C
R
I
M
I
N
INJU$TICE AL

INTRODUCTION

A COST ANALYSIS OF
WRONGFUL CONVICTIONS,
ERRORS, AND FAILED
PROSECUTIONS IN CALIFORNIA’S
CRIMINAL JUSTICE SYSTEM

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 1

INTRODUCTION

2 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

INTRODUCTION

CRIMINAL
INJU$TICE

A COST ANALYSIS OF
WRONGFUL CONVICTIONS,
ERRORS, AND FAILED
PROSECUTIONS IN CALIFORNIA’S
CRIMINAL JUSTICE SYSTEM

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 1

INTRODUCTION

Acknowledgements
This report was written by Rebecca Silbert, John Hollway, and Darya Larizadeh. The authors thank the
following research assistants and students who contributed to the report: Dan Georgatos, John Yap,
Liz Baroi, Scott Wallace, Rebecca Tublitz, Robert McCarthy, Jackie Aranda, Giselle Olmedo, Efan Wu,
Megan Sallomi, Jasmine Eucogco, Kevin McNellis, Candice Shikai, Travis Henry, Julie Furdella, Ray Kim,
Charlotte Kiaie, and Manny Thompson. Colleen Shirazi and Devin Jones in the Berkeley Law Information
Systems and Technology Department provided invaluable support in setting up and managing the
database. Valuable background information was provided by Maurice Possley and Sam Gross of the
National Registry of Exonerations at the University of Michigan Law School. Denise Foderaro, Barry
Krisberg, and Jeanne Woodford also provided much appreciated assistance.
The authors of this report are grateful to Steven Raphael, Ph.D. Professor of Public Policy at the Goldman
School of Public Policy, University of California Berkeley, for developing the methodology for costs of
trial, appeal, and counsel. Professor Raphael’s gracious contribution is deeply appreciated.
The Chief Justice Earl Warren Institute on Law and Social Policy is a research and policy center at
UC Berkeley School of Law. The Warren Institute is led by Faculty Director and Honorable William H.
Orrick, Jr. Distinguished Professor Christopher Edley, Jr. and by Executive Director Rebecca Silbert.
John Hollway is the Executive Director of the Quattrone Center for the Fair Administration of Justice at
the University of Pennsylvania Law School. The Center is a national research and policy hub created to
catalyze long term structural improvement to the US criminal justice system.
Copies of this report will be available electronically on The Opportunity Institute’s website,
http://theopportunityinstitute.org and on the Quattrone Center’s website,
https://www.law.upenn.edu/institutes/quattronecenter
©2015

2 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

INTRODUCTION

Table of Contents
4	INTRODUCTION	
6	

EXECUTIVE SUMMARY

8	

OVERVIEW, DEFINITIONS, AND METHODOLOGY

13	 Section 1: THE CHARACTERISTICS OF INJUSTICE
	

14	 Disproportionately More Errors Found in Convictions for Violent Crime as Compared to Property and Drug Crimes	

	16	
Taxpayers Paid for 2,186 Needless Years in Custody
	

17 	 Most of the Individuals Subjected to these Flawed Prosecutions Did Not Plea Bargain	

	

20 	 The Original Sentences Imposed in these Flawed Cases Were Disproportionately Severe

	

22 	 These Faulty Convictions Were Not Resolved Quickly

	

23	 Flawed Homicide Convictions Were the Most Expensive	

25	 Section 2: THE CAUSES OF INJUSTICE	
	

29	 Prosecutorial Misconduct	

	

36 	 Judicial Mistake During Trial	

	

39	 Problems with Eyewitness Testimony	

	

42	 Ineffective Assistance of Counsel (IAC)	

	

45	 Failure of Prosecutorial Discretion	

	

47	 Unreliable or Untruthful Official Testimony	

	

49	 Fourth Amendment Violations	

	

52	 Inadequate Police Practices Before Trial	

59	 Section 3: THE COSTS OF INJUSTICE	
	

62	 Incarceration, Trials, and Appeals: $148 million

	

