POLICE MISCONDUCT AS A CAUSE OF
WRONGFUL CONVICTIONS
RUSSELL COVEY
ABSTRACT
This study gathers data from two mass exonerations resulting from
major police scandals, one involving the Rampart division of the L.A.P.D.,
and the other occurring in Tulia, Texas. To date, these cases have received
little systematic attention by wrongful convictions scholars. Study of these
cases, however, reveals important differences among subgroups of
wrongful convictions. Whereas eyewitness misidentification, faulty
forensic evidence, jailhouse informants, and false confessions have been
identified as the main contributing factors leading to many wrongful
convictions, the Rampart and Tulia exonerees were wrongfully convicted
almost exclusively as a result of police perjury. In addition, unlike other
exonerated persons, actually innocent individuals charged as a result of
police wrongdoing in Rampart or Tulia only rarely contested their guilt at
trial. As is the case in the justice system generally, the great majority
pleaded guilty. Accordingly, these cases stand in sharp contrast to the
conventional wrongful conviction story. Study of these groups of wrongful
convictions sheds new light on the mechanisms that lead to the conviction
of actually innocent individuals.
I. INTRODUCTION
Police misconduct causes wrongful convictions. Although that fact has
long been known, little else occupies this corner of the wrongful
convictions universe. When is police misconduct most likely to result in
wrongful convictions? How do victims of police misconduct respond to
false allegations of wrongdoing or to police lies about the circumstances
surrounding an arrest or seizure? How often do victims of police
misconduct contest false charges at trial? How often do they resolve
charges through plea bargaining? While definitive answers to these
questions must await further research, this study seeks to begin the
 Professor of Law, Georgia State University College of Law. Special thanks are owed to Eric
Coffelt, who worked hard and traveled far to help me collect data for this study. I also wish to thank
Brandon Garrett and Sam Gross for their excellent comments on earlier drafts, and to Max Compton
for his research assistance.

1133

Electronic copy available at: http://ssrn.com/abstract=2349872

1134

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

inquiry. Specifically, this study attempts to improve our understanding of
the intersection of police misconduct and criminal justice, and, more
generally, to contribute to the ever-growing bank of knowledge about
wrongful convictions.
What we do know about wrongful convictions comes largely from
studies of cases terminating in exonerations. These exoneration studies
have produced a rich dataset from which several factors that contribute to
wrongful convictions have been identified. While critically important, this
dataset has significant limitations, chief of which is selection bias. The
vast majority of the exonerations studied to date occurred in rape cases
following DNA testing and murder cases often involving the death
penalty. Such cases, comprising a tiny sliver of the criminal justice system
workload, are relatively unrepresentative of the vast majority of felony
convictions. As a result, and as researchers compiling these datasets
acknowledge, the most closely analyzed data on wrongful convictions
does not capture a representative sample of the probable distribution of
wrongful convictions that occur.1
Drawing on new empirical data, this article adds to the wrongful
convictions dataset by assessing another group of exonerees—those
exonerated following revelations of systemic police and/or prosecutorial
misconduct. Specifically, this Article examines two high-profile scandals
that saw the wrongful conviction and later formal exoneration of large
numbers of persons. To date, little attention has been paid to such
exonerees. This, I argue, has affected perceptions of the scope and nature
of the wrongful conviction problem. The profile of persons exonerated
following revelations of major police misconduct varies dramatically from
that of the typical capital murder or DNA exoneree. The defendants in the
mass exoneration cases were convicted of different types of crimes, faced
less severe punishments, and were far more likely to plead guilty than
other exonerated defendants. Using extant data, earlier exoneration studies
have found that the primary cause of the wrongful convictions in those
studies is witness misidentification.2 Based on those findings, some
1. See Samuel R. Gross & Barbara O’Brien, Frequency and Predictors of False Conviction:
Why We Know So Little, and New Data on Capital Cases 11–12 (Mich. Law Sch. Pub. Law Res. Paper
No. 93, 2007), available at http://ssrn.com/abstract=996629.
2. See, e.g., D. Michael Risinger, Innocents Convicted: An Empirically Justified Factual
Wrongful Conviction Rate, 97 J. CRIM. L. & CRIMINOLOGY 761, 773 n.19 (2007) (finding, based on
analysis of sixty-two DNA exonerations as of 2000, that mistaken eyewitness identification is the most
common factor); Samuel R. Gross et al., Exonerations in the United States 1989 through 2003, 95 J.
CRIM. L. & CRIMINOLOGY 523, 542 (2005) (reporting based on study data that the “most common
cause of wrongful convictions is eyewitness misidentification”). Edwin Borchard reached the same

Electronic copy available at: http://ssrn.com/abstract=2349872

2013]

POLICE MISCONDUCT

1135

commentators and reform-minded organizations have drawn the
conclusion that witness misidentification is, in general, the leading cause
of wrongful convictions.3 While the former claim was correct at the time
those observations were made, the latter generalization likely was not
warranted. The data we currently have is simply too limited to permit any
accurate generalizations about how frequently wrongful convictions occur,
or which contributing factors generate them, in specific or even rough
proportion.4
Moreover, as one of the nation’s leading experts on exonerations,
Professor Samuel Gross, has frequently emphasized, the primary causes of
wrongful convictions are almost certainly crime-specific.5 That is, the
factors that tend to cause wrongful convictions in rape cases are different
from those that cause wrongful convictions in murder cases, and different
from the causes of wrongful convictions in burglary cases, assault cases,
and drug cases. The next generation of research must approach wrongful
convictions in a more fine-grained manner.
To that end, this study gathers data from two mass exonerations
resulting from major police scandals. These exonerations are starkly
different than most of the exonerations previously studied. In the “minerun” cases (if there is such a thing) resulting in individual exonerations,
often as a result of DNA testing, several contributing factors, ranging from
eyewitness misidentification to false confessions to faulty forensic
evidence and testimony, have been identified. In contrast, wrongful
convictions in the mass exoneration cases are tied together by a single
dominant causal factor: police misconduct. Prior exoneration studies have
not focused on this group of exonerees, nor, by and large, have they

conclusion in his 1932 study of wrongful convictions. See EDWIN M. BORCHARD, CONVICTING THE
INNOCENT: ERRORS OF CRIMINAL JUSTICE xiii (1932).
3. Drawing on its database of persons exonerated through DNA testing, the Innocence Project
also claims that mistaken witness identifications are the leading cause of wrongful convictions. See
Facts on Post-Conviction DNA Exonerations, INNOCENCE PROJECT, http://www.innocenceproject.org/
Content/351.php. See also Keith A. Findley & Michael S. Scott, The Multiple Dimensions of Tunnel
Vision in Criminal Cases, 2006 WIS. L. REV. 291, 292 (2006) (describing “mistaken eyewitness
identifications” as “the most frequent single cause of wrongful convictions”).
4. In any event, new data on wrongful convictions suggests that witness misidentification no
longer can claim the top spot as a contributing cause of known exonerations. See Samuel R. Gross &
Michael Shaffer, Exonerations in the United States, 1989–2012 40 (Mich. Law Sch. Pub. Law and
Legal Theory Working Paper Series, Working Paper No. 277, 2012), available at http://ssrn.com/
abstract=2092195 (identifying “perjury or false accusation” as the leading known cause of wrongful
convictions during time period studied).
5. See id. at 102.

Electronic copy available at: http://ssrn.com/abstract=2349872

1136

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

incorporated data from these cases into the datasets.6 This Article does just
that. The experiences of those wrongfully convicted as a result of police
misconduct differ from other exonerees in interesting and potentially
important ways. The exonerations resulting from the Rampart and Tulia
police scandals raise the profile of police misconduct as a known cause of
wrongful convictions. In addition, the tendency of exonerees in these cases
to plead guilty rather than go to trial confirms what many have long
suspected: that the problem of wrongful convictions is not limited to the
small number of cases in which innocent defendants unsuccessfully
contest their guilt in a jury trial.
The article proceeds as follows: Part II describes the Rampart and Tulia
cases in detail, and sets forth the circumstances leading to mass
exonerations in those cases. Part III provides a brief description of data
about wrongful convictions that has been generated in the literature to
date. Part IV begins by describing the data used in this study. It then
examines the Rampart and Tulia exonerations in more detail, identifying a
subset of “actually innocent” exonerees in these cases that can be
compared with exonerees in other studies. Part IV then undertakes a closer
analysis of the circumstances and mechanisms that led to convictions of
innocent Rampart and Tulia defendants. One of the most interesting
contrasts between the mass exoneration cases studied here and other
instances of wrongful conviction is that, as compared to other groups of
exonerees, the great majority here pleaded guilty. Part V takes a closer
look at this critically important phenomenon and offers some hypotheses
as to why innocent defendants plead guilty in these cases at such a high
rate. Disturbingly, the evidence suggests that the factors at work here—
coercive penalties for contesting guilt at trial, coupled with few effective
defense strategies and unsympathetic forums—may describe the prevailing
conditions for a large number of, perhaps even most, criminal defendants.
Part VI considers various types of police misconduct, documenting the
prevalence of both “procedural perjury” and “substantive perjury,” and the
fine line between them. Part VI then compares the experiences of actually
innocent exonerees with those who the evidence suggests, though
wrongfully convicted, were probably not actually innocent. Based on this
comparative data, it finds evidence that innocence does dissuade some
defendants from pleading guilty, but that any “innocence effect” has only a
minor impact on guilty plea rates. Part VII briefly concludes.

6. One major exception is a new project attempting to compile a comprehensive catalogue of
known exonerations, including mass exonerations, in a “National Registry of Exonerations.” See id. at
2.

2013]

POLICE MISCONDUCT

1137

II. THE RAMPART AND TULIA EXONERATIONS
On December 16, 1997, the L.A.P.D. arrested police officer David
Mack on charges of stealing $722,000 from a Los Angeles area Bank of
America. Three months later, the department fired two other officers,
Brian Hewitt and Daniel Lujan, for severely beating a handcuffed prisoner
in an interrogation room.7 The common thread was that all three officers
were either former or current members of the Rampart CRASH, or
Community Resources Against Street Hoodlums, unit.8 Rampart is an area
covering 7.9 square miles to the northwest of downtown Los Angeles.9 It
is the most densely populated portion of Los Angeles, with 36,000 people
per square mile, and is widely known as a locus of gang activity. 10 At the
same time, the Rampart CRASH unit had a reputation for operating in a
largely autonomous fashion with little to no oversight.11 The arrest of
Officer Mack and termination of Officers Hewitt and Lujan motivated
L.A.P.D. Chief of Police Bernard C. Banks to form a special task force to
investigate the Rampart CRASH unit.12
Then, on March 2, 1998, six-and-a-half pounds of cocaine disappeared
from an evidence room in Los Angeles.13 Within a week, the special task
force investigators honed in on Los Angeles police officer Raphael Perez,
a member of Rampart CRASH, as the primary suspect. A year later, trial
on the charge ended with a hung jury.14 Shortly thereafter, Perez cut a deal
with prosecutors, agreeing to cooperate with a government investigation of
police wrongdoing in the Rampart CRASH unit.15 Perez worked with
investigators over the next year, divulging over 4,000 pages of
interrogation transcripts.16 Perez’s testimony revealed police corruption on

7. BOARD OF INQUIRY INTO THE RAMPART AREA CORRUPTION INCIDENT, LOS ANGELES
POLICE DEPARTMENT PUBLIC REPORT 2 (2000).
8. Id. at 1.
9. Id. at 55.
10. Id.
11. Id. at 69 (“Separate roll calls from the patrol division, a unique patch, jackets, an emphasis on
narcotics enforcement, and an outward appearance of elitism were common CRASH traits that
Rampart shared with other CRASH . . . units.”); see also id. at 77–78 (discussing the lack of oversight
of the Rampart CRASH unit due to the physical separation of the unit from the rest of the Rampart
division and other affirmative acts taken by the unit to isolate itself).
12. Id. at 1.
13. Peter J. Boyer, Bad Cops, THE NEW YORKER, May 21, 2001, at 67.
14. Id. at 69.
15. CRASH is the acronym given to L.A.P.D.’s specialized anti-gang units. Terry McDermott,
Rafael Perez: The Road to Rampart, L.A. TIMES, Dec. 31, 2000, available at http://articles.latimes
.com/2000/dec/31/news/mn-6831.
16. RAMPART RECONSIDERED: THE SEARCH FOR REAL REFORM SEVEN YEARS LATER 53 (2006).

1138

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

an unimagined scale, implicating police officers in wrongful killings,
indiscriminate beatings and violence, theft, and drug dealing. Perez’s
testimony also implicated dozens of police officers in systematic acts of
dishonest law enforcement, exposing hundreds of instances in which
evidence or contraband was planted on suspects, false statements were
coerced or fabricated, and police officers offered perjured testimony in
court. Perez’s confessions prompted the L.A.P.D. to re-name its
investigative task force the “Rampart Task Force.”17 The Task Force was
charged with corroborating Perez’s allegations of corruption within
Rampart CRASH.18 What followed was, in the words of one independent
commission, one of the “worst police scandals in American history.”19
Ultimately, the District Attorney was able to corroborate hundreds of
Perez’s allegations and the L.A.P.D. entered into a consent decree with the
U.S. Department of Justice, submitting to federal oversight of
departmental operations.20 As a result of the scandal, more than three
hundred prisoners filed writs of habeas corpus seeking to overturn
allegedly tainted convictions, and approximately 156 felony convictions
were dismissed or overturned as a result of “Rampart related” writs,21 110
of which were either initiated or unopposed by the District Attorney.22
The extent of wrongdoing by the L.A.P.D., however, remains a
mystery to this day largely due to the overall ineffectiveness of the
L.A.P.D.’s internal investigation of the police force.23 Although Officer
Perez claimed that “ninety percent of the officers that work CRASH, and
not just Rampart CRASH, falsify a lot of information” and “put cases on
people,”24 no investigation or follow-up was ever undertaken to explore, or

17. Id.
18. Id.
19. REPORT OF THE RAMPART INDEPENDENT REVIEW PANEL 1 (2000), available at http://www.
ci.la.ca.us/oig/rirprpt.pdf.
20. RAMPART RECONSIDERED: THE SEARCH FOR REAL REFORM SEVEN YEARS LATER, supra
note 16, at 6–7, 37–38.
21. RAMPART RECONSIDERED: THE SEARCH FOR REAL REFORM SEVEN YEARS LATER App. C.
Fourteen Rampart related writs remained pending at the time these statistics were compiled. Id.
22. Id. To the best of my knowledge, all of the writs initiated or unopposed by the D.A.’s office
were granted. Courts were far more skeptical in writ cases initiated by defendants if the D.A. opposed
the writ. Although relief was granted in approximately forty such cases, the large majority of contested
writ applications were denied.
23. RAMPART RECONSIDERED: THE SEARCH FOR REAL REFORM SEVEN YEARS LATER, supra
note 16, at 47 (concluding that basic facts regarding Rampart corruption scandal remain unknown
seven years afterwards, including “[t]he extent of Rampart CRASH-like misconduct in the CRASH
units of other divisions, other specialized units and LAPD policing generally”); see also id. at 54 (“The
[L.A.P.D.] appeared to lack a clear and well-defined investigative approach and strategy and did not
establish a plan for interagency coordination.”).
24. Id. at 53.

