Skip navigation

Minority Practice, Majority’s Burden - The Death Penalty Today, Liebman and Clarke, 2012

Download original document:
Brief thumbnail
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
DRAFT: Please do not cite or distribute without
express permission.
Minority Practice, Majority’s Burden:
The Death Penalty Today
James S. Liebman
Simon H. Rifkin Professor of Law, Columbia Law School

Peter Clarke
J.D. expected 2012, Columbia Law School.

Ohio State Journal of Criminal Law, Forthcoming
Although supported in principle by two-thirds of the public and even more of the
States, capital punishment in the United States is a minority practice when the
actual death-sentencing practices of the nation‘s 3000-plus counties and their
populations are considered. This feature of American capital punishment has been
present for decades, has become more pronounced recently, and is especially clear
when death sentences, which are merely infrequent, are distinguished from
executions, which are exceedingly rare.
The first question this Article asks is what forces account for the death-proneness
of a minority of American communities? The answer to that question---that a
combination of parochialism and libertarianism characterizes the communities
most disposed to impose death sentences---helps to answer the next question
addressed here: Why so few death sentences end in executions? It turns out that the
imposition of death sentences, particularly for felony murder (a proxy for the
out-of-the-blue stranger killings that generate the greatest fear among parochial
communities), provides parochial and libertarian communities with a quick and
cheap alternative to effective law enforcement. And that alternative is largely
realized whether or not death sentences are ultimately carried out. This explanation
sheds light on two other criminal law conundrums---the survival of the most
idiosyncratic manifestation of the felony murder doctrine (which mysteriously
transmogrifies involuntary manslaughter into capitally aggravated murder) and the
failure of the death penalty to have a demonstrable deterrent effect (which is not
surprising if the death penalty operates as a weak substitute for, rather than a

Electronic copy available at:

powerful addition to, otherwise effective law enforcement strategies). The
explanation also reveals a number of costs the capitally prone minority imposes on
the majority of citizens and locales that can do without the death penalty, including
more crime, a cumbersome process for reviewing systematically flawed death
sentences whose execution is of less interest to the death sentences‘ originators
than their imposition, and a heightened risk---to the judicial system as well as
individual defendants---of miscarriages of justice.
These explanations, in turn, beg the most important and difficult question
considered here. Why do the majority of communities and citizens who can live
without the death penalty tolerate a minority practice with serious costs that the
majority mainly bears? With a bow towards Douglas Hay‘s famous explanation for
the survival over many decades of eighteenth century England‘s no less universally
vilified death-sentencing system---which likewise condemned many but executed
few---we offer some reasons for the minority‘s success in wagging the majority. In
response to recent evidence of a (thus far largely counterproductive) majority
backlash, we conclude by offering some suggestions about how the majority might
require the minority of death-prone communities to bear more of the costs of their
death-proneness without increasing the risk of miscarriage of justice.
I. A Local Institution
II. A Minority Practice
A. Death Sentences
B. Executions
III. An Explanation of the Death Penalty’s Localism
A. Parochialism
1. Parochialism Defined
2. Death Sentencing Localities as Parochial Communities
B. Libertarianism
1. Death Sentencing Localities as Libertarian Communities
2. The Death Penalty as a Libertarian Tool for Self-Protection
3. The Death Penalty as a Libertarian and Parochial Tool for
C. A Parochial and Libertarian Explanation of the Resilience of the
Felony-Murder Doctrine
1. Felony Murder’s Puzzling Resilience
2. Felony Murder’s Symbiotic Relation to the Death Penalty
3. Capital Felony Murder as a Tool for Parochial Self-Protection
4. Capital Felony Murder as a Tool for Parochial and Libertarian
D. The Death Penalty and the Illusion of Self-Protection
IV. An Explanation of the Death Penalty’s Frequent Imposition and

Electronic copy available at:

Infrequent Execution
A. The Paradox at the Heart of the System
B. A New View, from Without
C. The True Course of a Capital Case
1. The Origin of a Death Sentence: The Scarlet Letter Stage
2. The Death Row Stage
3. The Re-trial Stage
a. The More Parochial Response to Reversal
b. The More Libertarian Response to Reversal
D. A New Explanation of Why Reversals Have No Chastening Effect
V. The Costs of a Broken System: Imposed by Few, Borne By Many
A. Direct Costs Compared to Life Without Parole
1. Additional Costs Per Trial
2. Additional Costs Per Appeal
3. Additional Cost Per Execution
4. Overall Additional Cost
B. Indirect Costs
C. Other Externalized Costs
VI. An Exploration of Why the Majority Accepts the Costs the Minority
A. The Opaque Nature of the Costs
B. No Deterrence Dividend
C. The Death Penalty as a Back-Pocket Option
D. The Majority’s Fear of the Minority’s Reaction to Abolition
E. The Resonance of the Minority’s Parochial and Libertarian Values
F. Hay’s America?
VII. The New Millennium: Decline and Fall?
A. The Last Decade’s Death-Sentencing Decline
B. An Uneven Decline
C. A Smaller Tail Wagging a Larger Dog
VIII. Policy Options
A. The Insufficiency of Options Previously Proposed
B. Regulatory Strategies
1. Less, not More, Externalization of Costs
2. The Problems with Performance-Based Approaches
3. Local Improvements in Defense Representation
4. Managed Prosecution
IX. Conclusion
As Tip O‘Neill famously quipped about ―all politics‖ in the United States,

almost all there is to know about its death penalty is local, not national.1 Only
local differences can explain, for example, the simultaneous vilification of the
United States as the only Western nation to punish personal and civilian crimes
with death2 and celebration of the State of Michigan as the vanguard of abolition
in the Western world.3
Until relatively recently, the decision to impose death and carry out
executions was a local affair across most of the Western world.4 For example,
Douglas Hay‘s classic article Property, Authority and the Criminal Law vividly
depicts eighteenth century England‘s ―capitol assize[s]‖ as a distinctly local
spectacle.5 He explains that period‘s prodigious number of death verdicts as a
diabolically ingenious tool the small gentry class in each community used to
customize the maintenance of social order to local conditions and protect local
prerogatives against centralization of law enforcement in the hands of the Crown.6
By the twentieth century, however, most Western nations had witnessed a
transformative ―delocalization‖ of capital punishment, placing capital institutions
in the hands of national authorities who, for example, carried out executions at a
great distance from the site of the offense. 7
Typically, the effect of
nationalization was to rationalize the process, reduce regional capital-sentencing
disparities and the frequency of executions overall and eventually pave the way for
national elites to abolish the penalty despite the public‘s desire to retain it.8
Although, capital punishment in the United States has fitfully followed some
of these trends—especially since the Supreme Court began examining the
penalty‘s constitutionality in the late 1960s9—American federalism has mostly


ABOLITION 11 (2010) (noting that ―all other Western nations have decisively abandoned‖ capital
See Id.
Id. at 108.
Douglas Hay, Property, Authority and the Criminal Law, in ALBION‘S FATAL TREE: CRIME
AND SOCIETY IN EIGHTEENTH-CENTURY ENGLAND 17–18 (Douglas Hay et al. eds., 1975).
Id. at 18 (describing England‘s localized ―system of criminal law based on terror‖ that was
used in place of a national police force that ―the gentry would not tolerate‖).
GARLAND, supra note 2, at 108, 116.
Id. at 109.
See, e.g., Furman v. Georgia, 408 U.S. 238 (1972); Witherspoon v. Illinois, 391 U.S. 510


blocked any nationalizing tendency.10 In the United States, state, not federal
legislation, is the source of nearly all the nation‘s capital crimes,11 and nearly all
of its death sentences are imposed and executed in the name of a state.12 All but a
handful of the nation‘s condemned prisoners await their fate on death row in state
prisons;13 state governors have the power to commute their sentences or to pardon
them;14 and state courts are and have always been responsible for the vast majority
of court rulings affirming and reversing capital verdicts.15 Currently, 68% of all
States (34 of 50) retain the death penalty,16 which neatly matches up with polling
figures showing that public support for the death penalty has hovered between 65
and 70 percent over the last decade or so.17 Michigan‘s pioneering constitutional
expungement of the death penalty in 1846, as well as more recent legislation in
New Jersey, New Mexico and Illinois amply demonstrate that abolition is also a
state affair in this country.18

GARLAND, supra note 2, at 188.
Eileen M. Connor, The Undermining Influence of the Federal Death Penalty on Capital
Policymaking and Criminal Justice Administration in the States, 100 J. CRIM. L. & CRIMINOLOGY
149, 151 (2010) (―[The] number of federal capital prosecutions remains low, and the vast majority of
homicide prosecutions are undertaken by state criminal justice systems.‖)
Death Sentences in the United States From 1977 By State and Year, DEATH PENALTY INFO.
CTR., (last visited May 13,
2011) [hereinafter Death Sentences by State and Year] (listing the 112 individuals executed in the
United States in 2009, of whom 108 were executed by one of the several states); Execution List 2011,
DEATH PENALTY INFO. CTR.,, (last visited May
13, 2011) (listing the fifteen executions thus far in 2011, all of which were carried out by one of the
several states).
CTR., (last updated June 30, 2011) (listing 3,249
death row inmates in the United States as of October 1, 2010, of whom only 68 were under federal
sentence of death).
See, e.g., Richard A. Devine, The Death Penalty Debate: A Prosecutor‘s View, 95 J. CRIM.
L. & CRIMINOLOGY 637, 639 (2005) (criticizing Illinois Governor George Ryan‘s order commuting
the sentence of, or pardoning, every death row inmate in the state in the waning days of his
administration as a policy matter but acknowledging the governor‘s power to issue the order).
Andrew Gelman et al., A Broken System: The Persistent Pattern of Reversals of Death
Sentences in the United States, 1 J. EMPIRICAL LEGAL STUD. 209, 214, 222 (2004) (noting that,
despite frequent complaints about federal court interference with the death penalty via habeas corpus
review, state courts have been responsible for 90% of reversals of death verdicts).
See, DEATH PENALTY INFO. CTR., Facts About the Death Penalty, supra note 13, at 1.
Death Penalty, GALLUP, (last visited
April 6, 2011).
See GARLAND, supra note 2, at 36–38; Statement from Governor Pat Quinn on Senate Bill


None of this has gone unnoticed. Much superb recent commentary on the
death penalty has treated States and regional groupings of States as the key unit of
analysis.19 Summarizing and contributing to that commentary, Professor David
Garland identifies a four-fold typology of American States and the death
penalty—(1) abolishing States; (2) States such as Colorado, Connecticut and
Wyoming that have death penalty statutes but only rarely invoke them; (3) States
such as California, Pennsylvania and Tennessee that frequently impose death
sentences but rarely execute them; and (4) States such as Ohio, Oklahoma and
Texas that impose and execute death sentences fairly frequently.20 The historical,
political, demographic and other explanations these observers have offered for
differences in the use of the death penalty have also tended to be
state-centric—focused, for example, on States that did and did not practice slavery,
join the Confederacy or tolerate lynching.21
[hereinafter Governor Quinn‘s Abolition Statement].
AMERICAN CAPITAL PUNISHMENT 72–78 (2003); Carol S. Steiker & Jordan M. Steiker, A Tale of Two
Nations: Implementation of the Death Penalty in ―Executing‖ Versus ―Symbolic‖ States in the
United States, 84 TEX. L. REV. 1869, 1870 (2006) (noting that the United States is not monolithic in
its death penalty practices and dividing the nation into ―abolitionist‖ states, ―symbolic‖ states and
―executing‖ states). A sampling of the rich literature on the American death penalty includes DAVID
(2008); BEYOND REPAIR: AMERICA‘S DEATH PENALTY (Stephen P. Garvey ed. 2003); CHARLES L.
Arthur L. Alarcón, Remedies for California‘s Death Row Deadlock, 80 S. CAL. L. REV. 697 (2007);
Jack Greenberg, Capital Punishment as a System, 91 YALE L.J. 908 (1982); James S. Liebman, Slow
Dancing with Death: The Supreme Court and Capital Punishment, 1963–2006, 107 COLUM. L. REV.
1 (2007); Stephen F. Smith, The Supreme Court and the Politics of Death, 94 VA. L. REV. 283
(2008); Carol S. Steiker & Jordan M. Steiker, Sober Second Thoughts: Reflections on Two Decades
of Constitutional Regulation of Capital Punishment, 109 HARV. L. REV. 355 (1995); Robert
Weisberg, Deregulating Death, 1983 SUP. CT. REV. 305.
GARLAND, supra note 2, at 200.
See, e.g., GARLAND, supra note 2; ZIMRING, supra note 19, at 72–78; William S. Lofquist,
Putting Them There, Keeping Them There, and Killing Them: An Analysis of State-Level Variations
in Death Penalty Intensity, 87 IOWA L. REV. 1505 (2002); Steiker & Steiker, supra note 19, at 1916;
Andrew Ditchfield, Note, Challenging the Intrastate Disparities in the Application of Capital
Punishment Statutes, 95 GEO. L.J. 801 (2007). Although there is a robust scholarship on intrastate


As revealing as it is, this state-focused commentary misses important facets of
the death penalty‘s localism. For example, these analyses typically pit States like
Texas—the so-called ―Death Penalty Capital of the Western World,‖ with nearly
500 executions since 198222—against more politically liberal States like Maryland,
which rarely carry out executions and are on no one‘s list of the nation‘s
bloodiest.23 Yet, just under two-thirds of the counties in Texas did not carry out a
single execution in the past thirty five years.24 Texas‘s El Paso County—the

geographic disparities in the imposition of the death penalty, it focuses on the legal implications, if
any, of the fact that defendants committing essentially the same capital-eligible crime in different
counties of the same state often are predictably sentenced to the death penalty in some counties and to
a lesser sentence in others. See sources cited infra note 19. Recently, Adam Gershowitz has taken
the analysis further, identifying the death penalty‘s localism as a policy, as well as a legal, problem
that not only contributes to ―geographic arbitrariness,‖ but also to uneven procedures, poor legal
representation and high error rates. Adam M. Gershowitz, Statewide Capital Punishment: The Case
for Eliminating Counties‘ Role in the Death Penalty, 63 VAND. L. REV. 307, 311, 318–28 (2010)
[hereinafter Gershowitz, Statewide Capital Punishment]; see also Adam M. Gershowtiz, Pay Now,
Execute Later: Why Counties Should Be Required to Post a Bond to Seek the Death Penalty, 41 U.
RICH. L. REV. 861, 871 (2007); infra notes 404–412 and accompanying text. But see Stephen F.
Smith, Response: Localism and Capital Punishment, 64 VAND. L. REV. 105, 107, 109–15 (2011)
(responding to Gershowitz and concluding that county-by-county death sentencing disparities ―are
not troubling at all, but rather the inevitable effect of any system of nationally varying law
enforcement‖). In considering why counties have different approaches to the death penalty,
Professor Gershowtiz mainly blames differential spending.
Gershowtiz, Statewide Capital
Punishment, supra at 318–23; See cf. id. at 318 (―[I]t is hard to explain the wide variations in
counties‘ use of the death penalty.‖). As Professor Stephen Smith points out, however,
Gershowtiz‘s explanation boils down to a matter of differences in ―‗local preferences,‘‖ including
preferences for approaches to and levels of support for public safety. Smith, supra at 110–11. As
we develop below, understanding the reasons for local differences—what Professor Smith refers to
generically as ―the values, priorities, and felt needs of local communities‖ to which local criminal
justice strategies respond, see id. at 110—is crucial to resolving the question whether the penalty‘s
localism is a good or bad thing.
CTR., (last visited May 11, 2011).
David Michael Smith, The Death Penalty Capital of the Western World, 13 PEACE REV.
495, 495 (2001) (―‘When it comes to the death penalty, Texas has the dubious distinction of being the
bloodiest state in the U.S., outstripping all others in sheer numbers of executions‘‖ (quoting an
official of Amnesty International)); see also Corinna Barrett Lain, Deciding Death, 57 DUKE L.J. 1,
45 (2007) (―Texas [in 2000] was the undisputed capital of capital punishment among the states.‖);
Adam Liptak, At 60% of Total, Texas is Bucking Execution Trend, N.Y. TIMES, Dec. 26, 2007, at A1.
Professor Baumgartner identified the originating county of all executions in the United
States between 1977 and 2007. His data show that only 90 (35%) of Texas‘s 254 counties carried
out at least one execution between the Supreme Court‘s reinstatement of the death penalty in 1976
and 2010.
Frank R. Baumgartner, Spreadsheet of Executions by County, (last updated Oct.
2010) [hereinafter Baumgartner, Spreadsheet].


seventh-fifth most populous county in the nation 25 —carried out only four
executions during those three and one-half decades.26 By contrast, Baltimore
County, Maryland---the nation‘s seventy-first most populous county—registered
the twenty-third most death sentences of any county of any size in the United
States between 1973 and 1995.27
More generally, many of the most important capital decisions are exclusively
the domain of local actors—including whether to investigate, charge, convict and
condemn a suspect. Applying the same statute, moreover, and faced with similar
circumstances, local actors often make these decisions quite differently from
counterparts in neighboring locales.28 There even is evidence that state appellate
courts and federal habeas corpus courts exercise different levels of scrutiny of
death verdicts depending upon which locality generates them. 29 More than
anything else, therefore, it is the practices, policies, habits and political milieu of


Population figures are from the 2000 Census. , Resident Population Estimates for the 100
Largest U.S. Counties Based on July 1, 2009 Population Estimates: April 1, 2000 to July 1, 2009,
U.S. CENSUS BUREAU, (last visited
June 14, 2011)
Baumgartner, supra note 24. See U.S. CENSUS BUREAU, supra note 26. Compare
Texas‘s Potter and Brazos counties, each with about one-fifth the population of El Paso County,
which executed eleven and twelve individuals, respectively, during the same period.
See JAMES S. LIEBMAN ET AL., A Broken System, Part II: Why There Is So Much Error in
Capital Cases, and What Can Be Done About It, Table 11A, at B-2, (2002), [hereinafter BROKEN SYSTEM II]. Over
the 1973–1995 period, suburban Baltimore County accounted for over 55% of Maryland‘s death
sentences, despite having less than 15% of the state‘s population (1990 census) and being only its
third most populous county. See id. Baltimore County‘s proportion of Maryland‘s death row
population actually increased in subsequent years. Raymond Paternoster et al., Justice by
Geography and Race: The Administration of the Death Penalty in Maryland, 1978-1999, 4 MARGINS:
MARYLAND‘S L.J. ON RACE, RELIGION, GENDER & CLASS 1, 2 (2004) (noting that, as of 2004, fully
two-thirds of those on death row in Maryland were convicted by Baltimore County).
See, e.g., BALDUS ET AL., supra note 19, at 117–33; GARLAND, supra note 2, at 117; Leigh
B. Bienen, Capital Punishment in Illinois in the Aftermath of the Ryan Commutations: Reforms,
Economic Realities, and a New Saliency for Issues of Cost, 100 J. CRIM. L. & CRIMINOLOGY 1301,
1324–41 (2010); G. Ben Cohen & Robert J. Smith, The Racial Geography of the Federal Death
Penalty, 85 WASH. L. REV. 425, 433 (2010); Gershowitz, Statewide Capital Punishment, supra note
21, at 308–10, 314–28; Paternoster et al., supra note 27, at 2 ; Glenn L. Pierce & Michael L. Radelet,
Race, Region, and Death Sentencing in Illinois, 1988–1997, 81 OR. L. REV. 39, 62 (2002); Smith,
supra note 21, at 106–14.
See Gelman et al., supra note 15, at 247 (presenting statistical analysis indicating that state
courts are more likely to overturn death verdicts from urban than from rural and small-town counties,
while federal habeas corpus decisions exhibit the opposite trend); infra notes 192–94 and
accompanying text.


local prosecutors, jurors and judges that dictate whether a given defendant in the
United States—whatever his crime—will be charged, tried, convicted and
sentenced capitally and executed.
By focusing on county rather than state data, this Article paints a new and
surprising portrait of the death penalty in the United States. Unlike state-by-state
analyses or national polling numbers—which portray the penalty as a penological
tool of choice of a comfortable majority of Americans and American
jurisdictions 30 —county-level analysis reveals that the modern American death
penalty is a distinctly minority practice across the United States and in most or all
of the 34 so-called death penalty States. Indeed, analyzed in this way, the data
pose a modern analogue to the puzzle at the heart of Douglas Hay‘s examination of
England‘s eighteenth century capital punishment: how a small minority of capital
punishment partisans gained control over the nation‘s killing machine and
administered it for decades in a costly, even facially irrational manner, and why
everyone else went along.31
Part II discusses the raw numbers, which show that American counties
responsible for the vast majority of death sentences and executions represent a
minority of Americans. Moving from simple counts to an uneasy combination of
statistical analysis and cultural speculation, Part III identifies two threads that
appear to link communities that frequently use the death penalty—parochialism
and libertarianism—and shows how their intersection enhances the penalty‘s
appeal. Part IV uses the confluence of the same two forces to explain another
puzzle confounding death penalty practice both in the United States today and in
eighteenth century England: the disjuncture between a large number of death
verdicts and a small number of executions. In both periods, post-conviction
review—by clemency officials in England, by courts in the United States—is the
most common explanation, yet in neither period did centralized review have the
expected, chastening effect on local practice. Again analogizing to Hay‘s
solution to a parallel quandary, Part IV explains why this is so in the United States
The remainder of the Article discusses the implications of the minority‘s use
of the death penalty for the majority of jurisdictions that do not much use it. Part
V catalogues the myriad and escalating costs locally practiced capital punishment


See supra notes 16–18 and accompanying text.
See Hay, supra note 5, at 18–19.


imposes on residents of the majority of communities that rarely or never use it, and
asks why the majority continues to tolerate those costs. Part VI complicates the
analysis by distinguishing the last dozen years of capital punishment practice from
the two dozen preceding years, noting important changes that reveal that the
majority is fighting back. If so, however, the unintended effect has only been to
increase the disparity between the costs and benefits of the death penalty for the
majority. Part VII suggests how the majority might better align costs and
There are two obvious ways to measure usage of the death penalty: how often
it is imposed, and how often it is carried out. Although the former sets an upper
boundary on the latter, the correlation between the two can be less pronounced
than one would expect. Thus, although eighteenth century wags across the
Channel sneered that England might as well adopt the gallows as its national
symbol considering how many crimes received death sentences, in fact, even in its
capital punishment heyday, Georgian England executed only about 20% of those it
convicted of capital crimes and condemned to die.32 An even greater disparity
exists in the United States today, which—since the Supreme Court permitted
executions to proceed in 1976 after a ten-year hiatus—has executed only about
15% of those it has sentenced to die.33 In this Part, we examine the frequency of
both death sentences and executions in the United States in the recent past.
A. Death Sentences
The only comprehensive county-by-county analysis of modern death
sentencing in the United States is our own Broken System, Part II: Why There is So
Much Error in Capital Cases, and What Can Be Done About It, which examines
death sentences imposed in the United States between 1973 and 1995.34 During

England and Wales capitally sentenced about 580 people a year between 1770 and 1830,
but executed only about 115 a year. See V.A.C. GATRELL, THE HANGING TREE: EXECUTION AND THE
ENGLISH PEOPLE 1770–1868, 7, 618 (1994).
CTR., (last visited May 12, 2011).
See BROKEN SYSTEM II, supra note 27 (reprised in abbreviated form in Gelman et al, supra
note 15); see also James S. Liebman, Rates of Reversible Error and the Risk of Wrongful Execution,
86 JUDICATURE 78 (2002).


that period, thirty-four states sentenced at least one person to death, yet fully 60%
of the counties in those States did not impose a single sentence of death over the
twenty three year period—out of an estimated 332,000 homicides and 120,000
murder convictions occurring there during that time.35 Even in Texas, nearly
60% of its counties did not impose a single death sentence in the period.36
Not only have many counties de facto abolished the death penalty, but many
others have employed it only sparingly—once or twice a decade. Fairfax County,
Virginia, for example, with a population of nearly one million, imposed only five
death sentences between 1973 and 1995.37 Conversely, a relatively small number
of counties account for an extraordinary proportion of the nation‘s death verdicts.
During the same period, Seminole County, Georgia had the same number of death
sentences as Virginia‘s Fairfax County but had a population more than 100 times
smaller.38 Similarly, Hillsborough County (Tampa), Florida, with about the same
population as Fairfax, imposed over thirteen times more death sentences
(sixty-seven) than Fairfax County. 39 More than half of the death sentences
imposed nationwide over the twenty-three-year Broken System study period
originated in only sixty-six, or 2%, of the nation‘s 3,143 counties, parishes and
boroughs.40 16% of the nation‘s counties (510 out of 3,143) accounted for 90% of


See Gelman et al., supra note 15, at 214, 252.
See BROKEN SYSTEM II, supra note 27, at 246.
BROKEN SYSTEM II, supra note 27, at B-5. All county populations are from the 2000


Id. at B-1.
Id. at B-3, B-5.
See Peter Clarke, County-by-County Death Sentencing Spreadsheet (April 2011) (on file
with the authors) [hereinafter, Clarke Spreadsheet] (showing that from 1973 to 1995, 66 counties
imposed 2,569 of the 5,131 total death sentences imposed in the period). This spreadsheet reports
the death sentences per capita for all counties in the United States in which at least one death
sentence was imposed during one of two periods, 1973-1995 (the Broken System II period) and
2004-09. The death sentence data for the earlier period is from BROKEN SYSTEM II, supra note 27,
or the NAACP Legal Defense Fund‘s quarterly death row census, Death Row USA. Death Row
at, and may be obtained in printed form for earlier
years from the NAACP Legal Defense Fund in New York City. Broken System II is the source for
all 1973-95 county data for counties five or more death verdicts in the 1973-1995 period; Death Row
USA is the source for all county data for counties with one to four death verdicts during that period.
Death sentences by county for the later period are from a compilation created by Professor Robert J.
Smith of the Charles Hamilton Houston Institute, Harvard University Law School for research he is
conducting on regional disparities in death sentencing. Smith‘s compilation reports the number of
death sentences handed down by county, by year, between 2004 and 2009. In the Clarke
Spreadsheet, Smith‘s data for each of those years have been summed into a six-year total. We are


its death verdicts in the period.41
As the death-sentencing disparities just noted between the equally populous
Fairfax and Hillsborough Counties reveal, these numbers are not a result of the
heavy use of the death penalty by a small number of densely populated counties.
Between 1973 and 1995, the counties where only a fifth of all Americans lived
imposed two-thirds of its death sentences.42 Counties with 10% of the nation‘s
residents imposed 43% of its death sentences. 43 Even considering only
death-sentencing States, counties comprising around 10% of the population were
responsible for over 38% of the death sentences.44
In short, county-level data reveal something that state-level analyses do not:
notwithstanding broad public and statutory support, the vast bulk of death
sentences are imposed by on behalf of a small minority of Americans. A given
defendant‘s likelihood of receiving a sentence of death depends greatly on the
county in which he was tried.45
B. Executions
The figures are even more striking when we focus on counties where death
sentences end in actual executions. Recently, University of North Carolina
political scientist Frank Baumgartner found that only 454 (14% ) of the nation‘s
3,147 counties, parishes and boroughs carried out an execution between 1976,
when the Supreme Court permitted executions to go forward again, and 2007,
when Baumgartner‘s study ends.46 Six-sevenths of all American counties have
not carried out an execution in four and one-half decades.47
deeply grateful to Professor Smith for sharing his data with us. Population figures are from the
2000 census.
Clarke Spreadsheet, supra note 39.
See, e.g., Bienen, supra note 28.
Baumgartner Spreadsheet, supra note 24. Because of the moratorium on executions in the
nation between 1966 and 1976, the same figure holds if one commences the period in question in
PUNISHMENT 106, 106-125 (1973) (describing the moratorium). In this Article, we use 1976, not
1966, as the start date for our relevant analyses.
Baumgartner Spreadsheet, supra note 24. Baumgartner records 454 different county and
county-equivalents from which originated at least once sentence of death. There are 3,141 counties
and county-equivalents in the United States (3141-454 = 2687/3141 = 85.5%).


From this perspective, the localism and small proportion of American
communities that use the death penalty are startling. Fourteen counties—about
four-tenths of a percent of all U.S. counties, encompassing less than 5% of the
nation‘s population—carried out over half (53%) of its executions between 1976
and 2007.48 Nearly one-quarter (23%) of all executions came from only six
counties, with fewer than 2% of the country‘s population.49
Although jurisdictions with only a tenth of Americans account for 43% of its
death sentences, that same small slice of the population accounts for nearly 70% of
all executions.50 Between 1976 and 2007, half of all executions were carried out
on behalf of less than 5.5% of Americans.51 Twenty-eight percent of Americans
account for over 95% of its executions.52 A clear majority (57%) of Americans
live in counties that have not executed a soul for nearly a half century.53 Over
70% reside in counties that have executed one person or less during the period.54
Despite the comfortable majorities that profess to telephone pollsters that they
favor the death penalty, its actual imposition has been a minority practice in the
United States for decades.55 Moreover, the actual execution of the penalty is an
even rarer phenomenon. It is employed on behalf of a small fraction of
Americans who reside in a minuscule proportion of its communities.
The irregular pattern of modern American death sentencing cries out for a
local explanation. What common thread holds together the minority of locales
that regularly impose and execute death verdicts and separates them from the
majority of communities that do not use the penalty? Given that the rest of the


Id.; Clarke spreadsheet, supra note 39.
Frank R. Baumgartner, Race, Innocence, and the Death Penalty 19 (April 7, 2010),
pdf; Clarke spreadsheet, supra note 39.
Clarke spreadsheet, supra note 39.
See Clarke spreadsheet, supra note 39.
See supra notes 16–18 and accompanying text.