64	 California Victims Compensation and Government Claims Board (VCGCB): $5 million	

	

66	 Settlements: $68 million	

	

68	 Costs and Cases Not Captured in this Report	

75 	 Section 4: A NOTE ON GROUP EXONERATIONS	
79	 Section 5: CONCLUSION – NEXT STEPS AND RECOMMENDATIONS	
87	 Section 6: APPENDICES
	

88	 APPENDIX A – County Comparisons	

	

91	 APPENDIX B – Methodology for Cost of County Jail	

	

94 	 APPENDIX C – Methodology for Trial, Appeal, and Attorney Costs	

	

100	APPENDIX D – Lack of Access to County Court Records in California	

	

102	 APPENDIX E – Coding: All Errors Identified in Cases Reviewed for this Report	

	

104	 APPENDIX F – Total Cost By Type of Error, All Errors	

	

106	 APPENDIX G – Public Records Act Request for Settlements	

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 3

INTRODUCTION

Introduction
Criminal (In)justice examines 692 adult felony criminal cases where California missed the mark in
public safety by failing to prosecute the right person or by pursuing a flawed or unsustainable conviction.
The individuals subjected to these faulty proceedings endured hundreds of trials, mistrials, appeals, and
habeas petitions and served more than two thousand years in prison and jail, at a total cost to
California taxpayers of more than $282 million.1
The scope of Criminal (In)justice is, to our knowledge, unprecedented. We began by following the
well-trodden path analyzing wrongful convictions and exonerations. As we started to explore the
cases, however, we realized that criminal justice errors affect not only those who were convicted and
later declared innocent, but also those whose convictions were overturned without a declaration of
innocence. We could not ignore the many defendants subjected to years of detention, trials, retrials and
incarceration only to have their cases eventually dismissed without a formal finding of innocence, nor
could we ignore the victims left unserved when convictions against alleged perpetrators were dismissed
due to error. We were also working in California which, unlike many other states, imposes a particularly
burdensome process on individuals seeking a judicial declaration of innocence. As a result, innocent
individuals in California may rationally choose not to pursue judicial declarations of innocence after their
convictions are dismissed, especially if those individuals are indigent. For all of these reasons, we did not
have confidence that using “innocence” as a criterion for inclusion would provide a complete picture of
those who were not served by our criminal justice system.
In the end, we included every case we could find where
•	 the defendant was convicted of a felony,
•	 the conviction was reversed between 1989 and 2012,2 and
•	 the charges were subsequently dismissed or the defendant acquitted on retrial.
We evaluated cases regardless of the reason for reversal, using a methodology specifically designed to
avoid the subjective and often unknowable guilt or innocence of the defendant. Our cases came from
court records, the National Registry of Exonerations, and other public records. Regardless of source, we
relied on court records and we base our conclusions on those actual documents, not on extrapolations
and assumptions. As a result, we include over 160 wrongfully convicted individuals from the National
Registry, as well as others who may, or may not, be exonerated or “wrongfully convicted” as that term
has come to be commonly understood. What cannot be disputed is that each of them was convicted of
one or more felonies, that those convictions could not be sustained, and that the errors examined here
imposed a substantial cost on both the individuals subjected to them and all California taxpayers.

FOOTNOTES
1	

All dollar figures are adjusted to 2013 dollars.

2	

The year 1989 was chosen because it was the first year that a defendant was exonerated due to DNA evidence.