2013]

POLICE MISCONDUCT

1139

even clarify, those allegations.25 In speaking with an investigative panel,
some officers, who spoke anonymously out of fear of retribution,
expressed concerns that the department did not genuinely seek to uncover
the extent of the corruption.26 In fact, the L.A.P.D. failed to produce a
promised “after-action” report which, according to the department, was
going to include “the exact nature and disposition of each allegation.”27
Consequently, whatever the department may have uncovered about
widespread corruption throughout the force remains outside the public
domain.28
In the spring of 2003, while the Rampart story was winding down,
news of another major police scandal broke, this time not out of a major
metropolitan police force but instead in the tiny west Texas town of Tulia,
located in Swisher County. The Tulia operation began as a roundup of
suspected drug dealers in the summer of 1999, but transformed into what
some described as a wholesale assault on the black residents of Tulia. The
operation was spearheaded by a freelance agent named Tom Coleman. 29
Working undercover, Coleman claimed to have bought powder cocaine
from more than 20% of the adult black residents of Tulia.30 In all, nearly
fifty persons were convicted of selling drugs to Coleman, in most cases
based solely on Coleman’s uncorroborated testimony.31
The first several Tulia defendants fought the drug charges at trial and
were convicted and sentenced to draconian prison terms.32 After seeing the

25. Id. at 48, 54.
26. Id. at 65 n.95.
27. Id. at 62 (internal quotations omitted).
28. Id. at 63.
29. New York Times reporter Bob Herbert had covered the story even before Coleman’s
credibility was shattered and the cases fell apart. See Bob Herbert, Tulia’s Shattered Lives, N.Y. TIMES
(Aug. 5, 2002), http://www.nytimes.com/2002/08/05/opinion/tulia-s-shattered-lives.html?scp=7&sq=
tulia&st=cse.
30. NATE BLAKESLEE, TULIA: RACE, COCAINE, AND CORRUPTION IN A SMALL TEXAS TOWN 5
(2005).
31. Id.
32. Id. at 6. Joe Moore was charged with making two drug deliveries: a single gram of crack
cocaine and an eight ball of powder cocaine. With two prior drug felonies on his record, Moore faced a
potential sentence, if convicted, of twenty-five to ninety-nine years in jail. Id. at 44. At trial, Coleman
testified that he purchased the drugs at Moore’s house. Moore acknowledged that Coleman and
another individual had come by his house, but contended that he chased Coleman away and that
Coleman was lying about purchasing the drugs. Id. Apart from a small bag of cocaine that Coleman
claimed to be the fruits of the sale, no evidence corroborated either of the alleged transactions. Id. In a
contest of credibility pitting law-enforcement officer against convicted drug dealer, the jury banked on
the officer. Moore was convicted and sentenced to ninety years. Id. at 59.
The next two defendants, Chris Jackson and Jason Williams, were convicted based on virtually
identical evidence. Id. at 82–83. Jackson received a twenty year sentence for allegedly selling an eight
ball of cocaine. Id. at 83. Williams was convicted of multiple transactions and, based on the additional

1140

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

writing on the wall, however, most of the remaining defendants agreed to
plead guilty. In all, forty-seven persons were charged and thirty-five were
convicted. Of the twelve who were not convicted, several were placed on
deferred adjudication.33
As these cases were tried, however, it became increasingly evident that
Coleman’s testimony was not credible. Defense attorneys discovered that
Coleman had been arrested on theft charges in a neighboring county and
lied about it on his employment application to the task force.34 They also
learned that Coleman had a history of employment problems, mental
health problems, and significant unpaid debts.35 Worse still, it became
increasingly evident that Coleman’s bosses in Tulia, as well as the
prosecutor in the Tulia cases, knew of Coleman’s problems and lied about
them under oath in the course of the Tulia trials.36
After the Texas Court of Criminal Appeals remanded four of the Tulia
convictions for evidentiary hearings on claims that the prosecutor had
failed to turn over material exculpatory evidence as required by Brady v.
Maryland,37 hearings were conducted before a different trial judge. In the
course of the hearings, it became clear that Coleman had perjured himself
on numerous occasions during the Tulia trials, and that other law
enforcement officials may have done so as well.38 Ultimately, the state
agreed to a global settlement with defense attorneys in which it stipulated
that Coleman was not a credible witness, vacated every conviction
obtained as a result of the sting operation without seeking new trials
against any of the defendants, and provided $250,000 to be divided among
the defendants.39 In exchange, the defendants agreed not to sue the
county.40 The state judge who presided at the hearing found “that Mr.

allegation that the sales occurred in a drug-free zone, received a forty-five year sentence,
notwithstanding that he was only nineteen years old and had no prior convictions. Id.
Cash Love, one of the few white suspects arrested in the sting operation, was convicted of making
eight sales, some in a drug-free zone, and was sentenced to 361 years. Id. at 92. Another defendant,
Donnie Smith, was convicted of the first of several drug charges and sentenced to two years. Id. at
136. He pleaded guilty to the remaining charges and received a twelve and a half year sentence. Id. at
136–37. After Smith, two more defendants gambled on jury trials and lost. Freddie Brookins, Jr. was
sentenced to twenty years. Id. at 157. Kareem Abdul Jabbar White received a sixty year sentence. Id.
at 177.
33. Id. at 409.
34. Id. at 103–04.
35. Id. at 302–05.
36. Id. at 305–07, 388–89.
37. 373 U.S. 83 (1963). See BLAKESLEE, supra note 30, at 317.
38. BLAKESLEE, supra note 30, at 388–89.
39. Id. at 384, 386.
40. Id. at 385.

2013]

POLICE MISCONDUCT

1141

Coleman had engaged in ‘blatant perjury’ and was ‘the most devious,
nonresponsive law enforcement witness this court has witnessed in
twenty-five years on the bench in Texas.’”41 Coleman was eventually tried
and convicted of one count of perjury and sentenced to ten years
probation.42
Although the settlement was contingent on approval by the Court of
Criminal Appeals, when that approval was not immediately forthcoming,
the Texas legislature passed a bill “specifically authorizing” the judge “to
grant bond to the defendants.”43 Texas Governor Rick Perry then asked the
Texas Board of Pardons and Paroles to review the cases. Pardons were
granted to all thirty-five Tulia defendants convicted as a result of the sting
operation.44 Two more individuals later were exonerated by courts.45
Rampart and Tulia together account for nearly two hundred cases of
wrongful conviction and represent two large sets of exonerations
stemming from police corruption scandals. But these are not the only
major scandals that have recently beset law enforcement organizations in
the United States, or even in Texas. In Hearne, Texas, numerous cases in
2001 were dismissed following revelations that a drug task force was
systematically targeting black residents in an effort, allegedly, to drive
them from the community.46 As in Tulia, the evidence against the
defendants in these cases consisted solely of the uncorroborated assertions
of a single, unreliable, police informant. Although most cases were
dismissed prior to conviction, some innocent defendants pleaded guilty
before the police wrongdoing was exposed.47 A year later, in the so-called
“Dallas Sheetrock scandal,” at least thirty-nine criminal cases were
dropped or dismissed after it was discovered that white powder allegedly
recovered from criminal suspects and identified through field-tests as
cocaine was actually ground up sheetrock packaged to look like cocaine.

41. Steve Barnes, Rogue Narcotics Agent in Texas Is Found Guilty of Perjury, N.Y. TIMES (Jan.
15, 2005), http://www.nytimes.com/2005/01/15/national/15tulia.html (last visited, Sept. 30, 2011).
42. BLAKESLEE, supra note 30, at 408.
43. Id. at 389.
44. A lawsuit brought against Coleman and the “26-county Panhandle Regional Narcotics
Trafficking Task Force” by forty-five individuals caught up in the sting operation was settled in 2004
after the defendants in the lawsuit agreed to payment of six million dollars. See Barnes, supra note 41.
45. Janelle Stecklein, Judge reverses probation decision of Tulia brothers, AMARILLO GLOBENEWS (Feb. 5, 2011, 1:12 AM), http://amarillo.com/news/local-news/2011-02-05/decision-reversedtulians.
46. A.P., Texas: Suit Filed Over Drug Arrests, 11/2/02 N.Y. TIMES (Abstracts), 2002 WLNR
10823176.
47. See First Am. Compl. For Inj. Relief And Damages And Jury Trial at 5, Kelly v. Paschall,
No. 02-A-02-CA-702 JN (W.D. Tex. Apr. 2, 2003).

1142

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

All of the victims in the scandal were blue-collar Mexican immigrants
who spoke little or no English.48
Another Tulia-like scandal erupted more recently in St. Charles Parish,
Louisiana, where seventy narcotics cases made by a single undercover
officer were dismissed following revelations that the undercover officer
had lied under oath in a criminal investigation. Before the scandal broke,
at least twenty persons in cases made by the undercover officer had
already pled guilty.49 An even larger Rampart-style corruption case has
unfolded in Camden, New Jersey.50 Other incidents have also grabbed
recent headlines.51 Gross and Shaffer have identified twelve separate
incidents involving group exonerations based on police misconduct
involving exonerations of at least 1,100 people.52
48. See Paul Duggan, ‘Sheetrock Scandal’ Hits Dallas Police: Cases Dropped, Officers Probed
After Cocaine ‘Evidence’ Turns Out to Be Fake, WASH. POST (Jan. 18, 2002), http://www.washington
post.com/ac2/wp-dyn/A229-2002Jan18?language=printer.
49. See Cop’s Alleged Past Ruins 70 St. Charles Parish Drug Cases, FOX EIGHT LIVE (Sept. 20,
2010, 2:40 PM), http://www.projectnola.com/the-news/news/42-fox-8/106713; Lori Lyons, St. Charles
drug arrests in jeopardy after undercover cop discredited, authorities say, THE TIMES-PICAYUNE (Sept.
20, 2010, 8:15 PM), http://www.nola.com/crime/index.ssf/2010/09/st_charles_drug_arrests_in_jeo.
html. To the best of my knowledge, however, and unlike Tulia, none of those guilty pleas have been
overturned.
50. The full scope of the Camden scandal is still not clear. More than two hundred people have
had convictions vacated or charges dismissed as a result of confessed misconduct implicating at least
five officers in wrongdoing. The misconduct includes planting evidence on innocent persons and
providing false testimony to convict them of crimes they did not commit. See Barbara Boyer, Two
Former Camden Officers Face More Federal Charges, THE INQUIRER (Sept. 10, 2011), http://articles.
philly.com/2011-09-09/news/30135226_1_original-indictment-amount-of-illegal-drugs-special-operati
ons-unit.
51. In Tulsa, at least five Tulsa police officers have been charged with perjury and witness
tampering. One defendant faced fifty-eight counts of wrongdoing. At least eleven people were released
from prison as a result, with more cases under review. See Emory Bryan, Five Tulsa Police Officers
Indicted in Corruption Probe, THE NEWS ON 6 (July 20, 2010, 9:17 PM), http://www.newson6.com/
story/12840428/five-tulsa-police-officers-indicted-in-corruption-probe?redirected=true. In Denver, one
out of every seventeen police officers has been subject to administrative discipline for “‘departing
from the truth’” or similar conduct in matters related to their official duties. That figure counts only
those who have been formally sanctioned. It excludes those who are currently under investigation for
similar violations, those who were investigated but insufficient proof of wrongdoing was presented to
sustain a charge, and those whose misdeeds have not yet been detected. Christopher N. Osher, Denver
cops’ credibility problems not always clear to defenders, juries, DENVER POST (July 10, 2011),
http://www.denverpost.com/news/ci_18448755 (reporting that eighty-one officers still on the force out
of 1,434 are on a list citing violators “in at least one of the following categories: departing from the
truth, violating the law, making false reports, making misleading or inaccurate statements, committing
a deceptive act, engaging in conduct prohibited by law, engaging in aggravated conduct prohibited by
law, conspiring to commit conduct prohibited by law, soliciting or accepting a bribe, removing reports
or records, destroying reports or records or altering information on official documents”). This list only
includes the names of officers against whom violations have been formally substantiated. It does not
include officers who are under investigation, or who were investigated but not cited. Id.
52. Gross & Shaffer, supra note 6, at 84.

2013]

POLICE MISCONDUCT

1143

In short, Rampart and Tulia produced numerous exonerations and
received a significant amount of national attention, but they are not
unique. Revelations of large-scale police misconduct both preceded and
post-dated them, suggesting that police misconduct leading to the
wrongful conviction of innocent persons is a disturbingly common feature
of the criminal justice system.
III. WHAT WE KNOW ABOUT EXONEREES GENERALLY
When it comes to wrongful convictions, very little hard data exists.53
After all, it is extraordinarily difficult to systematically identify erroneous
convictions of innocent persons. Because the criminal justice process itself
is assumed to provide the definitive test of guilt or innocence, there are
few external Archimedean points from which its results may be tested.
Our knowledge about innocent persons who are wrongfully convicted,
therefore, is derived primarily from exonerations—that is, cases in which
some government official, acting in an official capacity, has made a formal
finding or declaration that a defendant is “not guilty of a crime for which
he or she had previously been convicted.”54
Wrongful convictions have been the subject of academic inquiry since
Edwin Borchard published his pathbreaking studies on the matter in the
early part of the twentieth century.55 Other studies, including an influential
article by Hugo Bedau and Michael Radelet, followed.56 Until quite
recently, however, the leading study of criminal exonerations has been
Gross, Jacoby, Matheson, Montgomery, and Patil’s analysis of
exonerations occurring between 1989 and 2003 (“Gross Study”).57 That
study has now been updated and greatly expanded in an examination of
exonerations through 2012.58 Brandon Garrett has also made major
contributions to the bank of knowledge of the exonerated through a series
of articles, and a book (collectively, the “Garrett Study”) on DNA

53. See Robert Carl Schehr, The Criminal Cases Review Commission As A State Strategic
Selection Mechanism, 42 AM. CRIM. L. REV. 1289, 1290–91 (2005) (describing the number of
wrongfully convicted persons as a “dark number, a statistical unknown”).
54. Samuel R. Gross et al., Exonerations in the United States 1989 Through 2003, 95 J. CRIM. L.
& CRIMINOLOGY 523, 524 (2005).
55. See Edwin M. Borchard, European Systems of State Indemnity for Errors of Criminal Justice,
3 J. AM. INST. CRIM. L. & CRIMINOLOGY 684 (1913); EDWIN BORCHARD, CONVICTING THE
INNOCENT: ERRORS OF CRIMINAL JUSTICE (Yale Univ. Press 1932).
56. See Hugo Adam Bedau & Michael L. Radelet, Miscarriages of Justice in Potentially Capital
Cases, 40 STAN. L. REV. 21 (1987).
57. Gross et al., supra note 54.
58. See Gross & Shaffer, supra note 6.

1144

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

exoneration cases.59 As the authors of both the Gross and Garrett Studies
concede, the fact that an individual has been exonerated does not
conclusively prove that the individual is actually innocent, although often
the nature and circumstances of the evidence will leave little doubt.
Nonetheless, formal exoneration is the best that our legal system is usually
capable of doing, and thus provides the best indicator we have of instances
in which an actually innocent person has been wrongfully convicted.
These studies have identified some commonalities in cases resulting in
exoneration. First, most exonerees were convicted of very serious crimes,
typically resulting in sentences of death or long terms of imprisonment.60
Second, the vast majority of exonerees contested their guilt at trial. Only a
tiny handful of exonerees, about 6%, pled guilty.61 This fact is particularly
striking because the vast majority of criminal convictions, upwards of
90%, are obtained through guilty pleas. Third, although many types of
procedural and evidentiary errors have been identified in cases of wrongful
conviction, earlier studies consistently pointed to eyewitness
misidentification as the leading cause of wrongful convictions, followed
closely by faulty forensic evidence.62 Fourth, the studies suggest that
persons of color are at far greater risk of false conviction than whites.63
The authors of these studies are quick to deny that the data is in any way
representative of the wrongfully convicted more generally. As Gross and
Shaffer observe in a more recent study, “[e]xonerations are unlikely,
uncommon and unrepresentative of the mass of invisible false
59. Garrett’s main findings regarding the profile of innocent persons who were wrongfully
convicted can be found in Brandon L. Garrett, Judging Innocence, 108 COLUM. L. REV. 55 (2008).
Those findings were updated through 2010 in his book. See BRANDON L. GARRETT, CONVICTING THE
INNOCENT (Harvard Univ. Press 2011).
60. Gross et al., supra note 54, at 535 (“With a handful of exceptions, everyone on our list of
exonerees was sentenced to death or to a long term of imprisonment. Ninety-three percent were
sentenced to ten years in prison or more; 77% were sentenced to at least twenty-five years. . . .”)
(footnote omitted).
61. Id. at 536 (finding that 5.8% of exonerees (20/340) pleaded guilty); GARRETT, supra note 59,
at 150 (reporting that 6%, or sixteen of 250 DNA exonerees, pleaded guilty).
62. Gross, supra note 2, at 542; GARRETT, supra note 59, at 48, 89 (reporting that eyewitness
misidentifications factored in 76% of DNA exoneration cases, and faulty forensic testimony or
evidence in 74%). Gross’s more recent study, which includes data from a broader source of
exonerations, including the mass exonerations, finds that perjury or false accusation is the leading
contributing factor to wrongful convictions overall, but that the prevalence of various contributing
factors turns heavily on crime of conviction. See Gross & Shaffer, supra note 6, at 40.
63. Of the first 250 persons exonerated by DNA evidence, 62% were black, 30% were white, and
8% were Hispanic. Asians constituted less than 1% of the total. See GARRETT, supra note 59, at 5.
Moreover, as Brandon Garrett has observed, although minorities are overrepresented in the prison
population, their numbers among exonerees, or at least DNA exonerees, are even greater. Garrett,
supra note 59, at 66.

2013]

POLICE MISCONDUCT

1145

convictions.”64 Nonetheless, because their data is the best and most
reliable that we have concerning wrongful convictions, there is an
inevitable tendency to generalize or draw inferences about the
characteristics, frequency, and causes of wrongful convictions from the
data.65
As this Article will show, however, the mechanisms that produce
exonerations in the police scandal cases differ substantially from those in
other sorts of cases where exonerations have been most common. In trying
to understand the role of police misconduct in causing wrongful
convictions, then, it is imperative to deploy a narrow lens.
It is impossible to know how frequently police misconduct of the type
uncovered in Rampart and Tulia occurs, or how many wrongful
convictions result from such misconduct. What happened in Rampart and
Tulia appears to have involved widespread misconduct by police officers
and prosecutors. In these cases, investigators discovered a culture of
corruption that fostered official misconduct. Even if instances in which
entire police departments, or at least entire units within a department,
succumb to such cultural corruption are rare, the type of misconduct that
led to the wrongful conviction of defendants in Rampart and Tulia could
just as easily be perpetrated by smaller groups of corrupt officers, or even
by officers acting on their own. Wrongful convictions resulting from
occasional police misconduct involving only a single officer, or a
relatively small group of corrupt police officers, scattered throughout the
nation’s police departments, would be almost impossible to detect. And
yet, the aggregate effect of such misconduct could easily generate a very
large number of wrongful convictions. It is also possible that such cases
may truly be rare. We simply do not have any way to know.
In either case, the lack of attention paid to date to the mass exoneration
cases has tended to reinforce some misconceptions about the causes and
characteristics of the convictions of innocent persons. The vast majority of
exonerations studied to date arose from murder and rape cases, in which
defendants received typically severe sentences—often long prison terms or
death sentences. The vast majority of these exonerations—some 94%—
involved defendants who contested their guilt at trial and who were, as a
result, able to pursue the full panoply of post-conviction remedies

64. Gross & Shaffer, supra note 6, at 9.
65. For example, based on this data some have concluded that the rate of false convictions is
much lower among guilty pleas than trials. See, e.g., Larry Laudan, Is It Finally Time to Put ‘Proof
Beyond a Reasonable Doubt’ Out to Pasture? 15 (U. of Tex. Public Law & Legal Theory Working
Paper Series, Paper No. 194, 2011), available at http://ssrn.com/abstract=1815321.