Western world has formally abolished the civilian death penalty—and the great
majority of Americans and their localities have largely managed without it for
decades—what is it that impels a narrow sliver of communities to cling to a
seemingly anachronistic institution?
At this point, the available hard data fail us. There is no comprehensive
study of the demographic, sociological, ideological or penological characteristics
that distinguish high death-sentencing counties from the rest, much less any that
support causal explanations. There are, however, some tantalizing clues in our
Broken System II study.56 In addition to other authors‘ suggestions about the
death penalty‘s localism, we rely on these inferences in offering hypotheses about
why these particular counties choose to practice capital punishment at all and at
such higher rates than other communities. Our premise is that localities that
―use‖ the death penalty—that execute it or at least impose it with any degree of
frequency—are not random idiosyncrasies, but share notable traits, tendencies and
traditions. Specifically, we propose that these localities are led to cling to the
death penalty and favor it over other potential responses to violent crime by a set
of common instincts that lie at the intersection of what we call ―parochialism‖ and
Broken System II is a comprehensive set of regression studies of the features
of states and counties that—and of possible explanations for why they—experience
high rates of serious, i.e., prejudicial and reversible, error in the capital verdicts
they impose. The study is only suggestive for present purposes because it does
not seek to explain the comparative use of the death penalty (the topic of this
Article), but instead comparative rates of judicially identified serious error when it
is used.
The reason we can use Broken System II at all for our current purposes is
because—at both the state and county levels—the strongest distinguishing feature
it associated with high error rates was high capital-sentencing rates. In other
words, states and counties that produced the most error-prone death verdicts were
also the ones that generated the most verdicts per 1000 homicides. At the state
level, capital-error rates increase from less than 15% to more than 75% as
death-sentencing rates per 1000 homicides move from the lowest to the highest end


See BROKEN SYSTEM II, supra note 27, at 82–145 (reprised in Gelman et al., supra note 15,
at 253–54). For the reasons emphasized in the next paragraph, even the Broken System II study
provides only clues, nothing more.


of the spectrum observed across states and years in the study.57 The same is true
at the county level; the more death verdicts per homicide a county imposes, the
more likely it is that each new death sentence will be infected with reversible
error.58 This county factor operates independently of, and in addition to, the
similar dynamic that operates at the state level.59 At both the state and the county
levels, the link between error rates and death-sentencing rates was powerful.
Holding all other factors constant, death verdicts imposed in jurisdictions with the
highest rates of death-sentencing per 1000 homicides were nine times more likely
to be reversed as a result of serious error than capital verdicts imposed in
jurisdictions with the lowest rates of death-sentencing per 1000 homicides.60
Broken System II also identified a variety of other features associated with
high rates of serious error in death verdicts at the state and county level. 61
Because those features are associated with high rates of capital-sentencing error,
which in turn are associated with high rates of capital-sentencing per se, we
hypothesize—fully acknowledging that we are nowhere near proving—that those
features are also characteristic of high death-sentencing jurisdictions.
A further caveat is that some of the associations Broken System II established
were only at the state, and not at the county level, because data on that feature are
available only at the state level.62 Although as to features for which we did have
data at both the state and county level, the county effects tended to be about as
strong as the state effects.63 We cannot say the same for certain as to features for
which there is no county-level data. In those cases, our hypotheses about the
attributes of high death-sentencing counties are doubly speculative, because they
are extrapolated based on features of jurisdictions with high rates of reversible
error in capital cases and are based on features of the states in which the counties
are located.
A. Parochialism


Id. at 342 (reprised in Gelman et al., supra note 15, at 242).
Id. at 349 (reprised in Gelman et al., supra note 15, at 254).
Id. at 349 (reprised in Gelman et al., supra note 15, at 254).
Id. at 224 (reprised in Gelman et al., supra note 15, at 241).
Id. at 239-41 (reprised in Gelman et al., supra note 15, at 243–44).
See, e.g., BROKEN SYSTEM II, supra note 27, at 96.
Id. at 269 (reprised in Gelman et al., supra note 15 at 252–53).


1. Parochialism defined
The term ―parochial‖ can have a pejorative connotation of closed-mindedness
or self-serving behavior.64 That is not our intent here. Rather, by ―parochial‖
we mean the attribution of innate importance and validity to the values and
experiences one shares with the members of—and thus to the security, stability and
continuity of—one‘s closely proximate community. We mean something like
―localism for its own sake‖—in much the way that the phrase ―parochial schools‖
is used to distinguish schools attractive to families with a common set of religious
beliefs from the majority of public and also other private schools.65
We also understand the concept to convey a sense of anxiety or threat.
Given the recent acceleration of national and worldwide cultural integration, we
understand ―parochialism‖ to include fears that prized local values and experiences
are embattled, slipping into the minority and at risk from modernity,
cosmopolitanism, immigration-driven demographic change, and a coterie of
―progressive‖ and secular influences, including permissiveness and crime. 66 We
assume that the perceived agents of this unwanted change include government
officials at the state and especially federal levels, including ―unelected,‖ ―elitist‖
judges and the outside lawyers who appeal to them.67 From this perspective, the

See, e.g., Jedediah Purdy, The Coast of Utopia, N.Y. TIMES, Feb. 22, 2009, at BR10,
available at (―It has never been
clearer that the country's best self is a global inheritance, its worst a parochial self-certainty.‖)
For an example of a largely favorable comparison between ―parochial‖ and public schools
(e.g., on grounds of the moral reasoning capacity, self-esteem and educational aspirations of
students), see Morality in Education: Morality in Parochial vs. Public Schools, UNIV. OF MICH. DEP‘T
OF PSYCHOLOGY, (last visited
May 4, 2011).
See GARLAND, supra note 2, at 158–61, 170–71 (discussing the ―localism‖ of American
politics generally and Americans‘ distrust of groups of people different from themselves who they
believe exercise excessive power over the national state).
Consider, for example, a local prosecutor's reaction to a federal habeas judge's reversal of a
death sentence:

Where's the fairness for the law-abiding citizens? How many appeals are you entitled
to? The constitution demands finality of punishment.
Us good folks here in Tennessee are able to take care of our business in the criminal
justice system. We're bright enough to understand the system. Once the Criminal
Court has spoken, and the Tennessee Supreme Court has spoken, that's enough.
We don't need this guy telling us we're wrong.


If he had to stand for election every

boundaries of the immediate community represent a bulwark against outside
influences that threaten to dilute or entirely dissolve the community‘s cohesion,
and parochialism compels a spirited defense of those frontiers.
2. Death-Sentencing Localities as Parochial Communities
There is evidence that the minority of localities that frequently impose the
death penalty are parochial in the sense described above. To begin with, studies
consistently find that localities with higher death-sentencing rates tend to have
traditional rather than modern political values, a resonance with masculine honor
codes and revenge, Old Testament and evangelical Christian religious beliefs, and
Republican Party affiliation.68 In addition, Broken System II found evidence that
communities with high rates of capital error—which are strongly correlated with
high death sentencing rates—are ones where influential citizens feel they are under
particular threat from crime.69 Interestingly, neither crime nor homicide rates
themselves predict high reversal rates and the high death-sentencing rates that go
with them. 70 What is predictive, however, is a high rate of homicide
victimization of white residents relative to the rate of homicides affecting black
residents.71 While in virtually all communities, the homicide rate experienced by
African-American citizens is greater than that experienced by whites, in heavy
death-sentencing communities that disparity is smaller. Other things equal, the
smaller the disparity is between white and black homicide victimization, the higher
the death-sentencing rate is.72 Heavy use of the death penalty thus seems to occur
when the worst effects of crime have spilled over from poor and minority
neighborhoods and are particularly salient to parts of the community that we can
predict have greater influence over local law enforcement, prosecuting and judicial
eight years or so and let the people speak, we wouldn't be seeing opinions like this. We
don't need a federal judge involved in our justice system. He needs to be held
accountable for what he's doing.
Tom Chester, Prosecutor Lashes Judge for Reversal of Death Sentence, KNOXVILLE NEWS-SENTINEL,
May 21, 1994, at A1.
See, e.g., Gelman et al., supra note 15, at 224.
Gelman et al., supra note 15, at 228.
BROKEN SYSTEM II, supra note 27, at iii (reprised in Gelman et al., supra note 15, at 217).
Id. (reprised in Gelman et al., supra note 15, at 244).
Id. (reprised in Gelman et al., supra note 15, at 242, 244); see also id. at 166–67 (noting that
higher error rates are correlated with higher death-sentencing rates); supra note 54 and accompanying
text (noting that we are extrapolating from factors associated with high error rates to ones associated
with high death-sentencing rates).
See Gelman et al., supra note 15, at 248.


In view of the importance of relative rates of white and black criminal
victimization, not crime rates themselves, we also hypothesize that it is this
cross-boundary, cross-class and cross-race spill-over effect of crime—or the
elevated fear of it—that disposes communities towards the harshly retributive
response of capital punishment. In other words, we hypothesize that a
―parochial‖ tendency to feel embattled from ―outside‖ influences, including crime,
disposes communities to use the death penalty.
There are at least four ways that the Broken System II data support this
hypothesis. First, of course, crime typically is, and is perceived to be, an
intrusive act of deviance from the accepted social order. In this sense, the
criminal is always an outside threat to a given community. Second, as we just
noted, the crime experienced in high-death-sentencing communities is statistically
more likely to be, or to appear to influential members of the community to be,
spilling over from the precincts where poor and minority citizens live into their
own communities. 74
Third, Broken System II strongly associated high
capital-error rates and usage of the death penalty with proximity to a large
population of either African-American citizens or citizens receiving government
welfare support. 75 This is particularly striking because African-American
communities and jurors themselves are far less likely than white ones to impose
the death penalty.76 Fourth, there is an ―interaction effect‖ between the second and
third factors above. In communities where both conditions are present—a large
African-American community nearby and a high rate of white homicide
victimization relative to black homicide victimization—the death-sentencing rate is
higher than one would expect by simply summing the effect of each of those
factors by itself.77 In other words, the combination of a large African-American
population nearby and a larger proportion of the risk of homicide being borne by
whites relative to blacks has a particularly strong association with more frequent
resort to the death penalty and higher rates of reversible error in death verdicts
Everything we know about the psychology of racial stereotypes and crime


BROKEN SYSTEM II, supra note 27, at 166.
BROKEN SYSTEM II, supra note 27, at 241 (reprised in Gelman et al., supra note 15, at 248).
See Gelman et al., supra note 15, at 227 & n.47 (citing sources).
BROKEN SYSTEM II, supra note 27, at 240 (reprised in Gelman et al., supra note 15, at 244).


suggests a dynamic that links this combination of high comparative crime
victimization in white communities and their close proximity to poor and minority
communities to a disposition to impose harsh punishment:
Ethnicity comes into play for two reasons. First, ethnicity is associated
with political power. Threats to the safety and security of the white
community tend to be taken more seriously by law enforcement officials
than threats to minority communities. Second, partially through the
unfortunate operation of racial stereotypes, citizens and officials tend to
use the ethnicity of suspects (particularly if the suspects are
African-American) and the size of nearby minority populations
(especially African-American populations) as proxies for the threat of
violent crime posed by particular individuals and present in particular
social environments.79
As the perceived80 threat rises, so does the likelihood of a punitive response.
We thus hypothesize that the relatively greater share of the risk of crime
falling on whites in high death-sentencing communities, together with the
misperceived threat from large nearby populations of poor African-Americans,
gives influential members of these communities a partly accurate, partly inflated,
sense of threat from crime emanating from outside. By ―outsider,‖ we certainly
mean transient individuals who do not reside in the community and are just passing
through. The association of capital crimes with gas-station and convenience-store
hold-ups and home invasion crimes along major interstate arteries is a
manifestation of this effect.81 But ―outsider‖ also includes individuals who cross


Gelman et al., supra note 15, at 228.
A crucial feature of this dynamic is the combination of actual threats, measured by
homicide rates, and inaccurately perceived threats, as measured by proximity to African-American
populations. The research literature, discussed in Gelman et al., supra note 15, at 248 & n.99 (citing
sources), demonstrates that white populations associate consistently proximity to African-Americans
with a much higher threat of crime than actually exists.
See, e.g., Houston v. Dutton, 50 F.3d 381, 383 (6th Cir. 1995) (affirming reversal of a
capital conviction for an out-of-towners‘ gas-station armed robbery and murder), cert. denied, 516
U.S. 905 (1995); Mitchell v. Hopper, 538 F. Supp. 77, 85–86 (S.D. Ga. 1982) (upholding death
sentences for men recently released from prison in Florida who crossed the border into southern
Georgia and killed a fourteen-year-old boy and shot his mother several times after marching them
into a meat cooler at the road-side grocery store where they worked), aff‘d, 827 F.2d 1433 (11th Cir.
1987), cert. denied, 483 U.S. 1050 (1987); Coleman v. State, 226 S.E.2d 911, 913 (Ga. 1976)
(affirming death sentence of one of four capitally sentenced defendants who broke into a rural mobile
home in southern Georgia in order to burglarize it, then successively killed, and in one case raped, six


racial and economic boundaries within communities in order to commit crimes.
The perceived threat posed from outsiders disposes parochial, more than other,
communities to invoke the death penalty in response to the threat—especially
when they can directly and powerfully communicate their fears to elected
officials.82 The importance of the last-mentioned element is suggested by the
Broken System II finding that excessive imposition of the death penalty increases
with the extent to which local judges are subject to political pressure—as measured
by the type, frequency and partisan nature of elections used to select judges.83
The use of the death penalty in response to perceived threats to influential
members of insular communities from cross-boundary crime helps explain the high
death-sentencing rate in communities that otherwise do not fit the capital
punishment stereotype. An example mentioned earlier is Baltimore County,
Maryland—the predominantly white, suburban donut that encircles the majority
African-American Baltimore City.84 The consistent pattern across the United
States of a two to five times greater chance of being sentenced to die for killing a
white victim than for the same killing of a black victim may be a more generalized
repercussion of the same dynamic.85
extended family members over the course of the day as each arrived home or came to check on the
family members who had already been killed), cert. denied, 431 U.S. 909 (1977).
BROKEN SYSTEM II, supra note 27, at 351 (reprised in Gelman et al., supra note 15, at 249).
See Gelman et al., supra note 15, at 256. This finding probably applies as well to district
attorneys, virtually all of whom in the United States, especially in death sentencing counties, are
locally elected. See William J. Stuntz, The Pathological Politics of Criminal Law, 100 MICH. L.
REV. 505, 533–35 & n.117 (2001) (noting that over 95% of district attorneys are elected, giving them
incentives to produce prosecutions and convictions the public favors).
See supra note 27 and accompanying text. Other similar localities with high death
sentencing rates include Cleveland, Ohio (Cuyahoga County); Gary, Indiana (Lake County);
Oakland, California (Alameda County) Philadelphia, Pennsylvania; and St. Louis County, Missouri.
See BROKEN SYSTEM II, supra note 27, at B-3 to B-6; Gershowitz, Statewide Capital Punishment,
supra note 21, at 314–18. See also infra notes 344–54 and accompanying text (discussing Los
Angeles and Phoenix).
See, e.g., McCleskey v. Kemp, 481 U.S. 279, 287 (1987) (discussing ―an extensive
analysis‖ of racial death sentencing patterns in Georgia, which found that ―even after taking account
of . . . nonracial variables, defendants charged with killing white victims were 4.3 times as likely to
receive a death sentence as defendants charged with killing blacks [and that] black defendants were
1.1 times as likely to receive a death sentence as other defendants,‖ indicating ―that black defendants
. . . who kill white victims have the greatest likelihood of receiving the death penalty‖); Samuel R.
Gross and Robert Mauro, Patterns of Death: An Analysis of Racial Disparities in Capital Sentencing
and Homicide Victimization, 37 STAN. L. REV. 27, 55. (―In Georgia, those who killed whites were ten
times as likely to be sentenced to death as those who killed blacks; in Florida the ratio was about to
eight to one, and in Illinois about six to one.‖); Glenn L. Pierce & Michael L. Radelet, Death
Sentencing in East Baton Rouge Parish, 1990-2008, 71 LA. L. REV. 647 (2011) (finding a 2.6 times
higher chance of being sentenced to death in East Baton Rouge Parish if the victim is white, rather


B. Libertarianism
1. Death-sentencing localities as libertarian communities
The evidence from Broken System II and elsewhere suggests that the counties
that frequently use the death penalty exhibit libertarian as well as parochial
impulses.86 By libertarian beliefs, we mean preferences for more rather than less
protection of acts of individual autonomy, for less over more frequent exercises of
state power, and for low taxes over high services. 87 Libertarianism also is
associated with a vigilante streak—a willingness to take the law into one‘s own
hands and out of the untrustworthy hands of government.88 In other words, the
uneasy combination of a desire for seriously retributive responses to
non-consensual acts interfering with another‘s autonomy, yet little or no spending
on police, prosecutors, investigators, courts, corrections, rehabilitative services and
other methods of combating crime may dispose libertarians towards self-help, even
CRIMINAL JUSTICE SYSTEM 133 (1999) (discussing McCleskey v. Kemp, supra); RANDALL KENNEDY,
RACE, CRIME AND THE LAW, NEW YORK 328 (1997) (discussing McCleskey v. Kemp, supra).
Consider also the prominent use of the death penalty by communities in the South to punish a
black man‘s rape of a white woman, even after the penalty had stopped being imposed there and
elsewhere for other crimes short of murder. See Charles Lane, The Death Penalty and Racism: The
2010, (noting that of the 455 men executed
for rape in the United States between 1930 and 1967, 90% were African American). Cf. Coker v.
Georgia, 433 U.S. 584 (1977) (barring the death penalty for rape of an adult woman); Kennedy v.
Louisiana, 554 U.S. 407, 413 (2008) (extending Coker‘s holding to bar the death penalty for rape of a
child). In its time and place, interracial rape constituted the quintessential cross-boundary, stranger
crime against members of a privileged community.
See infra notes 89–92 and accompanying text.
See, e.g., MILTON FRIEDMAN, CAPITALISM AND FREEDOM 199 (1964) (arguing that ―the
greater part of new ventures undertaken by government in the past few decades‖—what Friedman
calls ―the visible hand for retrogression‖—―have failed to achieve their objectives,‖ but that the
nation has progressed nonetheless because of ―initiative and drive of individuals co-operating through
(arguing that ―[t]he interventionist policies as practiced for many decades by all governments of the
capitalistic West have brought about . . . wars and civil wars, ruthless oppression of the masses by
clusters of self-appointed dictators, economic depressions, mass unemployment, capital consumption,
and famines,‖ which governments then have claimed ―demonstrate the necessity of intensifying
interventionism‖); JAN NARVESON, THE LIBERTARIAN IDEA 207 (1988) (―[L]ibertarians are
notoriously unhappy with the state . . . .‖).
See Angela M. Schadt & Matt DeLisi, Is Vigilantism on Your Mind? An Exploratory Study
of Nuance and Contradiction in Student Death Penalty Opinion, 20 CRIM. JUST. STUD., 255, 259
(2007) (―[P]ublic concerns for the costs and perceived effectiveness of the official criminal justice
system have served as the impetus for various forms of vigilante justice . . . .‖).


in the realm of criminal justice and law enforcement.
The Broken System II study found that states with high rates of reversible
error in capital cases also have three other features of interest here: high
death-sentencing rates, low clearance rates for serious crime and low levels of
expenditure on criminal courts.89 Holding other factors constant at their average
value, jurisdictions with the highest capital-error rates have seven times higher
capital-sentencing rates than jurisdictions with the lowest capital error rates; are
seven times less likely to capture, convict and incarcerate criminals for serious
crimes committed there; and spend a third as much on their criminal courts.90


See BROKEN SYSTEM II, supra note 27, at 352–54 (reprised in Gelman et. al, supra note 15,
at 240–43, 244–46). The study defined clearance rates as the number of prisoners incarcerated in
the state compared to the number of FBI Index Crimes committed. See id. at 352. Comparable
figures on state and county expenditures on capital defense costs are not available, but there is ample
evidence that high death sentencing counties do what they can to skimp on these costs. See, e.g.,
Stephen B. Bright, Counsel for the Poor: The Death Sentence Not for the Worst Crime but for the
Worst Lawyer, 103 YALE L.J. 1835, 1851 (1993–1994) (surveying indigent defense systems across
many states and finding that ―gross underfunding . . . pervades indigent defense‖ making it
impossible ―to attract and keep experienced and qualified attorneys because of lack of compensation
and overwhelming workloads‖); Matt Thacker, Clark County Prosecutor Supports Limits on
prosecutor‘s efforts to limit reimbursement of capital defense lawyers for their fees, experts and
investigators as a way to be able to afford to prosecute more cases capitally); Robert Weisberg, Who
Defends Capital Defendants?, 35 SANTA CLARA L. REV. 535, 536–38 (1995) (noting that ―whenever
a political official tries to attack expenditures for public defenders [handling capital cases], she is
likely to point to the absurdity of paying lawyers who are constantly attacking their employers‖).
Professor Weisberg provides a typical example, from the Mitchell case cited supra note 69 which he
a handled in federal habeas corpus proceeding, of the poor quality of local lawyers appointed to
represent defendants in capital cases:
[In the] second capital case I worked on . . . the defense lawyer [at trial] was paid $150
for the entire case, and believe me, he earned every penny of it. He only interviewed his
client once, as he was arranging a plea bargain for the guilt phase. He did not identify
or reach out to any of the several, excellent mitigation witnesses—teachers, coaches,
ministers, or whatever—who were only a few hours' drive away. The reason for the
omission, he said, was that he thought the case hopeless. He never attacked the
confession in the case, even though, as it later became obvious, the defendant was left
alone in the interrogation room with an off-duty police officer who was a cousin of the
victim, and who threatened the defendant with a gun. He did not try to interview the
officer, because he did not like the man.
Id. at 537–38.
See BROKEN SYSTEM II, supra note 27, at 352–53 (reprised in Gelman et. al, supra note 15,
at 240–43, 244–46).


Broken System II also found that the large numbers of capital verdicts awaiting
state court appellate review that are in jurisdictions with high death-sentencing
rates generally tend to depress reversal rates, further revealing an unwillingness in
those jurisdictions to provide sufficient resources to permit appellate courts to do
their work and identify the full range of serious error that high death-sentencing
rates cause.91 In short, residents of jurisdictions with high death-sentencing and
high capital error rates are far less disposed than others to empower state actors,
create effective government and professional structures, and pay taxes to
investigate, apprehend, prosecute, provide representation for, provide judges to
review the convictions and sentences of, and imprison perpetrators of serious
crimes.92 All of these are libertarian tendencies.
Curiously, Douglas Hay noted a similar dynamic in Georgian England.93
Hay marshals evidence showing that the peasantry‘s new-found mobility and
proneness to riot and crime on the eve of the Industrial Revolution greatly
frightened the rural gentry.94 Although the most efficient solution would have
been for Parliament to fund and empower the royal government to create a national
police force to maintain order, for the strongly localist and libertarian gentry,
escaping the frying pan of local anarchy into the fire of royal dominance had no
appeal.95 Instead, the lords relied on the most powerful local lever they had—the
criminal justice system and the power it gave them to strategically dole out death
sentences and pardons as a form of ―terror‖ to cower the lower classes.96
Whatever the merits of Hay‘s controversial explanation of the death penalty in
eighteenth century England, we conclude that analogous dispositions help explain
the uneven use of the death penalty in the United States today. In place of the
terroristic tool Hay describes, however, we see a device that parochial and


See id. at 370–72 (reprised in Gelman et. al, supra note 15, at 242–43).
See id. at iv (reprised in Gelman et. al, supra note 15, at 243). For a general discussion of
the historical relationship in the United States, and particularly the South, between underinvestment
in law enforcement, low clearance rates, high crimes rates and highly punitive sentences imposed on
criminals who are caught, see William J. Stuntz, Unequal Justice, 121 HARV. L. REV. 1969, 1984–85,
1992–93 (2008).
Hay, supra note 5, at 18.
Id. at 20–21.
Id. at 41; David Friedman, Making Sense of English Law Enforcement in the 18th Century,
2 U. CHI. L. SCH. ROUNDTABLE 475, 476 (1995) (― Eighteenth-century England viewed a system of
professional police and prosecutors, government-paid and –appointed, as potentially tyrannical and,
worse still, French‖ ).
Hay, supra note 5, at 25; see supra notes 5–6 and accompanying text.


libertarian communities use to ration the costly and state-empowering government
response to cross-cultural crime.
2. The death penalty as a libertarian tool for self-protection
In a number of respects, capital punishment fits well within a
libertarian—which is to say staunchly retributive97—theory of justice. In a world
where citizens are free of government influence in making individual decisions,
they also deserve to be fulsomely punished for making antisocial, unacceptable
choices.98 Libertarians are deeply skeptical of determinist explanations for crime.
In their view, crime is not predestined by nature, nurture or social inequality but is
a choice individuals freely make that merits severe punishment.99
Among possible punishments, moreover, the death penalty may have a
particular appeal to libertarian communities.
As we have noted, those
communities tend to be skeptical of government‘s ability to do anything very well,
and the longer and more complicated the task, the greater the skepticism. This
suggests a preference for a single act of state ―execution‖ over a lifetime of state
efforts to securely incarcerate a murderer. And researchers have consistently
identified this preference as a distinguishing feature of jurors most prone to impose
the death penalty.100 As these jurors tell researchers, they prefer death over life


(discussing ―the common libertarian insistence that there are principles of justice that must be
respected in the treatment of individuals even if violations of those principles were socially
(describing the ―libertarian conceptions of free will on which retributivism depends‖).
See, e.g., N. Stephan Kinsella, A Libertarian Theory of Punishment and Rights, 30 LOY.
L.A. L. REV. 607, 645 (1997) (arguing that libertarian ―justice requires that the aggressor be held
responsible for the dilemma he has created as well as the aggression he has committed‖).
Id. at 626 n. 46.
Studies showing that death prone jurors tend to have the least amount of faith in the
willingness and ability of the government to identify and continue incarcerating the most dangerous
prisoners and to apprehend them a second time if they are paroled include JOHN H. BLUME ET AL.,
Lessons from the Capital Jury Project, in BEYOND REPAIR? AMERICA'S DEATH PENALTY 144, 167
(Stephan P. Garvey ed., 2002); Ursula Bentele & William J. Bowers, How Jurors Decide on Death:
Guilt is Overwhelming; Aggravation Requires Death; and Mitigation is No Excuse, 66 BROOK. L.
REV. 1011, 1053 (2000); William J. Bowers et al., Foreclosed Impartiality in Capital Sentencing:
Jurors' Predispositions, Guilt-Trial Experience, and Premature Decision Making, 83 CORNELL L.
REV. 1476, 1499 (1998); William J. Bowers & Benjamin D. Steiner, Death By Default: An Empirical
Demonstration of False and Forced Choices in Capital Sentencing, 77 TEX. L. REV. 605, 675 (1999);
Craig Haney & Mona Lynch, Clarifying Life and Death Matters: An Analysis of Instructional
Comprehension and Penalty Phase Closing Arguments, 21 L. & HUM. BEHAV. 575, 579–80 (1997);


sentences for fear that incompetent parole officials will one day free the defendant
to offend again—a sentiment that might easily extend to the view that paying
prison and parole officials to exercise that discretion, not to mention paying to feed
and house killers, is the worst waste of tax dollars.101
We hypothesize that, taken together, these dispositions dispose libertarian
communities towards the death penalty. Suspicion of government and taxation
leads to weak and unprofessional law enforcement, which in turn generates low
clearance rates and a heightened fear of crime. Still not willing to pay and build
the institutions needed to respond systematically, libertarian communities turn to
the death penalty as an extreme, but only episodic and after-the-fact response that
can (indeed, constitutionally must102) be limited to the small number of offenses
that result in capture and conviction and involve a homicide. The death penalty
thus provides libertarian communities with a starkly expressive and retributive
substitute for routine, ongoing and professional surveillance, apprehension,
conviction and incarceration across the run of all serious crimes. Through the
death penalty, libertarian communities can express their abhorrence for the most
serious of transgressions against their and their fellows‘ autonomy without having
to pay the price for effective law enforcement, courts and corrections.
There are numerous risks and ironies here. By leaving criminals at large, the
reluctance of libertarian communities to pay for professional law enforcement
directly threatens the autonomy of law-abiding citizens that libertarians hold so
dear. The communities‘ reluctance to pay for incarceration and correctional
resources likewise leads directly to the wide-open parole practices that in turn
dispose libertarian jurors to doubt the government‘s will and ability to secure
dangerous criminals. Worse, the slapdash law enforcement practices and inferior
results that characterize the low levels of law-enforcement effort preferred by
libertarian communities may increase their disposition to impose the death penalty
in cases where it is not appropriate. As we said in Broken System II, public
pressures generated by ―[w]ell-founded doubts about a state‘s ability to catch
criminals may lead officials to extend the death penalty‖ to cases in which the
evidence of guilt or desert of the death penalty is weak.103 In contrast:
[S]tates with relatively more effective non-capital responses to
James Luginbuhl & Julie Howe, Discretion in Capital Sentencing Instructions: Guided or
Misguided?, 70 IND. L.J. 1161, 1177 (1995).
See e.g., Bowers & Steiner, supra note 86, at 675; Luginbuhl & Howe, supra note 86, at
See supra note 72 and infra notes 107, 110 and accompanying text.
BROKEN SYSTEM II, supra note 27, at iii.


crime—i.e. arrest, conviction and imprisonment—may be under less
pressure than states with weaker law enforcement records to use the
death penalty. That in turn may dampen the penalty‘s use in weak cases
in which the temptation to use unreliable procedures is high.104
3. The death penalty as a libertarian and parochial tool for self-protection
Perhaps the biggest irony, however, is the disposition of libertarian
communities to tolerate—indeed, to esteem—the most powerful civilian exercise
of state power known to our society.105 One might rather expect libertarians to
deny government this most powerful of tools and to fear state incompetence and
overreaching in administering it. And, indeed, a passel of libertarians, including
George Will, Richard A. Viguerie, Pat Robertson, Oliver North and Bob Barr,
have expressed exactly these reservations about the penalty.106 As we note above,
however, there are reasons why the death penalty might appeal to libertarians as
the best, because most episodic, way for the state to wield the power to punish.
As we discuss further below, moreover, it is not pure libertarianism of the George
Will sort, but a mix of it and parochialism that characterizes communities that
frequently use the death penalty. This section describes the interaction of the two
Consider first how parochial and libertarian tendencies can feed on one
another. Parochialism indicates a strong apprehension of serious outsider
crime;107 the ineffective law enforcement that attends the libertarian preference for
low spending, weak institutions and aversion to surveillance likely aggravates this


Id. at 168–69.
Noting the same irony in regard to penalties other than the death penalty is FRANKLIN
171–74, 231–32 (2001) (concluding that mistrust of government is associated, counter-intuitively,
with a greater willingness to use extreme penal sanctions, such as California‘s mandatory minimum
punishments for three-time offenders, even though the punishments trigger more government control
and a larger public sector).
For collections of these statements, see, New Voices—Conservative Voices, DEATH
PENALTY INFO. CTR., (last visited
June 16, 2011); New Voices: Prominent Conservative Calls for Death Penalty Moratorium, DEATH
(last visited July 11, 2011); BROKEN SYSTEM II, supra note 27, at 4.
See supra notes 57–70 and accompanying text.


fear.108 Indeed, a feature of libertarian communities in relatively close proximity
to each other may be the movement of offenders from one community to the other
as the simplest way to escape the unsophisticated law enforcement efforts that are
likely to prevail where they last committed a crime.
Indeed, this mixture of conspicuous fear of outsider crime, demand for harsh
punishment when the community is invaded and autonomy is threatened, and
doubts about government‘s intestinal fortitude when it comes to punishment can be
a toxic brew. Historically, one result in the United States has been lynching. In
the face of doubts about officials‘ willingness or ability to exact the retribution
demanded by parochial and libertarian under siege from outside forces, the public
stood ready to take matters in their own hands.109 The echo of this same impulse
in the disposition of modern death-prone jurors to doubt the willingness or ability
of officials to exact deserved punishment is, we suggest, not accidental but a
manifestation of the same parochial and libertarian proclivity for death as the surer
and more expressive punishment in a context of fear of outsiders and weak state
There are also reverberations in modern practice in high-death-sentencing
communities of lynchers‘ low regard for the administrative and professional
niceties of investigation, prosecution, trial and correction and the tendency to see
death verdicts and executions as a quick and unvarnished communal expression of
retribution for a heinous cross-boundary crime. Extending the parallel still
further, as lawless and vicious as lynching was, it often carried with it a patina of
democratic sanction, as a paroxysm of communal, and community-protecting
violence. 111 Albeit now with due process, judicial supervision and decorum
restored, a jury‘s public pronouncement of a death sentence has the same quality of
a democratic expression of community comeuppance of the offending outsider.
We are not the first to draw these parallels.