4 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

INTRODUCTION

An ideal criminal justice system should arrest only the individual who actually committed a crime,
deploy taxpayer resources responsibly, eliminate error, and impose consequences that are fair and just.
Each mistake in the criminal justice process can have serious and lifelong consequences on the affected
individual, and we should strive to eliminate all error. We nonetheless expect that some will note that
the system theoretically worked in these cases, because its established checks and balances ultimately
uncovered these wrongs and the convictions were eventually dismissed. We do not concede that this is
a successful result. A perfect system may never be realized, but it must always be pursued.
We also expect that some will point to the fact that California prosecutes and convicts over 200,000
people every year,3 and argue that the 692 faulty convictions described in this report reflects an
acceptable rate of error. We reject the proposition that an acceptable rate of error can apply to
proceedings that impact people’s lives in the way that criminal prosecution can. The cases described in
this report are actual events that impacted real people, with real families and real jobs, prosecuted in
real California courts. There is perhaps no greater interference government can make in a person’s life
than to deny that person his or her liberty, and for that reason alone turning a critical eye on our criminal
justice system is a valid and vital endeavor no matter what the rate of error. Just as with airline safety and
medical mistakes, the acceptable rate of error is zero and that should be the goal.
Moreover, there are undoubtedly more than 692 Californians who have suffered injustices in our state
and federal criminal justice system. California does not systematically collect information about reversed
convictions, and the cases in this report thus represent only those that could be found through various
public sources. We did not include juvenile cases, misdemeanor convictions, or cases in which an
individual was detained and prosecuted but never convicted. The 692 cases and $282 million in taxpayers
costs catalogued here are thus most certainly the tip of the proverbial iceberg.
Mistakes happen, and human error persists. But our criminal justice system serves not only those
who are facing criminal charges, but also anyone who may face the power of criminal prosecution in
the future and anyone who may become a victim of crime. This is all of us. Everyone involved, whether
legislators, prosecutors, defense attorneys, or judges must “strive to enhance and maintain confidence
in our legal system....to ensure the confidence of the public.”4 We ignore flaws within the system at
our own peril.
—June 2015

FOOTNOTES
3	

Office of the California Attorney General, “Crime in California 2013,”
http://oag.ca.gov/sites/all/files/agweb/pdfs/cjsc/publications/candd/cd13/cd13.pdf?.

4	

California Code of Judicial Ethics, Preamble (amended effective January 15, 2015),
http://www.courts.ca.gov/documents/ca_code_judicial_ethics.pdf.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 5

EXECUTIVE SUMMARY

Executive Summary
Criminal (In)justice examines 692 individuals who were prosecuted and convicted in California state
or federal courts, only to have their convictions dismissed because the government prosecuted the wrong
person, because the evidence was lacking, or because the police, defense, prosecutors, or court erred to
such a degree that the conviction could not be sustained. The 692 individuals subjected to these failed
prosecutions spent a total of 2,346 years in custody, and their prosecutions, appeals, incarceration, and
lawsuits cost California taxpayers an estimated $282 million when adjusted for inflation.5 Eighty-five
of these cases arose from a large group exoneration—the Rampart police corruption scandal—and are
discussed separately in a later section of this report.
The remaining 607 convictions, all of which were reversed between 1989 and 2012, illuminate a dark
corner in California’s criminal justice system. These 607 individuals spent a total of 2,186 years in
custody. They burdened the system with 483 jury trials, 26 mistrials, 16 hung juries, 168 plea
bargains, and over 700 appeals and habeas petitions. Many of the individuals subjected to these
flawed prosecutions filed lawsuits and received settlements as a result of the error, adding to the
taxpayer cost. Altogether, we estimate that these 607 faulty convictions cost taxpayers $221 million
for prosecution, incarceration and settlement, adjusted for inflation. This estimate is only a window
onto the landscape of possible costs, as it does not include the often unknowable costs suffered by those
subjected to these prosecutions.
The sections below provide a review of these 607 cases and offer some recommendations for change.
The first section, Characteristics of Injustice, paints a collective picture of the cases in our sample.
Compared to California’s average, the individuals subjected to these errors were disproportionately
prosecuted for violent crimes, especially homicide. This may be because prosecutions for violent crime
are more likely to generate error than prosecutions for other crimes, though that question was beyond
the scope of our research. Whatever the reason, failed prosecutions for violent crime account for a
greater percentage of the wasted $221 million than failed prosecutions for other crimes. Indeed, flawed
homicide convictions alone account for 52% of the $221 million, in part because these homicide cases
took an average of 11 years to resolve and generated more lawsuits and civil settlements.
The second section, Causes of Injustice, catalogs the multitude of errors, dividing them into eight
categories: eyewitness identification errors, prosecutorial misconduct, ineffective defense counsel,
judicial mistake during trial, Fourth Amendment search and seizure violations, inadequate police
practices before trial, unreliable or untruthful official testimony (officer or informant), and failure of
prosecutorial discretion.