1146

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

available to them. These defendants make up an unrepresentative pool in a
criminal justice system characterized by overwhelming rates of guilty
pleas.66 As a result, while the exoneration data currently available tells us
something about how innocent people can be convicted in rape and murder
cases at trial, it tells us very little about how innocent people might agree
to plead guilty in cases involving the more mundane criminal offenses that
make up the bulk of criminal courts’ daily workloads. Little research to
date illuminates this important corner of the wrongful convictions
problem.67
The present study seeks to partially rectify that problem in two ways.
First, the study examines, to the extent available data permits, the causes
and characteristics of the wrongful convictions identified in the mass
exonerations. Second, the study contrasts that data with the data gathered
in earlier exoneration studies to challenge some common assumptions
about wrongful convictions more generally. By examining data from
exonerations which arose in settings very different from the typical DNAbased exonerations, my goal is to provide more nuance to our
understanding of wrongful convictions, to debunk some suggested
inferences from earlier data sets, and to identify new avenues for
investigation and reform.
IV. ACTUALLY INNOCENT RAMPART AND TULIA EXONEREES
A. The Data
The data used in this study comes from two well-publicized incidents
of systematic police misconduct, the Rampart and Tulia scandals. Both
incidents involved dozens of criminal defendants. More than 150 persons
were exonerated as a result of Rampart, and thirty-seven as a result of
Tulia.68 Unlike some other recent police misconduct scandals,
66. See, e.g., United States Sentencing Commission’s 2011 Annual Report, Chapter Five:
Research 33 (2011), available at http://www.ussc.gov/Data_and_Statistics/Annual_Reports_and_
Sourcebooks/2011/2011_Annual_Report_Chap5.pdf (reporting that 96.9% of federal felony
convictions were the result of guilty pleas).
67. See Jon B. Gould & Richard A. Leo, One Hundred Years Later: Wrongful Convictions After
a Century of Research, 100 J. CRIM. L. & CRIMINOLOGY 825, 834–35 (2010) (observing that “research
offers few glimpses into errors” affecting “‘lesser’” felonies, and certainly misdemeanors”).
68. Thirty-five persons were pardoned by Texas Governor Rick Perry. Adam Liptak, Texas
Governor Pardons 35 Arrested in Tainted Sting, N.Y. TIMES (Aug. 23, 2003), http://www.ny
times.com/2003/08/23/us/texas-governor-pardons-35-arrested-in-tainted-sting.html. Two others were
exonerated by courts on petitions for habeas relief. See Ex parte Barrow, 2010 WL 2618851 (Tex.
Crim. App., June 30, 2010); Janelle Stecklein, Judge reverses probation decision of Tulia brothers,

2013]

POLICE MISCONDUCT

1147

investigations of alleged police misconduct did not reveal wrongdoing
until tens of defendants in Tulia, and hundreds in Rampart, had seen their
cases through to conviction. Accordingly, the Rampart and Tulia cases
provide an extensive set of data regarding not only how police misconduct
can lead to wrongful arrest or wrongful charges, but to wrongful
convictions.
The vast bulk of the data on the Rampart scandal relied on in this study
was extracted from files obtained from the Los Angeles District
Attorney’s Office. These files contained office memoranda tracking
developments in the Rampart investigation and most helpfully, writs filed
by the District Attorney and defense counsel seeking relief for wrongfully
convicted defendants. This data was supplemented with other information
gleaned from official reports, newspaper articles, and other articles on the
scandal appearing in the popular press and in academic commentary. With
respect to Tulia, I relied extensively on the facts and case descriptions
compiled by Nate Blakeslee in his thorough and engaging account of the
Tulia scandal.69 I have cross-checked Blakeslee’s data, to the extent
possible, with other published reports about Tulia, and with data made
available to me by attorneys involved in the Tulia cases.
Of the two, the Rampart material provides the greatest insight into how
police misconduct “on the ground” can trigger a disastrous chain of events
for innocent persons directly resulting in criminal convictions. Because the
writs filed on behalf of wrongly convicted Rampart defendants often
included narrative accounts of the circumstances of arrest, the Rampart
cases provide an illuminating glossary of the many ways that police
misconduct can lead to wrongful convictions. Study of these cases in the
aggregate provides a fairly detailed empirical picture of wrongful
convictions resulting from dishonest policing. The data pertaining to the
Tulia cases shows less variation in the factual circumstances surrounding
the charges, primarily because of the relatively uniform way in which the
Tulia convictions were generated: each Tulia defendant was convicted
based almost exclusively on the uncorroborated testimony of a single
corrupt undercover agent. However, the Tulia data permits useful
observations about the adjudicative procedures in such cases, and deepens
the data pool in this regard.

AMARILLO GLOBE-NEWS (Feb. 5, 2011), http://amarillo.com/news/local-news/2011-02-05/decisionreversed-tulians.
69. NATE BLAKESLEE, TULIA: RACE, COCAINE, AND CORRUPTION IN A SMALL TEXAS TOWN
409–17 (2005).

1148

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

The data breaks down as follows. Although more than 150
exonerations resulted from the Rampart scandal, the District Attorney’s
files contained case-specific data for only ninety-seven of those cases, and
detailed case data as to eighty-seven of these. None of the case
information was complete in any sense of the word.70 As a result, there is
more information about some cases than others, depending on the extent
of factual detail provided in the affidavits and writs for habeas corpus.
Not all of the individuals whose convictions were reversed or vacated,
however, were “actually innocent” of the crimes for which they were
convicted. Many defendants obtaining relief in Rampart did so because of
procedural misconduct on the part of the police, not because the police
were without evidence of wrongdoing. For example, many Rampart
defendants were exonerated when it became clear that the police officers
who had arrested them lied about the circumstances leading to the
discovery of contraband. Where evidence of this sort of police misconduct
surfaced convictions were rightly reversed, but there is no reason to
believe that these defendants were not in fact engaged in criminal conduct.
Prior exoneration studies have focused on cases involving what often is
referred to as “actual” or “factual” innocence.71 Actual innocence cases are
those in which either the wrong person was convicted of a crime
committed by another, or a person was convicted of a crime that did not
actually occur.72 In the first Gross study, all 340 exonerees had been
absolved through “an official act declaring a defendant not guilty of a
crime for which he or she had previously been convicted”73 premised on
“strong evidence of factual innocence” and no “unexplained physical
evidence of the defendant’s guilt.”74 The first Gross Study excluded from
its purview exonerations in cases where the evidence indicated that the
exonerees had been “involved in the crimes for which they were
convicted.”75 The Garrett Study similarly focused only on those who “did
70. The D.A.’s office has maintained files from the Rampart cases, but the information was not
well-organized and consisted mostly of the legal pleadings prepared in habeas corpus proceedings. We
did not have access to the investigative files in individual cases. Useful information that might further
clarify the circumstances in these cases no doubt could be found in such files, if they exist. In any
event, we did not have access to them.
71. See, e.g., D. Michael Risinger, Innocents Convicted: An Empirically Justified Factual
Wrongful Conviction Rate, 97 J. CRIM. L. & CRIMINOLOGY 761 (2007).
72. Bousley v. United States, 523 U.S. 614, 615 (1998) (“Actual innocence means factual
innocence, not mere legal insufficiency.”).
73. Gross, supra note 2, at 524.
74. Id. at 524 n.4.
75. Id. at 527.

2013]

POLICE MISCONDUCT

1149

not commit the charged crime,”76 adopting the criteria used by the
Innocence Project to identify its set of innocent exonerees.77
Accordingly, I subdivided the Rampart cases into three categories: the
“actually innocent” group, the “maybe innocent” group, and the “not
innocent” group. In making these divisions, I followed Gross and Garrett
in defining an actually innocent exoneree as one who was not involved in
the commission of the supposed crime, did not commit the charged crime,
or who was convicted of a crime that never occurred.
In assigning the Rampart exonerees to these various groups, I relied
heavily on the factual findings presented to reviewing courts by the
District Attorney’s Office in petitions for writs of habeas corpus filed by
the state, conceding the wrongful conviction and seeking the release of the
defendant. I also relied on the factual statements included in the
declarations and affidavits filed by investigating agents accompanying the
D.A.’s filings. Where habeas petitions were initiated by defendants, I
relied on factual allegations made by the petitioners only where those
allegations were conceded in the state’s response. I also only assumed the
truth of factual allegations made by defendants and their witnesses if the
D.A.’s Office affirmatively stated in its filings that prosecutors or
investigators had discovered evidence corroborating those accounts.
Based on my review of the files, misconduct unrelated to the factual
guilt or innocence of the defendant seems to have been the primary basis
for exonerations in forty-nine cases.78 In those cases, defendants were
exonerated because police officers lied about probable cause, about where
a search took place, or about whether the suspect consented to a search. In
thirty-eight cases, however, the police misconduct plainly did implicate
the guilt/innocence determination. In these cases, police planted drugs or
guns on suspects, lied about observing defendants committing crimes, or
coerced confessions from innocent individuals. Where convictions were
reversed based on reliable evidence of such misconduct, they constitute
“exonerations” in the fullest sense of the term and are consistent with the

76. Brandon L. Garrett, Claiming Innocence, 92 MINN. L. REV. 1629, 1645 (2008).
77. GARRETT, supra note 59, at 285–86 (explaining that the list of 250 DNA exonerations
matches that maintained by the Innocence Project, which conservatively defines DNA exonerations
by, for example, omitting “cases in which there has been no exoneration despite DNA evidence of
innocence,” and only includes cases in which there is no doubt that the “convicts are actually
innocent”).
78. In twenty-seven cases, defendants were exonerated for procedural misconduct by the police,
but there remained evidence of criminal culpability on the part of the defendants. Exonerees in these
cases were coded as “not innocent.” In another twenty-two cases, the guilt or innocence of the
exonerees was not clear from the record. These cases were coded as “maybe innocent.” For a further
discussion, see infra Part VI.

1150

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

criteria used by others, including Gross and Garrett, who have studied the
wrongful convictions of innocent persons. This study thus focuses
primarily on those thirty-eight cases.
Tulia supplies another thirty-seven cases. Thirty-five Tulia
exonerations resulted from pardons recommended by Texas Governor
Rick Perry. Two defendants were granted writs of habeas corpus vacating
their convictions. Several other defendants were also wrongly charged or
convicted in the scandal but were never pardoned. One such defendant was
a minor who was sent to boot camp and who had already completed his
sentence when Governor Perry recommended the Tulia pardons. Other
defendants negotiated deals for deferred adjudication. Because convictions
were never formally entered in those cases, pardons were not deemed
necessary. Although these defendants too were wrongfully convicted, they
were not included in the study because no official exonerations were ever
granted in their cases.79
There are those who continue to assert that at least some of the Tulia
defendants were, in fact, guilty. Indeed, about a half-dozen of the Tulia
defendants admitted that they helped undercover agent Coleman purchase
crack cocaine.80 None of the defendants, however, ever admitted
involvement in the sale of powder cocaine, and it was the powder cocaine
charges that provided the basis for the most serious sentences imposed on
the Tulia defendants. Differentiating among Tulia defendants is made
more difficult because the Tulia defendants were pardoned en masse,
based on the fact that the cases were uniformly predicated on the word of a
proven liar, and thus no formal individual findings of innocence were ever
made. However, what evidence we do have points strongly toward
innocence of virtually all of the Tulia defendants. First, in a sting resulting
in the arrest of forty-six individuals, where most arrests occurred in the
early morning hours, by surprise, at the suspects’ homes, not a single
suspect was caught in possession of cocaine or crack.81 Second, none of
the alleged drug transactions were recorded on audio or video tape.
Indeed, there was virtually no corroborating evidence presented to
implicate any of the defendants in the charged crimes. Third, the charges
were relatively implausible by their nature. The defendants were drawn

79. At least one person wrongfully charged in the Tulia drug sting, Etta Kelly, did not receive a
pardon because she pleaded guilty in exchange for deferred adjudication and thus a conviction was
never actually entered in her case. See John Reynolds, Pardons Urged in Drug Cases, LUBBOCK
AVALANCHE-J. (July 31, 2003), http://lubbockonline.com/stories/073103/reg_073103064.shtml.
80. BLAKESLEE, supra note 30, at 296.
81. See Herbert, supra note 29.

2013]

POLICE MISCONDUCT

1151

from Tulia’s poorest classes where the drug of choice was crack cocaine.
Yet all of the alleged transactions involved small amounts of powder
cocaine. These facts strongly undermine the credibility of the charges and
support the theory that Tom Coleman, the investigating undercover
officer, may have been scamming the Drug Task Force for money by
claiming to have engaged in fake sales and then logging into evidence
powder cocaine that he himself severely diluted.82 In short, while it is
possible that one or two of the Tulia defendants were in fact guilty,
substantial evidence demonstrates that the vast majority of the Tulia
defendants were innocent of any criminal wrongdoing, and the
exonerations granted them by the Texas governor based on this evidence is
sufficient to bring all of the Tulia defendants within the category of the
“actually innocent.”
B. Rampart and Tulia Exoneree Demographics
The vast majority of those wrongfully convicted in Rampart and Tulia
were persons of color. Although the data available for this study did not
specify race or ethnicity of the Rampart defendants, an informal review of
the surnames of the defendants strongly suggests that most, if not all, of
the Rampart exonerees were of Hispanic origin. That conclusion fits with
the population of the Rampart area, which is heavily Hispanic,83 and the
demographics of the Rampart area street gangs that the CRASH unit at the
center of the Rampart scandal policed.
More precise data exists with respect to the Tulia defendants, the
overwhelming majority of whom were persons of color. Of the thirty-five
Tulia defendants who received pardons, thirty-one were black, two were
Hispanic, and two were white.
The average age of the Tulia exonerees was 29.8 years. There was
insufficient data to determine the average age of the Rampart exonerees.84
With respect to gender, as is true in criminal law generally, the great

82. Coleman was subsequently convicted of perjury for false testimony given during hearings of
Tulia defendants. See Steve Barnes, Rogue Narcotics Agent in Texas Is Found Guilty of Perjury, N.Y.
TIMES (Jan. 15, 2005), http://www.nytimes.com/2005/01/15/national/15tulia.html?ref=tomcoleman.
83. One 2010 study of the area reported that 63% of Rampart’s residents are Latino, 26% are
Asian-American, and 5% are African-American. The dominant ethnic groups in Rampart are
Mexicans, Salvadorans, and Guatemalans. THE URBAN INSTITUTE, EVALUATION OF THE LOS ANGELES
GANG REDUCTION AND YOUTH DEVELOPMENT PROGRAM: RAMPART ZONE PROFILE 2 (2010),
available at http://www.urban.org/uploadedpdf/412274-Rampart.pdf. The two most violent gangs
operating in Rampart are 18th Street and Mara Salvatrucha, or MS-13. Both gangs recruit heavily, if
not exclusively, from Latinos. Id. at 3.
84. See Summary of Data (on file with author).

1152

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

majority of the exonerees from the Rampart and Tulia scandals were male.
Of ninety-seven Rampart exonerees, approximately 90% were male.85
Interestingly, among the Rampart exonerees determined in this study to be
“actually innocent,” all were male.86 Although there is no definitive
answer as to why there were no females among the actually innocent
Rampart exonerees, it is possible to speculate. The strong gender disparity
might reflect an array of factors, including the fact that street gangs are
predominantly, if not exclusively, male organizations; that the CRASH
unit’s specific mission was to target gang activity in Rampart; and that
females tend less often to be involved in the types of activities—streetlevel drug dealing and armed conflict—around which most of the false
allegations arose. It might also be the case that false uncorroborated
allegations of wrongdoing by males are more plausible than similar
allegations against females, and thus dishonest police trying to lie credibly
are more likely to make such allegations against male suspects.
A somewhat larger percentage—19% (7/37)—of exonerated
defendants in Tulia were female. An even larger percentage—24%
(11/47)—of the total Tulia defendants were female.87 The somewhat
smaller percentage of women who received pardons reflects the fact that
more women had their cases dismissed prior to prosecution, negotiated a
deferred prosecution, or otherwise avoided receiving the type of lengthy
prison sentence for which a pardon was needed.
C. Offenses of Conviction
The types of crimes leading to wrongful convictions in the mass
exoneration cases are strikingly different from those leading to
exonerations in other cases. Whereas most known exonerees typically
have been convicted of rape or murder, the vast majority of the exonerees
in the police scandal cases were convicted of relatively low-level drug
crimes. All thirty-seven of the Tulia exonerees were convicted of drug
crimes,88 while nearly half of the actually innocent Rampart exonerees
(18/38) were convicted of drug crimes.89 In addition, an almost equal
number of actually innocent Rampart exonerees (16) were convicted of

85. Id.
86. Id.
87. See BLAKESLEE, supra note 30, at 409–17 (providing list of defendants).
88. Drug crimes here refer to any narcotics offense, including possession, transportion, and
trafficking, of illegal narcotics, usually cocaine, crack cocaine, and heroin.
89. See supra note 84.