Professor Zimring, for example,

See supra notes 73–78 and accompanying text.
See GARLAND, supra note 2, at 172 (describing how, during the nineteenth century,
―Southern whites claimed this private power [over law enforcement represented by lynching] as a
means to maintain white supremacy‖).
See supra notes 69–70, 85–90.
See GARLAND, supra note 2, at 33 (noting that lynching was fuelled by ―the furious
demands of ‗popular justice‘ that emerge when local majority sentiment is outraged by crime‖ and
concluding that ―[t]he self-righteous power of ‗the people‘, emboldened by ideologies of popular
democracy and myths of self-rule, is an incendiary force in American politics‖).


has demonstrated a telling geographical overlap between today‘s high frequency
death-imposing counties and those that frequently carried out lynchings in the
past.112 Even more to our point, Professor Garland has noted the paradox that
lynching and the death penalty are simultaneously opposites and the same and has
pointed to the parochialism and libertarianism of communities that have favored
both practices as an explanation. In his words, the modern death penalty is
consciously designed as an ―antilynching,‖113 yet acts as ―a radical inversion of
form, a mirror image‖ of the lynchings of yore.114 On the one hand:
[e]xecutions, if they actually occur, take place not in the local town
square but instead at a great distance from the crime, both in time and in
space. Executions methods are avowedly ―non-violent,‖ designed to
minimize bodily injury and degradation. Bureaucratic protocols dictate
a dispassionate administrative routine with crowds, ceremony, and
cruelty reduced to a minimum.115
Yet, ―[c]ontemporary capital punishment continues to have many
substantive features in common with those lynchings,‖ that it tries to but
cannot disavow:
It continues to be driven by local politics and populist politicians. It
continues to be imposed by leaders and lay people claiming to represent
the local community. It continues to give a special place to victims‘
kin. It continues disproportionately to target poorly represented blacks,
convicted of atrocious crimes against white victims. . . . Finally, the
collective killing of hated criminals (or merely assertion of the right to do
so) remains one of the ways in which groups of people express their
autonomy, invoke their traditional values, and assert their local
We agree that, as different as the two practices are, lynching and capital
punishment are responsive to the same parochial and libertarian impulses towards
cataclysmic expression of a community comeuppance of perceived outside threats.


ZIMRING, supra note 19, at 89–118; see also OGLETREE & SARAT, supra note 19, at7–8.
GARLAND, supra note 2, at 34.
Id. at 35.


We simply add that when parochialism and libertarianism coincide, a
temptation—or even a felt necessity—arises to invoke the starkly retributive,
unvarnished and ostensibly deterrent force of the death penalty to compensate for
the crime-solving shortcomings of localities disposed towards meager spending on
law enforcement. The death penalty, that is, substitutes as a locally reassuring
and externally intimidating demonstration of the community‘s disposition to
defend itself against criminal invasions where pervasive and effective law
enforcement is considered unaffordable or unpalatable.117
C. A Parochial and Libertarian Explanation of the Resilience of the
Felony-Murder Doctrine
1. Felony murder‘s puzzling resilience
Further support for our thesis may be found in the puzzling resilience of the
Capital-Felony-Murder Doctrine in American criminal jurisprudence.118 Waves
of reform, typified by the Model Penal Code, have sought to constrain the
application of the criminal law, particularly the law of homicide, by using mental
states specified for each element of the crime to mark lesser and greater gradations
of culpability and desert of punishment.119 In this sea of precise mental states tied

Cf. James M. Inverarity, Populism and Lynching in Louisiana, 1889-1896: A Test of
Erikson‘s Theory of the Relationship Between Boundary Crises and Repressive Justice, 41 AM.
SOCIOLOGICAL REV. 262, 264 (1976) (―[M]any prominent white Southerners . . . viewed lynching as a
necessary and legitimate adjunct of the legal system . . . .‖).

[F]elony murder . . . is one of the most widely criticized features of American criminal
law. Legal scholars are almost unanimous in condemning [it] as a morally indefensible
form of strict liability, . . . an anomaly, [and] a primitive relic of medieval law that
unaccountably survived the Enlightenment and the nineteenth-century codification of
criminal law . . . [and most believe there is] no way to rationalize its rules . . . and no
reforms worth urging . . . short of its utter abolition. [The] author of the leading
criminal law textbook, called the felony murder doctrine ―rationally indefensible,‖ and
the American Law Institute's Model Penal Code Commentaries observed that
―[p]rincipled argument in favor of the felony-murder doctrine is hard to find.‖
Guyora Binder, The Culpability of Felony Murder, 83 NOTRE DAME L. REV. 965, 966–67 (2008)
(citations omitted). But see id. at 966, 1032–46 & n.4 (defending the felony murder rule as
expressive of the community‘s reaction to harm caused in the course of felonies and citing other
intermittent defenses of the rule). Felony murder has long been abolished in other Anglo-American
jurisdictions. See DAVID LANHAM, ET AL., CRIMINAL LAWS OF AUSTRALIA 180 (2006) (discussing
the felony murder doctrine‘s abolition through the United Kingdom‘s Homicide Act of 1957).
Dannye Holley, The Influence of the Model Penal Code‘s Culpability Provisions on State
Legislatures: A Study of Lost Opportunities, Including Abolishing the Mistake of Fact Doctrine, 27


to each element of a crime, the felony murder doctrine is an anomalous island,
given its omission of any culpability requirement as to the element of death. The
doctrine is controversial because of the alchemy it uses to manufacture a higher
culpability level and punishment than otherwise would be allowed.120 Its main
use is to transform an accidental killing in the course of a felony—which normally
would merit at most a manslaughter charge based on the inference that the felon
must have recognized the risk of death—into the equivalent of purposeful
murder. 121 If the predicate felony is one of the ―big four‖—robbery, rape,
kidnapping or burglary—the same alchemy transforms the manslaughter into
first-degree murder, which otherwise would require premeditation and
2. Felony murder‘s symbiotic relation to the death penalty
The Felony Murder Doctrine is associated with the broad—and
overbroad—use of the death penalty 122 and with high rates of error and
SW. U.L. REV. 229, 229–30 (1997) (noting that ―[o]ne of the greatest changes made by the Model
Penal Code was its redefinition and reformulation of culpable mental states . . . . The Code
obliterated ill-defined, confusing common law language and concepts and replaced them with four
specifically defined hierarchical levels of culpability…‖)
H. PILLSBURY, JUDGING EVIL 106–08 (1998); Charles Liebert Crum, Causal Relations and the
Felony-Murder Rule, 1952 WASH. U. L.Q. 191, 203–10; George P. Fletcher, Reflections on
Felony-Murder, 12 SW. U. L. REV. 413, 415–16 (1981); Martin R. Gardner, The Mens Rea Enigma:
Observations on the Role of Motive in the Criminal Law Past and Present, 1993 UTAH L. REV. 635,
706–08; Sanford H. Kadish, Foreword: The Criminal Law and the Luck of the Draw, 84 J. CRIM. L.
& CRIMINOLOGY 679, 695–97 (1994); H.L. Packer, Criminal Code Revision, 23 U. TORONTO L.J. 1,
3–4 (1973); Nelson E. Roth & Scott E. Sundby, The Felony-Murder Rule: A Doctrine at
Constitutional Crossroads, 70 CORNELL L. REV. 446, 451–52, 490–91 (1985); Stephen J. Schulhofer,
Harm and Punishment: A Critique of Emphasis on the Results of Conduct in the Criminal Law, 122
U. PA. L. REV. 1497, 1498–99 & n.2 (1974); see also Guyora Binder, Making the Best of Felony
Murder, 91 BOSTON UNIV. L. REV. 403, 404–05 & nn.1–4 (2011) (―The felony murder doctrine . . . is
part of the law of almost every American jurisdiction[, but] . . . is also one of the most widely
criticized features of American criminal jurisdiction‖) (citing authority).
See, e.g., State v. Middlebrooks, 840 S.W.2d 317, 336 (Tenn. 1992) (―The result of the
felony murder doctrine in Tennessee is thus to impose a rule of strict liability allowing the underlying
felonious intent to supply the required mens rea for the homicidal actus reus and to impose vicarious
liability for the acts of another.‖), cert. dismissed, Tennessee v. Middlebrooks, 510 U.S. 124 (1993);
Kevin Cole, Killings During Crime: Toward a Discriminating Theory of Strict Liability, 28 AM.
CRIM. L. REV. 73, 74 (1990) (―When a criminal kills a victim during the course of a robbery, for
example, the criminal is guilty of murder by operation of the felony-murder rule without any inquiry
into the killer‘s mental state respecting the killing.‖).
See, e.g., Tison v. Arizona, 481 U.S. 137, 158 (1987) (placing constitutional limits on
imposition of the death penalty on minor participants in felonies in which a homicide occurs);
Enmund v. Florida, 458 U.S. 782, 797 (1982) (setting somewhat stricter limits on the death penalty


arbitrariness in its administration.123 This is not surprising, given the doctrine‘s
use, in another inexplicable feat of legal magic, to remove constitutional obstacles
to a death sentence that otherwise would clearly bar it.
In Gregg v. Georgia, after a decade and a half of uncertainty about the
constitutionality of the death penalty,124 the Supreme Court concluded that ―when
a life has been taken deliberately by the offender, we cannot say that the
punishment is invariably disproportionate to the crime. It is an extreme sanction,
suitable to the most extreme of crimes.‖125 To assure, however, that the ―extreme
sanction‖ did not become ―disproportionate‖ through its application to defendants
for minor participants in felonies in which a homicide occurs, later modified by Tison, supra);
Lockett v. Ohio, 438 U.S. 586, 608 (1978) (plurality opinion) (holding that Ohio unconstitutionally
limited the aggravating circumstances capital jurors could consider, including by barring the
defendant from arguing in favor of a sentence less than death based on her limited participation in a
felony in which a man was killed); Woodson v. North Carolina, 428 U.S. 280, 301 (1976) (plurality
opinion) (concluding that the mandatory death penalty for all participants in a felony murder,
including the wheel person who was outside during the crime, violated the Eighth Amendment); see
also Richard A. Rosen, Felony Murder and the Eighth Amendment Jurisprudence of Death, 31
BOSTON COL. L. REV. 1103, 1128 n.64 (1990) (collecting studies finding that felony murders
constitute anywhere from 48% to 83% of capitally eligible or capitally punished crimes); ILLINOIS
For example, the Illinois Governor‘s Commission on the Death Penalty recommended
eliminating the felony-murder aggravating factor from the Illinois capital punishment statute.
―Commission members [concluded] that this eligibility factor swept too broadly and included too
many different types of murders within its scope . . . .‖ ILLINOIS DEATH PENALTY COMMISSION
REPORT, supra note 107, at 72 (2002). Because ―so many first degree murders are potentially death
eligible under this factor, it lends itself to disparate application throughout the state . . . [and is the]
eligibility factor . . . most likely subject to . . . discretionary decision-making‖ by prosecutors and
jurors. Id.; See id. at 3 (finding that even ―[a]fter eliminating those cases in which the ‗multiple
murder‘ factor and the ‗course of a felony‘ factor appear together, the ‗course of a felony‘ eligibility
factor accounts for just over 40% of the cases in which the death penalty has been imposed‖); see
also David C. Baldus et al. Arbitrariness and Discrimination in the Administration of the Death
Penalty: A Challenge to State Supreme Courts, 15 STETSON L. REV. 133, 138 n.14 (1986) (collecting
studies); David McCord, Should Commission of a Contemporaneous Arson, Burglary, Kidnapping,
Rape, or Robbery Be Sufficient to Make a Murderer Eligible for a Death Sentence?—An Empirical
and Normative Analysis, 49 SANTA CLARA L. REV. 1, 1 (2009) (finding that over 60% of defendants
committed murder in the course of one of five predicate felonies, triggering death eligibility and
criticizing the felony murder aggravating factor as overbroad as currently applied across the United
States); Chelsea Creo Sharon, Note, The ―Most Deserving‖ of Death: The Narrowing Requirement
and the Proliferation of Aggravating Factors in Capital Sentencing Statutes, 46 HARV. CIV.
RTS.-CIV. LIB. L. REV. 223, 234 (2011) (criticizing the felony murder aggravating factors in capital
sentencing statutes nationwide due to its ―extraordinary breadth‖).
See MELTSNER, supra note 45, at 149–287.
Gregg v. Georgia, 428 U.S. 153, 187 (1976) (plurality opinion) (emphasis added).


guilty of something less than ―the most extreme of crimes,‖126 the Court required
States to ―narrow‖ the circumstances in which the death penalty could be imposed
to killings more aggravated than the jurisdiction‘s base definition of capital
homicide (classically, premeditated and deliberate first-degree murder).127 To
satisfy this narrowing requirement, a jury must find at least one statutorily
enumerated aggravating factor that distinguishes the killing from the jurisdiction‘s
definition of capital homicide.128 An aggravating factor that applies to ―every
first-degree murder‖ does not suffice because it does ―not narrow the class of
defendants eligible for the death penalty.‖129
Although the Court has never expressly deviated from these principles,
inexplicably, it has never applied them to felony murder cases. To begin with,
felony murder deviates from the category of homicide to which Gregg ostensibly
limited the death penalty and from the type of culpability otherwise required for
first-degree murder—where ―life has been taken deliberately by the offender.‖130
The whole point of felony murder is to permit murder-liability though the killing
was accidental and not deliberate.131 Yet, the Supreme Court has not hesitated to
affirm non-deliberate felony murder as a sufficient basis for the death penalty.132


Id. at 196–97; see Arave v. Creech, 507 U.S. 463, 474–75 (1993) (noting that the Eighth
Amendment requires States to ―genuinely narrow the class of persons eligible for the death penalty‖
(quoting Zant v. Stephens, 462 U.S. 862, 877 (1983)); Godfrey v. Georgia, 446 U.S. 420, 430 (1980)
(overturning death sentence because Georgia failed to interpret the relevant statutory aggravating
circumstance to assure that ―[e]ach of the cases [in which death was imposed] is at the core and not
the periphery‖ of the category of death-eligible murders‖ (citations omitted)); Pulley v. Harris, 465
U.S. 37, 50 (1984) (describing the ―constitutionally necessary narrowing function‖); Gregg v.
Georgia, 428 U.S. 153, 194, 197–98.
See, e.g., Arave v. Creech, 507 U.S. 463, 474–75; Lewis v. Jeffers, 497 U.S. 764, 775
(1990); see also Ring v. Arizona, 536 U.S. 584, 606, 609 (2002) (describing the Court‘s Eighth
Amendment jurisprudence regulating the death penalty as having ―interpreted the Constitution to
require the addition of an element or elements to the definition of a criminal offense in order to
narrow its scope‖ and holding that the Constitution requires that element to be found by a jury
beyond a reasonable doubt).
Arave v. Creech, 507 U.S. 463, 475 (emphasis added).
Gregg v. Georgia, 428 U.S. 153, 187 (emphasis added).
See, e.g., People v. Brackett, 510 N.E.2d 877, 882 (Ill. 1987) (affirming conviction for
felony murder where, as a result of a rape, the victim refused to eat and accidentally choked to death
on a feeding tube inserted by her doctors); People v. Hickman, 297 N.E.2d 582, 586 (Ill. 1973)
(affirming conviction for felony murder where one police officer accidentally shot another during a
manhunt for the unarmed burglars).
See, e.g., Schad v. Arizona, 501 U.S. 624, 644 (1991); see also State v. Middlebrooks, 840
S.W.2d 317, 337 (Tenn. 1992) (―[T]he vast majority of states that have the death penalty permit it to


In addition, one of the most common aggravating factors enumerated in state
capital sentencing statutes is that the killing occurred in the course of one of the
―big four‖ felonies of robbery, rape, kidnapping or burglary.133 And a number of
states allow death sentences when the accompanying felony is the only statutory
aggravating factor present in the case. 134 In other words, after using felony
murder to raise an unintentional killing in the course of a felony that otherwise
would qualify as no more than manslaughter to murder, and then—if the felony is
one of the ―big four‖—using the same magical reasoning to raise the offense to
capital-eligible, first-degree murder, many high death-sentencing states allow the
prosecution to use the same felony as the ostensible additional factor that
aggravates the offense to a category ―narrow[er]‖ than first-degree murder.135
Through this sleight of hand, every killing in the course of a ―big four‖ felony
automatically qualifies for the death penalty.
A few state courts have ruled as a matter of local law that a felony cannot
provide both the sole basis for first-degree murder and the sole aggravating factor
in support of a death sentence. 136 But when the Tennessee Supreme Court
premised a similar ruling on the Federal Constitution‘s capital narrowing
requirement, the Supreme Court promptly granted certiorari, as if to reverse.137
be imposed in cases of felony murder under some circumstances.‖ (citations omitted)), cert.
dismissed, Tennessee v. Middlebrooks, 510 U.S. 124 (1993).
See Rosen, supra note 107, at 1120, 1125–29 & nn.41, 62, 63, 65 (noting that ―[a] large
majority of states that enacted new death penalty laws after Furman [v. Georgia, 408 U.S. 238
(1972)] retained both . . . one form or another of the felony murder rule[] as [a] bas[i]s for a
conviction of capital murder,‖ and many of them also recognize a killing in the course of robbery,
rape, burglary and kidnapping as an aggravating factor supporting the imposition of death (collecting
See Rosen, supra note 107, at 1127, 1134–35 nn.62, 63, 79, 83, 86 (citing decisions).
Arave v. Creech, 507 U.S. 463, 475 (emphasis added); see Rosen, supra note 107, at
1121–24, 1135 (noting that ―defendants continue to be sentenced to death solely because they
committed a murder during the course of a felony, that is, simply because they fit into a class of
murder defendants that, in some states, is no narrower than‖ before the Supreme Court imposed the
narrowing requirement in Gregg and subsequent cases).
See State v. Middlebrooks, 840 S.W.2d 317, 346 (―We hold that, when the defendant is
convicted of first-degree murder solely on the basis of felony murder, the [state‘s felony murder]
aggravating circumstance . . . does not narrow the class of death-eligible murderers sufficiently under
the . . . Tennessee Constitution because it duplicates the elements of the offense.‖); Engberg v.
Meyer, 820 P.2d 70, 89–92 (Wyo. 1991) (―We now hold that where an underlying felony is used to
convict a defendant of felony murder only, elements of the underlying felony many not again be used
as an aggravating factor in the sentencing phase.‖); State v. Cherry, 257 S.E.2d 551, 566–68 (No.
Car. 1979) (prohibiting use of a felony murder aggravating circumstance if the defendant had been
convicted solely of felony murder), cert. denied, 446 U.S. 941 (1980).
Tennessee v. Middlebrooks, 507 U.S. 1028 (1993) (per curiam) (granting certiorari to


The Tennessee Supreme Court subsequently made clear that it reached the same
conclusion under state law, and the Supreme Court dismissed certiorari as
improvidently granted.138 But the Court has repeatedly denied certiorari and
permitted states to execute death sentences premised on accompanying felonies
that were the sole basis for a murder conviction and the single statutory
aggravating circumstance.139 Felony murder cases thus are the lone category in
which the Supreme Court tolerates death sentences absent any ―narrowing‖ of
death-eligibility beyond the minimal threshold for first-degree murder.140
The extent of the anomaly here is revealed by comparing how existing law
currently applies to two hypothetical defendants: Bill, who purposefully robs Vic
at gunpoint; and Bob, who purposefully kills Vic with a gun. If no one dies in the
course of Bill‘s robbery, Bill may not constitutionally be executed despite his
purpose to rob Vic.141 But if Vic accidentally knocks into Bill‘s gun, causing it to
fire and kill Vic, then without more, Bill can be executed for the accidental killing.
The felony mysteriously elevates the manslaughter to murder, then to first-degree
murder, then to aggravated capitally punishable murder. If, however, Bob
interrupts the robbery in order to purposefully shoot Vic dead, Bob cannot without
more be executed for intentionally killing Vic.142 Although, unlike Bill, Bob
review State v. Middlebrooks, supra).
Tennessee v. Middlebrooks, 510 U.S. 124, 125 (1993) (per curiam).
See Rosen, supra note 107, at 1126–1127 nn. 62–63, 1134–35 nn. 79, 83, 86 (citing
Some observers understand the Supreme Court‘s decision in Lowenfield v. Phelps, 484 U.S.
231 (1988), to embody a holding that approves the questionable practice we describe the Court as
merely having tolerated by thus far refusing to forbid it. See, e.g., JOHN KAPLAN ET AL., CRIMINAL
LAW: CASES AND MATERIAL 472 (6th ed. 2008) (―[Under Lowenfield], as a matter of federal
constitutional law, the Supreme Court would permit the [accompanying-felony aggravating
circumstance] ‗bonus‘ . . . where the underlying murder charge was solely based on a felony murder
ground . . . .‖). Lowenfield is to the contrary. It clearly reprises the requirement that state capital
statutes ―narrow‖ death eligibility to a category smaller than every first-degree murder, see
Lowenfield, 484 U.S. at 244; finds that narrowing to be generally satisfied by the list of aggravating
factors that Louisiana uses to distinguish ―first-degree murder,‖ for which the death penalty is not
available, from ―capital murder‖ for which death may be imposed, id. at 246; and concludes that
narrowing occurred in the particular case as a result of the jury‘s finding that the killing not only was
premeditated and deliberate, thus establishing first-degree murder, but also that it was committed in a
manner that created a ―great risk of death to more than one person,‖ thus elevating the crime to
―capital murder,‖ id. at 245. What seems to have confused the authors of KAPLAN ET AL., supra, is
that the jury found the aggravating factor at the guilt phase, not the separate sentencing phase, but
that procedural detail in no way undermines the Court‘s usual requirement (except, inexplicably, in
felony murder cases), of proof of a circumstance that makes the offense more aggravated than bare
first-degree murder.
Kennedy v. Louisiana, 554 U.S. 407, 413 (2008) (strongly suggesting that interpersonal
crimes short of homicide may not constitutionally be punished with death).
This example gives the lie to the traditional ―transferred intent‖ explanation of felony


clearly meant to end Vic‘s life, no aggravating factor is present, and the law will
not magically manufacture one.
3. Capital Felony Murder as a Tool for Parochial Self-protection
What, then, is the clandestine ―compelling state interest‖ that trumps
fundamental legal values that otherwise would have buried felony murder long
ago, and that makes a purpose to commit a felony-less-than-homicide and an
accidental death more worthy of execution than a purpose to kill and a resulting
The solution to this legal conundrum, we believe, lies in the
Felony-Murder Doctrine‘s symbiotic relation to the death penalty and that
penalty‘s appeal to the minority of parochial and libertarian localities that
frequently use it.
The key clue to the felony-murder riddle is the ―big four‖ predicate
felonies—robbery, rape, kidnapping and burglary—particularly the last one,
burglary. The predicate felonies often are explained, and the felony-murder rule
sometimes is said to be justified, because those felonies are so dangerous in and of
themselves or so likely to be committed with a dangerous weapon.143 But neither
of these things is true of burglary.144 Other offenses, such as car-jacking and
various drug and gun-related crimes, are much more dangerous than burglary yet
are not standard predicates for the felony bump-up to first-degree murder or the
accompanying-felony bump-up to aggravated, death-eligible murder. What does
set all of the big four felonies apart from others is how effectively they serve as
proxies for ―stranger‖ or ―invasion‖ crimes. Both interpersonally in the physical
manner of their commission and sociologically in the way they manifest
themselves to members of privileged communities, the ―big four‖ offenses cross
boundaries, come frighteningly ―out of the blue‖ and are the handiwork of
murder as substituting the intent to rob for the intent to kill. See, e.g., KAPLAN ET AL., supra note
140, at 423 (listing justifications for the felony-murder rule). If an intent to kill and a resulting dead
body cannot, by themselves elevate a killing to first-degree murder and automatic capital eligibility,
how can an intent to rob and a resulting dead body do so? Nor for this same reason, and another
given at notes 128–132 and accompanying text below, can deterrence of felonies that are intrinsically
dangerous or committed in a dangerous manner account for the felony-murder alchemy that turns
accidental homicide into first-degree and capital-eligible homicide.
See, e.g., Binder, supra note 118, at 1045 (arguing that felony murder is justified as long as
it is limited to predicate felonies ―inherently involving violence or destruction,‖ which excludes
burglary and drug offenses but includes such crimes as robbery); Kenneth W. Simons, When Is Strict
Criminal Liability Just?, 87 J. CRIM L. & CRIMINOLOGY 1075, 1121–24 (1997) (defending felony
murder liability predicated on inherently dangerous felonies).
See, e.g., KAPLAN ET AL., supra note 140, at 424 (presenting data showing that deaths occur
in only fewer than 4 out of every 100,000 burglaries, compared to 10 deaths per 100,000 for auto
theft, 350 per 100,000 for rape, and just under 600 per 100,000 for robbery).


frightening outsiders.145
The surreptitiously compelling interest that explains the law‘s otherwise
inexplicable treatment of accidental deaths in the course of ―big four‖ felonies as
worse-than-purposeful, capitally aggravated murder is, we suggest, the same
impulse that leads parochial communities to value the death penalty itself: a
powerful fear of cross-boundary crime. Both tools—and in particular their
frequently conjoined use146—provide law enforcement officials with a powerful
device for assuaging fears and communicating how seriously they take and how
harshly they are prepared to punish outsider crime.147
Understanding Felony-Murder Doctrine as a response to cross-cultural,
invasion crimes explains a number of the doctrine‘s bizarre nuances. When an


EXPERIENCE 1960-1975, at 7–8, 10–14, 52–53 (1981) for this sociological insight. Of course, most
robberies, rapes, kidnappings and burglaries are committed by and against poor individuals from the
same poor neighborhood. But it should not surprise us to find, and the Broken System II results bear
out, that anomalies in the criminal law are most strongly influenced by the most politically powerful,
not the most numerous, of victims. See KENNEDY, supra note 72, at 3–4; Stuntz, supra note 78, at
1984–85; see supra notes 64–72 and accompanying text. It is also true that other so-called
―inherently dangerous‖ crimes, beyond the ―big four,‖ serve as predicates for second- and even
first-degree felony murder. See KAPLAN ET AL., supra note 140, at 439–42. But here, too, the
―cross-boundary‖ concept described in text helps explain, if not justify, the notoriously arbitrary
judgments states and courts make about what crimes do and do not qualify as ―inherently dangerous‖
for purposes of being recognized as predicates for felony murder. See id. at 441–42 (collecting
cases identifying the following offenses as not inherently dangerous enough to be a predicate for
felony murder: fraud by chiropractors inducing patients with curable cancer to forgo surgery in favor
of futile treatments; possession of a firearm and an illegal sawed-off shotgun by a felon; reckless
firing of a gun illegally possessed by a convicted felon; and ruling that following offenses do qualify
as ―inherently dangerous‖ predicates to felony murder: theft by tow-truck of an automobile from a
deserted used car lot (the intoxicated perpetrators crashed the tow truck into another car on the
highway) and furnishing heroin).
See supra note 108, 118, 120. This symbiosis between felony murder and the death
penalty is further evidenced by the United Kingdom‘s abolition of the felony-murder doctrine and the
death penalty around the same time—in 1957 in the former case and 1965 in the latter. See, Murder
(Abolition of Death Penalty) Act, 1965, NATIONAL ARCHIVES. (1965 Chapter 71),
To a substantial extent, we agree with other observers who recognize a deeply ―expressive‖
connotation to the Felony Murder Doctrine. See, e.g., Binder, supra note 103, at 1032. What we
add is a description of what is being expressed—a fear not of especially dangerous or violent crimes
but of outsider, cross-boundary crimes—a fear, we understand, but unlike the expressivist apologists
for felony murder, we do not consider sufficient to justify the doctrine‘s departure from
jurisprudential fundamentals and privileging of some communities over others.


innocent person, not a party to the felony, kills a co-felon in the course of the
crime, the trend is against finding the surviving felon guilty of felony murder.148
Some states exclude killings committed by anyone other than a felon.149 Other
states exclude all killings of co-felons.150 Only a handful of states make any
foreseeable killing of anyone by anyone during a felony a basis for felony murder
liability.151 Commentators struggle to explain these rules.152 The doctrine itself
refuses to acknowledge the standard reasons for exculpating defendants for harm
for which they were a ―but for‖ cause—that they did not advert to or proximately
cause the harm. Yet, through these, the doctrine does grant mercy based on
fortuities that typically make no difference and violate the usual rule that, formally
at least, the law treats all victims the same.153
Our analysis again solves the riddle. The trend is to impose excessive
punishment—measured by rules that apply in all other cases—only when the
victim is a member of the community and the killer is not. The compelling
interest that surreptitiously trumps the normal rules arises only when the identity of
the killer and victim make the death a cross-boundary offense against a member of
the local community. In truth, the Felony-Murder Doctrine does not target
dangerous felonies in the course of which someone dies; it targets stranger crimes
in the course of which an outsider kills a member of a privileged community.
Consider, as well, the so-called ―merger rule‖ limiting felony murder. If the
underlying felony is an aggravated assault in which the assailant intends to scare
the victim with a gun, which accidentally fires killing the victim, the crime does
not qualify as a felony murder in most states and is not first-degree murder or
capital-eligible.154 Ostensibly, this is because the act that killed the victim and the

See, e.g., WAYNE R. LAFAVE, CRIMINAL LAW 793–94 (5th ed. 2010) (―Although it is now
generally accepted that there is no felony murder when one of the felons is shot and killed by the
victim, a police officer, or a bystander, it is not easy to explain why this is so.‖).
Famous holdings to this effect include People v. Washington, 44 Cal. Rptr. 442, 445 (Cal.
1965), and Commonwealth v. Redline, 137 A.2d 472, 476 (1958).
See, e.g., State v. Canola, 374 A.2d 20, 30 (N.J. 1977).
See LAFAVE, supra note 133, at 793–96.
See supra note 133.
See, e.g., MODEL PENAL CODE AND COMMENTARIES § 3.02 (1985) (noting principle that, for
purposes of balancing of evils recognized by the criminal law, all victim‘s lives are valued the same).
Indeed, if the defendant was not grossly reckless, for example, because she thought the gun
was unloaded or a toy, the crime does not even constitute bare murder. For discussion of the merger
rules, see, e.g., KAPLAN ET AL. supra note 140, at 442–48; LAFAVE, supra note 133, at 803–05.


one constituting the predicate felony ―merge‖ into a single act. But so what? If
every burglary resulting in an accidental killing of the homeowner is the moral
equivalent of deliberate, first-degree, capitally aggravated murder, then why is not
every aggravated assault resulting in an unintended killing its moral equivalent as
well?155 In both cases, the defendant caused a death, and intending to frighten
someone with a gun is considerably more dangerous and evil than a cat burglar‘s
plan to break into a home and steal. And why, as we have already asked, does the
law treat an intent to kill constituting second-degree murder less culpable and
worthy of the death penalty, than an accidental killing in a cat burglary? Given
our analysis, the answer is simple: regardless of its potential dangerousness, assault
rarely is a stranger crime and is not expressive of cross-class or cross-cultural
invasion.156 Burglary is both.
4. Capital Felony Murder as a Tool for Parochial and Libertarian
Parochialism helps explain the Felony Murder Doctrine and why insular
communities demand extra punishment, especially death, for cross-cultural crime.
Yet, if punishing cross-cultural crime is the goal, why stop at felonies resulting in
death? Why not punish all robbers, rapists, kidnappers and burglars with death?
The answer, we think, is that felony murder, like the death penalty itself, is
also a libertarian rationing device. Libertarian communities can be expected to
resist the massive expansion government infrastructure and taxes that such a policy
would require. By using the Felony Murder Doctrine to enable it to reserve the
harshest, most expressive punishment for the occasional cross-cultural crime that
results in the most tragic result, these communities can make an example of an