FOOTNOTES
5	

All costs in this Report are adjusted for inflation to 2013.

6 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

EXECUTIVE SUMMARY

Prosecutorial misconduct and eyewitness identification were the most common errors in the flawed
homicide prosecutions. When broken down by type of error, prosecutorial misconduct accounted for
more of the cost in our sample than any other type of error. By contrast, the most common errors in
our sample were Fourth Amendment search and seizure errors, and judicial mistake. These errors were
resolved relatively quickly, however, and resulted in relatively little cost.
The third section, Costs of Injustice, walks through the cost analysis. It documents the many hurdles
raised by the California Victims Compensation and Government Claims Board. This section also
identifies many of the additional costs not captured by our methodology, including costs arising from
wrongful misdemeanor convictions, flawed juvenile convictions, and cases that resolved prior to
conviction, among others. These unaccounted costs highlight the fact that this report documents only
a portion of the vast unknown waste in California’s criminal justice system—but it is at least a beginning.
Criminal (In)justice ends with a section on Next Steps and Recommendations. The problems
presented in this report are undoubtedly complex, and each of them individually could be subject to its
own investigation. This report does not attempt to comprehensively define the universe of best practices
that will solve all of the issues raised. Instead, it identifies promising avenues for reform and highlights
practices and jurisdictions that are leading the way. In particular, in 2006 the California Commission
on the Fair Administration of Justice issued a report containing detailed recommendations regarding
eyewitness identification, false confessions, informant testimony, problems with scientific evidence, and
accountability for prosecutors and defense attorneys. The recommendations represent the unanimous
views of a diverse group of prosecutors, defense attorneys, judges, law enforcement, and other stakeholders.
To date, however, many of the substantive reforms have not been adopted, compromising public safety
and leaving our criminal justice system at risk of endlessly repeating the errors catalogued here.
(The report can be found at www.ccfaj.org.)

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 7

OVERVIEW, DEFINITIONS, AND METHODOLOGY

Overview, Definitions, and Methodology
Criminal (In)justice examines 692 flawed convictions, 607 of which are discussed in the body of this
report.7 Cases were included if:
•	 The defendant was convicted of a felony,
•	 The conviction was reversed in full, and
•	 The case was subsequently dismissed, or the defendant was acquitted on retrial.8
The term “error” as used in this report means any mistake, misconduct, or illegal or incompetent act that
caused the reversal of the conviction, whether such act was
•	 factual (e.g., the individual did not commit the crime; the evidence was inaccurate or untruthful),
•	 procedural (e.g., the police, prosecution, defense or court committed error ranging from
inadvertent mistake to intentional misconduct), or
•	 legal (e.g., the defendant’s rights were violated; the defendant’s acts did not violate the law).
We did not include any case for which a court record could not be located confirming both the reversal
of the conviction and its ultimate dismissal or acquittal. We have captured only facts and legal findings
reflected in official records. No errors, facts, or theories of law have been inferred or presupposed.
Each error catalogued in this report has been verified by the court of record, and our coding has been
limited to causes of action specifically found by the court on the record of the case. If a defendant
made a particular legal argument, it was coded and captured. If the defendant did not make a particular
argument, even if we believed that it might have been legally available (based, for example, on newspaper
articles), it was not included. If the court rejected the defendant’s argument or did not reach a decision
on a particular legal allegation, it was not included.
Notably, cases were not included or excluded on grounds of innocence, and our analysis was not
confined to those defendants who had been declared innocent by the court. Under our country’s
Constitution, a verdict of “not guilty” or the dismissal of a conviction means only that the government
could not prove that the person committed the crime. That person may have been innocent, or not.9
In addition, unlike some other states, California does not have an automatic or simple process by which

FOOTNOTES
7	

The remaining 85 cases arose from a group exoneration and are discussed separately in a later section of the report.