2013]

POLICE MISCONDUCT

1153

gun possession offenses.90 Of these, defendants were most frequently
charged as felons in possession of firearms, although California law
provides a variety of unlawful gun possession offenses, making it unlawful
for minors, gang members, probationers, and parolees to possess firearms
as well.91 Actually innocent exonerees include persons convicted of each
of these various offenses. A few were convicted of both drug and gun
offenses. A larger number were initially charged with drug and gun
offenses, but, through plea bargaining, negotiated a conviction on only one
offense. Four of the actually innocent defendants also were convicted of
assaulting police officers. Often, those convictions were enhanced with
false allegations that the assailant used a gun or another deadly weapon to
commit the assault. One Rampart exoneree was convicted of the offense of
“giving false information to a police officer.”92
What these offenses of conviction primarily have in common is that
they are all easily manufactured by the arresting officers. Drugs and guns
are easily planted and, once “found,” constitute completed offenses. To the
extent there were alleged victims in any of these cases, the victims
uniformly were police officers. In none of the cases was there a need to
obtain any corroborating evidence or eyewitness testimony from persons
other than police officers. As a result, it was easy for police to falsely
charge suspects with commission of these crimes, and extremely difficult
for defendants to defend against them.
Most of the actually innocent Rampart exonerees received relatively
light sentences. Several of the exonerees were sentenced only to terms of
probation. Most were sentenced to short prison terms ranging from 6
months to a few years. A few of the actually innocent defendants,
however, received quite severe sentences. The median sentence of the
actually innocent Rampart exonerees was two years. The average sentence
was a little more than three years, and the disparity between the median
and average sentence reflects the small number of severe sentences that
were imposed on a few of the defendants.
The most severe, and in many ways the most egregious, Rampartrelated sentence was imposed on Javier Ovando. In the case that probably
did the most to trigger the Rampart scandal, police officers shot Ovando,
then nineteen years old and a member of the 18th Street gang, four times

90. Id.
91. Id.
92. Pet. for Writ of Habeas Corpus at 2, In re James Edward Thomas, No. JJ04795 (Cal. Super.
Ct. May 22, 2000).

1154

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

in the neck and the chest.93 In official reports of the incident submitted by
police officers Rafael Perez and Nino Durden, the officers claimed that
Ovando had broken into a vacant apartment where Perez and Durden were
conducting surveillance. Ovando allegedly was armed with a
semiautomatic rifle and a “military-style ‘banana clip.’”94 The officers
said, and later testified under oath, that they shot Ovando after he refused
to comply with their order to put the weapon down.95 As a result of the
shooting Ovando was left paralyzed from the waist down.96 Despite the
severe injuries he suffered and the fact that he had no prior felony
convictions, Ovando was charged with two counts of assault with a
firearm on a police officer and one count of exhibiting a firearm in the
presence of a police officer.97 Firearm use enhancements were also
alleged. A jury convicted Ovando essentially as charged, and the court
sentenced him to twenty-three years and four months in state prison.98
The allegations regarding Ovando’s conduct, however, were pure
fiction. Officer Perez subsequently admitted in a deposition that Ovando
was unarmed at the time of the shooting, that the shooting was
unprovoked, and that he and Officer Durden had planted a gun on Ovando
to cover up that fact.99 According to Perez, the gun planted on Ovando had
been obtained during a “gang sweep” a few days prior to the incident,100
and the serial number had been filed off so that the officers could use the
weapon as a “throwaway.”101 Perez further stated that the gun was wiped
clean of prints by the officers before it was placed next to the injured
man.102 Thus, Ovando had the double misfortune of being shot and
paralyzed, and then convicted of a serious crime he did not commit.
Apart from the Ovando case, the longest Rampart sentences were
imposed in drug cases. Russell Newman was sentenced to twelve years,

93. For a news analysis discussing details of the case, see Lou Cannon, One Bad Cop, N.Y.
TIMES, Oct. 1, 2000, at SM32.
94. Pet. for Writ of Habeas Corpus, Ex. A at 2, Decl. of Richard A. Rosenthal, In re Javier
Francisco Ovando, No. BA139642 (Cal. Super. Ct. Sept. 16, 1999).
95. See Cannon, supra note 93.
96. Id.
97. Id.
98. Rosenthal Decl., supra note 94, at 2.
99. Id. at 3.
100. Id., Ex. B at 1–2. The reasons for the shooting remain unclear. Perez testified that Durden
shot Ovando during the course of an argument, and that he shot Ovando “reflexively.” Others have
speculated that after Durden shot Ovando, Perez “tried to finish off Ovando so that there would be no
witness to the crime.” Cannon, supra note 93, at SM37. Ovando himself has little recollection of the
shooting, and “still has no idea why the officers shot him.” Id.
101. Cannon, supra note 93, at *5.
102. Id.

2013]

POLICE MISCONDUCT

1155

Esaw Booker to nine years, and Walter Rivas to seven years, all for
allegedly dealing cocaine.103
In comparison, the Tulia cases resulted in substantially harsher
sentences. All of the Tulia defendants were convicted of selling small
quantities of powder cocaine. Despite the small quantities, many of the
defendants received draconian sentences, often as a result of prior felony
convictions. William Cash and Joe Moore both received prison sentences
in excess of ninety years. Kareem Abdul Jabbar White was sentenced to
sixty years. Jason Jerome Williams and Kizzie White received sentences
of forty-five and twenty-five years, respectively. The average sentence of
the exonerated Tulia defendants was 157.8 months or a little over thirteen
years. The median sentence, however, was thirty-six months, which was
well below the average sentence primarily because of the large number of
defendants who were sentenced to extended terms of probation in lieu of
prison.104 Two defendants, Mandis and Landis Barrow, spent ten years in
jail as a result of a probation revocation before a Texas court granted their
habeas writ and ordered them released.105
D. Causes of Wrongful Conviction
While the leading identified cause of wrongful convictions in past
studies of exonerations is witness misidentification, a very different
dynamic is at work in the police misconduct cases. Police misconduct
generally, and perjury in particular, was the primary cause of wrongful
convictions in every Rampart and Tulia case resulting in exonerations.
Witness misidentifications played virtually no role in any of the cases.
Police misconduct in these scandals took many forms. Police officers
filed false police reports detailing observations of criminal conduct the
defendants never engaged in, or describing circumstances that if true
would have established criminal conduct. In most of the cases, police
either physically planted drugs or weapons on the defendants and then lied
about how they found the contraband, or simply misstated that they had

103. See Resp. to Pet. for Writ of Habeas Corpus at 2, In re Russell Newman, No. BA050279-01
(Cal. Super. Ct. Apr. 27, 2000); Resp. to Federal Pet. for Writ of Habeas Corpus at 2, In re Esaw
Booker, No. BA057055-02 (Cal. Super. Ct. Feb. 15, 2000); Resp. to Pet. for Writ of Habeas Corpus at
2, In re Walter Rivas, No. BA165829 (Cal. Super. Ct. Feb. 15, 2000).
104. For purposes of calculating average and median figures, sentences of probation were treated
as the equivalent of zero jail time. Because a later violation can result in a probationer serving the
entire term in jail, the numbers understate actual punishment.
105. Janelle Stecklein, Judge reverses probation decision of Tulia brothers, AMARILLO GLOBENEWS (Feb. 5, 2011, 1:12 AM), http://amarillo.com/news/local-news/2011-02-05/decision-reversedtulians.

1156

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

found drugs or weapons when they had not. Police officers then testified to
these same false facts at preliminary hearings and at trial in those rare
cases that did not end in guilty pleas. For example, Emmanuel Chavez was
arrested and ultimately convicted of possession of a firearm by a minor.106
In the arrest report, Officer Perez stated that “he and his partners
observed” Emmanuel Chavez pass “a sawed-off shotgun” to another minor
named Sergio Salcido.107 According to later evidence gathered by
investigators, however, police never saw either minor handle a gun.
Instead, Chavez and Salcido were stopped because police knew them to be
members of a “tagging crew.”108 As Perez frisked Salcido, a gun dropped
down Salcido’s pant leg and struck the pavement. The officers then made
up a story that allowed them to charge Chavez as well as Salcido for
possession of the gun.109
Similar police misconduct led to the wrongful conviction of Diego
Barrios. Barrios and several others were socializing in the parking lot of a
Jack-In-the-Box fast food restaurant when a “police car drove up to the
group and shined its high-beam lights on the group.”110 The officers
ordered everyone in the group to kneel down. They then searched and
questioned each person. Four persons, including a juvenile by the name of
Raymond C., were placed into a police car and taken to the police station.
Unknown to the police, Raymond C. had a handgun in his possession at
the time which he deposited, during the ride, behind the back seat of the
squad car.111 Police discovered the gun after searching the car at the station
and demanded to know who had dropped the gun. Raymond C. admitted
the gun was his, but according to Barrios, “the officers said they did not
‘want’ a juvenile,” and instead “‘put the gun’ on Barrios.”112 Barrios pled
guilty to a charge of unlawful gun possession.
On a different occasion, police approached another group of youths in a
parking lot. After police recovered a handgun from underneath a parked
car, they arrested one of the youths and brought him to the station where
they asked him, among other things, who owned the gun. When he failed

106. Pet. for Writ of Habeas Corpus at 2, In re Emmanuel Chavez, No. FJ10183 (Cal. Super. Ct.
Feb. 23, 2000).
107. Id., Ex. A at 1.
108. Id.
109. Id.
110. Return and Answer to Pet. for Writ of Habeas Corpus, at 9–11, Decl. of Michael Gannon, In
re Diego Barrios, No. BA126209 (Cal. Super. Ct. July 13, 2000) (recounting statements made by
several witnesses).
111. Id.
112. Id. at 11.

2013]

POLICE MISCONDUCT

1157

to provide an answer, the interrogating officer “told him that he was going
to jail for the gun and rubbed it against Lobos’ fingers.”113 Lobos pled
guilty at his arraignment to a charge of unlawful possession of a firearm
by a felon.114
In these cases, contraband discovered by police in the possession of
one person, or an unknown person, was attributed to others in order to
permit an arrest to be made or to facilitate additional arrests. In other
cases, police planted guns or drugs obtained elsewhere on suspects, or
simply claimed that they found guns or drugs on suspects who in fact were
not in possession of them.115 This is precisely what officers did in the
Ovando case, where, after shooting Ovando, the officers planted a weapon
on him picked up elsewhere to falsely implicate him in criminal conduct
and cover up their own misdeeds.116
Other examples include the case of Ivan Oliver, who was charged with
unlawful possession of a gun after police raided a party at which he was
present. In Oliver’s case, police searched the residence where the party
was held and located several guns. One officer then, investigators
concluded, “arbitrarily decided who would be arrested for possessing
them,” while other “CRASH officers created scenarios accounting for the
recovery of each gun and . . . wrote the arrest report accordingly.”117 In
several cases, defendants did not even know what offenses they were
alleged to have committed until long after being arrested. One defendant,
who was charged with narcotics possession, stated that he “did not find out
why he was being arrested until he got to the jail and asked a jailer to tell
him what his ‘pink slip’ indicated.”118

113. Pet. for Writ of Habeas Corpus, Ex. B, Decl. of Brian Tyndall, In re Allan Manrique Lobos,
No. BA 131378 (Cal. Super. Ct. Jan. 31, 2000).
114. Pet. for Writ of Habeas Corpus, In re Allan Manrique Lobos, No. BA 131378 (Cal. Super.
Ct. Jan. 31, 2000).
115. A gun was planted on Jose Armando Lara, for instance, by Officer Durden, and the weapon
was booked into evidence only after Durden obliterated the serial number. See Pet. for Writ of Habeas
Corpus, Ex. A at 1, Decl. of Richard A. Rosenthal, In re Jose Armando Lara, No. BA145000 (Cal.
Super. Ct. Jan. 25, 2000).
116. See id. at 2.
117. Pet. for Writ of Habeas Corpus, Ex. A at 1, Decl. of Laura Laesecke, In re Ivan Oliver, No.
BA135752 (Cal. Super. Ct. Feb. 11, 2000).
118. Pet. for Writ of Habeas Corpus, Ex. B at 1–2, Decl. of Brian Tyndall, In re Carlos Guevara,
No. BA165829 (Cal. Super. Ct. Feb. 15, 2000). Another Rampart exoneree, Rene Matlong, similarly
stated that “he had no idea he had been arrested for possession of a gun until he was given his
paperwork several hours later at the jail division.” Pet. for Writ of Habeas Corpus, Ex. B, Decl. of
Brian Tyndall, In re Rene Barela Matlong, No. BA131589 (Cal. Super. Ct. Apr. 14, 2000). Two other
Rampart exonerees made almost identical allegations. Pet. for Writ of Habeas Corpus, Decl. of Brian
Tyndall, In re Juan Torrecillas, No. BA143145 (Cal. Super. Ct. Apr. 7, 2000) (reporting that
Torrecillas told investigators that “he had no idea he had been arrested for possession of cocaine” until

1158

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

Although present in every Rampart case, police perjury was not the
sole cause of the false convictions. In some cases, police coerced an
individual to make false statements inculpating the defendant. George
Alfaro, for instance, was arrested for violating a gang injunction based on
such evidence.119 According to police officers, Alfaro and two other
suspects were arrested after police recovered a baggie of rock cocaine at
the scene.120 The officers claimed that one of the suspects admitted that he
possessed the drugs for sale. Rampart investigators, however, concluded
that the drugs were planted at the scene, and the officers coerced the
admission.121 As a result of the incident, Alfaro’s probation was revoked
and Alfaro was sentenced to two years in state prison.122
In other cases, police simply falsely reported incriminating statements
made by others. This happened in the case of Gregorio Lopez.123 Lopez
and another man named Omar Alonso were arrested after police claimed
they saw Alonso in possession of a magnetic key holder containing
cocaine and Lopez attempt to discard a similar item. According to the
arrest report, police searched Lopez and recovered a gun from his
waistband. In fact, investigators found, the drugs said to belong to Lopez
were planted, and the gun was found in his car rather than on his person.
The prosecution’s case was also bolstered by inculpatory statements
allegedly made by Alonso. No such statements, it turns out, were ever
made, nor did the officers administer Miranda warnings as they claimed to
have done.124

informed of charges at jail); Answer to Pet. for Writ of Habeas Corpus at 9, Decl. of Barbara Moulton,
In re Alex Umana, No. BA144035 (Cal. Super. Ct. June 22, 2000). These accounts of being framed
match up with Officer Perez’s own statements about how police typically went about framing suspects.
“When we talk about planting or putting a case on someone, for some reason, some
investigators or some attorneys have thought that we actually—I go into the car, take the three
bindles [of drugs], lay it next to him and go, ‘See, that’s what you dropped.’
It’s not the way it works. I take them into custody, put them in the car, do whatever. And
you know, when it’s time to book evidence, we go and get the evidence from our car. There’s
no need to go and, you know, lay it on the ground next to them or put fingerprints on the
baggies or anything like that. That’s just not the way it works.
We get the evidence when it’s needed and move forward.”
McDermott, supra note 15 (quoting Rafael Perez at LAPD Board of Rights hearing on June 3, 2000).
119. See Pet. for Writ of Habeas Corpus at 2, In re George Kenneth Alfaro, No. BA159394 (Cal.
Super. Ct. Jan. 31, 2000).
120. Id. at 6.
121. Id. at 2.
122. Id. at 2.
123. Lopez is an alias. His real name is Leonel Ramos Estrada. Because court papers refer to
Estrada as Lopez, I use that name here.
124. Pet. for Writ of Habeas Corpus, Decl. of Brian Tyndall, In re Omar Ramos Alonso and
Leonel Ramos Estrada, No. BA148402 (Cal. Super. Ct. Apr. 10, 2000).