Or put the other way around, if an assault that accidentally turns deadly merges, why does
not an armed robbery in which the victim is killed merge, too, given that armed robbery is simply an
assault with a deadly weapon (which merges) plus larceny (which is not a recognized predicate
crime)? See People v. Burton, 491 P.2d. 793, 801 (Cal. 1971) (rejecting this argument on the
ground that the intended robbery is ―independent‖ of the killing, but an intended assault is not). The
illogic of the merger rule has been clear from the beginning. See State v. Shock, 68 Mo. 552, 574
(1878) (Norton, J., dissenting) (―If B starts out with a fixed felonious purpose to ‗inflict great bodily
harm‘ on A . . . without intending to kill but to stop with infliction . . . harm and death ensues, the
felony committed in inflicting the great bodily harm is no more merged in the killing than would a
rape perpetrated by B upon A, which resulted in the death of A . . . .‖).
The assault, battery and child abuse predicate crimes that are most commonly held to
―merge‖ with the killing, and thus to provide no basis for felony murder alchemy, are offenses that
classically and typically afflict (usually poor) acquaintances and family members. See cases cited in
KAPLAN ET AL. supra note 140, at 442–48; LAFAVE, supra note 133, at 803–05.


occasional stranger-criminal and express how seriously they take stranger crime,
without actually having taken it too seriously most of the time.
To be sure, from a libertarian perspective, both the Felony Murder Doctrine
and punishment itself impose costs—felony murder violates the principle that we
measure evil by the extent to which individuals choose to interfere with the
autonomy of others; punishment withdraws the criminal‘s autonomy while
empowering the state in ways that threaten everyone‘s freedom. Rationing is
important, therefore, as a way to balance the autonomy harms that crime and
punishment inflict on innocent and guilty individuals. Using the victim‘s death to
mark the rare occasions on which the law most harshly and retributively
intervenes, and using the criminal‘s death to express how seriously the community
regards the evil of intruding on a victim‘s autonomy, makes libertarian sense.
What libertarianism by itself cannot explain, however, is the use, not of any
victim‘s intended death, but of a particular victim‘s even accidental death in the
course of a ―big four‖ felony to mark the point where harshly retributive state
intervention, and the extinguishing of the offender‘s life, become tolerable. It
takes the parochial impulse to explain that aspect of the rationing device.
On the other hand, were parochialism the only important influence, one would
expect insular communities to erect other bulwarks than the death penalty against
depraved and dangerous outside influences. They might, that is, respond to their
fear of outsider crime with well-funded, professional police forces, high-powered
prosecutorial and judicial law enforcement mechanisms, and a disposition to do
whatever it takes to put all robbers, rapists, kidnappers and burglars behind bars for
life. Yet, as we have seen, the opposite is frequently the case: these jurisdictions
tend to suffer from some of the poorest quality, underfunded law enforcement
institutions. In this case, it takes the conjoining of the libertarian impulse to the
parochial one to explain the result.
D. The Death Penalty and the Illusion of Self-Protection
Based on the available data, we conclude that the small set of jurisdictions
that propel the modern American death penalty are driven by the combined
instincts of parochialism and libertarianism. These communities exhibit a fear of
outside influences that threaten the local values and experiences that set them off
from the national and global mainstream. Whites in these communities, who we
take to be a proxy for more privileged residents, tend to have high rates of
homicide victimization relative to the rates experienced by African-American
residents, and the white population tends to be located in close proximity to poor
and African American communities—factors we associate with a partly justified,
partly inflated apprehension of crime. These jurisdictions also tend to have more

populist electoral mechanisms through which fears of this sort can be powerfully
communicated to local officials. Taken together, these factors generate an
elevated fear of stranger crime on the part of privileged members of capitally prone
communities that is forcefully communicated to local officials whose jobs depend
on the adequacy of their response to the fear.
Ironically, however, the communities‘ crime fears are aggravated by their
parsimonious public spending habits and mistrust of government actors and
institutions. These factors lead to low-quality law enforcement, as indicated by
low rates at which the communities clear serious crimes by arrest, conviction and
incarceration; an inability to protect privileged residents from the kinds of
homicide rates that are common in poor and minority neighborhoods; and high
rates of serious error in the many capital verdicts the communities impose.
This peculiar combination of fear of cross-boundary crime and unwillingness
to support and systematize law enforcement leads these communities directly to
the death penalty. Capital punishment provides them with the seemingly
incongruent conditions they seek—a public, powerfully expressive and harshly
retributive—but, on the other hand, carefully rationed, inexpensive and
episodic—demonstration of the community‘s and officials‘ abhorrence for outsider
crime. Through the ultimate punishment, and the preference for it over
alternative responses to crime that rely more heavily on state institutions, the
communities have adapted and domesticated a vigilante tradition with deep roots
in their localities.
Given their parochial and libertarian proclivities, it is easy to see why
death-prone communities insist on preserving the anachronistic and doctrinally
dubious Capital Felony Murder Doctrine. Wielding this penological magic wand,
the communities can simultaneously make a lot more of and a lot less of the crimes
that particularly plague them. Using the proxy of the ―big four‖ felonies, the
communities can conjure successive strata of otherwise nonexistent culpability out
of the stranger crimes they abhor. Using the rare fortuity of a community
member‘s death in the course of one of those felonies, the localities can shrink the
body of crimes to which a concerted response is required to a number small
enough to avoid having to empower the state to respond systematically. Through
their retributive and expressive synergy, the Felony Murder Doctrine and the death
penalty enable parochial and libertarian communities to create the illusion of a
powerful response to the cross-boundary crimes that most frighten them without
having to empower or pay the state to provide it.
The death penalty‘s particular attraction to simultaneously parochial and
libertarian communities helps explain the penalty‘s markedly uneven use among

and within states and the puzzling resilience of the capital-felony-murder doctrine.
As of yet, however, we have shed little light on the other central paradox of the
modern American death penalty: why so few of the offenders we sentence to die
are actually executed? We turn next to that question.
A. The Paradox at the Heart of the System
The defining paradox of the American system of capital punishment is the
stark discrepancy between the number of people sentenced to die and the number
actually executed. As we demonstrated in the Broken System studies, an
American sentenced to die has about a two-to-one chance of having his death
sentence overturned on appeal or state post-conviction review.157 The decision
for community prosecutors and courts to ―go capital‖ is a gamble the house often
does not seem to win.
As high as the reversal rate is, it greatly overestimates the likelihood of
execution. After the jury imposes a sentence of death, the verdict enters a
multi-layered system of state and federal judicial review, which has proven
necessary to uncover the high number of reversible errors in capital cases.158 This
litigation takes, on average, twelve years before the appellate system is able to

See Gelman et al., supra note 15, at 216–17 (estimating the total error rate during the
1973–1995 study period as 68%); see also Michael O. Finkelstein & Bruce Levin, The Machinery of
Death, 18 CHANCE 34, 36 (2005) (―[The 68 percent figure in the Broken System] study has received
overwhelming academic approval and has been repeatedly cited by the federal courts and federal
judges, including Supreme Court justices.‖); Michael O. Finkelstein et al., A Note on the Censoring
Problem in Empirical Case-Outcome Studies, 3 J. EMPIRICAL LEGAL STUDIES, 375, 382 (finding after
recalculating the Broken System data that ―the complete case estimate of the reversal rate was 67.8
percent (which is quite close to the published estimate of 68 percent) and the self-consistent estimate
[after conducting additional statistical analyses to account for ―censoring‖ or the possibility of change
over time in regard to verdicts still under review when the study ended] was 62.2 percent‖).
See BROKEN SYSTEM II, supra note 27, at 69 (comparing the 40% reversal rate in the third
and last stage of review, after the preceding stages have already found serious error in and removed,
respectively, 41% of all capital verdicts entering the review process and 10% to 17% of verdicts that
survived the first round of review, and concluding that these numbers reveal both a high rate of error
and the need for multiple levels of inspection to remedy the problem); Liebman, supra note 33, at 82
(―The 41 percent-10 percent-40 percent pattern of reversal rates at the three successive review stages
does not exhibit the sharply downward trend of remaining flaws . . . that one expects in a fully
effective progression of inspections.‖)


sign-off on an execution.159 As a result, most death-row inmates are many years
away from being executed, and in the meantime, a number of them will die of
natural causes, commit suicide or be killed by other inmates.160
In Overproduction of Death, published in 2000, we illustrated this process by
tracking the outcomes of the 263 death sentences imposed nationwide eleven years
earlier in 1989. Of the 263 verdicts, 160 (61%) had not completed the review
process. Among the 103 verdicts that had reached the end of the appeal process,
78 (30% of those imposed; 77% of those finally reviewed) had been overturned by
a state or federal court based on a finding of serious legal error. Nine of the
condemned inmates died on death row while awaiting a decision. Only thirteen
were executed.161
The ordeal of state and federal review that every capital sentence must endure
in order to be cleansed of error and reach execution is a central feature of
American capital punishment today. A hefty majority of death sentences are
ultimately overturned due to serious flaws in the conviction or sentence. In the
process, all of them are suspended for well over a decade while the detection
process proceeds. The most poignant illustration of our broken system is that of
the 5,826 people sentenced to die between 1973 and 1999, only 313, roughly one
in eleven, were executed during that period.162 About four times as many had
their convictions overturned or were granted clemency. 163 Of these myriad cases
sent back for retrial at the second appeal phase (the only phase for which data are
available), 82% ended in sentences less than death, and 9% ended in not guilty
B. A New View, from Without
In deciding what to make of a legal regime that operates this way, there are
two perspectives an observer can take. The first is to analyze the regime on its


See BROKEN SYSTEM II, supra note 27, at 36.
See, e.g., Richard C. Dieter, Smart on Crime: Reconsidering the Death Penalty in a Time of
James S. Liebman, The Overproduction of Death, 100 COLUM. L. REV. 2030, 2055–56
See BROKEN SYSTEM II, supra note 27, at A–1.
Id. at 77–78.
See id. at i; Liebman, supra note 33, at 82.


own terms. From the law that governs the regime, one can discern its goals and
aspirations. With some empirical research, one then can determine whether the
regime‘s implementation accomplishes what is intended. The regime describes
its own logic, and the observer evaluates it based on how well it succeeds. This is
the view from within.165
The reality revealed by the Broken System findings, particularly the paradox
of an affinity for death sentences and aversion to executions, satisfies no one‘s
stated goals for the death penalty in America. No constitutional principle, statute,
judicial decision, politician, pundit or think-tank advocates a capital punishment
system in which the majority of verdicts are overturned and the average defendant
languishes on death row for over a decade while an exhaustive and expensive
review process grinds along to an uncertain outcome. 166 Yet, these conditions
more or less define the steady equilibrium that our capital system seems to have
reached.167 In light of the glaring discontinuity between policy and outcome, the
view from within fails to explain the stasis that characterizes our seemingly broken
A second perspective observes the regime from without. By taking a
birds-eye view of the pattern and practices through which the regime operates in
actual fact, as a whole, over an extended period of time, and without accepting
everything the system says about itself, this approach attempts to uncover a hidden
logic that cannot be discerned from within.168
In Albion‘s Fatal Tree, Douglas Hay famously took this second perspective,
puzzling over a capital system in which the ranks of those sentenced to death
swelled year upon year, but actually sent only one in five condemned prisoners to
the gallows.169 Stays and clemency were endemic to the system, triggering a


A good example of this approach is Steiker & Steiker, supra note 19.
See Liebman, supra note 19, at 4 n.3 (citing sources decrying the evident incoherence of
American death penalty doctrine and results); Governor Quinn‘s Abolition Statement, supra note 18
(giving Governor Quinn‘s reasons for signing legislation abolishing the death penalty in Illinois).
But cf. infra notes 332–359 and accompanying text (discussing changes in the nation‘s
death penalty equilibrium since 2000).
See, e.g., Greenberg, supra note 19.
See supra notes 5–6, 79–82 and accompanying text (discussing Hay); see also GATRELL,
supra note 32, at 7 (discussing outcomes of capital sentences between 1770 and 1830 and finding that
only about 20% were carried out).


glaring discrepancy between the penalty‘s imposition and its execution.170 From
all sides of the political spectrum, among courts and informed observers alike, the
capital system was an object of near universal derision for its seeming irrationality
and inability to accomplish its straightforward objectives.171 Yet, despite this, the
system proved exceptionally stable over time, resisting numerous calls for reform.
Although Hay wrote about the death penalty in eighteenth century England, his
vivid description of a seemingly broken capital system has eerie echoes in the
American death penalty of today.
Hay‘s view from without broke from previous scholarship of the regime by
noteworthy observers such as Blackstone, Radzinowicz and Romilly, who had
examined the system from within.172 Hay rejected the belief—unanimous from
the eighteenth century onward—that despite its inexplicable stability, the legal
regime had consistently failed to meet its own stated goals.173 Instead, he posited
a strong, but previously hidden, logic to the system that revealed it to be a
powerful tool for achieving its users‘ deepest objectives, namely, the terroristic
social control of a restive peasantry by the rural English aristocracy.174
Hay describes a capital system over which the members of the local gentry
held sway, through their power over the definition of capital crimes (via their
membership in Parliament), prosecutions (which they often initiated and
funded175), trial judges (who came from their ranks), executive clemency (by
officials acting on the gentry‘s advice), and public hangings (which the gentry‘s
clergy framed for the public in commanding sermons delivered to the assembled
masses as they waited for the trapdoor to drop).176 Using these powers, the gentry


Hay, supra note 5, at 40–49, 57.
Id. at 23–24, 56.
Id. at 22–24.
Hay, supra note 5, at 39.
Hay, supra note 5, at 25–26.
See, e.g., Friedman, supra note 73, at 475–76 (―England in the eighteenth century had no
public officials corresponding to . . . police or district attorneys. . . . A victim of crime who wanted
a constable . . . to apprehend the perpetrator was expected to pay the expenses of doing so. . . .
[T]he prosecutor was usually the victim . . . [who undertook] to file charges . . . present evidence to
the grand jury, and, if the grand jury found a true bill, provide evidence for the trial.‖).
Hay, supra note 5, at 27; see, e.g., Paul H. Robinson et al., The Disutility of Injustice, 85
N.Y.U. L. REV. 1940, 2016 (2010) (presenting findings from empirical research suggesting that
evidence injustices committed by prosecutors or the judicial system can undermine the justice
system‘s moral credibility, with crime-control costs as citizen‘s become reluctant to support, assist
and defer to the system).


obtained life or death power over huge swathes of the peasantry by making capital
virtually every invasive indiscretion through which criminals availed themselves of
the lords‘ silver or linens;177 strategically doled out prosecutions and mandatory
death sentences to the miscreants, and acts of slow-acting mercy (which often took
years of appeals to finalize) on behalf of members of their families who humbly
supplicated themselves to the lords in return for the favor—increasing the former
in times of riot and stress, and the latter as a form of mass plea bargaining to entice
the rabble to go back indoors;178 and managed the messages conveyed to the
public and lessons taught and learned through trial judges‘ statements upon
imposing death verdicts and the clergy‘s well-attended sermons when executions
took place.179 Through this power to take life (swiftly and demonstratively) and
give it back (excruciatingly slowly and quietly), Hay concludes, the gentry
exercised a virulent ―terror‖ over the masses, which, however, was carefully
cloaked in the calming and obscuring ideology of evenhanded due process and
clemency and the solemnity and spectacle of a judge‘s pronouncement of death
and the clergyman‘s and hangman‘s publicly dispatching the offender to the next
Hay argues that the death penalty particularly appealed to the gentry as a form
of social control because of its extreme and expressive, but only episodic and
after-the-fact, response to the demands of law and order. As such, it shielded
them from their existential fear: a strong central government bolstered by a
national police force, prosecuting corps and law enforcement regime.181
Clearly, Hay‘s Marxian explanation of England on the verge of the industrial
revolution does not translate well to the modern United States. Indeed, many
question the validity of Hay‘s interpretation as applied to the era he depicts.182
We find much to admire, however, in the model he provides of an analysis from


Hay, supra note 5, at 18, 21.
Id. at 40–49.
Id. at 17–18, 26–30.
Id. at 32–39.
Id. at 40–41.
See, e.g., John H. Langbein, Albion's Fatal Flaws, 98 PAST AND PRESENT 96, 120 (1983);
Robert B. Shoemaker, The ‗Crime Wave‘ Revisited: Crime, Law Enforcement and Punishment in
Britain, 1650–1900, (book review) 34 HISTORICAL J. 763, 763 (1991) (―Aibion‘s fatal tree attracted
considerable criticism, and a large body of published work now exists which calls into question . . . .
Douglas Hay‘s argument that criminal law was manipulated by the ruling class as a means of social


outside the system that aims to identify the hidden logic of a stable, yet seemingly
illogical, system of capital punishment.
Until now, our own writing has considered the death penalty largely from
within, arriving at a conclusion about the modern American death penalty that is
close to the starting point for Hay‘s explorations. The Broken System studies
described a modern American death penalty system so fraught with reversible error
that only 5% of the 5,826 death verdicts imposed were carried out during the
twenty-three years studied from 1973 to 1995.183 Coupled with evidence that
defendants are overwhelmingly sentenced to a penalty less than death on retrial
after reversal, the findings reveal a capital punishment system with little to be said
in its defense from within. Although Broken System begged, it did not answer,
the question of why a system like this would be allowed to persist.
In Overproduction of Death, we discovered a set of skewed incentives that
help explain why there is little supply-side constraint on the number of flawed
death sentences that localities produce. Local prosecutors stand to gain by
imposing as many death verdicts as possible, regardless of the verdicts‘ failure rate
on appeal, because they quickly realize the political gains, and the costs of review
and reversal are slow to materialize and shouldered by others.184 Across the aisle,
the modest resources of the anti-death-penalty bar require them to focus only on
those clients that face the most imminent threat of execution. With their cases
clustered at this narrower, post-conviction end of the capital appeals funnel, these
lawyers understandably value reversals for their immediate clients over a doubtful
promise of fairer, more reliable trial procedures for hypothetical, future capital
defendants whom the lawyers do not now represent.185 Again, however, although
Overproduction of Death explained how these incentives keep the systems‘ repeat
players from exposing and resisting the capital system‘s stable diseconomies, the
article did not consider why the legislatures that created the system, the courts that
regulate it or the taxpayers who underwrite it allow it to survive.
More recently, Slow Dancing with Death explored the Supreme Court‘s
tortuous and ambivalent regulation of the nation‘s capital punishment
architecture.186 The article describes how the Court‘s responsibility for directly


See BROKEN SYSTEM II, supra note 27, at 36.
Liebman, supra note 146, at 2097, 2101.
Id. at 2076.
Liebman, supra note 19.


superintending a system of blatant, court-administered state violence that often and
evidently lacks the trappings of law has compelled it to intervene in matters of
substantive criminal law and punishment that (with textual encouragement from
the Constitution) it has traditionally left to the States.187 Yet, the Court has ended
up being paralyzed by the responsibility—unwilling, on the one hand, to accept the
moral exposure entailed by the substantive review it originally promised,188 yet
fearful, on the other hand, of the public reaction if it tried to abolish a form of state
violence it has been unable to domesticate.189 In the latter regard, the parochial
and libertarian proclivities of the communities that support the penalty, and their
history of taking matters violently in their own hands when they doubt the state‘s
willingness or capacity to protect them, suggests that the Court has good reason to
Slow Dancing with Death finds the Court in about the same situation vis-à-vis
the American system of capital punishment as the nation at large—helplessly
watching the system churn out case after flawed case, seemingly without purpose
or direction. The article helps explain why the Supreme Court has not succeeded
in altering a situation it believes is untenable, but still leaves unanswered the
question of why the system arose in the first place and why it remains so stable,
despite the vilification it receives from death penalty proponents, detractors and the
agnostic alike. Following Hay, and informed by our above discovery of the
minority status of the death penalty and the characteristics of the relatively few
communities that value it, we are now prepared to take a fresh look from without.
C. The True Course of a Capital Case
One way to obtain this view is to write the story of the death penalty, not the
way a lawyer would, but as a documentarian might. Informed by the facts
revealed in the Broken System study, we undertake something like that approach in
this section.


Id. at 16 (noting, inter alia, that the Due Process Clauses of both the Fifth and Fourteenth
Amendments appear to bless the death penalty as a substantive matter).
Id. at 125 (arguing that the Court ―lost heart‖ and renounced many of the supervisory
responsibilities its earlier decisions regulating the death penalty promised).
Id. at 122 (arguing that the Court is at once tormented by sanctioning the continuation of
raw, state-imposed violence and fearful of the ―struggle with the political branches that banning the
violence would ignite‖).
See supra notes 59–102 and accompanying text.


1. The Origin of a Death Sentence: The Scarlet Letter Stage
A home invasion murder in a small town—the victim is well known, the
details disturbing. The community is stricken; a family cries out for justice. A
stranger to the community, the accused, has committed a senseless and despicable
act. He must be punished.
Amid public grief and cries for vengeance, the prosecutor goes for death.
That is what the community wants. It believes itself to be under siege from
outside and demands immediate steps to allay its fear and express its anger. In
another era, in a county of self-reliant citizens like this one, the accused might not
have made it this far. But today, the community trusts the prosecutor to carry the
case forward to the desired capital conclusion. The prosecutor‘s ability to repay
the public trust and remain in office depends on her success.
A jury is empanelled and delivers its verdict: guilty. But the community, the
victim‘s family and the state are not ready to exhale. A final matter remains for
the representative dozen: to choose between life and death.
At this stage of his long and uncertain journey to execution, or not, the
accused has few friends. The community is angry. Skilled lawyers willing to
represent the hated and hateful likes of this defendant are not available to him.
They concentrate their fire on those who have moved further along the conveyor
belt and are more imminently in danger of being executed. The accused must
make do with the less clearly competent and committed legal assistance the
victimized community is willing and able to provide.191
Facing facts, it is clear that none of the participants see the function of this
stage as conducting a perfect trial or constructing an air-tight case that will
withstand appeal. Indeed, truly facing the facts, it is clear that the goal of this
stage is not even a verdict with a fifty-fifty prospect of being upheld. Historical
error rates are simply too high to permit that assumption, particularly in
communities that use the death penalty the most.192 Instead, the evident function
of this stage of the proceeding is to visit upon the stranger-perpetrator an awful


See sources cited supra note 75.
See, e.g., BROKEN SYSTEM II, supra note 27, at 295–99, B-4, B-5 (listing counties with at
least five death sentences and 100% reversal rates: Baltimore Count, MD; Orange, CA; De Kalb,
GA; Tulsa, OK; San Bernardino, CA; Lake, IN; Richmond, GA; Camden, NJ; Pasco, FL; Jefferson,
AR; Calcasieu, LA; Knox, TN); infra notes 180–187 and accompanying text.


judgment, that proclaims the community‘s anger and abhorrence at his violation of
their sanctity and its resolve to deal immediately and harshly with those whose
insidiously invasive acts put the entire community in fear.
It is not lost on the community that there are other ways to improve public
safety. A more vigorous and professional law enforcement apparatus, for
instance, might catch more who offend and deter more who think about doing it.
Yet, true to its libertarian self-sufficiency and mistrust of government, the
community is unwilling to empower and invest in that amount of government.193
Instead, it prefers to wager its scarce law enforcement resources on the death
penalty—infrequently in the scheme of all stranger crime but frequently when the
cross-boundary offense fortuitously or intentionally kills. The penalty is not part
of a comprehensive strategy or penology. It is the community‘s unvarnished
expression of communal retributive anger that, although infrequent, is visible and
dramatic and signals reassuringly to itself and wrathfully to outsiders how
seriously it takes the invasive offense.
The jury‘s verdict is death. The sentence is pronounced. God‘s mercy on
the invader‘s soul is invoked. He is branded with the scarlet letter of being
worthy of death.
Whether he is actually executed will be for others to decide, years down the
road. But for now, communal catharsis occurs. The community has largely
achieved what it needs and desires. As Professor Weisberg has pointed out,
―[s]imply having many death sentences can satisfy many proponents of the death
penalty who demand capital punishment, because in a vague way they want the
law to make a statement of social authority and control.‖194
Of course, other things equal, police, prosecutors and the assembled public
want the sentence to be executed. If out-of-town judges overturn it years later,
those in the community who are still paying attention may curse the outsiders‘
insensitivity and elitism.195 But facing facts again, an unmistakable feature of the
scarlet letter stage, at least for most of the players, is that they do not deeply care if
the condemned man lives or dies. How else can we explain the fact that the
highest-frequency death sentencing counties have long coped with significantly

The high rates of serious error found when the community‘s death verdicts are reviewed on
appeal further attests to its unwillingness to invest in high-quality law enforcement and courts
systems. See Gelman et al., supra note 15, at 243.
See Weisberg, supra note 19, at 387.
See supra note 58; infra notes 195-204 and accompanying text.


higher capital-error and reversal rates than the already staggering national
two-thirds average? Indeed, until the Supreme Court put its foot down,
prosecutors frequently encouraged juries to abandon their hesitation to choose
death on the grounds that mistakes would be reversed on appeal.196 Though these
communities wish for eventual executions, the real dividend they receive evidently
comes from the death verdict itself, irrespective of its quality or aftermath.
Given a choice between narrowing the focus and securing higher-quality
death verdicts that stand up on appeal and, instead, sticking with the quick and
unvarnished verdicts the community has been imposing, it chooses the latter.
When faced with a trade-off between more sentences and more executions, the
community opts for the ceremony and spectacle, the expressive comeuppance and
revenge, attending each additional death verdict.
For the condemned man, however, things are different. And the same is true
for the family of the victim. If the defendant is properly advised by trial counsel
(a fifty-fifty proposition, at best) and if members of the victim‘s family are
properly informed by the police, prosecutor or judge (in our experience, this almost
never happens197), these parties with a keen interest in the final outcome will know
that the capital trial and verdict are so much sound and fury. For them, the
process is only beginning; the real outcome is in doubt and will not be clear for
2. The Death Row Stage
When the condemned man arrives on death row, he takes his place at the back
of the line of thousands of other death row inmates awaiting their turn for each of
three successive inspections of their conviction and sentence. First comes the
state direct appeal, usually lasting about three years.198
During the Broken
System study period (1973–95), the elected state high court judges who preside at
this stage found 41% of the verdicts they examined so prejudicially flawed that

See, e.g., Caldwell v. Mississippi, 472 U.S. 320, 329, 333, 334 & nn.4, 5 (1985) (imposing
constitutional ban on arguments, frequently made, that the jury should be less than fully concerned
about the life-or-death responsibility it bears, given that the outcome will be scrutinized on appeal).
See Liebman, supra note 146, at 2134 & nn.247, 248 (discussing the toll the review process
takes on the victim‘s family who typically are left uninformed by the other players).


reversal and a do-over was required.199 In other words, as a result of the first
stage of review, over two out of every five defendants is back at square one,
awaiting trial.200
The prisoners whose verdicts survive direct appeal (59% in the Broken System
period) trudge on to the next, state post-conviction stage of review, which
examines the case for a separate set of errors. Few of the decisions at this stage
are published, making it difficult to discern how many death verdicts that cleared
the first hurdle are overturned at this stage and how many simply languish awaiting
decision. The limited information established that no fewer than 10% of the
surviving verdicts were found lacking at this stage because of an error sufficiently
egregious that it probably affected the outcome of the trial—again requiring re-trial
if the defendant was to be executed.201 The figure could be as high as 78% if
every surviving case for which an outcome is unknown ended in a reversal.
Broken System II‘s very conservative estimate of the actual number is 18%.202
This is itself a remarkable number when it is considered that the inspector at this
stage is the same locally elected trial judge who imposed the verdict in the first
place and is reviewing her own handiwork, to which her judicial superiors already
gave their stamp of approval at the first stage of review.203
Publicly accessible decisions again become available at the third, habeas
corpus level of inspection, where federal judges review the half or so of the
original verdicts still standing for constitutional error.204 Even though federal
courts can only examine claims of error already rejected at one or both of the
previous state court stages, 205 during the Broken System period, they found


BROKEN SYSTEM II, supra note 27, at 8.
See BROKEN SYSTEM II, supra note 27, at 8 (reprised in Gelman et al., supra note 15, at


Id. (reprised in Gelman et al., supra note 15, at 216–17).
Id. at 8 & n.88, 17–18 & n.103. .
See Gelman et al., supra note 15, at 215.
The reversal rate at the second phase is no less than 10%, and a 10% reduction of the 59%
of cases that survived the first level of review (i.e., subtracting 5.9 percent of the original set), means
that 53% of verdicts survive the first two stages. See Gelman et al., supra note 15, at 216–17. A
more likely, although still conservative estimate, is that there is a 18% second-stage reduction of the
59% of cases surviving the first stage (subtracting 11% of the original set), see supra note 183 and
accompanying text, in which case only 48% of the original set survive the first two stages. See
BROKEN SYSTEM II, supra note 27, at 18 n.103.
See 28 U.S.C. § 2254(b) (1996) (requiring exhaustion of state remedies as a prerequisite to
federal habeas corpus review).


prejudicial violations in two out of five capital verdicts that survived both prior
stages of review.206 Reversals at this final stage occurred on average about
thirteen years after the prisoner was sentenced to die; final decisions approving
death verdicts typically occurred about a year earlier on average.207
Using the 10% underestimate of reversals at stage two and ignoring a handful
of reversals at an intermittent fourth level of review, the odds during the Broken
System period were better than two to one (68%) that a man sentenced to death in
modern America would have his case overturned and sent back to its originating
county. 208
In other words, the massive error detection process consumes over a dozen
years on average and considerable public treasure before a conclusion is reached.
Decisions in capital cases are highly visible and closely scrutinized by the
public, 209 and judges do not lightly reverse capital verdicts and order costly
do-overs. Nine out of ten of the nearly 2,400 reversals during the Broken System


BROKEN SYSTEM II, supra note 27, at 8.
See BROKEN SYSTEM II, supra note 27, at 91; see Nancy J. King et al., Executive Summary:
Habeas Litigation in U.S. District Courts: An Empirical Study of Habeas Corpus Cases Filed by
State Prisoners Under the Antiterrorism and Effective Death Penalty Act of 1996, U.S. DEPT. OF
(noting that capital federal habeas corpus cases resulting in reversal of death verdicts in the 2000–02
period took longer to complete than cases resulting in an affirmance of the death verdict). The figures
in text are of course averages. Individual cases take much longer. See, e.g., Alarcón, supra note
20 , at 711 & n.75 (showing that of the thirteen men executed in California between 1992 and 2006,
ten had been on death row for fourteen years or more, and of those ten, five had been there for twenty
of more years); See Gershowitz, supra note 21, at 346–47 (discussing a Tennessee death row inmate
sentenced to die in 1984, whose case was under review in the Supreme Court twenty-four years later,
at a point about midway through the third tier of review); infra notes 195–204 and accompanying text
(describing Carl Isaacs, who Georgia executed in 2003, thirty years after he first arrived on death
BROKEN SYSTEM II, supra note 27, at 8, 58 (reprised in Gelman et al., supra note 15, at
216–17). The full reversal-rate equation, reflecting the findings at all three phases of review is .41 +
.10(.59) + .40(.53) = .68. See id. Using a sophisticated methodology to project future changes in
reversal rates and apply them to the cases that were still under review when the Broken System study
was completed, Finkelstein and colleagues re-estimated the 68% figure as 62%. See supra note
142. Using the still conservative, but likely more accurate, 18% reversal rate at the second stage,
the 68% figure climbs to 71%. See BROKEN SYSTEM II, supra note 27, at 18 n.103 (presenting a full
reversal-rate equation, reflecting a more realistic estimate of reversals at the second stage of review
as: .41 + .18(.59) + .40(.48) = .71); supra note 183 and accompanying text.
See Gelman et al., supra note 15, at 231 n.64.