8	

We also include a small number of cases where the defendant’s most serious conviction or convictions were reversed, but the defendant remained
convicted of a less serious crime. These cases were included only when we could irrefutably establish that the individual was imprisoned for at least a
year longer than he would have been, if he had been convicted of only the lesser crime from the beginning.

9	

“Under the American system of justice, the high degree of certainty required in a criminal case, proof beyond a reasonable doubt, can sometimes result
in the guilty going free. While this allows the presumption of innocence to prevail, it does not mean that the trier of fact has concluded the defendant is
factually innocent. It does mean that guilt has not been proved by the standard required...Thus, under our system of justice, the guilty can go free in order
to ensure as much as possible that the innocent are not convicted. This is an accepted consequence of a system of justice that places a high price on
freedom—a system of justice for which this country has fought on numerous fronts and offered up innumerable lives to preserve and defend.”
In re Sodersten, 146 Cal. App. 4th 1163, 1170 (2007).

8 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

OVERVIEW, DEFINITIONS, AND METHODOLOGY

an individual can seek a declaration of innocence. The best option is a Petition for Finding of Factual
Innocence filed under California Penal Code Section 851.8, but these petitions are cumbersome and rare,
rendering these petitions a poor arbiter of innocence.10 Moreover, but for those few individuals who
endure the arduous process of seeking a judicial declaration of innocence, an individual’s innocence
may continue to be the subject of police, prosecution, or media speculation even after the defendant has
returned home and the conviction has been fully dismissed.
In 1992, 15-year old Francisco Carrillo11 was convicted of murder in Los Angeles County; the conviction
was based on eyewitness testimony of six eyewitnesses, all of whom knew each other. After the
conviction, a witness was located who said that Mr. Carrillo was not at the scene, and all of the
eyewitnesses subsequently recanted their testimony. One eyewitness later testified that the police had
pointed Mr. Carrillo out during the identification process, and another eyewitness admitted that he could
not have seen the shooter from where he had been standing. In 2011, after twenty years in custody,
Mr. Carrillo was released. On the issue of innocence, Mr. Carrillo stated, “There are some people I’m sure
I will never convince of my innocence, but I’m OK with it…There’s something that kind of takes you over
when you know it wasn’t you.”12

Our own sample demonstrates the limitations inherent in using innocence as a criterion for inclusion.
In our sample of 607 people, 131 (22%) argued to the court at some point in their proceedings that they
were factually innocent but only 28 (5%) subsequently obtained a Penal Code Section 851.8 declaration
of innocence. We include five cases where the state awarded compensation as a result of the wrongful
conviction, but there was never an 851.813 declaration of innocence. And, we include four cases where
state compensation was denied despite the prior finding of innocence under 851.8. In our cases where
DNA helped clear the individual, only 64% (9 out of 14) had successful 851.8 petitions.

FOOTNOTES
10	 In an 851.8 petition, the individual must demonstrate that no reasonable cause exists to believe that he or she committed the alleged crime. The “no
reasonable cause” standard that must be met by the petitioner is much higher than the “beyond a reasonable doubt” standard used to assess a person’s
guilt at trial, and the burden of proof is on the individual seeking the declaration, not on the government. There is no right to court-appointed counsel for
this type of petition, so the individual seeking the declaration of innocence must pay the often considerable costs of an attorney to litigate on his or her
behalf. The prosecution can contest the petition, meaning that a prosecutor could concede that a conviction should be dismissed, then argue that the
defendant should not receive a declaration of innocence. The entire process can take years.
11	 In this report, we use a person’s full name if they are included in the National Registry of Exonerations or another public innocence website such as the
Northern California Innocence Project, or if the person’s situation has been subject to press coverage that continues to be publicly available. If we became
aware of the person solely from court records or there is no indication that the person agreed to publicize their story, we use first name and last initial.
12	 “Francisco Carrillo, quoted in Jack Leonard, “Man goes free after murder conviction is overturned, Los Angeles Times, March 17, 2011,
http://articles.latimes.com/2011/mar/17/local/la-me-carrillo-20110317.
13	 The law regulating the California’s compensation board was amended in 2013 to prohibit the board from denying compensation if another court has
declared the petitioner innocent.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 9