2013]

POLICE MISCONDUCT

1159

Several of the Rampart exonerees falsely confessed, or were reported
to have done so. Clinton Harris, for example, was convicted of possession
of a firearm by a felon after police reported that they had observed Harris
wearing a gun in the waistband of his pants, and that he had admitted that
the gun was his, saying “[d]amn, I knew I shouldn’t have bought the
gun. . . .”125 In fact, Harris never made any such statement. At the time of
arrest Harris was seated on the couch of a friend’s apartment. Police
officers entered the apartment without consent and found a gun on a table.
According to Officer Perez, they decided to attribute the gun to Harris
“because he was an ex-con.”126
Delbert Carrillo was arrested after police officers allegedly “‘noticed a
large bulge in his front shirt pocket.’”127 The arresting officers explained
in the police report:
Knowing defendant to be on active parole and having a criminal
history, we asked him what he had in his pocket (to ensure that it
was not a weapon or narcotics). The defendant’s expression went
from that of being calm to nervous, and he hesitantly reached into
his pocket and removed a clear plastic baggy containing
approximately nine white paper bindles, the type routinely used to
package rock cocaine, and stated, “its [sic] rocks.” [W]e recovered
the bag and found it to contain nine paper bindles, each one,
containing approximately ten off-white wafers resembling rock
cocaine.128
After Carrillo was arrested, police obtained a signed statement reading,
with original misspellings, as follows:
I DelBert Carrillo contacted officer Cohan and BRehm to discouse a
matter at the time I had cocane in my posseion. and Because I new
them I thought It would not Be a proBlem. Officers then overed it in
my Shirt pocket. DEC. I make this statement freely.129
Carrillo was charged with possession for sale of cocaine base and
ultimately pleaded guilty to an amended complaint that charged him with

125. Pet. for Writ of Habeas Corpus, Decl. of Barbara Moulton, In re Clinton Harris, No.
BA140224 (Cal. Super. Ct. 2000) (alterations in original).
126. Id., Ex. A, Decl. of Natasha S. Cooper, at 1. Harris allegedly admitted that he was in
possession of a gun at the time of the arrest.
127. Pet. for Writ of Habeas Corpus at 3, In re Delbert Carrillo, No. BA169722 (Cal. Super. Ct.
Sept. 8, 2000) (internal quotation marks omitted).
128. Id.
129. Id. at Ex. C (statement form attached to police report).

1160

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

possession for sale of a controlled substance. He was sentenced to the
statutory minimum term of two years. Carrillo later alleged that the drugs
were planted and police officers coerced him into signing the statement by
threatening to file additional charges against him if he refused.130 His
conviction was vacated after the state discovered evidence corroborating
Carrillo’s account of the incident.131
In short, then, the primary “cause” of false convictions in the Rampart
and Tulia scandals was police perjury, some form of which was present in
100% of the cases. Innocent defendants who won exonerations primarily
had been convicted in the first instance on the basis of the false reports and
false testimony of corrupt police officers. That same police misconduct,
however, was also responsible for the generation of other types of false
evidence, including false witness statements and false confessions that
supported the police officers’ false reports and perjurious testimony in
court.
After police perjury, the most common “causes” of false convictions
were the false confessions generated through police misconduct. False
confessions were present in about 13% of the Rampart cases. Interestingly,
that figure is consistent with findings by Gross and Garrett on the
approximate frequency of false confessions in wrongful conviction
cases.132 While a substantial amount of commentary has focused on the
problem of false confessions, and commentators have probed how
innocent defendants might be induced to confess to crimes they did not
commit, very little discussion exists regarding the problem of entirely
fabricated confessions. Yet, as the Rampart cases show, some false
confessions “occur” simply because police lie about what suspects actually
said.
When the mass exoneration data is added to the existing data regarding
the causes of wrongful convictions, there is ample room to doubt the claim
that witness misidentification is the leading cause of false convictions.
Indeed, when the Rampart and Tulia cases are combined with the data
gathered by Gross in his first study (which intentionally excluded these
cases), perjury dislodges witness misidentification as the most prevalent
cause of known wrongful convictions during the time period covered in

130. Answer to Pet. for Writ of Habeas Corpus at 2, In re Delbert Carrillo, No. BA169722 (Cal.
Super. Ct. Nov. 9, 2000).
131. See id. at 3.
132. Gross found false confessions in 15% of the 340 exonerations examined in his first study.
Gross et al., supra note 54, at 544. Gross’ more comprehensive second study also found false
confessions in 15% of the 873 cases. See Gross & Shaffer, supra note 4, at 40. Garrett found false
confessions in 16% of the 250 DNA exoneration cases he studied. GARRETT, supra note 59, at 18.

2013]

1161

POLICE MISCONDUCT

the study. In the combined data set, perjury is a factor in 221 of the 415
exonerations of innocent defendants occurring between 1989–2003.
Witness misidentifications are a close second, factoring into 219 cases
during this same period. Gross’ more comprehensive study, which
includes mass exonerations in the data set, confirms that perjury and false
accusation, and not witness misidentification, is known to be the leading
factor contributing to wrongful convictions.133
TABLE 1: CAUSES OF FALSE CONVICTIONS FOR EXONERATIONS

Eyewitness
Mis-i.d.
Reported
Perjury
False
Confession

Gross Study
(340)
64% (219/340)

Mass Exon’s
(75)
0% (0/75)

Combined
Total
53% (219/415)

43% (146/340)

100% (75/75)

53% (221/415)

15% (51/340)

13% (5/38)134

15% (56/378)

While there of course is no way to know how generalizable these
numbers are, the data does suggest that efforts to reform the criminal
justice system in order to prevent wrongful convictions should include
greater focus on the prevention of police misconduct. During the last
decade, a major effort has been made to improve the reliability of lineup
identification procedures. The revised data set suggests that those
concerned with decreasing the incidence of wrongful convictions should
devote similar attention to enhancing the integrity and reliability of police
officer statements and testimony.
E. Method of Conviction
Perhaps the most striking insight to be drawn from the mass
exoneration data concerns the high rate of guilty pleas seen in these cases,
which provides strong evidence that the wrongful conviction problem
extends to defendants who plead guilty as well defendants who contest
guilt at trial. Earlier studies of exonerations found only a negligible
number of innocents who were exonerated after pleading guilty. In Gross’

133. Gross & Shaffer, supra note 4, at 40.
134. Excludes Tulia data because information about those investigations was not sufficient to
make a determination.

1162

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

first study of 340 exonerations, only twenty of the exonerees, or
approximately 6%, pled guilty.135 The vast majority, about 94%, were
convicted after a trial. Garrett’s data tell the same story. In Garrett’s study
of 250 DNA exonerations, sixteen, or 6%, pled guilty. The rest were
convicted at trial. This data has been interpreted by some to mean that
innocent people generally do not plead guilty, or if they do, they do so
only under extraordinary circumstances.136
The accuracy of guilty pleas is a major determinant of the scope of the
problem of wrongful convictions. After all, the vast majority of criminal
convictions, upwards of 90%, are a result of guilty pleas.137 If innocent
people do not plead guilty but rather insist on going to trial, then the upper
estimate of wrongful convictions is bounded by the small proportion of
persons overall who are convicted at trial. In other words, even if 100% of
defendants who were convicted at trial were actually innocent, the
wrongful conviction “rate” would still be only about 5%, since
approximately 95% of all defendants plead guilty. If, on the other hand,
the fact that a defendant pleads guilty provides no guarantee that the
defendant is not actually innocent, then the potential magnitude of the
wrongful conviction problem is many times greater. Even if the rate of
false guilty pleas is low, the far-greater size of the guilty plea pool ensures
that it adds up to a quantitatively large problem.
1. Evidence that the Innocent Do Plead Guilty138
It has long been apparent that the innocent do, on occasion, plead
guilty.139 The more important question, however, is how often false guilty
pleas occur, and how false guilty plea rates compare with false trial

135. Gross, supra note 2, at 536.
136. See, e.g., Ronald J. Allen and Larry Laudan, Deadly Dilemmas, 41 TEX. TECH L. REV. 65, 71
(2009) (concluding that false guilty plea rate is much lower than false conviction rate at trial based on
evaluation of Garrett data).
137. See Sourcebook of criminal justice statistics online (2006), http://www.albany.edu/source
book/pdf/t5462006.pdf.
138. See Pet. for Writ of Habeas Corpus at 11, Decl. of Brian Tyndall, In re Gerald Peters, No.
BA131401 (Cal. Super. Ct. May 9, 2000) (reporting that “Peters plead[ed] guilty to the charges on the
advice of his attorney because he believed he would face a stiffer penalty if he chose to fight the
charges in a trial and lost”); Scott Glover & Matt Lait, 10 More Rampart Cases Voided, LA TIMES
(Jan. 26, 2000), http://www.streetgangs.com/topics/rampart/012600more10.html (“Davalos, 41, an
upholstery worker who served 91 days in jail. He said he only agreed to a plea bargain because he was
threatened with eight to 16 years in prison.”).
139. For a list of sources discussing the problem of innocent persons pleading guilty, see Gabriel
J. Chin & Richard W. Holmes, Jr., Effective Assistance of Counsel and the Consequences of Guilty
Pleas, 87 CORNELL L. REV. 697, 740 n.305 (2002).

2013]

POLICE MISCONDUCT

1163

conviction rates. If there ever was any real doubt that false guilty pleas can
occur in large numbers, the Rampart and Tulia data put those doubts to
rest, indicating that at least in some types of cases, innocent defendants are
far more likely to be convicted through a guilty plea than at trial. In the
Rampart cases not involving alleged probation violations, twenty-five of
thirty-two exonerees pled guilty. In Tulia the numbers were about the
same: twenty-seven of thirty-four. Overall, fifty-two of the exonerees, or
81%, were convicted through guilty pleas, and twelve, or 19%, were
convicted after trial. Those numbers represent a far more typical
distribution of guilty pleas and trial convictions than was seen in the Gross
and Garrett data, and provide strong reason to believe, notwithstanding
prior exoneration studies showing a low incidence of guilty pleas among
exonerees, that the problem of wrongful convictions is not contained to
those who contest their guilt at trial.140 Indeed, the mass exoneration cases
make clear that, at least with respect to the types of charges at issue in the
Rampart and Tulia cases, the method of conviction makes very little
difference to the reliability of the conviction. In Rampart and Tulia,
wrongful convictions resulted from guilty pleas and trials alike, and as is
true in the criminal justice system generally, guilty pleas accounted for the
majority of the convictions.
TABLE 2: METHOD OF CONVICTION OF EXONERATED DEFENDANTS

Guilty plea
Trial conviction

Gross Study

Mass Exon’s

6% (20/340)
94% (320/340)

81% (52/64)
19% (12/64)

Combined
Total
18% (72/404)
82% (332/404)

As Table 2 suggests, in comparison with other exonerees, the Rampart
and Tulia exonerees pled guilty at much higher rates. The percentage of
convictions obtained through guilty pleas, however, still falls short of the
average. What is the significance of these numbers? On one hand, a trial
rate approaching 20% in low level drug cases might seem remarkable.
These are typically simple cases to prosecute and the vast majority of such
cases undoubtedly would normally be resolved through guilty pleas. On
the other hand, the evidence in many of these cases, especially the Tulia

140. See, e.g., Ronald J. Allen & Larry Laudan, supra note 136, at 71 (calculating wrongful
conviction rate by guilty plea based on Garrett study data showing that 9/200 exonerees pled guilty).
Michael Risinger critiques that reasoning in a responsive essay. See D. Michael Risinger, Tragic
Consequences of Deadly Dilemmas: A Response to Allen and Laudan, 40 SETON HALL L. REV. 991,
995 (2010).

1164

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

cases, was extremely weak. These were cases built on the testimony of a
single undercover cop, with no electronic recording of the transactions or
corroborating evidence, and in some cases in the face of significant alibi
defenses. Given this, it is difficult to tell whether a 20% trial rate is high or
low. Regardless, the fact that so many mass exoneration cases were
resolved by guilty pleas should erode any perception that actually innocent
defendants almost uniformly refused to plead guilty.
Of course, it is possible that the Rampart and Tulia cases, rather than
the exoneration cases studied by others, represent the outlier. There are
several reasons, however, to believe that with respect to the frequency of
false guilty pleas, the Rampart and Tulia cases provide a more typical
distribution, and that wrongful convictions, like convictions generally, are
usually the end product of a guilty plea rather than a trial verdict. First, as
both Gross and Garrett acknowledge, cases resulting in exonerations are
the beneficiaries of a phenomenally rare confluence of events that are
simply not present in typical cases.141 For a DNA exoneration to occur, for
example, the crime must have been one involving biological evidence,
where that evidence is dispositive of the defendant’s guilt. That evidence
must have been gathered but not tested, or not tested properly, preserved
for years or decades, and located in quantities and in sufficient quality to
permit testing, and defendants must have preserved the means to launch a
legal challenge against their conviction once the evidence is discovered.
The preconditions for exoneration after a trial conviction are only rarely
satisfied; rarer still will they exist where the defendant pleads guilty.
In guilty plea cases, the state is less likely to preserve evidence for later
testing, and because no trial record exists, even where such evidence was
preserved, it is difficult to assess the significance of exculpatory test
results. Defendants who plead guilty typically waive their rights to appeal
and to post-conviction review.142 As a result, innocent people who plead
guilty almost invariably lack a viable procedural mechanism to prove their
innocence in a post-conviction proceeding, at least absent the type of
extraordinary circumstances that occurred in Rampart and Tulia. To make
matters worse, many statutes governing access to post-conviction DNA

141. See, e.g., Gross & Shaffer, supra note 6, at 4–5 (explaining how exonerations tend to be the
product of “blind luck” or “improbable chains of happenstance”).
142. See Nancy J. King, Appeal Waivers And The Future Of Sentencing Policy, 55 DUKE L.J. 209,
209 (2011) (finding in her empirical study that defendants who pleaded guilty waived their right to
appeal in nearly two-thirds of cases).

2013]

POLICE MISCONDUCT

1165

testing specifically preclude defendants who plead guilty from obtaining
testing.143
There is good reason, moreover, to view the mass exoneration cases as
far more typical of garden-variety wrongful convictions than the cases
included in the earlier Gross and Garrett studies, a very large percentage of
which (100% in the Garrett study) involved post-conviction DNA testing.
In the Rampart and Tulia cases, most defendants were convicted of drug or
gun crimes, which are far more common than the rape, murder, or rapemurder convictions making up the vast majority of the earlier studied
exoneration cases. Although some sentences were draconian, especially in
Tulia, most sentences were relatively modest in severity, as are most
felony sentences imposed on typical felony convicts. As noted above,
most of the Rampart exonerees received relatively light sentences, with the
average sentence being approximately three years, and the median
sentence less than two years. These figures are consistent with national
averages for state felons.144 In contrast, exonerees in the first Gross study
had almost uniformly received harsh sentences for the most serious
crimes. This was especially true among the non-DNA exonerations in the
data pool, of which 85% (166/196) were serving sentences for murder or
manslaughter, and 22% among all of the exonerees (74/340) were
sentenced to death.145
Moreover, the exonerations in Rampart and Tulia were largely the
product of happenstance. The Rampart exonerations in particular involved
run-of-the-mill drug and gun cases that never would have received even
passing interest from the outside world had it not been for the cooperation
deal struck by Rafael Perez. Unlike typical DNA exonerations, the
exonerations in Rampart came about without the intervention of Innocence
Projects or big-firm pro-bono advocacy. There were few trial transcripts,
physical evidence, or other compelling evidence from which a defendant’s
actual innocence could be determined.146 The Tulia exonerations did
benefit from substantial pro-bono advocacy, but one suspects that none of
the events leading to the uncovering of misconduct in Tulia would have
been uncovered had the extent of the misconduct not been as sweeping,
the sentences not as draconian, and the racial component not as overt as it

143. See, e.g., FLA. STAT. ANN. § 925.11(1)(a)(2) (West 2006). But see TEX. CODE CRIM. PROC.
ANN. art. 64.03(b) (West 2007).
144. See supra note 137.
145. Gross, supra note 2, at 531.
146. This, in part, was a necessary byproduct of a set of convictions obtained largely through plea
bargains.

1166

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

was. Setting aside the extraordinary manner in which the police
misconduct was discovered, the kinds of convictions at issue in Rampart
and Tulia were far more typical, substantively and procedurally, than those
that have eventuated in DNA exonerations.
V. EXPLAINING WRONGFUL PLEAS
In addition to providing an empirical basis for the claim that innocent
people plead guilty, the mass exoneration cases vividly illustrate how and
why actually innocent defendants plead guilty. In general, there appear to
have been three main factors driving innocent Rampart and Tulia
defendants to plead guilty: an outsized trial penalty, a lack of viable
strategies to contest the charges, and presumptively or actually
unsympathetic forums. Each is considered briefly below.
A. New Data on the Trial Penalty
Without a doubt, the overwhelming reason that innocent Rampart and
Tulia exonerees pleaded guilty to crimes they did not commit was that
they feared that they would do much worse at trial if they did not plead
guilty. Typical are the sentiments expressed by one innocent Rampart
exoneree who on advice of his attorney pleaded guilty in exchange for a
three-year term of probation, believing that “he would face a stiffer
penalty if he chose to fight the charges in a trial and lost.”147 That exoneree
likely was not wrong. The existence of a trial penalty has been long
acknowledged, albeit bemoaned by many.148 It is an institutionalized
feature of contemporary criminal justice. Nonetheless, the coercive impact
of the trial penalty is unmistakable, and is plainly evident in the Rampart
and Tulia cases.
Tulia provides an extreme example of the coercive impact of the trial
penalty. Of the thirty-seven innocent Tulia exonerees, seven went to trial
and were convicted, twenty-seven pleaded guilty, one did both, and two

147. In re Gerald Peters, supra note 138. Peters also alleged that he was physically abused by
officers in an interview, but that “he never made a complaint regarding this incident because he felt ‘it
would do no good.’” Id. Similarly, two months after pleading guilty, Ruben Rojas had second thoughts
and wrote a letter to the judge who had sentenced him. In the handwritten letter, Rojas explained that
“I was informed that I was facing 25 years to life by my defense counsel and that there was no way I
could have won my case because I was up against a police officer.” He added: “I never did what I was
charged for. . . I’m not guilty.” Matt Lait, Another Inmate Set to Be Freed in Police Probe, L.A.
TIMES, Nov. 17, 1999, at A1.
148. See, e.g., Candace McCoy, Plea Bargaining as Coercion: The Trial Penalty and Plea
Bargaining Reform, 50 CRIM. L.Q. 67 (2005).