period were ordered by state judges elected to office by voters who profess to
support the death penalty, remove judges who they believe do not,210 and so
chasten the governors they elect that state chief executives have all but stopped
awarding capital clemency even in egregious cases.211 Most of the remaining
10% of reversals were ordered by federal judges appointed by ―law and order‖
Presidents Nixon, Reagan and Bush (the elder)—judges whose reversal rate in
non-capital habeas corpus cases is a few percent or less.212 The sensitivity of
elected state judges to the political risks from reversals is suggested by another
Broken System II finding. State judges are less likely, on average, to overturn
error-laden death verdicts originating in rural and small-town communities—where
reversals are likely to be more visible and controversial—than from cities, leaving
it to their life-tenured federal counterparts to weed out the rural and small-town
bad apples.213
The willingness of state and federal judges to overturn flawed capital verdicts
reveals something important about them, however. On the whole,214 and despite
their conservative and pro death penalty proclivities and susceptibility to political
pressures, these judges are different from denizens of the communities that impose
most death sentences. They are less parochial and libertarian. By definition,
they have absorbed at least some cosmopolitanism and respect for layers of
government. They have sojourned at college and law school and work in state
capitals and urban hubs—at a distance from the personally and communally
invasive reality of the crimes they review. They listen attentively to the
out-of-state, big-firm, big-city lawyers who prosecute the appeals of many death
row inmates and have the best record of success of all capital lawyers.215 Their


See BROKEN SYSTEM II, supra note 27, at 37–38; see Liebman, supra note 143, at 82. The
rare state judges who have reversed all or nearly all death sentences they reviewed have been
unceremoniously voted out of office and often have been replaced by judges even more loath to
overturn any death sentence—more than counterbalancing the effect of the former. See BROKEN
SYSTEM II, supra note 27, at 39, 65 & nn.161, 209.
See Liebman, supra note 43, at 2117 & n.211 (citing sources).
BROKEN SYSTEM II, supra note 27, at 38–39; see also King et al., supra note 186, at 10
(reporting on district court (i.e., non-final) outcomes of a sample of federal habeas corpus petitions
filed in a sample of federal courts between 2000 and 2002, in which the reversal rate for capital cases
was 35 times higher than that in non-capital cases).
BROKEN SYSTEM II, supra note 27, at 336 (reprised in Gelman et al., supra note 15, at 247).
For an exception, see infra note 204and accompanying text.
BROKEN SYSTEM II, supra note 27, at 318 (reprised in Gelman et al., supra note 15, at
250–52) (noting that capital prisoners are more likely to have their convictions reversed in federal
habeas proceedings when they are represented by out-of-state, big-firm lawyers than when they are


work lives are governed by, and their decisions turn on, standard legal operating
procedures. In some basic way, they have made universalizing professional
commitments—sworn oaths—to uphold the law, making national law supreme
over local edicts, and practice a craft that requires them to hear out both sides and
base a decision on only the law and the facts.
Whether advised by counsel from the start or by the acuity of hindsight, the
condemned man comes to see that if he can stick it out for a decade or more, he
has a better than even chance of surviving his death verdict. Likewise, the family
of the victim comes to see that the passions that drove the trial to its
conclusion—the sense of outsider violation, the communal and expressive
retribution and the episodic singularity of their loved one‘s case—have little grip
on appeal. Something very different, and probably unexpected and ill-explained,
takes place. The divergence between these reactions and those of the community
itself become clear when the matter arrives back in the county for retrial.
3. The Retrial Stage
Upon reversal of the death verdict, the capital prisoner typically remains on
death row while the prosecutor decides whether to try him again. His ―case‖
comes home to the county where it originated. Years have elapsed since the
crime and original verdict. The modest time and resources the community is
willing to expend, even on cases like this, have been depleted and the passage of
time and scattering of witnesses make a second go-round costly. How does the
community react? In fact, there are two responses, one dominated by the
parochial impulse, the other by libertarian disposition.
i. The More Parochial Response to Reversal
The first response is a rather rare situation. It arises when the reaction to the
reversal fuses with a still palpable sense of outsider violation from the original
crime or a more recent one. An example is the notorious Alday case in Seminole
County, Georgia. Then Governor Jimmy Carter dubbed the 1973 rural home
invasion murders over many hours of the entire extended Alday family ―the most
heinous crime in Georgia.‖216 Despite unequivocal evidence of the codefendants‘
guilt, the federal court of appeals overturned the convictions, a decade and a half
later, due to ―inflammatory and prejudicial pre-trial publicity [that] so pervaded the

represented by other counsel).
Jerry Schwartz, 1973 Georgia Murders Back in Courts, N.Y. TIMES, Jan. 4, 1988, at A13,
available at 1988 WLNR 1285127.


community as to render virtually impossible a fair trial before an impartial jury.‖ 217
Rousing the same sentiments again, the reversal prompted 100,000 Georgians to
send Congress a truckload of petitions calling for the federal judges‘
impeachment.218 Unsurprisingly, the county prosecutor opted to go capital again
for two of the four defendants.219 A jury again sentenced the ringleader, Carl
Isaacs, to death, but imposed a life sentence on his accomplice Wayne Carl
Coleman.220 Even under these circumstances, the District Attorney decided not to
pursue a death sentence against a third accomplice, George E. Dungee, because he
was mentally retarded.221 The fourth accomplice, Isaacs‘ younger brother Billy,
who was fifteen at the time of the crime, testified against his brother in exchange
for a reduced sentence and has since been paroled.222 Carl Isaacs was executed in
2003, by which point he had become the longest serving death row inmate in the
United States.223
As to Isaacs (although, interestingly, not as to the other three killers), the
dynamic is clear. When the magnitude of the communal violation and fears it
engendered remain fresh in mind—or are rekindled by a new crime—the
outlander‘s cross-boundary offense merges with the faraway judges‘ secular and
universalizing assault on community safety and values, magnifying the parochial
sense of threat from outside. Although the localist instinct dominates, libertarian
disgust at the judges‘ impeachable inability to protect the community also comes
into play, triggering a desire to take matters once again in the community‘s hands.
Professor Garland describes this dynamic when explaining why abolitionists
gain so little traction in even strongly religious communities that value the death
Support for capital punishment came to be seen as an integral part of the


See Bill Montgomery, What Ever Happened to . . . George Elder Dungee: Man Convicted
of Role in Alday Family Murders, ATLANTA J.-CONST., March 19, 2007, at B1 available at 2007
WLNR 5136443.
See Jay Apperson, After 20 Years, Freedom Nears, BALTIMORE SUN, Oct. 21, 1993, at I13,
available at 1993 WLNR 790515; Schwartz, supra note 216.
See Ariel Hart, National Briefing South: Georgia: Longtime Inmate Executed, N.Y. TIMES,
May 8, 2003, at A30, available at 2003 WLNR 5204956.


―traditionalist‖ worldview, just as opposition to it became standard for
liberal ―progressives.‖ Depriving people of the right to impose capital
punishment—like depriving them of their guns or their right to school
prayer, or their right to ban abortion—came to be viewed as a kind of
elite contempt for common people, for their faith, and for their way of
Illustrating the response, Garland quotes Alabama judge Tom Parker‘s response to
a Supreme Court decision banning the execution of juveniles:
[T]he liberals on the U.S. Supreme Court already look down on the
pro-family policies, Southern heritage, evangelical Christianity and other
blessings of our great state. We Alabamians will never be able to
sufficiently appease such establishment liberals, so we should stop trying
and instead stand up for what we believe without apology.225
When the distant judges‘ and the stranger criminal‘s assaults on the
community‘s way of life fuse, the result is to confirm the community‘s need for the
death penalty and re-create the conditions for its episodic use as an expression of
the seriousness with which the community takes the incursions on its sanctity.
ii. The More Libertarian Response to Reversal
Libertarian impulses dominate the second, more usual response to reversals,
which is to let the death penalty go, and sometimes let the defendant go. Broken
System II‘s study of the aftermath of reversals at the second review stage shows
that 82% of the do-overs end in sentences less than death, including 9% ending in a
decision to release the defendant for insufficient evidence that he committed the
crime.226 In some cases, the jury reaches this conclusion after the prosecutor
brings new charges. In other cases, the prosecutor chooses not to re-prosecute at
all. More times than not, however, the prosecutor and defendant agree to the
result out of court, in a plea bargain for a life sentence—typically, with both sides
doing everything they can to make the outcome as low-key as possible, with no
news coverage.227


See GARLAND, supra note 2, at 251.
Id. at 251.
BROKEN SYSTEM II, supra note 27, at i (reprised in Gelman et al., supra note 15, at 221).
See Liebman, supra note 146, at 2119, 2127.


In other words, most members of the community hear little or nothing about
the reversal and subsequent—usually non capital—result. From their perspective,
the matter reached a satisfying conclusion years before, when a jury of their
fellows issued a very public, very expressive and merciless rebuke of the invader
and warning to others via their verdict of death. The one exception is the victim‘s
family, who are quietly asked to acquiesce in the bargain. This they often are
willing to do, after years of anguish at each unexpected and unexplained twist and
turn in the appeals process, the hard won recognition of what a new appeals
process would look like, and a promise that the man they (unlike the rest of the
community) continue every day to fear will remain in prison until he dies.228
We now have a first important insight from our view from without—from our
resolve to face the facts and be honest about what we see: the main value the
capital-prone community gets from the death penalty is its imposition, not its
execution. Most of what the community wants comes with the spasm of
retributive anger it expresses contemporaneously with its experience of the
homicidal invasion by publicly condemning the killer to die. It is then that it
exorcizes its fear and economically signifies its anti-crime mettle. In all but rare
situations, when the case returns after reversal, it has none of the attributes of the
expressive proxy for systematic law enforcement that the libertarian death penalty
is there to provide. By now, the case is little more than an expense the
community need not bear.
D. A New Explanation of Why Reversals Have No Chastening Effect
A second insight follows close behind, once we acknowledge that neither
response the death-prone community gives to reversals is the one the logic of
appellate inspection assumes. According to that logic, the community will be
chastened by the error its officials made and educated by the decision explaining
the reversal. Voters will consider dismissing the offending district attorney or
judge, in hopes of more certain and timely executions in the future. To head off
that result, those officials will look to sanction their employee or the appointed
defense lawyer who let the county down. Everyone will resolve to do better in
the future. Or so the reversing judges hope. But, as we saw, those judges and
the denizens of the local community are different,229 and so is their view of


See id. at 2134 n.247.
See supra notes 193–194 and accompanying text.


Consider, for example, San Bernardino County, California. Between 1973
and 1995, the county imposed thirty death sentences, more than one a year, at one
of the highest rates in California.230 All thirty death verdicts were overturned on
appeal.231 In contrast, neighboring and much more populous Riverside County
imposed about the same number of capital verdicts during the period
(twenty-seven), albeit at a much lower rate per homicide, but only 31% were
reversed.232 If execution was the sincere goal of San Bernardino County, one
might expect it to align its trial practices more fully with the legally ordained rules
for securing a valid capital conviction. But the facts do not bear out the
assumption that the county will learn from its mistakes. As of May 2011, only
one of San Bernardino‘s many dozens of death sentences since the late 1970s had
matured into an execution. 233 Yet, despite this abysmal record, the county
continues to churn out death sentences at one of the highest rates in the
nation—fourteen between 2000 and 2007 alone, costing an estimated $15.4
million—or the equivalent of thirty-two additional homicide investigators each
In The Overproduction of Death, we offered an explanation for the lack of an
effective feedback loop between appellate reversals and county level trials. We
noted, for example, that the local officials who secured the flawed verdict are not
required to defend it on appeal. That task falls to bureaucrats in the state attorney
general‘s office—lower status officials who are in no position to take the district
attorney to task for generating losing verdicts the state must defend. 235 The

See BROKEN SYSTEM II, supra note 27, at B-7. The County grew quickly during the
period, averaging about a million residents. County of San Bernardino, Demographics, available at
See BROKEN SYSTEM II, supra note 27, at B-7.
See id.
CTR., (revealing that since 1973, California has executed
thirteen people, only one of whom was convicted by San Bernardino County); Kevin Fagan et al.,
Condemned Murderer Executed at San Quentin for 1980 Slaying,SFGATE.COM (Jan. 29, 2002),
See, Death by Geography, A County by County Analysis of the Road to Execution in
nty_analysis_of_the_road_to_execution.shtml (last visited May 13, 2011).
Liebman, supra note 146, at 2121.


case-specific and claim-specific nature of review also militates against effective
feedback, because no single judge or panel of judges observes the county‘s overall
pattern of corner-cutting and shoddy miscues;236 the focus is on procedure (often
read as ―technicalities‖), not the substantive desert of punishment;237 the remedy
ordered is a do-over directed at no one in particular and calling for neither the
aberrant office to be revamped or the offending official to be punished;238 and in
any event, given the many years that elapse between a verdict and reversal, the
responsible official is often long gone by the time the reversal occurs.239 Finally,
even if the error does somehow get connected to the responsible official, the onus
is unlikely to outweigh the political and career capital the official accrued at the
time of the original conviction.240
From our bird‘s eye view here, however, it is clear that the problem is larger
than flaws in the feedback mechanism for informing local citizens of the errors and
the actors responsible for them, so they can demand steps to assure the scrupulous
proceedings and executable verdict everyone wants. As this suggests, feedback
assumes a common goal—to learn how to do better next time the task tried and
overturned the first time. As we have seen, however, what communities that most
use capital punishment want from death verdicts, and what the appeals process
reviews the verdicts for, are two different—even irreconcilable—things.
The appeals process seeks to let executions go forward as long as they are
based on professionally acceptable law enforcement and judicial procedures, and
there is no ―reasonable probability‖ of unreliability in the determinations of guilt
beyond a reasonable doubt and of sufficient aggravation net of mitigation to
warrant death. 241 The law enforcement science, standard procedures and


Id. at 2129.
Id. at 2126–27.
Id. at 2121. The rhetorical convention is to name no names, and the wrongdoers are
immune from damages in any event. Id. at 2126. See Thompson v. Connick, 131 S.Ct. 1350,
1357–1358 (2011) (citing immunity as the basis for reversing a $14 million damages award against a
district attorney whose lack of supervision allowed an assistant to withhold forensic evidence he
knew exonerated the capital defendant who subsequently was convicted); see also Bienen, supra note
28, at 1363 (noting that sanctions are rarely meted out to state‘s attorneys in cases of wrongful capital
convictions); Liebman, supra note 146, at 2121 (similar).
Liebman, supra note 146, at 2120.
Id. at 2127.
See, e.g., Kyles v. Whitley, 514 U.S. 419, 435, (1995) (defining a Brady violation as a
―showing that the favorable evidence [that the prosecution failed to disclose to the defense] could
reasonably be taken to put the whole case in such a different light as to undermine confidence in the


professional norms these protocols entail for evaluating verdicts lie at oblique
angles, or are positively anathema, to the parochial and libertarian values that
prevail in communities that most use the death penalty. What those communities
wanted—and what they more or less already got—from the verdict is different.
Their goal is a simple, unyielding expression, with no institutional or professional
frills, of what they think of cross-boundary invaders; a straightforward message
about how harshly they mean to punish such killers now and in the future, leave
aside the fine points of how to catch and try them.
When the community confronts the reversing court‘s directive to spend a little
more, tighten up on standards and process, engineer a more professionally
reasonable and probabilistically reliable product, the result is not the corrective
dialogue the appellate apparatus assumes,242 but dissonance. At worst, the order
and its aftermath are the mirror image of feedback—not a neutral outside
evaluator‘s advice from which to learn about how to improve, but a confirmation
of the pernicious outside threat against which the community imposed the flawed
penalty in the first place—a red cape before angry eyes. More often, however,
the order simply falls on the uncomprehending ears of people who cannot make
any sense of it and have already moved on.
The indifferent reaction to reversals years later is not the only evidence of
how unimportant actual executions are. Also indicative is where and how
executions are conducted—not, as Hay described, in the central square on market
day, with a fiery object lesson projected from a makeshift pulpit about to scaffold a
public hanging. Instead our executions are in an isolated penitentiary separated
for the public by brick walls, barbed wire and an inviolable TV blackout, in the
darkest of night, with a few last words murmured by a man on a gurney about to be
injected with drugs designed to put beloved pets to sleep.243 The death-prone
community, that is, having long since taken matters in its own hands and projected
the stern and unforgiving message it desired, has little concern for the sanitized
way the carceral state carries out what Professor Garland calls the ―anti-lynching‖

verdict‖); Strickland v. Washington, 466 U.S. 668, 687–88, 694 (1984) (providing that the assistance
of counsel violates constitutional norms if (1) ―counsel's representation fell below an objective
standard of reasonableness‖—as to which ―[p]revailing norms of practice as reflected in American
Bar Association standards and the like . . . are guides‖—and (2) ―there is a reasonable probability
that, but for counsel's unprofessional errors, the result of the proceeding would have been different‖;
a ―reasonable probability is a probability sufficient to undermine confidence in the outcome‖).
See, e.g., Robert M. Cover & T. Alex Aleinikoff, Dialogic Federalism: Habeas Corpus and
The Court, 86 YALE L.J. 1035, 1044 (1977) (characterizing federal habeas corpus review of criminal
convictions as a dialogue between federal and state courts).
Liebman, supra note 146, at 2058 & n.104.


of modern execution.244
Capital punishment is a minority practice in the United States today; a relative
handful of counties drive a huge proportion of death sentences. These
communities get a lot from the death penalty, but in unexpected ways that are far
more bound up with the verdict than the execution. For them, the visceral
message about invasive crime that the verdict—and in the rare cases where it
occurs, the execution—convey outward and especially inward substitute for
systematic state efforts to protect the community from crime. Because the death
penalty substitutes for systematic crime prevention and penology, one might
expect these localities to lose something in terms of the objectives traditionally
associated with criminal enforcement, such as incapacitation, deterrence and
retribution. 245 In fact, there are a number of costs associated with these
communities‘ use of the death penalty. This section identifies those costs and
who bears them.
There is no national figure for the cost of the death penalty,246 which of
course no doubt varies from state to state and county to county. Nor is there
consensus on how to measure the cost. On one point there is agreement,
however. Capital punishment in the United States always costs more than
non-capital proceedings and penalties.247
When a prosecutor seeks the death penalty, additional layers of investigative,
evidentiary, procedural and legal complexity automatically arise.248 To begin with,


GARLAND, supra note 2, at 34–35.
Given these communities‘ indifference between life without parole and the death penalty
once a death verdict has initially been imposed, their use of the death penalty provides little in the
way of additional incapacitation—even assuming the death penalty has an incapacitative advantage
over life without parole. Deterrence is likewise questionable, given the improbability that the death
penalty will be carried out and for the reasons discussed infra notes 289–293 and accompanying text.
Finally, although these communities are good at expressing capital retribution, they are not so good at
exacting it.
For one take on the issue, see Dieter, supra note 145.
See, e.g., Robert M. Bohm, The Economic Costs of Capital Punishment: Past, Present, and
Future, in AMERICA‘S EXPERIMENT WITH CAPITAL PUNISHMENT 592 (James R. Acker et al. eds. 1998);
other authority cited supra note 213; infra notes 228–270.


virtually every capital case goes to trial (because the defendant has nothing to
lose), while the vast majority of non-capital cases, including murder cases, are
resolved without trial, on a plea to a lesser crime.249 Additionally, the body of
legal doctrine and case law governing capital cases is vastly more complicated than
the law that applies in non-capital cases.250 Most importantly, capital cases entail
two full-blown trials, one on guilt and the other on sentence, in contrast to other
criminal proceedings which require only a guilt trial and a summary sentencing
Each component of the typical criminal trial process—for example, pretrial
investigation252 and motions practice—also takes far longer and is more costly in
capital trials. An example is jury selection, which in capital cases not only must
examine the usual questions of juror competency and impartiality, but also must
delve into a prospective juror‘s opinions and beliefs on capital punishment.253
Because of its sensitivity, this ―death qualification‖ process takes longer than other
selection steps, as each juror is questioned separately while sequestered from other
prospective jurors.254 Death qualification leads to significantly more prospective
jurors being excused ―for cause‖ than in non-capital cases, requiring many more to
be summoned and put through the lengthy selection process. In addition, the
prosecutor and defense typically have twice as many preemptory challenges in
capital than in non-capital cases, leading to more excusals and the need for more
prospective jurors. 255
The pretrial publicity, length and sequestration
requirements that typically attend capital trials require still more excusals and
prospective jurors. Although 100 potential jurors are sufficient to permit a panel
of twelve to be selected for most non-capital criminal trials, it is not uncommon for
(―Of felony convictions nationwide, 94 percent are obtained by guilty plea.‖); Liebman, supra note
146, at 2099 (noting frequency with which capital cases go to trial).
See Liebman, supra note 146, at 2100 n.171; Washington State Bar Association, Final
Report of the Death Penalty Subcommittee of the Committee on Public Defense 14–17 (2006),
available at [hereinafter Washington State Bar
See Liebman, supra note 146, at 2108.
See Gary Goodpaster, The Trial for Life: Effective Assistance of Counsel in Death Penalty
Cases, 58 N.Y.U. L. REV. 299, 303 (1983); Washington State Bar Report, supra note 228, at 17.
See Goodpaster, supra note 230, at 344; infra note 381 (citing Supreme Court decisions
discussing the comprehensive investigations required in capital cases to satisfy constitutional
See, e.g., Wainwright v. Witt, 469 U.S. 412, 424 (1985) (permitting states to excuse jurors
for cause if the trial judge finds that their views on the death penalty will keep them from imposing
the death penalty though the evidence warrants it).
See, e.g., Washington State Bar Report, supra note 228, at 16.
See id.


courts to summon over ten or even twenty times that many in capital case.256
Instead of the hours or day or so that suffices in other criminal cases, jury selection
in capital cases can take weeks.257
There is an even greater discrepancy in the cost of capital and non-capital
appeals. All three levels of appeals discussed above are automatic or nearly
automatic in every capital case. Because most non-capital cases are disposed by
plea bargain, which substantially limits the defendants‘ bases to appeal, the strong
tendency in those cases is not to appeal at all.258
A. Direct Costs Compared to Life Without Parole
The usual way to estimate the cost of a death penalty case is to compare it to
the cost of a comparable murder case in which the next most serious
punishment—life without parole—is imposed. In this section, we make that
comparison per trial, per appeal, per execution and on an annual statewide basis.
1. Additional Costs per Trial
Indianapolis prosecutors estimate that ―a death penalty trial . . . demands the
resources of five normal murder cases.‖259 Because of these additional costs, a
single capital case can cripple a county‘s criminal justice budget.260 For example,
county costs for defense lawyers alone (a fraction of all costs) in a 2005 murder
case in Yakima County, Washington exceeded $2 million, even though the
prosecutor eventually reached a plea bargain with both defendants.261 Given this


See id. .
See, e.g., New Jersey Death Penalty Study Comm‘n, Report 31 (2007), final.pdf [hereinafter New Jersey Death Penalty
Report] (noting that jury selection takes four to six weeks in a capital case as opposed to one or two
days in an otherwise comparable noncapital case).
See Liebman, supra note 146, at 2053 n.90.
Tim Sparks, Cost of Death Penalty Trial Can Tip Decision, FORT WAYNE J. GAZETTE, Oct.
25, 2001, at 1.
See Gershowitz, supra note 21, at 319–23 (collecting examples of prosecutors deciding not
to purse capital sentences in egregious cases in which the cost of doing so may, or admittedly did,
factor into the decision).
Chris Bristol, Death Penalty: The Cost Is High, YAKIMA HERALD-REPUBLIC, March 19,
2011. A plea deal was reached in one co-defendant‘s case a year before the deal in the other case,
explaining why the defense costs in the former case were $500,000 and in the latter case were three
times that much. See id.


history, when Yakima County Prosecutor Jim Hagarty recently considered
capitally indicting an alleged triple murderer, a dust-up broke out between him and
a county court administrator. Hagarty said cost would not affect his decision; the
administrator complained about ―overspend[ing] in a capital case‖ at a time when
―we have no reserves left.‖262 Statewide, a Washington State Bar Association
study estimated prosecution, defense and court costs at just the trial phase of a
capital case to be roughly $520,000 more than if same case were prosecuted
Our main concern, however, is not the trial costs counties cover for
themselves but those that the rest of the state bears as a result of subsidies and
state-funded services such as courts and public defender offices.264 In that regard,
a Maryland study estimates that each additional death penalty trial costs the state
$1.9 million—or about $1 million more than a non death penalty murder trial.265
A report the New Jersey Death Penalty Commission issued before the state
abolished the death penalty in 2009 estimated that the state‘s public defender
service, which in that state provides defense representation for local prosecutions,
cost the state $1.5 million more a year because of the death penalty. Although the
report did not break these costs cost down per case, there were only nine inmates
on the state‘s death row at the time.266 A Kansas study, found that trial costs for
death cases were sixteen times greater for capital than for non-capital murder

Id.; see also Bienen, supra note 28, at 1308 (―[A]t a time when state governments are not
meeting their most basic obligations, how can the state‘s policy of maintaining capital punishment
alone be immune from considerations of cost and relative value?‖); Smith, supra note 21, at 113
(describing a similar dust-up recently between an Ohio prosecutor and county board chair on the one
hand and a trial judge on the other, who all agreed that ―seeking the death penalty would have a
catastrophic effect on the county‘s budget,‖ but disagreed over whether the prosecutor should be
permitted to proceed capitally; in the end the trial judge took the ―extraordinary [step of]
invok[ing]cost concerns as grounds for precluding the prosecutor from seeking the death penalty‖);
id. at 113–14, nn.24–28 (citing instances of counties raising taxes, even multiple times, taking out
loans, and cutting police and fire budgets to be able to afford a single capital prosecution).
Washington State Bar Report, supra note 228, at 18. (estimating that pursuing the same
case capitally as opposed to non capitally costs, on average, an additional $217,000 for prosecutors,
$246,000 for public defenders, and between $46,640 and $69,960 in trial court operations).
See Gershowitz, supra note 21, at 353–54 (describing arrangements in a number of states
for reimbursing localities for substantial portions of the costs of capital prosecutions).
J. Roman et al., The Cost of the Death Penalty in Maryland THE URBAN INSTITUTE, 2
(March, 2008).
New Jersey Death Penalty Report, supra note 235, at 36.
See, Performance Audit Report: Costs Incurred for Death Penalty Cases: A K-GOAL Audit


2. Additional Costs per Appeal
That same Kansas study estimated that appeal costs for capital cases are about
twenty-one times greater than comparable non death cases.268 States and the
federal government bear these additional costs, because states fund the judges and
typically the lawyers for both sides who are provided at the first two levels of
appellate review, and the state (vis-à-vis state‘s attorneys) and the federal
government (vis-à-vis judges and defense lawyers) entirely fund the third level of
review.269 Greatly increasing the non-capital/capital differential is the fact that, in
non-capital cases, prisoners have no right to state-funded defense counsel at the
second and third stages of appeals and very few pursue those appeals.270 In
contrast, capital prisoners have such a right for capital prisoners under state law in
many states and federal law governing all states, and capital prisoners typically
pursue all levels of review.271 The upshot is that the dramatically higher appellate
costs instigated by a decision to proceed capitally are mainly triggered by the small
set of counties that impose most death sentences and are largely subsidized by state
and federal taxpayers who themselves make do with life without parole, at a 95%
plus savings.272
Analyzing only the first, direct appeal stage of review and only the cost of
defense representation, the Washington State Bar Association estimated that each
death penalty case cost the state $100,000 more on average than a comparable
non-capital case.273 A 2010 Indiana study found that the average cost to the state
of a capital trial and only the first stage of review is ten times greater than the
comparable cost of the average life-without-parole case.274 These comparisons of
of the Department of Corrections, KANSAS LEGISLATIVE POST AUDIT COMMITTEE 13, 41 (App. B)
(2003), (comparing estimated average
trial cost of a murder case resulting in a death sentence ($508,000) to estimated cost for a murder trial
where a death sentence is not sought ($32,000); noting that a substantial portion of these costs,
especially for defense counsel, are borne by state taxpayers).
Id. at 13 (comparing average estimated cost for capital direct appeals ($401,000) to cost of
appeal of similar case in which the death penalty was not imposed ($19,000)).
See Liebman, supra note 146, at 2048, 2051, n.85.
PROCEDURE §§ 7.2, 12.2–12.5 (6th ed. 2011).
See id.
See, e.g., supra note 243 and accompanying text.
See Washington State Bar Report, supra note 228, at 32.


costs per filed appeal vastly underestimate the scale of additional capital costs,
given that every death sentence is automatically reviewed on appeal, while only a
small fraction of non-capital sentences are appealed even at the first stage and even
fewer are challenged at the second two stages.275 Because it is a rare capital
defendant who can pay for his own defense,276 this additional cost, again, is borne
by state taxpayers most of whose own communities impose no such costs on the
Security considerations affecting the conditions of incarceration during
lengthy appeals also drive up the cost of incarcerating capital inmates—so much
so, that even the incarceration costs of the death penalty are greater than the cost
of imprisoning a murderer until he dies. For example, the New Jersey
Department of Corrections estimated that state taxpayers would save about $1
million per inmate over the life of the prisoner if (as eventually occurred) the death
penalty was abolished and capital sentences were converted to life without
parole. 277 The California Commission on the Fair Administration of Justice
found that the per-prisoner cost of death row incarceration is four times greater
than if the same man or woman were sentenced to life without parole—costing the
state $63.3 million more per year to maintain its (as of then) 670-person death row
than it would cost to incarcerate them under sentences of life without parole.278
Although this figure is annual, not over the life of the prisoner, the fact that
California now has over 700 inmates incarcerated on death row, and has executed
only thirteen over the last thirty-plus years, suggests that the life expectancy of
death-row inmates and of prisoners serving life without parole is not very
Many of the additional costs are caused by the high rates of serious legal error
that afflict capital verdicts, particularly from localities with high capital sentencing (estimating state capital costs as $449,887, compared
to $42,658 for each non capital case).
See supra notes 228, 236, 271 and accompanying text.
See Weisberg, supra note 75, at 535 (―The State virtually always pays for the defense of
those whom it seeks to execute.‖). §§ 7.2[f], 12.3[b]; Liebman, supra note 146, at 2053–54 n.90.
See Bienen, supra note 28, at 1385–86.
(June 30, 2008),, June 30, 2008 [hereinafter CALIFORNIA
See DEATH ROW USA, NAACP LEGAL DEFENSE FUND, 12, 38 (Spring, 2010), [hereinafter Death Row USA].


rates.280 The costs are even greater when the errors lead to the conviction of the
innocent. Between 2000 and 2009, alone, Illinois taxpayers (statewide) shelled
out $65 million in damages to innocent men whom local communities had
sentenced to die as a result of egregiously flawed investigations and trials.281
3. Additional Cost per Execution
From the standpoint of a taxpayer assessing the burden of different
law-and-order strategies, the sole distinguishing feature of a costly capital case and
a much less expensive life-without-parole case is the execution. That, in other
words, is what the taxpayer ―buys‖ for the additional cost. Because essentially all
capital verdicts cost anywhere from five to twenty-one times more than a
life-without-parole alternative, but (as a result of high reversal rates) fewer than
15% of death verdicts nationwide eventuate in an execution,282 excess costs per
execution are extremely high.
In Florida, for example, the Miami Herald estimated that, between 1973 and
1988, a capital case cost the state $3.2 million from indictment to execution, or six
times more than the overall cost to the state of seeking and carrying out a sentence
of life without parole. 283 But noting that few capital cases actually end in
execution, the Palm Beach Post in 2000 estimated that the excess cost of the death
penalty to the state per execution that actually does occur was $23 million.284
Florida citizens may be getting a bargain. A Maryland legislative commission
found that its taxpayers bore an additional expense of $186 million between 1978
and 1999 to obtain fifty-six death verdicts, of which five were carried out. It
estimated the incremental capital cost to the state per execution as $37 million.285
Executions cost considerably more than that in California.
4. Overall Additional Cost