OVERVIEW, DEFINITIONS, AND METHODOLOGY

These cases demonstrate that judicial declarations under 851.8 are poor arbiters of innocence. For this
reason, and because without such a petition an individual’s claim of innocence could be subject to
dispute, we have not attempted subjective determinations about who was “innocent enough” to be
included in our analysis. Rather, we eliminated innocence entirely as a criterion for inclusion, focusing
instead on felony convictions that were reversed and ultimately dismissed. Of course, many individuals
in our sample were innocent, including those who appear in the National Registry of Exonerations.
We take our investigation further than the National Registry, however, and include both innocent people
who were tried, convicted, and imprisoned as well as those who may or may not have committed the
crime for which they were prosecuted. These other cases represent a different but equally valuable point
of discussion, since releasing a guilty person can be just as problematic as convicting an innocent one.
By discussing all of these cases in terms of systemic error and costs, rather than in terms of unknowable
guilt or innocence, we hope to inspire a more thorough and thoughtful review of potential improvements
to our administration of criminal justice.

METHODOLOGY
California has no systematic way to effectively and comprehensively locate failed prosecutions such as
those examined here, which makes a true audit impossible. The cases studied in this report represent
only a subset of the unknown number of actual cases, and we have surely missed a number of potential
cases and costs. We have nonetheless attempted to be as accurate as possible with as many cases as
possible, if only to provide a window into the issues that should be subjected to further review and study.
Case selection for this report began with the California appellate courts, which provided records
reflecting all appellate court reversals between 2002 and 2012. This review generated a list of over 10,000
cases that potentially fit our criteria.14 Each case was then reviewed to ensure that (a) the appeal was of
the conviction, not of the sentence or any other action in the case, and (b) the appeal was of the entire
conviction, not just some of the counts on which the defendant had been convicted. If either criterion
could not be met, the case was excluded from further review. These factors reduced our list to 980 cases.
To rely only on those appellate court records would have meant excluding hundreds of convictions
that were reversed by the appellate courts prior to 2002, or that were reversed in the California trial
courts, the California Supreme Court, or a federal court.15 Therefore, we used the National Registry

FOOTNOTES
14	 Appellate court reversals from 2002 to 2012 were obtained from two sources: case lists maintained by five out of the six appellate defense panels in
California, and case numbers provided by the Administrative Office of the Courts (AOC) in response to our request. The AOC records the case numbers for
petitions filed in the California appellate courts by criminal defendants, and records when the appellate petition is granted in full. We requested the list
of case numbers for successful appeals filed by the defendant and the AOC graciously provided the list.
15	 In certain circumstances a California state conviction can be reversed by a federal court. We counted these as California state cases, because the original
conviction came from a California court and the case was ultimately resolved in that same state court even if the reversal was ordered by a federal court.
In addition, about 3% of the cases in our sample originated in a California-based federal court. These cases were included only if the federal court was
sited in California.
10 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

