsburg, VA: National Center for State Courts.
________. (1998). Resources: The key to determining time on appeal. Court Review
35:34.

________. (1996). Time on Appeal. Williamsburg, VA: National Center for State Courts.

Harrington, C.B. and D. S. Ward (1995). Patterns of appellate litigation 1945-1990. In L.

Epstein, ed., Contemplating Courts. Washington, DC: CQ Press.

Hoffman, R. B. and B. Mahoney (2002). Managing caseflow in state intermediate

appellate courts: What mechanisms, practices, and procedures can work to reduce delay?

Indiana Law Review 35:467-555.
Karmen, A. (2001). Crime Victims - An Introduction to Victimology. Fourth Ed.
Belmont, CA: Wadsworth.
King, N. J. and R. L. Noble (2004). Felony jury sentencing in practice: a three-state
study. Vanderbilt Law Review 57:885-962.
Kozinski, A. and S. Gallagher (1995). Death: The ultimate run-on sentence. Case
Western Reserve Law Review 46:1-32.
Latzer, B. (2001). The failure of comparative proportionality review of capital cases
(with lessons from New Jersey). Albany Law Review 64:1161-1244.
Latzer, B. (2002). Death penalty cases. Second Ed. Boston, MA: ButterworthHeinemann.
Latzer, B. and J. N. G. Cauthen (2000). Capital appeals revisited. Judicature 84:64-71.
Latzer, B. and J. N. G. Cauthen (2001). Another recount: appeals in capital cases. The
Prosecutor 35:25-28.
Liebman, J. S., J. Fagan, A. Gelman, V. West, G. Davies and A. Kiss (2002). A broken
system, part II: Why there is so much error in capital cases, and what can be done about
it. http://www.law.columbia.edu/brokensystem2/index2.html.
Liebman, J. S., J. Fagan and V. West (2000). A broken system: Error rates in capital
cases, 1973-1975. http://www.law.columbia.edu/instructionalservices/liebman/.
Lofquist, W. S. (2002). Putting them there, keeping them there, and killing them: An
analysis of state-level variations in death penalty intensity. Iowa Law Review 87:1505
1557.
Martin, J.A. and E. A. Prescott (1981). Appellate court delay: Structural responses to the
problem of volume and delay. Williamsburg, VA: National Center for State Courts.
Mason, M. P. (1978). Courting reversal: The supervisory role of state supreme courts.
Yale Law Review 87:1191-1218.
National Center for Health Statistics (2001). National Vital Statistics Reports. Deaths:
Preliminary data for 2000. (Vol. 49, No. 12)
http://www.cdc.gov/nchs/products/pubs/pubd/nvsr/49/49-13.htm.
National Center for State Courts. State Court Statistics, 1985-2001: [UNITED STATES]
[2001 DATA] [Computer file]. ICPSR version. Williamsburg, VA: National Center for

State Courts [producer], 2002. Ann Arbor, MI: Inter-university Consortium for Political
and Social Research [distributor], 2003.
________. State Court Statistics, 2002: [UNITED STATES] [Computer file]. ICPSR
version. Williamsburg, VA: National Center for State Courts [producer], 2003. Ann
Arbor, MI: Inter-university Consortium for Political and Social Research [distributor],
2004.
Powell, L. F., Jr. (1989). Committee Report and Proposal from the Judicial Conference
of the United States Ad Hoc Committee on Federal Habeas Corpus in Capital Cases.
reprinted at Congressional Record 135:S13482 (October 16, 1989).
Radelet, M. L. and M. Mello (1992). Death-to-life overrides: Saving the resources of the
Florida Supreme Court. Florida State University Law Review 20:195-228.
Radin, M. (1980). Cruel punishment and respect for persons: Super due process for death.
Southern California Law Review 53:1143-1185.
Rose, W. J. and Spangenberg, R. J. (1988). Time and Expense Analysis in PostConviction Death Penalty Cases in North Carolina. West Newton, MA: The
Spangenberg Group.
Shepherd, J. (2004). Murders of passion, execution delays, and the deterrence of capital
punishment. Journal of Legal Studies 33: 283-321.
Spangenberg, R. J. (1991). Final Report, Survey of Compensation Rates for Counsel and
Experts in Capital Cases on Direct Appeal, State Post-Conviction and Federal Habeas
Corpus. West Newton, MA: The Spangenberg Group.
Spurr, S. J. (2002). The future of capital punishment: Determinants of the time from
death sentence to execution. International Review of Law and Economics 22:1-22.
Stimson, J. A. (1985). Regression in time and space: A statistical essay. American
Journal of Political Science 29:914-47.
Thompson, M. P., F. H. Norris and R. B. Ruback (1998). Comparative distress levels of
inner-city family members of homicide victims. Journal of Traumatic Stress 11:223-242.
Traut, C.A. and C.F. Emmert (1998). Expanding the integrated model of judicial decision
making: The California justices and capital punishment. Journal of Politics 60:1166
1180.
Uelmen, G. F. (1988). The Lucas court: First year report. California Lawyer 8:30.
U.S. President (1996). Statement on signing the antiterrorism and effective death penalty
act of 1996. (April 24, 1996). Weekly Compilation of Presidential Documents 32:17.

Vick, D. W. (1995). Poorhouse justice: Underfunded indigent defense services and
arbitrary death sentences. Buffalo Law Review 43:329-460.
Vickrey, W. C. (1994). Opinion filings and appellate court productivity. Judicature
78:47-52.
Wilson, J. Q. and R. J. Herrnstein (1985). Crime & Human Nature. New York: The Free
Press.