See supra notes 50–52,181–187 and accompanying text.
See Bienen, supra note 28, at 1326.
See supra notes 32, 237–257 and accompanying text; infra notes 291 and accompanying

David Von Drehle, Bottom Line: Life in Prison One-Sixth as Expensive, MIAMI HERALD,
July 10, 1988, at 12A.
S.V. Date, The High Price of Killing Killers, PALM BEACH POST, January 4, 2000, at 1A.
INSTITUTE 1,3 (2008),


In California, it takes over two decades on average for a death verdict to run
the appellate marathon. Keeping cases moving at this snail‘s pace costs state
taxpayer‘s $137 to $184 million extra every year, according to the 2008 report of a
state commission.286 To reduce that average lapse of time to around the national
average of twelve years, the commission concluded that California would have to
spend approximately twice this amount. 287 The commission estimated that
commuting the sentence of everyone currently on death row to life in prison
without parole would save the state over $125 million annually.288 Although the
commission did not analyze cost increases over time, they are substantial. Its
estimate of an annual additional statewide cost of the death penalty of $125 million
is about a 40% increase over a similar estimate made fifteen years earlier.289 Nor
did the commission calculate a per-execution cost. But the state‘s track record
over the last twenty years of executing a prisoner on average about once every year

Compare CALIFORNIA DEATH PENALTY COMM‘N, supra note 254, at 6, 83–85 with Judge
Arthur L. Alarcón & Paula M. Mitchell, Executing the Will of the Voters?: A Roadmap to Mend or
End the California Legislature‘s Multi-Billion-Dollar Death Penalty Debacle, 44 LOYOLA L.A. L.
ting_the_Will_of_the_Voters.pdf (concluding based on a comprehensive cost analysis that ―$4 billion
of state and federal taxpayer money has been expended administering the death penalty in California
since 1978, with a cost in 2009 of approximately $184 million above what taxpayers would spend
without the death penalty,‖ and noting that this analysis is more complete than the California Death
Penalty Commission‘s analysis, including because it ―incorporates the costs associated with federal
habeas litigation,‖ which were omitted from the commission‘s cost estimate of $137 million per
year). See generally Gil Garcetti, California's Death Penalty Doesn't Serve Justice, L.A. TIMES,
ce.html (noting that author ―was the Los Angeles County district attorney for eight years and chief
deputy district attorney for four years‖ during which he ―was responsible for my office's decision to
seek the death penalty in dozens of cases‖ and concluding based on more recent evidence that
California‘s death penalty is ―an incredibly costly penalty, and the money would be far better spent
keeping kids in school, keeping teachers and counselors in their schools and giving the juvenile
justice system the resources it needs‖).
CALIFORNIA DEATH PENALTY COMM‘N, supra note 254, at 83.
See id. at 84. The $125 million estimate is derived by subtracting the annual cost to the
state of the same number of life-without-parole inmates as there are death row inmates—$11.5
million—from the state‘s $137 million annual death penalty cost to state taxpayers. The comparable
savings to North Carolina from substituting life without parole for the death penalty is about $11
million a year. See Philip J. Cook, Potential Savings From Abolition of the Death Penalty in North
Carolina, 11 AM. L. & ECON. REV. 498, 522–25 (2009).
See Stephen Magagnini, Closing Death Row Would Save State $90 Million a Year,
SACRAMENTO BEE, March 28, 1988, at 1.


and a half generates a cost to the state of $187.5 million per execution.290
Estimates from other states are also sobering. Conservatively, between
Governor George Ryan‘s January 2003 commutation or pardon of all 167 men and
women on death row291 and Governor Pat Quinn‘s March 2011 decision to sign
abolition legislation,292 Illinois spent over $150 million on the seventeen men
sentenced to die between 2000 and 2010—on average, about $20 million overall
per year, and $880,000 per condemned inmate per year.293 Between 1994 and
2003, before giving up the endeavor, New York State spent about $170 million
imposing seven death verdicts, executing none—an average of about $17 million
overall per year and $2.4 million per condemned inmate per year.294
The anecdotal quality of available cost information makes averages and
comparisons difficult.
There is, however, ample evidence that a local
community‘s decision to initiate a capital prosecution, and particularly to impose a
death sentence, foists tens of millions of dollars per execution on the rest of the
state that would be saved if the sentence instead were life without parole.
Because a relatively small number of communities in so-called ―capital‖ states
instigate a disproportionate—and disproportionately error-laden—share of these
costly verdicts, and because most of the costs are borne by the majority of
taxpayers whose communities don‘t much use the death penalty, the majority
might well ask whether the subsidy is worthwhile. That question becomes more
pressing when we consider the magnitude of the indirect costs the death-prone
minority inflicts on the majority.
B. Indirect Costs


See Death Row USA, supra note 255, at 13–37.
Jodi Wilgoren, Citing Issue of Fairness, Governor Clears Out Death Row in Illinois, N.Y.
See supra note 18 and accompanying text.
Bienen, supra note 28, at 1338.
2005), (―In the past ten years, the state
and local governments have spent over $170 million administering the [death penalty] law. Yet, not
a single person has been executed. Only seven persons have been sentenced to death.‖)


The communities that most use the death penalty do not simply choose to
impose a surfeit of death sentences, each costly in itself. In addition, their
parochial and libertarian dispositions lead them to use the death penalty—often for
felony-killings295—as an episodic, if deeply expressive, rationing device, in lieu of
systematic effective law enforcement. The result is that, compared to less
death-prone communities, these localities tend to spend considerably less on law
enforcement, courts and legal representation for capital defendants296 and have
substantially lower clearance rates for violent crimes.297
When a state or county uses the death penalty to broadcast its disposition to
deal harshly with stranger crimes affecting some neighborhoods, while in fact
declining to fund and empower government to enforce the criminal law
comprehensively in all neighborhoods—and when the result is low clearance rates
for serious crime—inevitably, more dangerous criminals go free. The lucky
criminals no doubt continue to prey on the same communities, perhaps
contributing to their high homicide victimization rates among whites relative to
blacks.298 But the notorious transience of criminals, and the opportunity to leave
rather than get caught for a past crime or for a later one that turns deadly and gets a
death sentence, also puts neighboring communities at risk. There is a triple irony
here. First, communities whose parochial tendencies make them particularly
fearful of outside invaders stoke their own fear by doing a bad job—as a result of
their libertarianism—of catching, convicting and incarcerating criminals. Instead,
they focus only on the tiny proportion of cross-boundary felonies in which a
privileged victim dies.299 Second, and as a result, communities that live in fear of
criminals invading their communities end up propelling more than their share of
local criminals into other communities. Third, to the extent that the criminals‘
new victims are in communities that prefer systematic law enforcement to the
death penalty, those communities end up subsidizing their death-prone neighbors
yet again, by expending their law enforcement dollars to catch criminals that their
libertarian neighbors did not bother to catch.
There is no way to know how many violent crimes non death prone
communities suffer because of low clearance rates in death prone neighboring


See supra notes 107–141 and accompanying text.
See supra note 76 and accompanying text.
See supra notes 75–78 and accompanying text.
See supra notes 59–63, 66 and accompanying text.
See supra notes 103–141 and accompanying text.


communities, or how much their crime control costs increase.
to think the costs are substantial.300

But there is reason

C. Other Externalized Costs
We already have shown how death prone counties‘ quick and unvarnished use
of the death penalty generates exceptionally high rates of serious error301 and risks
the conviction and execution of innocent individuals.302 And we have shown how
their parochialism and libertarianism blinds them to the corrective feedback from
the many reversals their verdicts generate.303 These conditions are costly to
taxpayers statewide304 and sew disrespect for law and legal institutions.305 They
also take a heavy toll on the victims‘ families.
The parochial and libertarian community is prepared to minimize law
enforcement and skimp on trial process, while ostentatiously imposing the death
penalty for the occasional stranger crime that turns deadly. It is resigned, as well,
to the length and likely disappointing results of the inevitable appeals—which the
majority of less death prone communities subsidize in any event—and to take a
pass on most capital re-trials. 306 Regrettably, however, no one prepares the
relatives of the victim for this aftermath.
The loved ones of homicide victims and others disagree on what ―closure‖
means, and whether executions provide it.307 But whatever else closure means, it
surely includes knowing the final result, one way or another. For the community,
the pronouncement of the verdict is the final, satisfying result.308 But not so for

Charles M. Blow, High Cost of Crime, N.Y. TIMES, Oct. 8, 2010, at A21 (discussing
research estimating the victim, lost productivity and criminal justice system of each murder ($17.25
million), rape ($448,532), robbery ($335,733), and burglary ($41,288)).
See supra notes 73–90, 181–187 and accompanying text.
See Steve Mills & Maurice Possley, I Didn't Do It. But I Know Who Did: New Evidence
Suggests a 1989 Execution in Texas Was a Case of Mistaken Identity, CHI. TRIB., June 25, 2006, at 1;
supra notes 149, 256–257 and accompanying text.
See supra notes 221–223 and accompanying text.
See supra notes 245–270 and accompanying text.
See, e.g., Robinson et al., supra note 159, at 2016.
See supra notes 205–207 and accompanying text.
See Michelle Goldberg, The ―Closure‖ Myth, SALON, Jan. 21, 2003, available at (reprising the ―closure‖ debate).
See text following note 88; supra notes 95, 175–179, 205–207, 220–223 and accompanying


the victim‘s family, for whom each milepost in the appellate marathon is a fresh
reopening of wounds, and for whom reversal and retrial mean a replaying of the
ghastly crime itself.309 As dozens of murder victims wrote Illinois Governor Pat
Quinn before he signed abolition into law in the state in 2011, ―‗to be meaningful,
justice should be swift and sure. The death penalty is neither. [It] drag[s]
victims‘ loved ones through an agonizing and lengthy process, which often does
not result in the intended punishment.‘‖310
Quinn addressed victims directly in his signing statement:
To those who say we must maintain a death penalty for the sake of the
victims‘ families, I say it is impossible not to feel the pain of loss that all
these families share or to understand the desire for retribution that many
may hold. But, as I heard from family members who lost loved ones to
murder, maintaining a flawed death penalty system . . . will not bring
closure to their pain. . . . We must instead devote our resources toward
the prevention of crime and the needs of victims‘ families, rather than
spending [the enormous sums of] money [needed] to preserve a flawed
Notice that Quinn responds in the voice of the majority of communities for whom
spending public resources on prevention and social services for victims is the
modal response to the problem of crime. It is the antithesis of the parochial and
libertarian voice of the communities that pronounce the lion‘s share of verdicts and
infuse them with the largest proportion of prejudicial error. What Quinn could
not say, is that a good portion of the costs he described—the anguish of victims in
the face of error and delays, and extra crime prevention costs on taxpayers—are
foisted on victims and the majority by the parochial and especially libertarian
choices capital prone communities make.
To be sure, the ultimate blame for all these costs belongs to the killer
himself—assuming he was the killer and was not wrongly condemned.312 But
here, too, our view from without requires us honestly to face facts. Since the

See Goldberg, supra note 283.
Governor Quinn‘s Abolition Statement, supra note 18 (quoting victims‘ statement); see
Goldberg, supra note 283 (―Death penalty advocates claim victims' families need [closure]—and
deserve it—in order to move on. But some of those family members say dealing with death row
issues for years only prolongs their pain.‖).
Governor Quinn‘s Abolition Statement, supra note 18.
See supra notes 149, 256–257, 278 and accompanying text.


reinstatement of the death penalty in 1976,313 capital states have executed death
sentences for only a vanishingly small fraction of homicides (about one-tenth of
1%) and of murders resulting in convictions (about three-tenths of 1%).314 Even
among the small fraction of convicted murders for which a death penalty is
imposed (less than 5% of the total), capital states have only executed about 15% of
the sentences since 1976.315 The few executions that do take place, moreover, are
typically a dozen years after the crime.316 As a matter of honestly faced fact,
therefore, the death penalty is not the punishment for murder in the United States;
the penalty instead is life without the possibility of parole, but with a small chance
of execution a decade later.317
This too has a cost. A precursor is Fyodor Dostoevsky‘s famous mock
execution in St. Petersburg‘s Semyonovsky Squarea terrifying event, staged by
Czar Nicholas I himself, that caused another prisoner to go mad. 318 Perhaps
channeling his existential forbearer, Albert Camus based his abolitionism on ―[t]he
devastating, degrading fear that is imposed on the condemned for months or years
is a punishment more terrible than death.‖319 Based on the same psychological
torture, the Judicial Committee of the Privy Council—the highest court of the
United Kingdom—unanimously ruled that a fourteen year delay between a trial
and execution rendered a pending Jamaican execution illegally ―cruel‖ and


Compare Furman v. Georgia, 408 U.S. 238 (1972) (striking down all death penalty statutes
then in effect in the United States for arbitrary and capricious selection of condemned) with Gregg v.
Georgia, 428 U.S. 153, 187 (1976) (plurality opinion) (upholding the constitutionality of ―guided
discretion‖ capital statutes adopted in the wake of Furman).
See Gelman, et al., supra note 15, at 214 (noting the during the 1973-1995 Broken System
study period, 331,949 homicides in capital states generated an estimated 300,257 arrests and 118,992
murder convictions but only 5826 capital verdicts (1.8 percent of homicides; 4.9 percent of murder
convictions) for which the judicial reversal rate was 68 percent and the resulting number of
executions was 326 (.1 percent of homicides; .3 percent of murder convictions; 5.6 percent of death
verdicts); supra notes 34–35and accompanying text.
See supra note 32 and accompanying text.
See supra notes 144–145 and accompanying text.
In a state like Pennsylvania, where only three death row inmates out of hundreds have been
executed since 1976, and all three chose to end their appeals prematurely and volunteered to be
executed, see infra notes 386–373 and accompanying text, the penalty for murder is perhaps better
characterized as ―life without possibility of parole but with the option of suicide-by-state.‖
See FYODOR DOSTOEVSKY, THE BROTHERS KARAMAZOV x (Constance Garnett transl. 1976)
(describing the mock execution); see also FYODOR DOSTOEVSKY, NOTES FROM THE HOUSE OF THE
DEAD xx (David McDuff transl. 1985).
173, 200 (1961).


―inhumane‖.320 Such an inordinate delay, the court noted, would never have
been permitted under English common law.321 Although before retiring, Justice
Stevens failed to convince a majority of his colleagues on the Supreme Court to
consider whether even greater capital delays violate the United States
Constitution‘s Cruel and Unusual Punishment Clause, 322 our bird‘s eye view
compels us to treat that as a foregone conclusion. From that perspective, we must
assume that a system that strongly tends to operate in a particular way for many
years is meant to operate in that fashion. And no one doubts that a choice, like
Czar Nicholas I‘s, to exact an extra increment of psychological torture and
retribution by threatening many, but carrying out only a few, executions violates
our Constitution.
Our view from without also compels us to ask why we continue employing
this (from all perspectives) torturous and facially irrational system. We have
already explained why the minority of death prone, parochial and libertarian
communities does so—and does so almost literally with abandon, given how
thoroughly they discard the costs into others‘ laps.323 The question begged, then,


Pratt v. Attorney General of Jamaica, 4 All E.R. 769, 783 (P.C. 1993) (collecting decision
by other courts).
See Lackey v. Texas, 514 U.S. 1045, 1045 (1995) (Stevens, J., respecting the denial of
certiorari) (concurring in the denial of certiorari but suggesting that the Court eventually would have
to consider the question presented there, ―whether executing a prisoner who has already spent some
17 years on death row violates the Eighth Amendment's prohibition against cruel and unusual
punishment‖); Id. at 1047 (noting that ―Justice Breyer agrees with Justice Stevens that the issue is an
important undecided one.‖). Since Lackey, Justice Breyer has repeatedly urged the Court to
consider the issues. See Allen v. Ornoski, 126 S.Ct. 1139 (2006) (Breyer, J., dissenting from denial
of certiorari); Foster v. Florida, 537 U.S. 990 (2002) (Breyer, J., dissenting from same); Knight v.
Florida, 528 U.S. 990, 993 (1999) (Breyer, J., dissenting from same); Elledge v. Florida, 525 U.S.
944 (1998) (Breyer, J., dissenting from same). For arguments that lengthy death row incarceration
raises serious penological concerns or is unconstitutional, see, e.g., Alarcón, supra note19, at 711,
n.75 (arguing that ―[e]xtraordinary delays are indeed unacceptable‖ and showing that of the thirteen
men executed in California between 1992 and 2006, ten had been on death row for fourteen years or
more, and five had been there for twenty of more years); Jeremy Root, Cruel and Unusual
Punishment: A Reconsideration of the Lackey Claim, 27 N.Y.U. REV. L. & SOC. CHANGE 281, 283
(2001–2002); Jessica Feldman, Comment, A Death Row Incarceration Calculus: When Prolonged
Death Row Imprisonment Becomes Unconstitutional, 40 SANTA CLARA L. REV. 187, 219 (1999).
See supra Parts III–IV.


is why the majority of jurisdictions and taxpayers that do not employ the death
penalty, or employ it only rarely and judiciously, are willing to pick up the tab?
A. The Opaque Nature of the Costs
The most obvious reason that non death imposing communities allow their
high death prone neighbors to reach into their wallets and impinge on their safety
is that the costs are hidden and thinly spread. Much may be at stake for the
majority in the aggregate, but very little of the cost is apparent to any one person at
any given time. Modern death sentencing practices and costs thus are a standard
public choice situation in which concentrated minorities with a clear sense of what
they want can fleece ―anonymous and diffuse‖ majorities with a less clear sense of
interests they have at stake.324
The way both abolitionists and advocates typically frame the death penalty
debate also obscures the penalty‘s costs. This is because the debate so often and
heatedly focuses on the practice‘s moral aspects—whether it is right, or instead is
essential, to respond to killing by killing. On these terms, the death penalty is a
matter of personal choice, which Americans like to think everyone gets to make for
herself, then majority rules. The importance of how the death penalty debate is
framed is highlighted by polling results and their treatment in the debate. It is
commonplace that just under two-thirds of Americans say they favor the death
penalty (64% vs. 29%).325 In fact, however, Americans are about evenly split on
whether they prefer death to life without parole (49% vs. 46%).326 The former,
morally salient question and its two-thirds statistic frame the death penalty debate.
But the latter, pragmatic question and its fifty-fifty statistic actually control
behavior: The only choice prosecutors and jurors have in real capital murder cases,
in every state in the nation, is between death and life without parole.327


Bruce A. Ackerman, Beyond Carolene Products, 98 HARV. L. REV. 713, 724 (1985).
325 See Gallup, supra note 17 (showing 64% for, 29% against and 6% no opinion when
respondents were asked, ―Are you in favor of the death penalty for a person convicted of murder?‖;
the corresponding results were 49%, 46% and 6%, when the question was, ―If you could choose
between the following two approaches, which do you think is the better penalty for murder -[ROTATED: the death penalty (or) life imprisonment, with absolutely no possibility of parole]?‖).
See Ashley Nellis, Throwing Away the Key: The Expansion of Life Without Parole
Sentences in the United States, 23 Fed. Sentencing Rep., 27, 28 (2010) (noting that ―only Alaska
[which has no death penatly] provides the possibillity of parole for all life sentences‖); LIFE WITHOUT
visited May 14, 2011).


The naiveté of logic and rationality also probably interfere with the majority‘s
protection of its self-interest. Given the usual assumption that irrational systems
do not persist for decades, most members of the public may be excused for
believing the death penalty system we have probably deters crime, even in
communities that do not much use it, and that it costs less than incarcerating a
prisoner for life.328 In other words, we can excuse the majority of non—or
infrequent—users of the death penalty for assuming their death prone neighbors
are doing them a deterrent and fiscal favor, rather than fleecing them.
Just how far to push these excuses for the majority‘s credulity is unclear.
The usual antidote to public-choice bamboozlement is for groups to form with the
express purpose of pooling and publicizing the individually small but collectively
large interests of their constituents. The Death Penalty Information Center, which
we repeatedly cite in this article, fills exactly that role—rather effectively, we
might add—focusing at least as much of its attention and information-gathering on
the actual cost and other pragmatic considerations, as on moral arguments.329 We
suspect that other explanations are also at play.
B. No Deterrence Dividend
First, we consider whether localities that do not use the death penalty do in
fact receive a ―deterrence dividend‖ from locales that do use it. That is, one
community‘s investment in a death sentence might deter crime
region-wide—protecting surrounding localities and lowering their enforcement
costs, even if they do not reciprocate with their own death sentences.
In fact, as we note above, high death imposing counties are not net
suppressers of crime. On the contrary, they are inordinately bad at catching,
convicting and incarcerating serious criminals.330 This indeed appears to be by
their parochial and libertarian choice. They use the death penalty—expressively,
but only episodically—to demonstrate that they really mean business about


See infra notes 305–310 and accompanying text (offering explanation of why the death
penalty as practiced in the United States today does not deter crime ); supra notes 224–270 and
accompanying text (documenting the far greater cost of the death penalty compared to life without
The Center‘s website is; see supra notes 12–13, 22, 32,
91, & 303; infra notes 336, 339, 361, & 377.
See supra notes 75–78 and accompanying text.


cross-boundary crime, when in fact the penalty is a substitute for engaging in the
real business of systematically fighting crime.331
The result is a decline in deterrence. As criminologist have theorized and
shown empirically, differences in the probability of capture and sureness and
swiftness of punishment are likely to have more of an effect on deterrence than
differences in the amount of punishment once an offender is apprehended,
convicted and sentenced.332 This is especially likely to be so when the choice is
between punishments that all are extremely harsh, as is true of the exclusive life
without parole and death penalty options for capital murder.333 Given these
principles, we can expect that death prone localities‘ libertarian aversion to funding
effective law enforcement, trials and corrections, and their resulting low clearance
and high reversal rates, more than offset any deterrent advantage the death penalty
otherwise has over life without parole. In other words, more costly, competent
and scientific investigation leading to prison sentences—and more professional
trials ending in rarely-reversed verdicts—are likely to be a more effective deterrent
than sloppy investigation and trials, followed by oft-reversed death verdicts and
only a few executions years later.
This dynamic helps explain a conundrum that has vexed capital punishment
scholars for decades. Why doesn‘t heavy use of the death penalty generate a
measurable decline in predicted crime?334 The question has vexed researchers

See supra text following note 88; supra notes 95, 175–179, 205–207, 220–223 and
accompanying text.
See, e.g., Jeffrey Fagan, Death and Deterrence Redux: Science, Law, and Causal
Reasoning on Capital Punishment, 4 OHIO ST. J. CRIM. L. 255, 273 (2006); Daniel S. Nagin, Criminal
Deterrence Research at the Outset of the Twenty-First Century, 23 CRIME & JUST. 1, 21 (1998);
Daniel S. Nagin & Greg Pogarsky, Integrating Celerity, Impulsivity, and Extralegal Sanction Threats
into a Model of General Deterrence: Theory and Evidence, 39 CRIMINOLOGY 865, 866 (2001);
2010), (citing
See Fagan, supra note 332, at 271–72. The fact that 11% of all modern executions in this
country have taken place after prisoners elected to give up their appeals and be executed, rather than
exercise their better than even chance of living in prison for the rest of their lives, see infra notes 316,
372–373 and accompanying text, suggests the harshness of life without parole, though not its
equivalence to death.
To date, no convincing empirical support for the proposition that the death penalty deters
murders has been provided. Each time a new study suggesting a deterrent effect is published,
contrary analyses appear to knock it down. See, e.g., Gregg v. Georgia, 428 U.S. 153, 184–85
(1976) (plurality opinion) (―Statistical attempts to evaluate the worth of the death penalty as a


because it is the wrong question. Given modern death sentencing practices in the
United States, the comparison is not between the death penalty and life without
parole, everything else equal. The comparison is between the death penalty as a
self-conscious substitute for systematic law enforcement, professional adjudication
and scientific penology and, on the other hand, life without parole under conditions
closer to those sought by modern penology. The increment in deterrence from
imposition of the death penalty is cancelled out by the decrement in deterrence
from the predominant use of that penalty to starve law enforcement, in the process
punishing fewer crimes and delaying imposition of sentence.
An actual deterrence dividend thus cannot explain the majority‘s willingness
to subsidize the minority‘s heavy use of the death penalty. Imposing more death
verdicts correlates with less, not more, law enforcement. We need other
explanations for the majority‘s forbearance.
C. The Death Penalty as a Back-Pocket Option
Perhaps a better explanation for the non death prone majority‘s willingness to
pay, even dearly, for others to use the death penalty is that they want to keep the
penalty in their own back pocket. They may sense that there is a murderer out
there somewhere on whom even they would want to impose the penalty. They
preserve it ―just in case,‖ and pay for others to use it.
The recent Stephen Hayes trial in Connecticut is illustrative. In 2007, Hayes
and another man broke into the suburban home of Dr. William Petit. After
terrorizing the household for hours, beating Petit with a baseball bat and raping his

deterrent to crimes by potential offenders have occasioned a great deal of debate [but t]he results
simply have been inconclusive . . . .‖); Hashem Dezbakhsh et al., DOES CAPITAL PUNISHMENT HAVE A
344, 368 (2003) (arguing that authors‘ econometric model demonstrates that executions reduce the
murder rate); John J. Donahue & Justin Wolfers, Uses and Abuses of Empirical Evidence in the
Death Penalty Debate, 58 STAN. L. REV. 791, 841 (2005) (conducting exhaustive statistical analysis
concluding that none of the recent deterrence studies—e.g., by Dezbakhsh et al., supra, and
Shepherd, infra—―suggested that the death penalty has large effects on the murder rate‖); Fagan,
supra note 332, at 261 (identifying methodological ―flaws and omission‖ in recent empirical studies
claiming to show a deterrent effect from capital punishment, which ―render [them] unreliable as a
basis for law or policy‖ and concluding that there is ―no, reliable, scientifically sound evidence . . .
identify[ing] whether [the death penalty] can exert a deterrent effect‖); Kenneth Jost, Death Penalty
Debates: Is the Capital Punishment System Working?, 20 CONG. Q. RESEARCHER 965, 969–71 (2010)
(citing studies reaching opposite conclusions); Joanna M. Shepherd, Murders of Passion, Execution
Delays, and the Deterrence of Capital Punishment, 33 J. LEGAL STUD. 283, 307–08 (claiming that
statistical analysis reveals that each additional execution in the United States in 1999 resulted in
approximately three fewer murders, and that each death sentence resulted in approximately 4.5 fewer


wife, Hayes and the other man set Petit‘s two daughters, then the home on fire.335
Only Petit survived.336 In the midst of the case in June 2009, the state legislature
approved a bill to abolish the death penalty.337 After then-Republican Governor
Jodi Rell vetoed the measure, the Democratic candidate to replace her in the 2010
election, Dan Malloy, said he would sign similar legislation.338 In response, Petit
and others argued for keeping the penalty for the rare crimes like Hayes‘ triple
homicide.339 Almost simultaneously with Malloy‘s election as Governor, the
Hayes jury deliberated for many hours before sentencing him to die.340 Even
Hayes himself expressed relief, not wanting, he said, to live any longer with the
crimes he committed.341 Then, in May 2011, as the legislature seemed poised to
enact and the governor to sign abolition legislation, state Senator Edith Prague, a
capital punishment opponent whose vote would make a majority for repeal,
withheld her vote at Dr. Petit‘s request.342 ―‗I actually believe in repealing the
death penalty,‘‖ Prague, said.343 But ―‗[f]or Dr. Petit, for me to do one more
thing to cause him some kind of angst, I can't do it.‘‖344 The eighty-five-year-old
Senator asked to delay the vote for a year, evidently so Hayes‘ co-defendant could
first be tried capitally.345

See Jill Lepore, Rap Sheet, THE NEW YORKER, Nov. 9, 2009, available at; Alison Leigh
Cowan, & Christine Stewart, Suspect in Connecticut Killings Left Long Trail of Lawbreaking, N. Y.
at and 2007 WLNR 14967187.
See William Glaberson, Reliving Horror in a Test for the Death Penalty, N.Y. TIMES, Jan.
19, 2010, at A1.
See id.
Id. (quoting Dr. Petit telling Connecticut legislators that ―[m]y family got the death penalty,
and you want to give murderers life. That is not justice.‖).
Susan Campbell, Death Penalty Support Must be Absolute, HARTFORD COURANT, Nov. 10,
2010, at B1.
Steven Hayes, Statement, HARTFORD COURANT, Dec. 2, 2010, at xx, available at
Mark Pazniokas, At Petit's Request, Two Senators Stop Repeal of Death Penalty, THE


Although Connecticut currently has a dozen men on its death row, it has only
executed one since the reinstatement of the death penalty in the mid-1970s, and
that man voluntarily ended his appeals and asked to be executed.346 Still, the
chance to brand men like Hayes and his codefendant as death-worthy—the only
action that can explain Senator Prague‘s simultaneous request to delay the matter
and intention to vote to abolish a year later—may explain the majority‘s
willingness to subsidize counties and states that impose the death penalty with
greater frequency.
It is hard to assess the strength of this explanation for the majority‘s
willingness to pay for a costly penalty they do not use. The preference of half of
Americans for life without parole over the death penalty,347 even when the abstract
nature of the question invites them to imagine the back-pocket case for which they
might want to impose death, suggests that it is not decisive. In any event, this
explanation casts the death penalty as at best a luxury—something to we might be
able to afford in flush times, but not necessarily when budgets are tight.
D. The Majority‘s Fear of the Minority‘s Reaction to Abolition
There may be a darker reason for the majority‘s passive acquiescence: fear.
In this regard, the Supreme Court‘s deeply ambivalent behavior towards the death
penalty348 may be a microcosm of the majority‘s.
In our federal system in which crime and punishment are mainly the domain
of states and localities, subject only to the Constitution, it falls to the Court to set
national policy and limits on the death penalty. As Professor Garland has shown,
the Court‘s actions in this regard reveal a keen attention to the public mood.349
Consider, for example, the fierce late-sixties, early-seventies backlash against the
antiwar and civil rights movements‘ perceived assaults on ―states‘ rights‖ and
mainstream values.350 Those concerns helped drive the nearly unanimous state
response to the Court‘s 1972 decision in Furman v. Georgia,351 rejecting the death

David Owens, Ross Execution First Death Penalty Imposed in New England in 45 Years,
HARTFORD COURANT, Dec. 22, 2009, at B1.
See supra notes 301–303 and accompanying text.
See Ackerman, supra note 299, at 719.
Garland, supra note 2, at 206–30.
Id. at 232.
408 U.S. 238 (1972). See GARLAND, supra note 2, at 209.


penalty as then applied. In turn, the alacrity and vehemence with which the states
re-upped for the death penalty no doubt contributed to the Court‘s own decision
four years later in Gregg v. Georgia352 to protect its credibility and acknowledge
the primacy of the states by leaving the new capital statutes in place.353 After
years of incendiary decisions inflaming passions in the South and elsewhere, to
strike a blow at another revered institution might have threatened a national
schism. The Court opted not to find out.354
There are several reasons to think that the Court‘s discretion was the better
part of valor. As Professor Garland notes, in many areas of the country, support
for the death penalty soon became a proxy—or our analysis might suggest, a
substitute—for public officials‘ support for a suitably ―tough-on-crime‖ agenda.355
We also have noted the association between communities‘ commitment to the
death penalty and their desire for strict populist electoral constraints of judges and
district attorney,356 as well as a correlation between that commitment and state
appellate courts‘ timidity in reversing their capital verdicts.357
Another hint of the fear the Supreme Court felt is found in a sentence in
Justice Stewart‘s plurality opinion in Gregg that has long puzzled the elder one of
us. The sentence is in the crucial passage where Stewart, a key member of the
Furman majority who had long been a focus of abolitionist hopes,358 explained
why the Court had decided to validate capital punishment. The death penalty,
Stewart wrote, is ―essential in an ordered society that asks its citizens to rely on
legal processes rather than self-help to vindicate their wrongs.‖359 Understanding
Justice Stewart to be worrying that an inflamed family member might take the law
into his own hands, what Justice Stewart deemed ―essential‖ seemed to me to be a
make-weight, not worthy of a decision to allow the state systematically to kill its
own people. The perspective introduced here, however, clarifies that what


428 U.S. 153 (1976).
See GARLAND, supra note 2, at 221–23; Liebman, supra note 19, at 32–33.
See Liebman, supra note 19, at 33.
See GARLAND, supra note 2, at 246 (noting that despite its narrow application, the rarity of
its use and the utter implausibility of the idea, capital punishment became the ―solution‖ to crime that
dominated public debate, and that voicing unquestioning support for the death penalty was necessary
to show that one was for law and order).
See supra notes 70 and accompanying text.
See supra notes 29, 70, 192 and accompanying text.
See MELTSNER, supra note 45, at 157, 270.
Gregg v. Georgia, 428 U.S. 153, 183 (1976).