2013]

POLICE MISCONDUCT

1167

others had their probation revoked.149 The first defendant to go to trial, Joe
Moore, was convicted and sentenced to ninety years in prison for allegedly
dealing 4.5 grams of cocaine.150 Moore had been offered an opportunity to
plead guilty in exchange for a twenty-five-year sentence (the minimum
available given the charges and Moore’s prior record), but he declined.151
Six more defendants stood trial, and were convictedand sentenced to
prison terms ranging from 20 to 361 years.152 In light of this precedent,
and with cases substantively indistinguishable in terms of the nature of the
charges and the strength of the evidence, the remaining defendants all
chose to plead guilty.153 Although the sentences imposed on those who
pleaded guilty in Tulia were often quite harsh, the harshness of their
sentences paled in comparison to those who were convicted at trial. On
average, Tulia defendants who pleaded guilty were sentenced to
approximately four years in prison.154 The Tulia defendants who contested
their guilt at trial received an average sentence of 615.4 months, or 51.3
years.155 Trial sentences at Tulia, in other words, were nearly thirteen
times harsher than sentences imposed following guilty pleas.
The trial penalty evident in Tulia might be attributed, at least in part, to
an apparently intentional prosecutorial strategy to frighten defendants into
foregoing trial. Such an express strategy was made easier in small-town
Tulia, where word of harsh sentences quickly spread among Tulia’s small
defense bar and the defendants themselves.
These dynamics were noticeably absent in Rampart. Unlike Tulia, there
is no indication that prosecutors were aware of the defects in the cases
they brought against innocent defendants. Indeed, after the scandal broke,
the Los Angeles District Attorney’s Office took affirmative steps to
investigate the scope of wrongdoing and to vacate convictions resulting
from police misconduct.156 In terms of size and population, the L.A. justice
system also obviously dwarfs Tulia’s. There is far less reason to believe

149. Donald Wayne Smith was charged with seven drug trafficking offenses, and the prosecutor
elected to try the cases separately. See BLAKESLEE, supra note 69, at 117. After Smith was convicted
in the first case and sentenced to two years in prison, he accepted a plea offer to resolve the remaining
charges in exchange for a 12.5 sentence to run concurrently with his other conviction. Id. at 136–37.
150. Id. at 59.
151. Id. at 48.
152. Cash Love was sentenced to 361 years by the trial court. Id. at 92.
153. See id. at 160–61.
154. See Summary of Data (on file with author).
155. Id. In calculating this figure, I counted Cash Love’s sentence as 99, rather than 361, years. I
also omitted Smith’s case. Smith’s two-year sentence was based on the least serious of only one of
seven charges.
156. See generally CONSTANCE L. RICE ET AL., RAMPART RECONSIDERED: THE SEARCH FOR
REAL REFORM SEVEN YEARS LATER (2006).

1168

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

that prosecutors sought to send any messages to specific classes of
defendants by seeking harsh trial sentences. Any implicit threat inherent in
the harsher trial sentences would seem to be endemic to the justice system
in general.
Nonetheless, the observable trial penalty in the Rampart cases, though
not on the same order as the average trial penalty in Tulia, was still quite
large. On average, actually innocent Rampart defendants who were
convicted at trial were sentenced to 101.25 months, or nearly 8.5 years.157
Actually innocent Rampart defendants who pleaded guilty were sentenced
to an average term of 18.5 months, or just over 1.5 years.158 Defendants
who contested their cases at trial, in other words, received sentences on
average more than five times harsher than those who agreed to plead
guilty.159 The trial penalty for the larger sample of all Rampart exonerees,
including those who did not appear to be actually innocent, was even
bigger. For this group, the average plea sentence was 20.3 months.160 The
average trial sentence was 136.3 months.161 Trial sentences were therefore
on average 6.7 times longer than plea sentences, with no apparent
qualitative differences among the types of crimes charged or the criminal
history of the defendants.162
The longest sentence imposed on any Rampart exoneree was a term of
fifty-four years to life, later reduced on appeal to twenty-nine years to life,

157. See Summary of Data (on file with author).
158. Id.
159. Arguably, one could object that these numbers are skewed by the inclusion of the Ovando
case—by far the harshest sentence imposed in any of the Rampart cases. There are several reasons,
however, to include that case in calculating the numbers. First, Ovando was not the only innocent
Rampart exoneree to be charged with a crime of violence. Jose Perez was charged with assaulting a
peace officer with a firearm, the same crime charged against Ovando, and pleaded guilty. Perez
received a sentence of three-years probation. Raul Munoz and Cesar Natividad were also (falsely)
charged with assaulting a peace officer with a deadly weapon—in their case—allegedly attempting to
run over a police officer with a car. Both settled the cases by guilty plea. Munoz was sentenced to
three-years prison, and Natividad was sentenced to a three year term of probation. None of these
defendants suffered the kinds of serious injuries that Ovando did. Ovando’s trial sentence was thus
about eight times harsher than the harshest plea sentence for comparable conduct, and if one treats a
term of probation as equivalent to about one-third of a prison term, his trial sentence exceeded the
average plea sentences in the three cases by 1400%. In Ovando’s case, prosecutors offered Ovando a
thirteen-year deal to resolve the case, but Ovando’s lawyer thought the offer was “way too severe” and
turned it down. Lou Cannon, One Bad Cop, N.Y. TIMES MAG., Oct. 1, 2000, available at
http://www.nytimes.com/2000/10/01/magazine/one-bad-cop.html.
160. See supra note 159.
161. Id.
162. In calculating these figures, there were sixty-one plea sentences and twelve trial sentences
included in the data. Sentencing data was unavailable in some of the cases. In other cases, defendants
were never sentenced because they failed to appear, and a bench warrant issued. Terms of probation
were not counted as punishment. Some defendants were deported as a result of their convictions.
Deportation was also not counted as punishment in calculating average sentences.

2013]

POLICE MISCONDUCT

1169

for Lorenzo Nava. Like most of the other Rampart exonerees, Nava was
convicted of drug and gun offenses and contested the charges at trial.163
After conviction, Nava received an initial fifty-four-year sentence under
California’s three-strikes law.164 Nava’s case, however, can be compared
to Joseph Jones, another Rampart defendant, to show that the long trial
sentence imposed on Nava was not simply a function of the three-strikes
law or other factors unique to his case. Like Nava, Jones was charged with
multiple drug counts and was potentially subject to prosecution under the
three-strikes law. According to Detective Chris Barling, who interviewed
Jones in the Rampart investigation, “Jones believed that he was facing a
life term,” and notwithstanding his contention that he was innocent,
decided to plead guilty on the advice of counsel.165 Pursuant to the plea,
Jones was sentenced to a prison term of eight years.166 The disparity in
sentence outcome between Nava and Jones is roughly consistent with the
average trial penalty evident in the Rampart cases, amounting to at least a
seven-fold penalty increase based on Nava’s initial trial sentence, and a
four-fold penalty increase based on Nava’s reduced sentence on appeal.167
This data provides further evidence that the real trial penalty could be
far larger than estimated in some studies.168 With trial sentences ranging
anywhere from four to thirteen times longer than plea sentences, the costs
of contesting a typical felony charge are prohibitive. Few defendants can
afford to run the risk. The experience of those wrongly convicted in the
Rampart and Tulia scandals demonstrates that the coercive power of the

163. See Summary of Data (on file with author).
164. Id.
165. Pet. for Writ of Habeas Corpus, Ex. B at 1, Decl. of Chris Barling, In re Joseph Jones, No.
BA154853 (Cal. Super. Ct. Nov. 10, 1999).
166. Pet. for Writ of Habeas Corpus at 2, In re Joseph Jones, No. BA154853 (Cal. Super. Ct. Nov.
10, 1999).
167. These numbers represent minimums because they ignore the upper end of the sentencing
range (life) and are based on the minimum sentence that Nava was required to serve.
168. Hans Zeisel estimated that trial sentences were 42% more severe than guilty plea sentences.
See Hans Zeisel, The Disposition of Felony Arrests, 1981 AM. B. FOUND. RES. J. 407, 444–49 (1981).
Other scholars have reached varying estimates. See, e.g., Nancy J. King et al., When Process Affects
Punishment: Differences in Sentences After Guilty Plea, Bench Trial, and Jury Trial in Five
Guidelines States, 105 COLUM. L. REV. 959 (2005) (finding that trial penalties varied from state to
state, and offense to offense, ranging from 13% to 461%); Jeffery T. Ulmer & Mindy S. Bradley,
Variation in Trial Penalties Among Serious Violent Offenses, 44 CRIMINOLOGY 631, 652 (2006)
(finding that defendants convicted at trial received sentences 57% longer than defendants who pleaded
guilty). The Federal Sentencing Guidelines provide a 35% discount for defendants who “accept
responsibility.” Because of the variety of charge bargains that typically accompany federal plea
bargains, actual discounts are almost always far larger than 35%. See Russell Covey, Reconsidering
the Relationship Between Cognitive Psychology and Plea Bargaining, 91 MARQ. L. REV. 213, 224–30
(2007).

1170

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

trial penalty causes innocent defendants as well as guilty ones to plead
guilty.
B. Lack of Effective Trial Strategies for Falsely Accused Defendants
Time and again, actually innocent defendants asked by investigators to
explain why they pled guilty repeated a common mantra: it was their word
against that of the police, and who were the prosecutors, judges, or jurors
going to believe?169 While this dynamic is present in most cases, it is
especially likely to have an effect where defendants have reason to believe
they will not be treated fairly. Most of the Rampart exonerees were gang
members, some with criminal histories. They were likely correct in
believing that few middle-class jurors would give credence to their claims
of police misconduct. In Tulia, racial dynamics clearly affected the
calculations of the black defendants, who assumed (correctly, given the
trial outcomes) that their protestations of innocence would be ignored. In
these cases, innocent defendants often had little except their own word to
prove their innocence, and their word was demonstrably not enough. In
part because of the nature of the cases, and in part because of their lack of
resources, the defendants were typically unable to amass any credible
exonerating evidence. Given that the police already had demonstrated a
willingness to testify falsely,170 many defendants realized that a successful
trial defense was unlikely and simply decided to cut their losses.
C. Unsympathetic Forums
A third reason so many innocent Tulia and Rampart defendants pled
guilty, even in cases where the evidence was flimsy, was an undoubtedly
accurate perception that the system itself was not constituted in a way
likely to give them much chance of prevailing. As one blue ribbon panel
observed after investigating the Rampart scandal, the Los Angeles County
criminal justice system is characterized by “assembly-line” justice.171
Many actors are complicit in pressuring innocent defendants to plead

169. See Pet. for Writ of Habeas Corpus, Decl. of Michael J. Hansen, In re William Zepeda and
Argelia Diaz, No. BA156980 (Cal. Super. Ct. Mar. 17, 2000) (reporting that Zepeda “decided to plead
guilty to the charge after he realized it was just his word against the officers”).
170. For example, Rafael Zambrano, who was charged with violating probation for unlawful gun
possession after police planted a gun on him, claimed that he “decided to plead guilty to the charge
after Officer Rafael Perez testified at his preliminary hearing.” See Pet. for Writ of Habeas Corpus, Ex.
B, Decl. of Brian Tyndall, In re Rafael Zambrano, No. BA138148 (Cal. Super. Ct. Feb. 11, 2000).
171. RAMPART RECONSIDERED, supra note 16, at 49.

2013]

POLICE MISCONDUCT

1171

guilty to crimes they did not commit, including prosecutors who “pressure
defendants to accept plea deals in extremely short time frames,”
“overworked public defenders” who counsel their clients to accept those
pleas, and judges who are quick to impose “draconian” sentences on those
who “drain[] judicial resources by demanding a trial.”172 Moreover,
prosecutors are reluctant to doubt the credibility of the police officers with
whom they work daily, and judges “are unwilling or unable to pursue their
own suspicions of police perjury or misconduct.”173 As a result, a falsely
accused defendant has little reason to believe that he will fare well by
going to trial, and has great reason to believe that he will be much worse
off by refusing to take a plea and cut his losses.
Ample evidence also suggests that judges often are biased toward the
prosecution.174 A large part of the bench is populated by former
prosecutors. These former prosecutors often have difficulty shedding their
former roles. Regardless of background, judges often form relationships
with prosecutors who appear regularly in their courtrooms, and many think
of themselves as part of a “law-enforcement” team. In addition, electoral
politics drive many judges to more pro-prosecution positions. Some judges
even campaign overtly on being “tough on crime” or “hard on
criminals.”175 Actually innocent defendants tried before such judges likely
are often led to believe, probably correctly, that they will not get the
benefit of the doubt should they go to trial.
Arguably, pro-prosecution judges played an especially prominent part
in many of the cases in which actually innocent defendants were convicted
at trial. The two judges presiding over the Tulia prosecutions initially
barred defense lawyers from impeaching undercover agent Tom
Coleman’s character. After defense counsel discovered that Coleman had

172. Id.
173. Id.
174. See, e.g., Susan D. Rozelle, Daubert, Schmaubert: Criminal Defendants and the Short End of
the Science Stick, 43 TULSA L. REV. 597, 606 (2007) (arguing that judges admit dubious forensic
science far more often on behalf of prosecutors than defendants); Rodney J. Uphoff, On Misjudging
and Its Implications for Criminal Defendants, Their Lawyers and the Criminal Justice System, 7 NEV.
L.J. 521, 529 (2007) (noting based on personal observation that “a significant number of judges with
prior prosecutorial experience bring a decidedly pro-prosecution attitude to the bench, and that attitude
invariably influences their decisionmaking”); Keith Swisher, Pro-Prosecution Judges: “Tough on
Crime,” Soft on Strategy, Ripe for Disqualification, 52 ARIZ. L. REV. 317 (2010) (arguing that elected
judges are biased against defendants and “tough on crime” prosecutors should recuse themselves in
criminal cases under ethics rules).
175. Swisher, supra note 174, at 328–29 (quoting numerous expressions of overt bias in judges
electoral campaigns, including one pledge from a Texas Court of Appeals judge who asserted “I’m a
prosecution-oriented person” who “see[s] legal issues from the perspective of the state instead of the
perspective of the defense” (internal quotation marks omitted)).

1172

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

been charged with theft in a neighboring county during the same time
period in which the undercover operation was conducted, the judges still
barred the defense from putting any of that evidence before a jury, thereby
effectively precluding defendants from presenting their theory of the
case.176 In the Rampart trial of Javier Ovando, the presiding judge made a
similar ruling against the defense, precluding Ovando’s attorney from
attempting to impeach Officers Perez and Durden with discrepancies
between contemporaneous statements given by them to investigators and a
new account of events proffered at trial.177 Decisions by trial judges to
preclude defendants from introducing evidence calling into question the
honesty and integrity of the police or challenging substantial
inconsistencies in the prosecution case were a major part of the Rampart
and Tulia stories, and a clear contributing cause to many of the wrongful
convictions that occurred.
D. Innocence as a Minor Factor in Plea Bargaining
Although the empirical evidence from the mass exoneration cases
leaves no doubt that innocent defendants plead guilty, the question
remains: does innocence have a measurable impact on whether a defendant
will hold out for trial? Anecdotally speaking, we know that some innocent
defendants turn down favorable plea bargains because of innocence. One
Rampart exoneree to do so was Alex Umana. Umana was returning from a
barbecue with his daughter and her mother when police stopped him and
placed him among a group of four to six people who had been detained by
police in the lobby of an apartment building. Although Umana was not in
possession of any drugs at the time, he was nevertheless charged with
possession of cocaine. Prosecutors offered a plea bargain for a
probationary sentence, but Umana rejected the offer “because he was
innocent and wanted to fight the charges.”178 Umana was convicted at trial
and sentenced to five years in state prison.179 Several Tulia defendants also
refused to plead guilty to drug charges because they were innocent. Take
the case of Freddie Brookins, Jr., for example, who was accused of selling
an eight-ball of powder cocaine to Coleman. Before trial, the prosecutor

176. See BLAKESLEE, supra note 30, at 100.
177. See Tamar Toister, Rampart Hasn’t Changed How Criminal Courts Do Business, L.A.
TIMES, Sept. 21, 2000, available at 2000 WLNR 8421802.
178. Answer to Pet. for Writ of Habeas Corpus at 9, Decl. of Barbara Moulton, In re Alex Umana,
No. BA144035 (Cal. Super. Ct. June 22, 2000).
179. Pet. for Writ of Habeas Corpus at 2, In re Alex Umana, No. BA144035 (Cal. Super. Ct. Apr.
6, 2000).