OVERVIEW, DEFINITIONS, AND METHODOLOGY

of Exonerations,16 newspapers, attorney outreach, and other online sources to locate an additional
454 potential cases. Cases were included if the reversal stretched back to 1989, the year that the first
individual in the United States was exonerated through DNA testing. The resulting list of 1,434 cases was
our starting point.
Public records from each case were collected on-site and by hand in their respective county courthouses
to determine the ultimate result in the case, as many defendants are re-convicted after a reversal and
we were including cases only if the convictions were ultimately dismissed.17 If the ultimate result was
dismissal or acquittal, the records were scanned and coded into a database. In a few counties this quest
could be done online, but in most counties the courthouse had to be visited in person, the paper files had
to be pulled by the court clerk, and the file in its entirety had to be reviewed by hand. This cumbersome
process, which took hundreds of hours and miles of travel, varied from county to county (the difficulties
encountered in obtaining these files including lost files and missing records are more fully described
in Appendix D); the counties of Napa and Riverside should be commended for being more accessible,
responsive and organized than most.
Almost half (692) of the 1,434 potential cases were included in our sample because the ultimate result
was dismissal or acquittal on retrial, including the eighty-five group exoneration cases that are discussed
separately in this report. Ninety-seven case files could not be located or the file was so incomplete that
a resolution could not be determined. The remainder (645) were not appropriate for our sample because
the defendant ultimately pleaded guilty or was reconvicted in a later court proceeding. In the end, not
including the eighty-five group exoneration cases, 376 (62%) of the 607 cases profiled in the main body
of this report were gathered through our systematic review of the appellate records, and 231 (38%) came
from other sources. All have been verified through court records.
Our cost methodology is fully described in the “Costs of Injustice” section as well as in Appendices B
and C. We coded each error based on the legal ground for reversal as stated by the court; the complete
list of all errors for which we coded is in Appendices E and F. Some of the calculations in this report
break down cost by type of error. For those individuals with more than one ground for reversal (i.e., more
than one validated error), we divided the total cost by the number of grounds. So if a particular individual
accounted for $100,000 in our sample, for example, and that individual’s conviction was reversed
because the prosecutor failed to disclose exculpatory evidence (a Brady violation) and because the
judge erroneously admitted evidence that should have been excluded, we assigned $50,000 to the Brady
category and $50,000 to the erroneous admission of evidence category.

FOOTNOTES
16	 The National Registry of Exonerations, based at the University of Michigan Law School, was founded in 2012 in conjunction with the Center on Wrongful
Convictions at Northwestern University School of Law. The Registry provides detailed information about every known exoneration in the United States since
1989, defined as cases in which a person was wrongly convicted of a crime and later cleared of all the charges based on new evidence of innocence.
http://www.law.umich.edu/special/exoneration/Pages/about.aspx.
17	 We also include a small number of cases where the defendant’s most serious conviction or convictions were reversed, but the defendant remained
convicted of a less serious crime. These cases were included only when we could irrefutably establish that the individual was imprisoned for at least a
year longer than he would have been, if he had been convicted of only the lesser crime from the beginning.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 11

OVERVIEW, DEFINITIONS, AND METHODOLOGY

We recognize the limitations of our case collection methodology in particular, as there is no way to
determine whether the cases we examine are a representative sample from which generalization would
be appropriate. We are certain, however, that we have not found all the cases that would have fit our
criteria, and at the very least we are therefore undercounting both the numbers and the cost of errors
within the California criminal justice system. We are equally certain that the failed prosecutions profiled
in this report utilized significant taxpayer resources and severely affected the lives of hundreds of
individuals, and that we must diligently work towards reducing these errors in the future.

12 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

SECTION 1:

THE CHARACTERISTICS
OF INJUSTICE

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 13

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

This section provides a collective overview of all 607 cases, including the types of crimes for
which these individuals were prosecuted, how long they were in custody, and how long the cases took
to resolve. 18

Disproportionately More Errors Found in Convictions for
Violent Crime as Compared to Property and Drug Crimes
Although the faulty convictions highlighted in this report range from simple drug possession to murder,
they are notable in that they do not follow the pattern set in the rest of the state. Indeed, 30% of the
individuals in our sample were improperly convicted of violent crimes, including homicide, rape,
assault, and robbery, while only 18% of California’s felony convictions were for violent crime in 2013.19
The inverse was true for property crimes: 26% of California’s 2013 convictions were for property crimes,
including burglary, theft, and embezzlement,20 but only 12% of the faulty convictions in this report
were for property crimes. The proportion of drug crimes in our sample mirrors that of convictions in
California in 2013—31% in our sample and 30% in California (see Figure 1).
The faulty convictions in our sample were disproportionately convictions for violent crime.