Stewart feared was not individual self-help but the vengeance of entire, close-knit,
self-sufficient communities. In the wake of Furman, these communities had
demanded new capital statutes to keep punishment of violently invasive outsiders
in their own hands. And the same communities had a history of violent and grisly
communal vengeance when they sensed that outside government actors with alien
values were compromising their ability to defend themselves.
If politicians, judges and even Supreme Court Justices live in fear of
inflaming parochial and libertarian communities‘ passions by taking away their
death penalty, perhaps the majority of non—or infrequent death penalty users fears
the same reaction if they stop subsidizing the penalty. Going further, we might
conclude that symbolic self-help is a rational underpinning of our capital
punishment system that our current capital system is well-tailored to serve, and
that the majority is wise to subsidize appellate review to take some of the vengeful
edge off of it. The death imposing community gets to demonstrate where it
stands on stranger crime; most defendants eventually get their reversals, albeit after
years of retributive torture; and in the end most cases are quietly re-tried to the
non-capital conclusion the majority would have imposed in the first place. This
analysis harkens back to Professor Robert Weisberg‘s hypothesis that the Supreme
Court manipulates death penalty doctrine with an eye on the number of executions
it believes a small portion of the populace demands and the temperance others
prefer.360 Although Weisberg‘s modus vivendi is between hotheads on both sides,
his analysis easily accommodates our view of a death prone parochial and
libertarian minority and a more pragmatic majority. The goal, he wrote, is to
―have some executions, but not very many. A small number of executions offers
a logical, if crude, compromise between the extreme groups who want either no
executions or as many as possible‖; it allows enough executions to ―keep the art
form alive, but not so many as to cause excessive social cost.‖361
Still, we wonder. Facing facts, a system that notoriously lacks ―face
validity,‖ that massively churns lives (those of the family of the victim and of the
condemned) and three levels of judicial proceedings, that exacts psychological
torture, that risks executing the innocent, that imposes heavy monetary and public
safety costs on the majority to serve little more than the symbolically expressive
needs of a parochial and libertarian minority, may not deserve the acquiescence
Professor Weisberg imagines.


Weisberg, supra note 19, at 386.


E. The Resonance of the Minority‘s Parochial and Libertarian Values
It may be, as well, that the ―us against them‖—or ―us grudgingly
accommodating them‖—tenor of our discussion misses the point. It is true that
most Americans have made peace with a modern world that sometimes requires us
to be each other‘s keeper, flaunts our dependence on others around the world
whose values we abhor, and threatens us with powerful forces against which we
need the government to protect us. But many Americans may still wish it were
not so. They thus may resonate with the minority of parochial and libertarian
communities that refuse to capitulate to the onslaught of these outside, entangling
forces. The death penalty, and the values that inform its peculiar use in death
prone communities, may appeal to a powerful sense of nostalgia for a bygone era,
akin to the romanticizing of the frontier and Old West.
This brings us back to Professor Garland‘s account of the 1970s reinvention
of the death penalty after Furman as a populist response to the perceived
overreaching of the ―soft-on-crime‖ liberal politicians and judges of the 1960s.362
―In post-1970s America,‖ Professor Garland writes, the death penalty became ―not
so much a policy or a penal sanction as a commitment, a symbolic badge that
declared the wearer‘s position on ‗law and order‘ issues—and on so much else
besides.‖363 Even without throwing in their lot with communities that base their
entire law enforcement strategy on a version of this death penalty projected
symbolism, members of the majority might feel that a refusal to subsidize capital
punishment would constitute a repudiation of traditional values with which they
strongly resonate.
F. Hay‘s America?
Why, in sum, does the majority quietly allow an idiosyncratic minority of
death prone communities to impinge so substantially on its aggregate, if not visibly
any single individual‘s financial and personal welfare? And why does the
majority continue to abide the taint capital punishment practices imposed on the
integrity of a judicial system on which all rely? Although our explanations may
not be entirely convincing, they do reveal in our times what Douglas Hay thought
he discovered about capital punishment in England a few centuries ago: that it has


See GARLAND, supra note 2, at 234–44.
See GARLAND, supra note 2, at 244–45.


been ingeniously designed to serve the needs of a few while imposing substantial,
if hidden, cost on the many.364
America‘s current capital practices parallel Hay‘s description of Georgian
England more fully than first appeared. Akin to the landed gentry, wary of the
traditional order‘s destabilization by incipient urbanization and a national
leviathan, death prone communities feel themselves under siege from disruptive
outside forces and the entanglements of the modern state. In England, the gentry
responded by using its control of the levers of criminal justice to dole out strategic
increments of terror and mercy in lieu of ongoing surveillance provided by a
national police force. 365 In the United States, parochial and libertarian
communities use capital felony murder to ration episodically ferocious responses
to stranger crime in lieu of systematic law enforcement. In both times and places,
actual executions yield to the benefits of the death verdict itself. And in both,
revered moral principles—due process and mercy in eighteenth century England,
close-knit communities and self-reliance in modern America—are used to blind the
majority to the advantage being taken of them by the minority.
Hay suggested that if the peasants had not been so blinded by ―ideology,‖
they would have risen up and overturned the system and its ―terror.‖366 We don‘t
espouse such a drastic response for the United States today. Nor do we think the
majority has been so thoroughly hoodwinked. In the next Part, we identify
evidence of a majority fighting back, albeit ineffectively. In the following Part,
we suggest ways for the majority to strengthen its response.
Although the basic operation of the nation‘s death penalty has remained the
same for decades, the number of death verdicts and (less so) of executions has
declined recently. In this Part we consider whether the decline modifies the
uneven distribution of death sentences and executions.
In 1996, American jurisdictions sentenced 315 men and women to death
nationwide. 367 No year since has witnessed as many, and, with two minor


See supra notes 5–6, 79–82 and accompanying text.
See supra notes 79–82 and accompanying text.
See supra note 82 and accompanying text.


exceptions, the count has declined every year since then.368 New death sentences
reached a post-Gregg low of 112 in both 2009 and 2010, representing a 64%
plunge since the mid-1990s.369 The largest declines occurred in the two years
between 1999 and 2001, when the number of death sentences fell precipitously to
159 from 277 (a 43% decline). 370 More recently, the decline has been
modest—from 123 in 2006 to 112 in both 2009 and 2010.371
The number of executions each year also has declined but not as much.
Executions peaked at ninety-eight in 1999, falling since then to about fifty each in
the last two full years—about where they were in the mid-1990s.372 Owing to
declines in death sentencing and the huge backlog of the condemned on death rows
around the country—3,259 inmates as of 2010 373 —the annual number of
executions has even exceeded annual new death sentences in a small number of
states, including Ohio and Virginia. 374 Nationally, however, death sentences
continue to be more than double the number of executions each year.375 As a
result of declining death sentences and ongoing appellate reversals leading to
lesser sentences—but not so much as a result of executions which declined
somewhat during the period—the nation‘s death row population has modestly
fallen. After peaking at 3,593 in 2000, and remaining close to 3,550 from 1999 to
2002, the death row census fell to a low of 3,207 in 2008 and has been close to that


Id. The year 2002 was up slightly from 2001, and 2010 had the same count as 2009.
CTR, (last visited May 14, 2011) (showing that annual
executions averaged 81 between 1997 and 2000, peaking at 98 in 1999; averaged 50.5 in 1995–1996;
and averaged 46 in 2009–10).
See Death Row USA, supra note 40, at 1.
Since 2004, Virginia has executed about twice as many people as it has sentenced to death.
See DEATH SENTENCES BY STATE AND YEAR, supra note 12; Washington Post, Virginia Executions
Ohio executed fourteen individuals between 2006 and 2009 but sentenced only eleven to die in the
same period.
AND YEAR, supra note 12.
YEAR, supra note 336.


figure over the last six years.376
A. The Last Decade‘s Nationwide Death Sentencing Decline
Observers have offered a variety of reasons for the decline in death sentencing
over the last fifteen years, including falling crime rates, a fear of executing the
innocent, broader availability of ―life without parole‖ options and improved capital
defense representation.377 Our analysis suggests a different reason: a rebellion of
sorts by the majority of communities and citizens in capital states who rarely or
never use the death penalty but for years have been subsidizing its profligate use
by the minority of jurisdictions that often impose it.378
To examine this hypothesis, we first must determine how the decline in death
sentencing has been distributed across counties. One possibility is that the nation
has reached a new, more sensible capital punishment equilibrium, with fewer,

14, 2011).
Possible explanations for the decline in death sentencing include: (1) a drop in violent
crimes, starting a few years earlier than the drop in death sentences and declining by a comparable
STATISTICS, (last visited May 14, 2011); (2) concerns, fueled by
the Broken System studies and the exoneration of more than a dozen inmates on death row in Illinois,
see supra notes 149, 256–257, 278 and accompanying text, about the condemnation and possible
execution of the innocent, see Adam Liptak, Number of Inmates on Death Row Declines as
Challenges to Justice System Rise, N.Y. TIMES, Jan. 11, 2003, available at
ustice-system-rise.html (―[L]egal experts across the political spectrum agreed that public discomfort
with the administration of the system [including DNA exonerations] has played a significant role [in
the decline].‖); supra note 278 [DeLuna]; (3) the adoption of life without parole as the only
alternative to a death sentence for capital murder, see supra notes 301–303 and accompanying text,
together with a Supreme Court decision requiring trial judges to inform capital jurors of that
alternative if the prosecution contends that execution is required to keep the defendant from
offending again, see Simmons v. South Carolina, 512 U.S. 154 (1994); and (4) improvements in the
quality and sophistication of capital defense lawyers, including as a result of the Supreme Court‘s
stricter line on ineffective assistance of counsel between 2000 and 2005, see infra note 381 and
accompanying text. See generally Steve Mills, Death Sentences, Executions Decline, CHICAGO TRIB.
Dec. 31, 2004, at 1, available at
(discussing various theories); Neil A. Lewis, Death Sentences Decline, and Experts Offer Reasons,
at (last visited May 14, 2011) (similar).
See supra notes 242–270 and accompanying text.


more evenly distributed death verdicts imposed only for ―the worst of the worst‖
crimes.379 In this scenario, the death sentencing disparity between the minority of
death prone counties and others would decline, as would rates of reversible error
associated with high death sentencing rates as well as the majority-to-minority
subsidy associated with curing the errors.380 Another possibility, consistent with
the generalized decline in violent crimes nationwide since the 1990s, is that death
sentencing has dropped in roughly equal proportion across all counties. In this
event, a minority-majority disparity and majority-to-minority subsidy would
remain, but be smaller. A final possibility is that the drop in death sentences has
been the result of an abandonment of the field altogether by jurisdictions that
previously used the death penalty only sparingly. A result of this quiet series of
county-by-county abolitions would be an increase in the concentration of death
sentences in a minority of death prone counties and, ironically, an increase in the
extent to which communities that are trying to wean themselves from the penalty
end up subsidizing its heavy use by a shrinking minority of death prone localities.
A smaller tail would be wagging a larger dog.
B. An Uneven Decline
To determine the current distribution of death sentences in the United States,
we examined the counties responsible for the 768 new inductees to death row
during the years 2004 through 2009. 381 Our results support the last of our
hypotheses, that death sentencing is retreating to its bastions, as less frequent users
abandon the practice altogether. 382 Between 1973 and 1995, the most death

A proponent of this view is Oregon prosecutor Joshua Marquis, who has been a leader of
the capital litigation committee of the National District Attorney‘s Association. See Liptak, supra
note XXX (―Mr. Marquis said there were other reasons for the decline in death sentences.
‗Prosecutors in America are far more discriminating in the kinds of cases they submit to juries,‘ he
said. ‗There is a recognition that the death penalty should be reserved for the worst of the worst.
If you look back 20 years, there clearly were jurisdictions where some prosecutors overused the death
See supra notes 50–53, 181–187, 246–257 and accompanying text.
Data for death sentences by county for the 2004-09 period were provided by Robert J.
Smith of the Charles Hamilton Houston Institute, Harvard University Law School. See supra note
That hypothesis has not been entirely proven, however, because not enough years have
elapsed to be certain that the communities in question have stopped using the death penalty
altogether, as opposed to keeping it in their ―back pocket‖ for the very worst cases. See supra notes
311–317 and accompanying text. For evidence from Texas of a sharp and an uneven decline in
death sentencing in recent years, and for support for our thesis that marginal death-sentencing
counties are giving up on capital punishment, see David McCord, What‘s Messing with Texas Death


prolific 1% of Americans accounted for 9% of death sentences; the top eighth
(12.5%) were responsible for half of all death sentences.383 In our more recent
study, counties representing 1% of Americans had nearly doubled their share to
16% of all new death sentences, and counties containing one-eighth (12.5%) of
Americans now originate two-thirds of all death sentences (67%). 384 High
frequency death sentencing communities are now responsible for an even greater
share of the national total of condemned inmates than in the Broken System period
from the early 1970s to the mid-1990s.
Further corroborating the third hypothesis, many major death sentence
generators continue on much as before or are picking up speed. As Table 1
illustrates, between 1973 and 1995, the counties of Los Angeles, Maricopa
(Phoenix),385 and Oklahoma (Oklahoma City) by themselves accounted for 6.5%
Sentences, 43 TEX. TECH L. REV. 601, 603–05, 607–09 (2010-2011) (discussing less populous Texas
counties that are ―opt[ing]-out‖ of the death penalty); see also id. at 609–11 (discussing the ―plunge‖
in death sentencing in Harris County (Houston)). See also Roger C. Barnes, Is the Death Penalty
About to Die?, SAN ANTONIO EXPRESS-NEWS, June 14, 2011, available at
6.php (noting recent drop in number of death sentences in Texas and concluding that ―Texas is
having second thoughts about death sentences, and executions‖).
Clarke spreadsheet, supra note 39.
How do cosmopolitan Los Angeles and Maricopa (Phoenix) Counties fit our model of high
death sentencing communities with parochial tendencies? The answer in the former case is gangs
and in the latter is illegal immigration. To begin with, Los Angeles County has roughly three times
the population of its more cosmopolitan flagship City. See STATE & COUNTY QUICKFACTS, 2009
May 15, 2011) (showing that LA County has 9.85 million residents while LA city has 3.85 million)
Moreover, Los Angeles County has been plagued for decades by ethnic gangs—a classic case of
threatening ―outsiders.‖ See, e.g., National Public Radio, Race Distinguished Gang Violence in Los
Angeles, available at (March 11,
2008) (summarizing a report that found 700 active gangs in Los Angeles with a combined
membership of more than 40,000, mainly composed of ethnic minorities). Los Angeles County has
been by far the largest contributor to California‘s massive death row (the largest in the nation, at over
700), see Clarke spreadsheet, supra note 39, and California has the highest percentage of gang-related
homicides in its pool of capital-eligible cases as a result of a statutory aggravating circumstance that
allows a single police officer‘s testimony that a defendant fits the profile of a gang member
(regardless of any actual connection to gang activity or gang violence) to establish death-eligibility.
See, e.g., Alarcón, supra note 19, at 699, n.8 (identifying the gang-member aggravating circumstance,
among others, as contributing to the excessive ease of imposing a death sentence in California and
supporting proposals to ―decrease the number of crimes punishable by death and the circumstances
under which death may be imposed so that we only [capitally] convict . . . [those] ‗we truly have the
means and the will to execute‘‖ (citation omitted)). We are indebted to our colleague Jeffrey Fagan
for these insights on Los Angeles, based on research he is conducting into the relationship between
alleged gang membership and imposition of the death penalty in Los Angeles and California
generally. As for Phoenix, consider a recent article comparing Maricopa (Phoenix) and Pima


of the death sentences nationwide.386 More recently, however, their share has
almost doubled.387 Against the backdrop of a steep (64%) drop in the nation‘s
annual death sentencing rate, these counties‘ previously high rates remained stable
or, in Phoenix‘s case, increased substantially, accounting for the rise in their share
of the national total.388

(Tucson) Counties, in the former of which death sentencing is high and increasing and in the latter of
which, it is relatively low and decreasing substantially. Arizona Divided: A Tale of Two Counties,
THE ECONOMIST, March 31, 2011, at 28, available at
(discussing stark differences in the two counties‘ reaction to an influx of Hispanic immigrants;
Maricopa County, led by Sheriff Joe Arpaio, considers itself to be under siege from illegal
immigration and has responded harshly; Pima County has taken a more moderate line; also noting
how Pima County residents distinguish themselves from what they consider to be their more
parochial peers in neighboring Maricopa County); infra Table 1, text following note 348 and infra
Table 4, text accompanying note 359. (documenting Maricopa County‘s increasing and Pima
County‘s decreasing use of the death penalty). The boom in death-sentencing throughout all of
southern California, see infra 349 and accompanying text, may well be in reaction to both gangs and
illegal immigration.
The above comparison between Tucson and Phoenix, and our generalizations about a county as
large and diverse as Los Angeles, prompt two caveats to our overall analysis. First, the parochial
and libertarian forces we describe are neither immune to, nor necessarily more powerful than, a
variety of contingencies relating to such things as the priorities and political ambitions of a given
county‘s district attorney or the commission there of a particularly wrenching crime. Those
contingencies may account for wide discrepancies in death-sentencing practices among counties, and
within the same county over time, that cannot be explained by differences in the extent to which the
counties are, or over time have become, more or less parochial or libertarian. Without denying the
importance of such contingencies, our argument is simply that parochial and libertarian tendencies
exert considerable, if not always decisive, influence over death-sentencing practices in the ways we
have suggested. Indeed, one such influence may be a county‘s susceptibility to contingencies that
tend to generate death sentences—e.g., the election of a prosecutor with an unusually strong
disposition to use the death penalty, including as a springboard for higher office. Second, the fact
that Los Angeles County encompasses neighborhoods ranging from some of the most, to some of the
least, ―progressive‖ or ―permissive‖ in the nation requires caution in generalizing about the effect
there and in other populous jurisdictions of any single tendency. Even so, we think it is reasonable
to suggest that perceived threats from, for example, gangs and illegal immigration can trigger strong
enough reactions in particular segments of the population of a county to influence policies, even if
many or most other residents take a more nuanced view of the supposed threat.

Clarke Spreadsheet, supra note 39.
The data in Table 1 are from Clarke Spreadsheet, supra note 39.
Clarke Spreadsheet, supra note 39.


Table 1


Death Sentences
Year, 1973–1995

Per Death Sentences
Year, 2004–2009

Los Angeles, CA



Maricopa, AZ (Phoenix)



Oklahoma, OK




As Table 2 illustrates, three of Los Angeles County‘s fellow southern California
jurisdictions—Orange, Riverside and San Diego Counties—also are part of a death
sentencing boom in regions close to the border with Mexico.389
Table 2
Orange, CA

Death Sentences
Year, 1973–1995

Per Death Sentences
Year, 2004–2009

Riverside, CA



San Diego, CA




While the status quo ante continues with little interruption in these major
urban capital contributors and many smaller, rural counties,390 huge swaths of
communities either slowed down their death sentencing rates dramatically in
2004–2009 or did not use the death penalty at all.391 New Jersey and New


The data in Table 2 are from Clarke Spreadsheet, supra note 39.

Examples of small counties experiencing increases in death sentencing rates between
1973-95 and 2004-09 include Caddo Parrish, LA (death verdicts increased from 2 in 1973-95 to 6 in
2004-2009); Cherokee, TX (death verdicts increased from 0 in 1973-95 to 2 in 2004-2009); Houston,
AL (death verdicts increased from 6 in 1973-95 to 9 in 2004-2009); Madison, NE (death verdicts
from 1 in 1973-95 to 3 in 2004-2009); and Seminole, FL (imposed 6 death verdicts in both 1973-95
and in 2004-2009). See Clark spreadsheet, supra note 39.
Examples of less populous counties experiencing sharp declines in death sentencing
between 1973–95 and 2004–09 include Bay, FL (death verdicts dropped from 13 in 1973–95 to 3 in
2005–09); Glynn, GA (death verdicts dropped from 8 in 1973–95 to 2in 2005–09); Lexington, SC


Mexico abolished the death penalty in 2009, and before doing so, did not register a
single capital sentence in the 2004–2009 period.392 Quietly, neither did New
Hampshire and South Dakota, and neither did conservative Montana and Utah.393
Washington, Wyoming and Maryland each registered only one, and Colorado had
only two.394 Perhaps most surprising is Virginia—which after Texas has carried
out the second most executions in the United States since Gregg.395 Since 2005,
Virginia has imposed only one death sentence a year.396
Outside of these states, there were eight counties, collected in Table 3, with
over 500,000 people that handed down at least six death sentences in the Broken
System period, but none between 2004 and 2009.397 These counties represent a
sizable portion of the nation‘s citizens who in the earlier period were significant
death penalty consumers, measured by their localities‘ death sentencing practices,
yet they stopped imposing capital sentences in the six-year window of our later
study.398 They again are representative of many smaller counties that likewise
have seen death sentences plunge to zero since 2000.399 These communities are
(small city; death verdicts dropped from 18 in 1973–95 to 5 in 2005–09)); Montgomery, PA
(suburban; death verdicts dropped from 12 in 1973–95 to 3 in 2005–09)); Randall, TX (death verdicts
dropped from 8 in 1973–95 to 2 in 2005–09); and Russell, AL (rural/small city; death verdicts
dropped from 6 in 1973–95 to 3 in 2005–09). See Clarke spreadsheet, supra note 40.
Clarke Spreadsheet, supra note 39.
See Clarke spreadsheet, supra note 39; infra note 364 and accompanying text (citing
legislative action on abolition in Montana).
See Clarke spreadsheet, supra note 39.
Measured by its high number of executions, Virginia has long been thought of as a death
penalty heavyweight. But as Broken System II discovered, Virginia acquired that status not by being
a heavy death-sentencing state---in fact, it imposed rather few death sentences per capita compared to
states like Alabama, Oklahoma and Texas. Virginia managed a high number of executions, because
its low death-sentencing rate translated into a low reversal rate, giving it the nation‘s highest rate of
executions per death sentence. See BROKEN SYSTEM II, supra note 27, at xxx. In the recent period,
Virginia has moved from being a moderate death sentencing state to being a marginal death
sentencing state.
The data in Table 3 are from Clarke Spreadsheet, supra note 39.
These are large urban areas, generating many homicides each year, so the absence of death
sentences in the recent six-year period is more reliably indicative of ongoing trends than it is in
counties with fewer people and homicides.
Examples of less populous counties experiencing declines in death sentencing to zero
between 1973–95 and 2004–09 include Bowie, TX (death verdicts dropped to 0 in 2004–09 from 8 in
1973–95); Creek, OK (death verdicts dropped to 0 in 2004–09 from 7 in 1973–95); Cole, MO (death
verdicts dropped to 0 in 2004–09 from 12 in 1973–95); Columbia, FL (death verdicts dropped to 0 in
2004–09 from 10 in 1973–95); Meriweather, GA (death verdicts dropped to 0 in 2004–09 from 8 in
1973–95); and Rockingham, NC (death verdicts dropped to 0 in 2004–09 from 8 in 1973–95). See
Clarke spreadsheet, supra note 39.


widely dispersed throughout the country, defying any state-centric or
region-centric explanation for the decline in death sentencing.
Table 3
Muscogee, GA (Columbus)

Death Sentences Per Death Sentences
Year, 1973–1995
Year, 2004–2009

Palm Beach, FL



Marion, IN (Indianapolis)



Santa Clara, CA (San Jose)



Gwinnett, GA (suburban
Davidson, TN (Nashville)





Franklin, OH (Columbus)



San Francisco, CA




Finally, part of the drop-off in death sentences occurred in counties—several
of the largest in the nation—that used to be but no longer are in the high frequency
category, yet still impose some death sentences. Together, the seven counties
listed in Table 4 accounted for about a tenth of the nation‘s death sentences during
the 1973–1995 Broken System II period. Between 2004 and 2009, their share fell
to 3.6%—less than that of either Maricopa or Los Angeles County by itself.


Table 4
Cook, IL (Chicago)400

Death Sentences
Year, 1973–1995

Per Death Sentences
Year, 2004–2009

Miami-Dade, FL



Hillsborough, FL (Tampa)



Pima, AZ (Tucson)401



Baltimore County, MD








The process evidently continues. In the past year, the legislatures of
Connecticut, 402 Kansas, 403 Kentucky, 404 Indiana, 405 Montana, 406 Maryland, 407
Nebraska, 408 Ohio 409 Pennsylvania, 410 and of course, Illinois, 411 have given


Illinois abolished capital punishment in March 2011. See supra note 18.
See supra note 347 (comparing Pima (Tucson) and Maricopa (Phoenix) Counties.
See supra notes 311–315 and accompanying text.
CTR., [hereinafter Recent Legislative Activity]
(last visited July 15, 2011) (discussing House Bill 2323, which would abolish the death penalty and
replace it with life without parole).
Id. (discussing House Bill 292, to abolish the death penalty).
Id. (discussing Senate Bill 344, to abolish the death penalty).
Id. (discussing Senate Bill 185, to abolish the death penalty, which was defeated in the
House Judiciary Committee on March 18, 2011).
Id. (discussing House Bill 1075 and Senate Bill 837, to abolish the death penalty).
Id. (discussing Legislative Bill 276, to replace the death penalty with life without parole,
which passed the Judiciary Committee in March 2011).
Id. (discussing a bill introduced in the Ohio General Assembly in March, 2011 to abolish
the death penalty and replace it with life without parole).
Id. (discussing Senate Bill 423, to abolish the death penalty introduced on March 24, 2011);
see also Governor Rendell Urges General Assembly to Review Effectiveness of Pennsylvania's Death


significant attention to abolition proposals.412
C. A Smaller Tail Wagging a Larger Dog
The distribution of recent death sentencing declines in the United States does
not reveal a more sensible death penalty equilibrium in which capital sentencing
proclivities nationwide have converged around the penalty‘s use only for the worst
crimes, and error rates and the majority-to-minority subsidy have faded. On the
contrary, the data demonstrates that the problems we discuss above have
sharpened. An increasingly isolated core of counties continues to impose the vast
majority of death sentences. And they continue to do so at high—probably
error-inducing413 —rates, making it likely that only a small proportion of the
See supra notes 18 and accompanying text.
Evidence from California and Florida reveal how deceiving statewide data can be.
Although, as we suggest above, southern California counties (Imperial, Los Angeles, Orange,
Riverside, San Bernardino and San Diego) are frequent and increasing users of death sentences
(driving state totals up), see supra notes 347–349 and accompanying text, some northern California
counties that previously used the death penalty have more recently stopped doing so (Santa Clara and
San Francisco). Similarly, some of the biggest drops in the nation were in western and southern
Florida (Dade, Hillsborough, Palm Beach and Pinellas Counties), while some of the most persistent
capital-sentencing counties were in the northern and eastern parts of that State (e.g., Brevard,
Broward, Duval and St. Lucie Counties). See Clarke spreadsheet, supra note 39.
Our data do not prove that high capital sentencing rates continue to generate high rates of
serious error or high rates of reversal by state or federal courts, and no study comparable to Broken
System has addressed the question since Broken System II was completed in 2002. The fact that
only around fifty executions occur each year from among a death row population above 3200, see
supra notes 336–337 and accompanying text, reveals that capital review procedures are not speeding
up, which in turn suggests that state and federal courts continue to find claims of error to be deserving
of review. Indeed, the Justice Department‘s annual capital punishment reports indicate that the time
between death sentences and executions has risen steadily over the past decade. Compare, CAPITAL
PUNISHMENT 2009 – STATISTICAL TABLES 1 (DECEMBER 2010), BUREAU OF JUSTICE STATISTICS, (last visited July 15, 2011) (calculating the
average time from death verdict to execution of those executed in 2009 as 14 years and 1 month)
with, CAPITAL PUNISHMENT 2005, BUREAU OF JUSTICE STATISTICS, at 1 (December 2006), (calculating the average time from death
verdict to execution of those executed in 2005 was 12 years and 3 months) with CAPITAL
at 1 (December 2002), (indicating that the average time from death
verdict to execution of those executed in 2009 was 11 years and 10 months); see generally King et al,
supra note 186 (showing of sample of cases showing that capital federal habeas corpus petitions were
taking twice as long to complete as of 2000-06 than was true before 1996). Likewise, the fact that the
nation‘s death row population has come down somewhat over the last decade, notwithstanding a
decline in executions during the same period, see supra notes 336–339 and accompanying text,
suggests that reversals leading to non capital sentences on retrial continue to occur. Given
Congress‘s adoption of legislation in 1996 reducing federal prisoners‘ access to federal habeas corpus


verdicts will ever end in executions.
Instead of a harmonizing of death sentencing rates nationwide, or a smooth
decline across still disparate counties, we are witnessing a retreat of the death
penalty to a small bastion of counties where its use remains high or is being
fortified. In contrast, many jurisdictions that previously used it sparingly are
losing their stomach for it entirely. As capital punishment becomes more and
more of a minority outcome, and the discrepancy between the small capital
sentencing core of jurisdictions and the majority of less death prone communities
sharpens, so does the system‘s irrationality from the perspective of the latter.
They have to subsidize practices by their parochial and libertarian neighbors in
which they are no longer disposed to indulge themselves—and to continue bearing
public safety and judicial integrity costs that spill over from their neighbors‘ risky
The picture gets even more disturbing when we consider what Carol and
Jordan Steiker call ―symbolic [capital] states.‖ 414 Over 25% of the death
sentences handed down today originate in states that are profoundly unwilling or
unable to carry out most capital sentences unless they have the consent of the
condemned.415 Faced with the facts set out in Table 5, it is fair to say that a
death verdict in California, Connecticut, Kansas, Maryland, Nevada, Pennsylvania
and Tennessee is little more than a cruel joke. It is cruel because the victims‘
family members are not told that the punch line they anxiously await almost
certainly will never occur, and because of the psychological terror the system
inflicts on condemned men given an involuntary execution does occur now and
then.416 It is a joke—at least to the majority that pays for the system without
review, it is likely that the reversal rate in federal court has declined recently. That was the finding
of a recent study of non-final (i.e. only district court not court of appeals and Supreme Court)
decisions in a non-representative sample of federal habeas corpus cases filed in 2000-02. See King
et al, supra note 186, at 2, 3, 10 (finding reversal rate of 13% for sample of cases filed in 2000-02,
which is 35 times higher than the reversal rate in non capital habeas corpus cases; predicting that the
actual capital reversal rate is higher because 25% of the cases in the study were still under review
when the study ended and because cases in which reversals occur tend to take longer that
affirmances, but predicting that the rate would not reach the 40% level in the Broken System II study).
Given Broken System II‘s finding that state courts account for 90% of reversals of capital verdicts,
and that state and federal reversal patterns tend to be compensatory—with state rates increasing if
federal reversal rates decline, and vice versa, and with state courts focusing more on urban cases and
federal courts focusing more on rural ones, see BROKEN SYSTEM II, supra note 27, at 65, 218–19), it
is unlikely that trends in the current federal reversal rate, whatever it may be, are indicative of trends
in the state courts or the overall reversal rate.
Steiker & Steiker, supra note 186, at 1870.
See supra notes 288–293, 316 and accompanying text.