2013]

POLICE MISCONDUCT

1173

offered Brookins a plea of five years.180 The maximum sentence for the
offense was twenty years. Brookins discussed the offer with his father, and
the following colloquy reportedly occurred: “Did you do it,” his father
asked him. “No, I didn’t,” Brookins replied. “Well then,” his father
responded, “don’t take the deal.”181 Against the advice of counsel,
Brookins declined the plea offer. He went to trial, was convicted, and was
sentenced to the maximum term of twenty years.182
Although we know that some defendants decline plea offers because of
innocence, it is also possible that other innocent defendants plead guilty at
equal or higher rates to avoid draconian trial penalties. The question
therefore remains: does innocence materially alter guilty plea rates? The
Rampart data sheds some additional light. For the purposes of this
analysis, I identified three groups of Rampart cases resulting in
exonerations. The first group, discussed above, consisted of those who
were actually innocent of the crimes of conviction. There were thirty-eight
such defendants in the data set. Of the remaining forty-nine, twenty-seven
were identified as clearly “not actually innocent.” This group consisted of
defendants who in fact were in possession of contraband or who admitted
that they were engaged in criminal conduct at the time of arrest, but whose
convictions were reversed based on procedural violations. The remaining
group of twenty-two consisted of defendants whose guilt or innocence
remains unclear given the record evidence. I identify this group as the
“may be innocent” group.
Although the numbers are small, they are large enough to permit some
tentative comparisons. With respect to plea rates, the data shows that
innocence does appear to make some difference. Twenty-five actually
innocent Rampart exonerees pleaded guilty, while seven were convicted at
trial.183 Actually innocent exonerees thus pleaded guilty at a rate of 77%.
In comparison, twenty-two of those who were not actually innocent pled
guilty while three were convicted at trial. In other words, 88% of those
who were not innocent pleaded guilty. Finally, of the remaining group of
“may be innocents,” seventeen pled guilty while two were convicted at
trial, providing an 89% guilty plea rate.184

180. BLAKESLEE, supra note 30, at 138, 148.
181. Id. at 148 (internal quotation marks omitted).
182. Id. at 157.
183. The other seven had their probation revoked, or were minors who were adjudicated
delinquent.
184. Of the rest, two admitted probation violations and 1 had his probation revoked.

1174

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

It thus appears from the data that actual innocence does induce some
defendants to refuse a guilty plea and hold out for trial, but that the
incentive has only a marginal effect, leading the innocent to contest their
cases at trial at an approximately 10% greater rate than those who are
actually guilty. Nonetheless, the data underscore that the vast majority of
the actually innocent resolve false charges against them by pleading guilty.
Very few held out for trial, and, as the numbers above documenting the
size of the trial penalty demonstrate, those who did and lost paid a heavy
price for that decision.185
VI. THE WRONGFULLY CONVICTED VS. THE ACTUALLY INNOCENT: DOES
THE DISTINCTION MATTER?
In the national dialogue about wrongful convictions, definitions of
terms like “innocence,” “exonerated,” and “wrongfully convicted” have
been contested. Has a person, convicted on the basis of unconstitutionallyobtained evidence, been “wrongfully convicted”? The answer, technically,
is yes, but commentators typically use terms like “legal innocence” to
describe defendants whose convictions resulted from significant
procedural error but who are not factually innocent, or at least cannot
establish their factual innocence.186 Legally innocent defendants were
“wrongfully convicted,” but typically are treated as occupying a lesser
status in the wrongful conviction debates than those who are “factually” or
“actually” innocent.
The term “actually innocent” has tended to be reserved for those who
succeed in establishing not only that their conviction was legally flawed,
but that they did not engage in any significant criminal wrongdoing.
Accordingly, both Gross and Garrett limited their datasets to those
defendants who were both formally exonerated by official act declaring
the defendant not guilty of the crime of conviction and who were “actually
innocent.” By “actually innocent,” Gross and Garrett mean that the
exoneration was based on evidence that the defendants “had no role in the
crimes for which they were originally convicted.”187 Defendants who were

185. Of course, whether trial was a good or bad decision for the average innocent defendant
falsely charged by corrupt Rampart officers is impossible to determine without information regarding
acquittals and dismissals of such defendants, which is unavailable. The data does show that those who
gamble on trial and lose fare far worse than those who plead guilty.
186. See Margaret Raymond, The Problem with Innocence, 49 CLEV. ST. L. REV. 449, 456 (2001)
(distinguishing between legal and factual innocence); Emily Hughes, Innocence Unmodified, 89 N.C.
L. REV. 1083 (2011) (critiquing the distinction).
187. Gross et al., supra note 2, at 524.

2013]

POLICE MISCONDUCT

1175

not guilty of the convicted offense, but who were guilty of committing
some lesser crime based on the same conduct, are not considered actually
innocent and are typically excluded from any count of exonerations.188
All of the individuals who were included in the Gross and Garrett
studies were thus persons whose convictions were formally vacated, either
through pardon or court order, and who were able to produce strong
evidence not only that the convictions in their cases were unreliable, but
that they were affirmatively innocent of wrongdoing. Other commentators
have also urged the importance of distinguishing between actually
innocent and procedurally innocent defendants, because convictions of
actually innocent people represent far more serious breakdowns in the
truth-seeking function of the criminal process.189 Still, police misconduct
remains troubling even where the victim of that misconduct is engaged in
unlawful behavior. Such misconduct undermines the effectiveness of
constitutional rules established to protect the bodily integrity, privacy, and
autonomy of citizens from incursion by the state. When police evade these
rules by lying about their conduct, they undermine those mechanisms and
weaken the protections safeguarding the innocent and the guilty alike.
A. Wrongful Convictions Resulting from Unconstitutional Police Conduct
The primary aim of this article has been to use the Rampart and Tulia
exonerations as a means to understand how police misconduct causes
wrongful convictions. Accordingly, until now the article has focused on
the Rampart and Tulia cases that meet the actual innocence criteria used
by other researchers in studying known wrongful convictions. As
discussed above, thirty-eight Rampart exonerees and thirty-seven Tulia
188. See id. at 524 n.4. Garrett’s study employed similar criteria, generally adopting the same
screening mechanism—affirmative proof of innocence—used by the Innocence Project to identify
potential clients. See GARRETT, CONVICTING THE INNOCENT, supra note 59, at 285–86 (explaining that
list of 250 DNA exonerations does not include “cases in which there has been no exoneration despite
DNA evidence of innocence” and only includes cases in which there is no doubt that the “convicts are
actually innocent”).
189. See Jon B. Gould & Richard A. Leo, One Hundred Years Later: Wrongful Convictions After
a Century of Research, 100 J. CRIM. L. & CRIMINOLOGY 825, 833 (2010). At least one commentator
has taken issue with the narrowness of the definition. In a forthcoming paper, Keith Findley argues
that the criteria used to define the “actually innocent” is too narrow, at least where proof of innocence
rather than absence of proof of guilt is demanded. Findley thus contends that all persons whose
convictions are formally vacated based on evidence of innocence should be considered innocent. See
Keith A. Findley, Defining Innocence, 74 ALB. L. REV. 1157 (2011).

1176

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

exonerees meet these criteria.190 A list of all of the innocent Rampart and
Tulia exonerees appears in the Appendix to this article.
Forty-nine persons—all Rampart defendants—were excluded from the
dataset notwithstanding the fact that their convictions were reversed or
vacated by a court with the affirmative consent of the Los Angeles District
Attorney on the basis of police misconduct. These defendants did not meet
the actual innocence criteria because there either remained affirmative
evidence of criminal wrongdoing or insufficient evidence of innocence in
the records available for this study. Under the Gross criteria, these
individuals were not “exonerated” in the relevant sense. There is no doubt,
however, that they were wrongfully convicted. Their cases illustrate
various ways in which police officers circumvent constitutional
protections and then lie about their conduct in order to convict criminal
defendants. I refer to this type of police misconduct as “procedural
perjury.”
B. Types of Police “Procedural Perjury”
Procedural perjury occurs when police lie about the circumstances of
an encounter in order to ensure that evidence obtained during the
encounter is not excluded or excludable. Procedural perjury is a common
enough problem that a word—“testilying”—has been coined to describe
the phenomenon.191 In one survey, insiders in the criminal justice system
estimated that police perjury occurs in 20% to 50% of all Fourth
Amendment suppression hearings.192 Seventy-six percent of police officers
also believed that police misrepresented the facts relevant to probable
cause determinations.193 In general, procedural perjury arises in three main
guises: lies about consent, lies about probable cause, and lies about
compliance with other constitutional rules of criminal procedure, most
commonly, the rules governing custodial interrogation set forth in
Miranda v. Arizona.194 As I use the terms, procedural perjury differs from

190. See infra Part IV.A. One Tulia defendant, Jonathan Loftin, was a minor at the time of his
wrongful conviction and did not receive a pardon because he had already served out his camp
sentence. Although his case is indistinguishable from the other Tulia cases in every other respect, I
have not included him in the dataset.
191. See generally I. Bennett Capers, Crime, Legitimacy, and Testilying, 83 IND. L.J. 835 (2008).
192. See Christopher Slobogin, Testilying: Police Perjury and What to Do About It, 67 U. COLO.
L. REV. 1037, 1041 (1996); see also Myron W. Orfield, Jr., Deterrence, Perjury, and the Heater
Factor: An Exclusionary Rule in the Chicago Criminal Courts, 63 U. COLO. L. REV. 75, 107 (1992).
193. See Myron W. Orfield, Jr., The Exclusionary Rule and Deterrence: An Empirical Study of
Chicago Narcotics Officers, 54 U. CHI. L. REV. 1016, 1050 (1987).
194. 384 U.S. 436 (1966).

2013]

POLICE MISCONDUCT

1177

substantive perjury in that when police commit procedural perjury, they lie
to circumvent procedural rules that otherwise would prevent them from
prosecuting apparently guilty suspects. Procedural perjury is a form of
whitewashing that is intended to facilitate what most police officers likely
perceive to be the most essential aspect of their jobs: to punish those who
are believed to be committing or to have committed crimes. Substantive
perjury, in contrast, occurs when police lie to incriminate innocent
persons. From the perspective of the criminal justice system’s
guilt/innocence sorting mechanism, substantive perjury is a far more
destructive practice than procedural perjury, although both forms of
perjury undermine the integrity of the criminal justice system and diminish
the credibility and the legitimacy of the police.
A New York state commission headed by Judge Milton Mollen issued
a report in 1994 documenting the “commonplace” types of procedural
perjury routinely committed by New York police officers in their day-today duties, which included lying about observing unlawful conduct or
incriminating facts to justify a search and seizure, lying about where
contraband was found to cover-up plainly unconstitutional conduct, and
lying about compliance with various rules of constitutional criminal
procedure.195 The commission found that perjury was a particular problem
in drugs and weapons cases, a finding that is consistent with the pattern of
police misconduct evident in the Rampart scandal.196 Indeed, all of the
types of “testilying” identified by the Mollen Commission are on vivid
display in the Rampart cases.

195.
For example, when officers unlawfully stop and search a vehicle because they believe it
contains drugs or guns, officers will falsely claim in police reports and under oath that the car
ran a red light (or committed some other traffic violation) and that they subsequently saw
contraband in the car in plain view. To conceal an unlawful search of an individual who
officers believe is carrying drugs or a gun, they will falsely assert that they saw a bulge in a
person’s pocket or saw drugs and money changing hands. To justify unlawfully entering an
apartment where officers believe narcotics or cash can be found, they pretend to have
information from an unidentified civilian informant or claim they saw the drugs in plain view
after responding to the premises on a radio run. To arrest people they suspect are guilty of
dealing drugs, they falsely assert that the defendants had drugs in their possession when, in
fact, the drugs were found elsewhere where the officers had no lawful right to be.
See Capers, supra note 191, at 869 (quoting COMM’N TO INVESTIGATE ALLEGATIONS OF POLICE
CORRUPTION AND THE ANTI-CORRUPTION PROCEDURES OF THE POLICE DEP’T, CITY OF NEW YORK,
COMM’N REPORT 36 (1994) (Milton Mollen, Chair)).
196. See Capers, supra note 191.

1178

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

1. Lies about Consent
One common form of procedural perjury on display in the Rampart
cases is a false claim that a suspect “consented” to a search. Consent is
simple to manufacture. Police need only claim that a suspect orally
consented to a search to establish the existence of consent, although police
in some Rampart cases went further and either forged a suspect’s signature
on a written consent form or tricked or coerced a suspect into signing such
a form.
In several cases, police falsely claimed to have obtained consent to
justify a forcible warrantless entry into a home. For example, police stated
that they had received information from unnamed sources that a woman
named Laura Villatora was storing cocaine and marijuana in her
apartment. According to the arrest report, police sought and Villatora
consented to a search of her apartment that turned up approximately seven
pounds of marijuana.197 Rampart investigators later concluded, however,
that Villatora, who was home with her daughter when the police arrived,
did not consent to a search. When it appeared that police efforts to force
open the door would cause serious damage to it, Villatora’s teenage
daughter, Laura Reyes, opened the door. Officers then “grabbed Reyes by
the hair and arm and dragged her to the living room” where they
“demanded to know where the drugs and money were located.”198 They
then began searching the apartment, eventually finding both drugs and
money.199 One of the officers later testified to the false version of events at
the preliminary hearing.200 As a result, Villatora pleaded guilty to a charge
of unlawfully possessing marijuana with intent to sell and was sentenced
to two years in prison.201
Similarly, in several cases police falsely claimed to have obtained
consent to search a vehicle. For example, in the course of accosting
Villatora and her daughter, police officers also learned that Villatora’s
marijuana supplier was a person named Porfirio Acosta. Police then
induced Villatora to arrange for Acosta to deliver drugs to her home.
When Acosta arrived, police detained Acosta and searched his car without

197. See Pet. for Writ of Habeas Corpus at 5, In re Laura Villatora, No. BA153152 (Cal. Super.
Ct. Jan. 31, 2000).
198. Id. at 10.
199. Pet. for Writ of Habeas Corpus at 13, Decl. of Michael Gannon, In re Porfirio Acosta, No.
BA153198 (Cal. Super. Ct. June 7, 2000) (detailing interviews with Laura Villatora and her daughter,
Laura Reyes).
200. In re Laura Villatora, supra note 200, Ex. B (Decl. of Brian Tyndall).
201. Id.

2013]

POLICE MISCONDUCT

1179

obtaining consent. Officers also searched his home. The officers then
falsely stated in the arrest report that Acosta had consented to the searches.
They even manufactured a false consent to search form.202 Acosta was also
charged with possession of marijuana for purposes of sale. He pleaded
guilty and was sentenced to three years’ probation and 120 days in county
jail.203
Another Rampart case involved Charles Harris, who was arrested after
police allegedly recovered 203 grams of rock cocaine and a handgun in a
supposedly consensual search of his vehicle.204 Investigators later
concluded that, in fact, Harris never consented to the search.205 Likewise,
police claimed to have obtained consent to search Juan Rojo’s car when,
according to witnesses the D.A.’s office concluded to be credible, “Rojo
was taken out of his car at gunpoint.”206 Officers testified falsely about
these events at the preliminary hearing and suppression hearing, at which
one officer testified that he merely “asked Rojo to step out of his vehicle
so that he could speak with him.”207 The officers also falsely claimed that
during this encounter Rojo consented to a search of his residence. Gricelda
Orellana was in the residence when police turned up seeking to search it.
According to Orellana, she “tried to lock the door” to prevent the police
from entering.208 Notwithstanding those efforts, the officers entered and
found cocaine in her bedroom. Orellana said that the cocaine belonged to
“some guy,” but told police officers that “Rojo was innocent.”209
Nonetheless, Orellana and Rojo both eventually pleaded guilty to one
count of possession for sale of cocaine and each served two years in
prison.210
2. Lies about Probable Cause
Perhaps the most common sort of lies told by the police are those used
to establish probable cause for searches and seizures that otherwise are

202. In re Porfirio Acosta, supra note 202, at 6–7. It appears that the officers not only lied about
obtaining consent from Acosta, but also either forged his signature on a consent to search form or
coerced or tricked him into signing it. Id.
203. Id. at 2.
204. See Pet. for Writ of Habeas Corpus at 5, In re Charles Edward Harris, No. BA157278 (Cal.
Super. Ct. Mar. 6, 2000).
205. Id.
206. Pet. for Writ of Habeas Corpus at 6, In re Juan Cerna Rojo and Gricelda Orellana, No.
BA156027 (Cal. Super. Ct. May 22, 2000).
207. Id.
208. Id. (Tyndall Decl.).
209. Id.
210. Id. at 2.

1180

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

constitutionally unjustifiable.211 Numerous Rampart cases involved such
misconduct.
So-called “dropsy” cases are one well-known form of “testilying.”212 In
a dropsy case, police claim that suspects in possession of drugs or guns
“drop” the contraband before any Fourth Amendment seizure takes
place.213 Since the contraband has been “abandoned” and is in “plain
view,” the evidence is admissible. That police often resort to this type of
perjury has been apparent for decades. Researchers observed a surge in
dropsy cases shortly after Mapp v. Ohio was decided,214 in which the
Supreme Court extended the exclusionary rule to the states.215
Dropsy cases were well-represented in the Rampart scandal. In one
case, officers discovered a bag of cocaine after they had seized two
suspects and then conducted a warrantless forty-five minute search of an
apartment building. The two men were arrested and charged as a result. In
their arrest report, the officers falsely claimed to have seen one of the men
hold and then drop the bag of cocaine.216 In another case, a police officer
saw a suspect suspiciously stuffing an object under his car seat. After
searching the vehicle, the officer discovered cocaine. He then stated in the
arrest report, and testified at the probation revocation hearing, that he had
observed the suspect drop the cocaine on the ground.217 This was later
determined to have been false.218 In a third case, police had information
that the narcotics were located in one of the rooms in a hotel.219 After
entering the room without a search warrant or consent, police found
cocaine and the two defendants inside. Instead of these facts, the arrest
report stated that police encountered the defendants in the hallway and saw
them drop canisters of rock cocaine to the ground.220 In a fourth case,
police seized a suspect and then, apparently after searching him and

211. See Jennifer Hunt & Peter K. Manning, The Social Context of Police Lying, 14 SYMBOLIC
INTERACTION 51, 56 (1991).
212. See Capers, supra note 191, at 868.
213. Id.
214. Id.
215. See, e.g., Sarah Barlow, Patterns of Arrests for Misdemeanor Narcotics Possession:
Manhattan Police Practices 1960–62, 4 CRIM. L. BULL. 549, 549–50 (1968).
216. See Pet. for Writ of Habeas Corpus at 10, Decl. of Olivia Rosales, In re Aristide Vanegas and
Rodolfo Arevalo, No. BA146324 (Cal. Super. Ct. Mar. 28, 2000). One of the men admitted that they
were delivering drugs to an apartment in the building at the time. Id. at 12 (Decl. of Brian Tyndall).
217. Resp. to Pet. for Writ of Habeas Corpus at 3, In re Oscar Lafarga, No. GA024373 (Cal.
Super. Ct. Nov. 19, 1999).
218. Id.
219. Pet. for Writ of Habeas Corpus, Decl. of Natasha S. Cooper, In re Carlos Martinez Pena and
Manuel Espinoza Ferrera, No. BA145491 (Cal. Super. Ct. Apr. 28, 2000).
220. Id.

2013]

POLICE MISCONDUCT

1181

finding no contraband, searched the area where the seizure occurred. The
officers located a firearm. Instead of reporting these facts, the officers
falsely claimed to have actually seen the suspect discarding the weapon.221
The suspect, Michael Williams, was charged with possession of a firearm
by a felon, contested the charge at trial, and was convicted. Williams was
sentenced to serve twenty-five years to life.222
Arguably more egregious than dropsy cases are cases where police
falsely claim to have seen the suspect actually engaged in criminal
conduct. For instance, Rampart officers detained and searched two
individuals without probable cause.223 After finding cocaine, the officers
stated in the arrest report that they had seen the suspect “engaging in the
sale of narcotics prior to her arrest” in order “to establish the necessary
probable cause for the detention and search.”224 Reports involving false
claims of direct observation of criminal conduct represent a potentially
more serious form of procedural perjury because the false statement not
only insulates the recovery of the contraband from suppression but
provides direct affirmative (albeit false) evidence of the suspect’s guilt,
increasing the chances that an innocent person will be convicted. This risk
is apparent in the case of Edward Villanueva, who was convicted of
possession of a firearm by a felon.225 According to the arrest report,
Officer Perez was manning an observation point when he personally
observed Villanueva with the firearms. Those statements turned out to be
false. Officer Perez later admitted that he stopped and searched Villanueva
based only on a report from a surveillance helicopter team who claimed to
have observed Villanueva with the guns.226 Officer Perez decided to report
the facts differently out of an apparent concern that the true facts left some
room to doubt whether probable cause existed for the search, or indeed,
whether Villanueva ever actually possessed the guns. Some cases combine
dropsy testimony and false claims of observed criminal activity. This
occurred in the prosecution of Octavio Fernandez.227 Officers searched
Fernandez and discovered drugs. To justify the search, police falsely

221. Pet. for Writ of Habeas Corpus, Ex. A, Decl. of Richard A. Rosenthal, In re Michael
Williams, No. BA128788 (Cal. Super. Ct. Jan. 31, 2000).
222. Id. at 2.
223. Id.
224. Pet. for Writ of Habeas Corpus, Ex. A at 1–2, Decl. of Richard A. Rosenthal, In re Sonia
Castro, No. BA128771 (Cal. Super. Ct. Jan. 31, 2000).
225. Pet. for Writ of Habeas Corpus at 2, In re Edward Yumol Villanueva, No. BA135887 (Cal.
Super. Ct. Apr. 28, 2000).
226. Id. Decl. of Natasha S. Cooper at 8.
227. Pet. for Writ of Habea[s] Corpus at 5, In re Octavio Fernandez, No. BA136807 (Cal. Super.
Ct. Apr. 10, 2000).

1182

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

claimed both that they witnessed Fernandez selling drugs and that he
dropped them prior to being seized.228
Another common type of procedural perjury involves misstatements
regarding the location in which contraband was found. In one case,
officers searched a suspect’s car and found a pouch of heroin-filled
balloons above the car’s rearview mirror.229 Worried that they lacked
probable cause for the automobile search, the officers falsely reported that
the balloons had been found in one of the suspect’s socks.230 In another
case, officers discovered marijuana during a search of a suspect’s
residence and reported instead that it was found either on his person or in
his car.231 In a third case, officers falsely stated they recovered a gun near
the front door of the defendant’s apartment and ammunition in his pocket.
These facts formed the basis for charging the defendant as a felon in
possession of a gun, for which he served four years in state prison.
Credible evidence later revealed that police had found the gun inside the
defendant’s apartment, under a bed, in a search of doubtful constitutional
validity. The gun, moreover, likely belonged to someone else.232
3. Lies about Miranda Compliance
While police engage in procedural perjury most frequently to avoid
Fourth Amendment suppression concerns, police also commit perjury to
evade other constitutional rules. This is especially true with respect to
compliance with the Miranda rules. Again, the willingness of police to lie
about their compliance with Miranda in order to ensure that incriminating
admissions or confessions made by suspects under interrogation are
admissible has been noted by other scholars. While conducting
observational studies of police interrogations, Professor Richard Uviller

228. Id.
229. Pet. for Writ of Habeas Corpus at 4, In re Blanca Sahagun and Carlos Carranza, No.
BA147177 (Cal. Super. Ct. Jan. 24, 2000).
230. Id. at Ex. A, Decl. of Laura Laesecke. One or both suspects may have been innocent.
Carranza’s account of the incident confirmed what Perez testified to, which is that Carranza never
possessed heroin. However, Carranza admitted that “he helped Sahagun arrange the sale of heroin on
the day of the incident,” which would seem to make him an accomplice. Sahagun, however, denied the
charges, alleging that “the entire arrest report was fabricated because she refused to be an informant
for Perez.” Sahagun claimed “that an officer at Rampart station displayed a clear baggie containing
brightly colored objects and told Carranza in her presence that he would ‘put this in [Carranza’s]
shoes.’” Id. at Ex. B (Decl. of Brian Tyndall (alterations in original)).
231. See Answer to Pet. for Writ of Habeas Corpus, Decl. of Vesna Maras, In re Gene Serrano,
No. BA150628 (Cal. Super. Ct., June 9, 2000).
232. See Pet. for Writ of Habeas Corpus, Decl. of Natasha S. Cooper, In re Salvador Luis Arias,
No. BA090914 (Cal. Super. Ct. May 12, 2000).

2013]

POLICE MISCONDUCT

1183

noticed that officers often “advance slightly the moment at which the
Miranda warnings were recited to satisfy the courts’ insistence that they
precede the very first question in a course of interrogation.”233 This type of
shading of the truth occurred while police were under the known
observation of an outsider. When police have no reason to believe they are
being observed, even more egregious deceptions are sometimes attempted.
For instance, many of the Rampart cases included false statements by
police that they had complied with Miranda. In the Charles Harris case
mentioned above, not only did police falsely claim that Harris consented
to a car search, they also falsely claimed to have advised him of his
Miranda rights.234 These false claims of Miranda compliance are often
only the necessary precedents to further false claims about incriminating
statements falsely attributed to the suspect. This too was true in the Harris
case, where not only did police lie about Mirandizing Harris, they also
attributed incriminating admissions to Harris that he never made,235
casting his actual guilt into doubt. The same pattern appears in the arrest
and prosecution of Carlos Romero. In that case, police conducted a
warrantless and nonconsensual search of a residence. During the search,
police threatened to arrest Romero’s sister, who in response “identified her
brother as a narcotics dealer and directed them to a stash of cocaine.”236
The officers then arrested Romero. They sanitized the search and seizure
by falsely claiming to have been directed to the stash of drugs by Romero
himself after they had advised him of his Miranda rights.237
C. The Blurred Line Between Procedural and Substantive Perjury, and
Other Forms Of Police Corruption
In some cases, it is impossible to determine whether perjury was
committed merely to secure a shortcut to conviction of a guilty suspect or
instead to convict an innocent man. Such was the case with Julian Lopez
Hernandez. According to the arrest report, Hernandez was arrested after
police found eleven balloons of heroin during a consensual search of his
233. H. RICHARD UVILLER, TEMPERED ZEAL: A COLUMBIA LAW PROFESSOR’S YEAR ON THE
STREETS WITH THE NEW YORK CITY POLICE 116 (Contemporary Books 1988) (cited and discussed in
Slobogin, supra note 192, at 1043 (speculating that “lying about events in the interrogation room may
be routine”)).
234. See In re Charles Edward Harris, supra note 205, Decl. of Laura Laesecke at 1.
235. See id.
236. Pet. for Writ of Habeas Corpus, Ex. A at 1–2, Decl. of Richard A. Rosenthal, In re Carlos
Romero, No. BA154270 (Cal. Super. Ct. Nov. 10, 1999).
237. Id.

1184

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

apartment. After his arrest, police claimed that they Mirandized
Hernandez, questioned him, and obtained incriminating statements.
Hernandez pleaded guilty and was permitted to take advantage of
diversion.238 During the subsequent Rampart investigation, Officer Perez
admitted under oath that Hernandez did not consent to the search and that
the officers never read Hernandez his Miranda rights prior to questioning
him.239 Perez thus admitted to serious procedural flaws in the search and
arrest of Hernandez, but appears to have maintained a belief in
Hernandez’s substantive guilt. Hernandez, however, asserted an entirely
different story. “[H]e claimed that the officers escorted him to an
unfamiliar apartment building and used a set of keys owned by a man
known as ‘Gerardo’ to open the apartment door. Inside, they found
numerous colored balloons on a windowsill in the living room.”
Hernandez “denied living at the location or knowing anything about the
narcotics,” but stated that he “pled guilty to avoid going to prison.”240
Either way, Hernandez was wrongfully convicted, but whether he was
actually innocent is impossible to determine on the scant available record.
Procedural perjury also goes hand-in-glove with other forms of police
corruption. William Zepeda and Argelia Diaz were convicted of
possession of cocaine for purposes of sale.241 Both served two year
sentences after agreeing to plea bargains.242 In making the arrest, the
officers lied about seeing Zepeda and Diaz selling drugs, and falsely
claimed to have obtained consent to search their apartment. While there,
the officers “stole a large sum of money from their apartment.”243 They
included these false facts in the arrest report, and then repeated the lies at
the preliminary hearing. All of the misconduct occurring in the
prosecution of Charles Harris was accompanied, similarly, by the
unreported appropriation of $6,000 from Harris’s residence and the theft
of at least $500 by the officers.244 In the Romero case, Officer Durden

238. See Pet. for Writ of Habeas Corpus at 2, In re Julian Lopez Hernandez, No. BA160279 (Cal.
Super. Ct. Jan. 31, 2000).
239. Id., Decl. of Laura Laesecke.
240. Id., Decl. of Brian Tyndall.
241. Pet. For Writ of Habeas Corpus at 2, In re William Zepeda and Argelia Diaz, No. BA156980
(Cal. Super. Ct., Mar. 17, 2000).
242. Id.
243. Id., Decl. of Olivia Rosales.
244. In re Charles Harris, supra note 193, at Decl. of Laura Laesecke. According to Rafael Perez,
the $6000 along with three guns seized from Harris’s house were given to Harris’s sister in exchange
for information about drug dealers. Id.

2013]

POLICE MISCONDUCT

1185

reportedly stole several pieces of jewelry from a suspect’s residence, as
well as $1,500 in cash.245
All of the false convictions in the Tulia cases, similarly, may have been
a collateral consequence of the efforts of a corrupt undercover officer to
steal drug “buy” money from the Drug Task Force.246 At least some people
believe that Officer Coleman pocketed the buy money, lied about buying
powder cocaine from the Tulia defendants, and then evidenced his
purported buys by turning in small amounts of white powder that he mixed
himself, each containing only enough cocaine to trigger a positive reading
on a lab test.247 In these cases, police lies about compliance with
constitutional rules were concomitant with, or in service to, other acts of
corruption.
VII. CONCLUSION
Police misconduct, when it occurs, is a major source of wrongful
convictions. The profile of those most at risk of such wrongful convictions
likely differs in some respects from that of other wrongfully convicted
persons. The offenses are generally less serious, and the sentences less
severe, than those involved in the DNA exoneration cases. These cases
involve drugs and guns, assaults on police officers, charges of disturbing
the peace, resisting arrest, or other allegedly violent or aggressive conduct
directed at the police. Hundreds of thousands, perhaps millions, of people
have been convicted of such crimes. There is simply no way to know how
many persons convicted of such offenses were actually innocent, but both
Rampart and Tulia provide stark evidence that police misconduct can, and
does, result in wrongful convictions.
Comparison of the mass exoneration data with prior exoneration
studies suggests that two important adjustments to the empirical picture of
wrongful convictions may be in order. Although earlier studies of
wrongful convictions found only a small number of cases involving guilty
pleas, in the mass exoneration cases, guilty pleas provided the main
procedural vehicle to criminal conviction. In more than 80% of the
combined Rampart and Tulia cases, innocent defendants pleaded guilty.
While innocence did seem to provide a marginal incentive to some
defendants to reject guilty pleas, actually innocent Rampart exonerees held

245. See Romero Pet., supra note 236, Ex. A at 2 (Decl. of Richard A. Rosenthal).
246. See BLAKESLEE, supra note 30, at 88.
247. Id. at 88–89.

1186

WASHINGTON UNIVERSITY LAW REVIEW

[VOL. 90:1133

out for trials only slightly more frequently than their guilty counterparts.248
The Rampart and Tulia exoneration data thus provides strong reason to
suspect that guilty pleas are not insulated from the risk of wrongful
convictions.
Consideration of this data should also raise the profile of perjury
among the causes of wrongful conviction. Although eyewitness
misidentification has received a substantial amount of attention as one of
the main identified contributing factors in wrongful convictions, the mass
exoneration cases make clear that the “causes” of wrongful convictions
vary significantly by crime. These exonerations show that police
misconduct is a potentially significant cause of wrongful convictions in its
own right. Procedural reforms that reduce the incidence of police
misconduct, therefore, should be high on the list of priorities among those
working to reduce wrongful convictions.

248. See infra Part V.D.

2013]

POLICE MISCONDUCT

APPENDIX
“Actually Innocent” Rampart Exonerees
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.

Alfaro, George Kenneth
Bailey, Samuel Joseph
Barrios, Diego
Booker, Esaw
Candido, Roberto
Carrillo, Delbert
Chavez, Emmanuel
Davalos, Octavio Gonzalez
Escobar, Edgar
Estrada, Leonel Ramos (aka Gregorio Ramos Lopez)
Flores, Luis Manuel
Gomez, Alfredo
Guardado, Manuel
Guevara, Carlos
Harris, Clinton
Hernandez, Miguel
Lara, Jose Armando
Lobos, Allan Manrique
Madrid, Jose Hugo
Matlong, Rene Barela (aka Rene Mationg)
Montes, Roy
Munoz, Raul Alfredo
Natividad, Cesar aka Danny Banuelos
Newman, Russell
Oliver, Ivan
Ordonez, Felipe Enriquez
Ovando, Javier Francisco
Perez, Jose
Peters, Gerald
Rivas, Walter
Rodriguez, Raul
Rojas, Ruben
Tapia, Daniel
Thomas, James
Torrecillas, Juan
Umana, Alex
Wesley, Mohammed Wayman
Zambrano, Rafael

1187

1188

WASHINGTON UNIVERSITY LAW REVIEW

Tulia Exonerees
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.

Allen, Dennis Mitchell
Barrow, James Ray
Barrow, Landis
Barrow, Leroy
Barrow, Mandis Charles
Benard, Troy
Brookins, Freddie Wesley
Cooper, Marlyn Joyce
Ervin, Aremnu Jerrod
Fowler, Michael
Fry, Jason Paul
Fry, Vickie
Hall, Willie B.
Henderson, Cleveland Joe
Henry. Mandrell L.
Jackson, Christopher Eugene
Kelly, Denise
Kelly Sr., Eliga
Klein, Calvin Kent
Love, William Cash
Marshall, Joseph Corey
Mata, Laura Ann
McCray, Vincent Dwight
Moore, Joe Welton
Olivarez, Daniel G.
Powell, Kenneth Ray
Robinson, Benny Lee
Shelton, Finaye
Smith, Donald Wayne
Smith, Yolanda Yvonne
Strickland, Romona Lynn
Towery, Timothy Wayne
White, Kareem Abdul Jabbar
White, Kizzie Rashawn

[VOL. 90:1133

2013]

POLICE MISCONDUCT

35. Williams, Alberta Stell
36. Williams, Jason Jerome
37. Williams, Michelle

1189