hoping to get much from it—because of the resources spent to achieve a minuscule
and capriciously determined number of involuntary executions.
Table 5


Since 1976



as % of
Death Row

as % of
Death Row

































In this sense, the irrationality of the system is getting worse. California and
Pennsylvania—prototypical ―symbolic states‖—accounted for 15% of death
sentences during the 1973–95 period.417 In our more recent 2005–2009 study
period, the two states imposed a quarter of all US death sentences.418 The
implication is clear, and from the subsidizing majority‘s perspective, troubling.
As majority communities and taxpayers take steps to curb the costs of their own,
previously modest, use of the death penalty, the subsidy they continue to pay is
emboldening the minority of death prone communities to intensify their profligate
ways. Where the Steikers found ―no obvious correlation between securing death
verdicts and carrying out executions,‖ 419 there now seems to be a negative
correlation. The more improbable executions become, the more death sentences
capital prone communities impose. The smaller the chance that the system will
generate the only outcome—executions—that distinguishes it from a system in
which life without parole is the most severe sentence, the greater the subsidy,
public safety and judicial integrity burden the majority of death avoiding


Clarke Spreadsheet, supra note 39.
Clarke spreadsheet, supra note 39.
Steiker & Steiker, supra note 19, at 1872.


communities and taxpayers must bear.420
Table 6 bears out this concern. It identifies the top six States in the nation
based on the number of death sentences imposed between 2004 and 2009. Then
it reports those States‘ ranking (out of thirty-two total states) based on the
historical probability that a death sentence each state imposes will actually result in
an execution.421 Five of the six top death sentencing States in the nation—all but
Texas—are ranked in the bottom two-fifths of States when it comes to how likely
it is that their death verdicts will be carried out.
Table 6

Death Sentences

Probability of Execution
(Rank out of thirty-two)



















Above, we show how for decades the majority of communities and taxpayers
that rarely use the death penalty have borne a huge share of the costs of the quick

Although the Steikers focus on ―symbolic states,‖ our analysis shows that the source of the
problem is a subset of death prone communities in the state. California and Florida, for example, are
made up both of counties that are increasing and those that are decreasing their use—but are not
decreasing their subsidy—of the death penalty. See supra Tables 1–4; supra notes 254–255,
259–260, & 262–266.
The number of death sentences imposed is from DEATH SENTENCES BY STATE AND YEAR,
supra note 12. The historical probability of an execution is calculated by dividing each state‘s
cumulative number of executions from 1977 to 2007 by its cumulative number of death sentences in
the same period. The data produced by this calculation and the rankings it generates are from,
CTR, (last visited July 15, 2011). Only states
that currently have the death penalty and carried out at least one execution between 1977 and 2007
are included.


and unvarnished death verdicts that a minority of parochial and libertarian
communities prefer over systematic law enforcement. We puzzled over the
majority‘s willingness to continue subsidizing the minority‘s profligacy. In this
Part we discovered evidence that the majority has recently tried to fight back by
decreasing its own use of the penalty. But far from chastening the minority, the
result has been to embolden it. Whether because of increasingly tight budgets in
communities that already underfund law enforcement, or to capture space cleared
in statewide capital punishment budgets by their neighbors‘ withdrawal from the
field, the minority have expanded their reliance on capital sentences in lieu of
effective prevention and adjudication, even as the likelihood has declined that their
error-laden death verdicts will ever be executed. On the assumption that the
majority, having already given the shirts off their backs, is increasingly unwilling
to continue being flayed, the next Part suggests some steps the majority might take.
A. The Insufficiency of Options Previously Proposed
This Article is the fourth in a series chronicling a ―broken‖ system of capital
punishment and proposing solutions.422 In Overproduction of Death and Broken
System II, we proposed a set of individual policy options designed to reconfigure
incentives and force the cost of effective death sentencing to be borne at trial.423
To some extent, state and federal institutions listened or figured it out for
themselves. One proposal was to improve the quality of counsel in capital
trials.424 In three rulings between 2000 and 2005, the Supreme Court required
more comprehensive mitigation investigations in the run-up to capital trials.425

See BROKEN SYSTEM II, supra note 27 (published in abridged form in Gelman et al., supra
note 15); Liebman, supra note 146; Liebman, supra note 19.
See BROKEN SYSTEM II, supra note 27, at 391–418 (reprised in GELMAN ET AL., supra note
15, at 254–60) (proposing policy options to narrow the death penalty to the most aggravated cases);
Liebman, supra note 146, at 2129 (suggesting strategies to get trial level officials to internalize the
costs of their errors in capital cases).
See Liebman, supra note 146, at 2147 (proposing that states adequately compensate
defense counsel and support services for indigent capital defendants).
See Rompilla v. Beard, 545 U.S. 374, 377 (2005) (holding that ―even when a capital
defendant‘s family members and the defendant himself have suggested that no mitigating evidence is
available, his lawyer is bound to make reasonable efforts to obtain and review material that counsel
knows the prosecution will probably rely on as evidence of aggravation at the sentencing phase of
trial‖); Wiggins v. Smith, 539 U.S. 510, 523 (2003) (requiring counsel to gather substantial


Another proposal was to adopt the alternative of life without parole.426 All death
penalty states have now done so.427 Together these changes dissuade jurisdictions
from pursuing marginal capital prosecutions out of fear of heavy defense costs and
given the availability of a stout alternative to death.428 In the process, the reforms
improve the reliability of the capital sentences that are imposed.
Our analysis reveals, however, that the effect of these steps is limited.
Although they almost certainly have contributed to declining use of the death
penalty in the majority of jurisdictions that already were fairly judicious capital
sentencers, we have produced substantial evidence that the minority of
jurisdictions that produce the lion‘s share of death sentences are not much affected
by these reforms. Recent patterns suggest that those communities are not pulling
back from the death penalty and may even be increasing the number of flawed
verdicts they impose.429 Our analysis also explains why. These communities‘
parochialism and libertarianism leads them to value death verdicts—particularly
for felony murder—for two reasons that in neither case are affected by court
reversals or modern carceral options like life without parole. First, they value
death verdicts for the clear message sent, then and there, about how the community
views and intends to respond to invasive outsider crime.430 Second, capital felony
murder lets them ration the few resources they are willing to spend on law
enforcement on the stranger crimes they most abhor, while retaining a sense of the
law in their own hands and while minimizing entanglement with the modern
administrative state.431 More generally, these jurisdictions value the verdict more
than the execution, so discipline tied to whether or not the verdict will be carried
out has little effect on them.432
In Slow Dancing with Death, we suggested that each state use comparative
proportionality review of murder sentences to define a statewide ―going rate‖ for
the amount of aggravation (once discounted by the amount of mitigation present in
mitigating evidence prior to the sentencing phase); Williams v. Taylor, 529 U.S. 362, 392 (1999)
(ordering relief where defendant‘s attorneys failed to investigate and present substantial mitigating
evidence at the sentencing phase of trial that might have influenced the jury‘s appraisal of the
defendant‘s moral culpability).
BROKEN SYSTEM II, supra note 27, at 404.
See supra note 303and accompanying text.
See supra notes 239–245, 301–303 and accompanying text.
See supra notes 347–350 and accompanying text.
See supra text following note 88; supra notes 95, 175–179, 205–207, 220–223 and
accompanying text.
See supra notes 128–141 and accompanying text.
See supra notes 95, 175–179, 205–207, 220–223 and accompanying text.


the case) that is needed for a death sentence, and to expunge outliers.433 This
strategy is partially responsive to the analysis here, because it would use the
application of the death penalty in the majority of jurisdictions that implement it
sparingly to discipline more death prone communities and orient the death penalty
towards only extreme cases.434 As Slow Dancing with Death describes, however,
after first seeming to require comparative proportionality review, the Supreme
Court pulled back—evidently because of the unwanted substantive responsibility
this placed on the Court to develop a national ―going rate‖ by comparing state
―going rates.‖435 The only state willing to implement the strategy in earnest was
New Jersey, and its failure to carry through with it was part of the justification for
the state‘s 2009 abolition of capital punishment. 436 We may never know,
therefore, whether comparative review would have enabled courts to manage the
dialectic between parochial/libertarian communities and other communities within
a state—or instead would have skewered them on one or the other of the dialect‘s
In any event, mechanisms are required to help the majority rein in the high
frequency death imposing minority. Repeal is the most obvious solution, but one
that, for reasons set out above and below,438 many courts and legislatures do not


Liebman, supra note 19, at 129 (arguing that the Supreme Court could have used returns
from juries, courts and legislatures nationwide to identify a national ―going rate‖ for death sentencing
and, on that basis, adopted a more defensible interpretation of the Cruel and Unusual Punishment
Clause and applied it to overturn outlying, excessively broad statutes, sentencing patterns and
individual death verdicts).
Id. at 129–30.
See NEW JERSEY DEATH PENALTY COMM‘N, supra note 235, at 49–50; supra note 18 and
accompanying text.
See Liebman, supra note 19, at xx; see also James S. Liebman & Lawrence C. Marshall,
Less is Better: Justice Stevens and the Narrowed Death Penalty, 74 FORDHAM L. REV. 1607, xx
(2006). Note a related dialectic that the Supreme Court indeed never managed to resolve, between
Justice Stewart‘s desire to avoid caprice by narrowing death verdicts to a small set of the worst of the
worst cases and Justice White‘s desire to increase the number of death verdicts to enhance the
deterrent and retributive power of the death penalty. See id. at 1608–48. Although Justice White
was far too rationalistic and scientific in his aim for additional deterrence and retribution to be a
perfect proxy for death prone parochial and libertarian communities, his never resolved battle with
Justice Stewart has elements of the dialectic between the minority and majority communities that we
discuss in this Article.
See supra notes 7–18, 59–141, 301, 304, 311–31 and accompanying text; infra notes
405–06 and accompanying text.


yet seem to be ready to adopt. 439 Accordingly, we limit ourselves here to
strategies in the shadow of the possibility of abolition, rather than abolition itself.
B. Regulatory Strategies
Death prone communities‘ suspicion of outside forces and distaste for
government complicate the political economy and the design of steps to regulate
local capital decisions. Enforcement is challenging, too, because the usual
appellate strategy for regulating local court action has little effect on these
communities, given the greater importance they assign to the verdict, which they
control, over its execution, which appellate courts control.440 Finally, the courts
and the death indifferent majority seem to be wary of crossing death prone
communities, out of respect for their traditions and self-reliance and fear of their
affinity for self-help.441
We take as a given, however, that the majority—which is a majority, after
all—has resolved to fight back.442 Moreover, the majority holds the top card; it
can repeal the death penalty—as legislatures in red and blue states alike are
currently considering.443 Under these circumstances, the threat of repeal provides
a classic ―penalty default.‖ It is an outcome about which both ―sides‖ may have
enough doubts (as to whether it will be accomplished) and fears (that it will be
accomplished and engender a frightening response) that they will agree to
cooperate and share secret information to stave off the uncertainty and feared
effects.444 Or course, if the majority is disposed to seek abolition, it should do so
and outright end the death penalty‘s fiscal, public safety and judicial integrity
harms. But if the majority is unsure it can achieve abolition or that it wants to, it


See supra notes 360–368 and accompanying text (discussing abolition proposals in several
states, thus far unsuccessful).
See supra notes 95, 175–179, 205–207, 220–223 and accompanying text.
See supra notes 94–102 and accompanying text.
See supra notes 332–372 and accompanying text.
See supra notes 360–370 and accompanying text.
See, e.g., Ian Ayres & Robert Gertner, Filling Gaps in Incomplete Contracts: An Economic
Theory of Default Rules, 99 YALE L.J. 87, 91–93 (1989) (describing how contract law uses onerous
and unpredictable requirements to encourage contracting parties to ―bargain around‖ the rule, thus
creating an incentive to redress information asymmetries in order to develop solutions better tailored
to the particular conditions than any single solution the law could prescribe); Bradley C. Karkkainen,
Information Forcing Regulation and Environmental Governance, in LAW AND NEW GOVERNANCE IN
THE E.U. AND THE U.S. 301–05 (Grainne de Burca & Joanne Scott eds. 2007) (citing authority and
extending Ayers & Gertner‘s analysis to public regulatory regimes).


can use the threat that rising costs will drive it to embrace that ―nuclear‖ option to
induce the minority‘s cooperation.
1. Less, Not More, Externalization of Costs
As we have noted, the majority‘s main goal should be to get death prone
communities to internalize more of the costs of their capital proclivities.445 This
principle may seem obvious, but it rules out a number of oft-proposed capital
reform strategies that have the counterproductive effect of increasing the state
(majority-to-minority) subsidy of local capital sentencing practices. Examples
include statewide public defender systems and state prosecutors on loan to
localities;446 state mechanisms for sharing the cost of capital prosecution and
representing the defendant; 447 and Professor Gershowitz‘s recent proposal to
require ―[a]ll aspects of death penalty cases—charging, trial, appeal, and
everything in between—[to] be handled at the state level by an elite group of
prosecutors, defense lawyers, and judges.‖ 448 Professor Smith‘s trenchant
response to Gershowitz‘s proposal applies to the other suggestions as well:
If states take over the enforcement of the death penalty . . . or underwrite
the cost of local prosecutions . . ., prosecutors will have less incentive to
show . . . restraint . . . . Given the greater resources that are available to
states, [they] are better positioned than localities to bear the financial
costs of seeking the death penalty. Consequently, . . . the most likely
result [is] more death sentences.449


See Liebman, supra note 146, at 2142 (―To succeed, reform efforts must realign the
incentives of all key players . . . to make it costly for police and prosecutors to pursue marginal, and
to obtain undeserved, capital judgments‖); supra notes 378–395 and accompanying text.
Representation, (last visited May 17, 2011) (noting that ―in 2009, the
U.N. Special Rapporteur on extrajudicial, summary or arbitrary executions conducted an official visit
to the United States to examine the administration of the death penalty in Alabama and Texas‖ and
―called for the two states ‗to establish well-funded, state-wide public defender services‘‖).
Proposals of this sort have recently been made, for example, in California and Washington
State. See California Death Penalty Comm‘n, supra note 254, at 20 (recommending that counties be
―fully reimburse[d]‖ by the state for payments for defense services and that the backlog of appeals be
addressed by state funding of an expanded state public defender); Washington State Bar Report,
supra note 228, at 33 (arguing that, in order to keep ―financial restraints‖ from interfering with
charging decisions, ―the state [should] assume the prosecution and defense costs of all aggravated
murder cases‖).
See Gershowitz, supra note 21, at 310.
Smith, supra note 21, at 120.


To be sure, Professor Gershowitz aims to dampen local excesses by putting
the capital charging decision in the hands of elite state officials.450 But for local
communities, this new effusion of state power and withdrawal of the sacred
prerogative of local communities to define the ―worst of the worst‖ for themselves
would certainly be a bridge too far, tantamount to repealing the death penalty.
For Gershowitz‘s proposal to go forward, therefore, the decision to proceed
capitally almost certainly would end up in the hands of local prosecutors and grand
juries, or of state officials whose capital sentencing dispositions mirror those of the
death prone communities. In either event, Professor Smith‘s objections are on
The more difficult question is whether the majority should remove state
subsidies that already exist.451 The problem is that removing existing supports,
such as statewide public defenders, could aggravate the problem by generating
more error-prone verdicts as a result of under-funded local representation.452 The
test for whether to remove existing subsidies is, therefore, straightforward: if the
local action being subsidized helps the locality directly achieve the one thing it
wants—death verdicts—removing the subsidy is a good idea. An example is
state sharing of the cost of the prosecution itself. Otherwise, however, the goal
should not be to remove the existing subsidies but to forbear adding new ones.
2. The Problems with Performance-Based Approaches
To get localities to stop incurring, or externalizing, capital sentencing costs,
states might directly aim (1) to limit death verdicts or (2) cost them out and impose
a corresponding charge on the sentencing community. Professor Gershowitz has
proposed a version of each of these strategies.
Most simply, Gershowitz has proposed to cap the number of capital
prosecutions a community can bring each year at ―two capital prosecutions for
every 100 murders‖ and for smaller jurisdictions with fewer murders, a maximum
of ―one capital prosecution per year‖ 453 This approach directly attacks the


See Gershowitz, supra note 21, at 338–39.
See supra notes 242–276, 402–403 and accompanying text.
See supra notes 75, 174, 308–389, 402–403 and accompanying text.
Adam M. Gershowtiz, Imposing a Cap on Capital Punishment, 72 MO. L. REV. 73, 78–79



problem of overuse of the death penalty. But its rigidity is enough to turn almost
anyone into a regulation hating libertarian. Prosecutors will complain that no
―right‖ number of death verdicts can apply to all jurisdictions and that
―heinousness‖ is not evenly distributed across murders and years, so prosecutors
would be forced to treat like crimes, defendants and victims differently; that a
county with only ninety-nine murders in a year will be senselessly disadvantaged
relative to an identical county with 101; that cities with many murders will be
disadvantaged in relation to towns with fewer than fifty, which nonetheless get
their one capital prosecution; and that counties should be able to trade ―death
permits,‖ creating a market in lives.454 Defendants will complain that creative
charging practices will let prosecutor manipulate the number of ―murders‖ in their
jurisdiction to increase their cap; that, in order to fill the quota in a ―down‖ year for
murders, less egregious cases will be selected; and that the remedy in any event
will not deter ―quick and dirty‖ procedures in cases that are chosen to meet the
A different approach analogizes the modern death penalty to a public service
that only a minority of people value, justifying a hefty ―user fee‖ each time a
county initiates a capital case. Charging the same fee in every case, however,
will miss the distinct uses counties make of the death penalty and the different
levels of risk their verdicts create. Even worse, whatever good cause is the
beneficiary of the funds from the fees will have an incentive to urge expansion of
the death penalty to generate more funds—as public schools in many states now
cheer on gamblers and purchasers of regressive lottery tickets. The temptation
then would be to set the fee, not to cover the subsidy, but to generate maximum
income. This would be an especially dangerous outcome if the court system
receives the funds, because it would encourage judges to forbear reversing cases
for fear of decreasing revenue. Charging more for post-reversal retrials would
create the opposite bias.
Professor Gershowitz tackles these problems in the last of his proposals by
analogizing capital cases not to boutique public services but to dangerous
nuisances.456 To get counties to use capital punishment sparingly, Gershowitz


Gershowitz rejects a ―cap and trade‖ strategy that would let prosecutors buy and sell
authorizations to proceed capitally. Gershowtiz, supra note 409, at 110, n.112. See generally
Bruce A. Ackerman & Richard B. Stewart, Reforming Environmental Law, 37 STAN. L. REV. 1333,
1334–41 (1985) (claiming that tradable permits ―at one stroke, cure many of‖ command and control's
―basic flaws‖).
Gershowitz, supra note 409, at 110–16.
Gershowitz, supra note 21, at 890.


proposes that state legislatures ―requir[e] county prosecutors to post a cash bond
and transmit the money to the state treasury before filing capital charges.‖457 If
prosecutors get a death sentence and preserve it on appeal, ―the bond would be
returned to the county with interest.‖458 But ―if the case is reversed the county
would forfeit the bond to the state.‖459 This approach addresses the right problem
(overuse) at the right moment (time of filing capital charges) and avoids
inflexibility by letting local prosecutors ―determine which capital cases they want
to pursue while forcing them to take full responsibility for those decisions‖ in the
event of reversal.460
But there is a problem, which also afflicts more traditional approach of
assessing costs or the other side‘s attorney‘s fees in the event an adverse verdict.
All these strategies would make state judges hesitant to overturn flawed death
sentences. Broken System II shows that elected state judges already are more
reluctant to overturn otherwise similar death sentences imposed in rural areas,
evidently due to the stronger adverse reaction there.461 Fear of provoking that
same reaction after every reversal could lead judges to rubber-stamp death verdicts,
encouraging more corner-cutting in trials and increasing appeal costs by delaying
deserved reversals until the case completes both state court levels of review and
moves to federal court.
Most of our objections to this point focus on the difficulty of measuring death
penalty costs and assuring they are borne by the communities that generated them.
From those communities‘ perspective, however, the bigger problem is the failure
to gauge the benefits they derive from the death penalty—benefits, they would
argue, that are greater in some communities than others due to different value
choices. The user ―fee‖ and ―cap‖ approaches directly impinge on communities‘
ability to be the arbiter of the benefits they get from death verdicts because those
strategies give the same answer for all communities to the questions of ―How
much is each case worth?‖ and ―How many is enough?‖ Even worse from this
perspective is Gershowitz‘s alternative proposal to give the capital charging
decision to state officials,462 because that takes away localities‘ longstanding and


Id. at 863.
Id. at 863–64.
Id. at 865.
See BROKEN SYSTEM II, supra note 27, at 65, 218–19.
See supra notes 404–06 and accompanying text.


highly prized ability to make the cost-benefit judgment in response to each crime.
The ―bond‖ and ―costs‖ strategies are less pernicious from this perspective, but
they have a similar odor insofar as they attach a punitive message to communities‘
capital sentencing decisions.
3. Local Improvements in Defense Representation
A more promising approach is to require trial procedures that improve the
quality of death verdicts and deter communities from seeking them except when
they meet statutory and constitutional standards. A good example is the
2000–2005 series of Supreme Court decisions requiring better capital defense
representation.463 In theory, at least, this approach can increase the reliability of
individual verdicts and decrease appeal costs. It also may discourage future
prosecutions in marginal cases out of a concern that when skilled defense lawyers
obtain not guilty or life verdicts in high profile cases, the message communicated
is the opposite of the fear-allaying and retributive message that parochial and
libertarian communities want death verdicts to convey. By diminishing the value
of the Capital Felony Murder Doctrine for rationing those communities‘ response
to outside crime, this reform might even trigger more professional and systematic
law enforcement techniques to allay community crime fears.
The problem, however, is that the enforcement mechanism for judicially
imposed requirements of this sort—appellate reversal of the death verdict and
remand for re-trial—works poorly in the current context. To be sure, reversals do
chasten communities that rarely use the death penalty. But those communities do
not impose the bulk of the error costs. Instead, those costs are imposed by death
prone communities for which the remedial dialectic—that the appellate court will
not let an execution go forward until the locality cures the systemic defects that
cause the errors—has no clout. Those communities value the initial verdict far
more than the execution, and they do so out of a strong desire not to engage in the
systematic and professional practices needed to cure the error.464
An important lesson, therefore, is that interventions will not be effective
unless they occur before capital prosecutions are brought, not after verdicts are
imposed. This constraint rules out exclusive reliance on rights enforced on
appeal. It does not, however, make all procedural solutions ineffective. Recall


See supra notes 380–381 and accompanying text.
See supra text following note 88; supra notes 95, 175–79, 205–07, 220–23 and
accompanying text.


our premise that cooperative steps are possible in the shadow of the penalty default
created by the repeal option. Under this impetus, it might be possible to secure
legislation making the death penalty a local option, which counties invoke by
agreeing to create a mechanism for qualified, well-paid capital defense
Legislation to this effect might include standards that county plans for defense
representation must meet, a review process for deciding whether the plans satisfy
the standards and post-approval monitoring of opt-in counties‘ track record of
reversals and other factors. The law might assign an administrative arm of the
state courts or the state attorney general to review and monitor plans or give the
task to a panel with judicial, prosecutorial, defense and taxpayer representatives
aligned with the majority interest in professional proceedings that moderate error.
The law should require the reviewing agency to benchmark each county‘s
proposed plan against those from similarly situated counties to create a race to the
top. Later, the agency should publicize counties‘ appellate track records and
compare counties‘ records to determine whether each continues to meet the opt-in
criteria. 466 In this way, the law would use results from appellate and
post-conviction review to discipline use of the death penalty without relying on
judges to enforce the discipline through ineffective directives to ―get it right or we
won‘t execute,‖ and without using fee options or bond forfeitures that create
incentives for too many death sentences or too few reversals.
The same outcome might be achieved less formally through negotiations with
capital prone counties that commit, as a way to stave off repeal, to create and fund
a qualifying mechanism for improved representation and to monitor and publicize
counties‘ comparative appellate track records. The goal in either event is to use
death prone counties‘ desire to avoid repeal and their freedom to choose how to
meet the representation standards and achieve a viable record on appeal to induce
cooperation. Of course, by increasing the cost of death verdicts, these approaches
would diminish the benefit death prone communities derive from death verdicts.


An example of this type of mechanism—albeit one available to states, not localities, and
that few states have thus far opted into—is the option the federal habeas statute gives states to obtain
more expedited and favorable habeas proceedings in capital cases if they agree to create systems for
providing qualifying state post-conviction representation to condemned prisoners. See 28 U.S.C. §§
2261, 2265 (2006); HERTZ & LIEBMAN, supra note 248, § 3.3.
For an analogous regime, see Katherine R. Kruse, Instituting Innocence Reform:
Wisconsin's New Governance Experiment, 2006 WIS. L. REV. 645, 670–73 (describing Wisconsin
legislation under which counties must adopt plans, reviewed by the state attorney general, to assure
the reliability of eyewitness identifications).


What distinguishes this strategy from the cap, fee and bond approaches is that it
enables each community to make its own judgments about how best to evaluative
the relative benefits and costs and how to maximize the former and minimize the
latter across all cases and in each individual case.467
4. Managed Prosecution
Along with improved defense representation at capital trials, Broken System II
proposed making the decision to prosecute capitally a more deliberative
process.468 The aim was to improve judgments often made hastily under the
influence of emotions generated by the crime and invite defendants to inform
prosecutors about competing considerations. 469 The United States Justice
Department, for example, requires U.S. Attorneys (some of whom have advisory
committees of their own) to secure approval for capital prosecutions from a central
Capital Case Review Committee that also hears from defendants.470 A number of
commentators and death penalty commissions have made similar proposals.471
This Article explains why prosecutors in death prone communities resist
procedures of this sort. Precisely because they are deliberative and take time,
these procedures undermine a central feature of the quick and unvarnished death
penalty that parochial and libertarian communities value. They diminish the
communities‘ and prosecutor‘s ability, as a substitute for systematic crime fighting
strategies, to make immediately and emotively plain to fearful citizens, the
perpetrator and prospective miscreants how seriously they take the invasion it has
suffered and how harshly it is disposed to respond.
Still, in the context of the majority‘s use of the possibility of repeal to induce
cooperation, there may be a place for proposals for more deliberatively managing
the decision whether to charge capitally. Facing facts, the majority might want to
acknowledge the dialectic between the parochial and libertarian dispositions of the
minority of death prone communities and their own more modern views on


See supra notes 406–12 and accompanying text.
See BROKEN SYSTEM II, supra note 27, 409–10.
See id.
See Rory K. Little, The Federal Death Penalty: History and Some Thoughts about the
Department of Justice‘s Role, 26 FORDHAM URB. L.J. 347, 404–40 (1999); Benjamin Weiser,
Pondering Death by Committee: What Is a Capital Crime. Federal Panel Decides Case by Case,
N.Y. TIMES, June 26, 1997, at B1.
See Gershowitz, supra note 21, at 338–42, n.163 (collecting proposals and making one).


penology and government and their frustration at having to subsidize the
minority‘s risky practices. If that dialectic is to be resolved in some fashion, the
best time to get agreement to have the conversation is in the shadow of the threat
of repeal, and the best time and a very good way to have the conversation is on a
contextualized, case by case basis as each possibly capital case presents itself.
Agreement, then, might be reached to predicate counties‘ opt-in to the death
penalty—or to predicate some other compromise that heads off a cataclysmic vote
on repeal—on the minority‘s agreement to make death sentencing decisions only
after discussion with a committee of sober representatives of the majority.
Questions to be addressed would be those at the heart of the dialectic: why is the
death penalty appropriate and necessary in the case? What risks do a death
prosecution and verdict pose for the community, the victim‘s family, the rest of the
state and its taxpayers and the defendant? And (in order to provoke a pretrial
version of the comparative proportionality review the Supreme Court has declined
to require after the fact472) what does the community‘s and other communities‘
appellate track record in similar cases suggest about the wisdom of going forward
The forum and format for such a discussion are matters of choice. Certainly
the views of the victim‘s family and defendant should be heard, and
representatives of the majority that foots the externalized portion of the capital
punishment bill should help do the probing. A local committee might serve,
although it might have too little diversity to assure that the leavening questions are
sincerely asked and answered. A statewide standing committee would be better,
to avoid those risks and gradually reveal—and perhaps impel—a rough, statewide
―going rate‖ that does indeed resolve the dialectic.473
As a matter of the facts on the ground, the death penalty in the United States
today is a minority institution. Even in avidly pro death penalty states, only a
small fraction of local jurisdictions with a minority of the population impose the
great bulk of death verdicts. The death penalty offers these parochial and
libertarian communities a passionately expressive, almost violent, response to the
outsider, cross-boundary incursions on their security that most terrify them. At


See supra notes 389–391 and accompanying text.
See supra notes 391–393 and accompanying text.


the same time, the penalty lets them severely economize on the freedom and fruits
of their labors that they have to surrender to the modern administrative state.
As spare and unvarnished as this response is locally, it is costly to the
majority of communities and their residents who do not much use the death
It leaves crime unattended that spills over into neighboring
communities. It generates verdicts fraught with costly error. It misleads the
family of murder victims into thinking death verdicts will be executed, not just
expressed. It tortures the condemned with a punishment positively Czarist in its
cruelty, of life without possibility of parole, but with a chance of execution a dozen
years later. It heaps scorn on court proceedings that churn but never conclude.
It costs $20 to $125 million extra for the rare, random execution that now and then
does occur. Although hidden and diffuse, the costs are so great that one has to
wonder why the majority puts up with them.
Indeed, even as the majority of death indifferent communities and citizens
recently have moved towards a quiet repeal of the death penalty in their own
locales, their death prone neighbors have maintained, even intensified, its use. A
diminishing set of communities is responsible for a larger share of death verdicts.
Ironically, however, death prone communities have grown increasingly indifferent
to executions. A quarter of the country‘s 3,200 condemned sit on a death row
where they are more likely die of illness or old age than by execution.474 The
seemingly senseless costs imposed on the many by the few increase.
There is evidence, however, that the majority is restive. Whether abolition is
in the cards is uncertain, both because of the majority‘s fear of the minority‘s
violent reaction to repeal and the majority‘s respect for the minority‘s traditions
and self-reliance. But a disposition to make their cost externalizing neighbors
pay their own freight does seem to exist.
We promised at the outset that we would ourselves acknowledge the
minority‘s traditions and self-reliance.475 And although we have not minced
words about the crude advantage the minority takes of the majority, we will
perhaps be criticized in the end for respecting the minority and its values too much.
Still, short of the majority‘s most logical response—repeal—there are


George Skelton, Repeal the Death Penalty, L.A. TIMES, July 14, 2011 (―A condemned man
in California is more likely to die of old age than an execution. Although 13 have been executed, 78
have died of natural or other causes.‖).
See supra notes 55–56 and accompanying text.


accommodations to be had to reduce the burden the minority foists on the majority,
and we propose two of them.