FOOTNOTES
18	 This section does not include an analysis of the race of the individuals subjected to these failed prosecutions, because accurate information is
unavailable. Race is not generally recorded in public documents and, even if it were available in a court record, it is not always clear whether the notation
is one that has been confirmed by the defendant, or one that was inferred by court or police staff.
19	 Office of the California Attorney General, “Crime in California 2013,” Table 40, http://oag.ca.gov/sites/all/files/agweb/pdfs/cjsc/publications/candd/cd13/
cd13.pdf?; Office of the California Attorney General, “Crime in California 2011,” Table 40 (2006 through 2011), http://oag.ca.gov/sites/all/files/agweb/
pdfs/cjsc/publications/candd/cd11/cd11.pdf? Office of the Attorney General, “Crime in California 1996,” Table 41 (1991 through 1996) http://ag.ca.gov/
cjsc/publications/candd/cd96/cd96obts.pdf. In 2013, from a total of 213,390 crimes, violent crimes were 38,034; drug offenses were 64,249; and property
offenses were 55,342. The remaining crimes were categorized as “all other offenses.” The comparisons above are not exactly alike, as the convictions in
this report stretch over more than twenty years and the Attorney General’s Report covers convictions in one calendar year. The distribution of convictions,
however, has not changed substantially over time. Violent crime has fallen slightly over time (from 20% in 1991, to 24% in 1996, 17% in 2006, 19%
in 2011, and 18% in 2013). Property crimes have also dropped (from 35% in 1991, to 30% in 1996, 27% in 2006, 26% in 2011, and 25% in 2013).
Drug crimes have remained fairly consistent (27% in 1991, 30% in 1996, 28% in 2006, 28% in 2011, and 30% in 2013). The proportion of “other crimes”
has risen from 17% in 1991 to 27% in 2011. This report uses the Attorney General’s definitions of “violent”, “property”, and “drug” convictions, which in
turn are derived from the California Penal Code.
20	 In collecting data, California follows the federal FBI guidelines for Uniform Crime Report (the UCR). The UCR does not include fraud and embezzlement in
its definition of “property crime” therefore fraud and embezzlement cases are not included in the state percentages. In our sample we grouped fraud and
embezzlement with theft. However, most of the cases were theft cases.
14 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

Figure 1: All 2013 California Convictions v. Unjust Convictions in This Report
35%

2013 Convictions in California

30%

Convictions in this Report

25%
20%
15%
10%
5%
0%
Violent Crimes

Property Crimes

Drug Crimes

The precise reasons for this disparity are unknown. We speculate that the higher frequency of error
in prosecutions for violent crime could be a reflection of increased emotion at play in those cases, or
increased public pressure to achieve a conviction, or perhaps a result of the increased complexity often
seen in prosecutions for violent crime. This topic is deserving of further research, as failed prosecutions
for violent crime have the greatest potential impact on victims and defendants alike (see Figure 2 for
a breakdown of all reversed cases in this report, by type of crime).
Figure 2: All Reversed Convictions in Report, Type of Crime

31% Possession of Drugs

3% Failure to Register as a Sex Offender
4% Forgery / Fraud / Embezzlement
4% Rape

15% Homicide

N = 607

4% Robbery
7% Possession of Weapons

9% Other
8% Burglary / Theft / Stolen Property

7% Assault
8% Lewd Act on a Minor21

FOOTNOTES
21	 Lewd Act on a Minor is the name given to California Penal Code Section 288(a), defined as an act upon a person under 14, on skin or through clothes,
with the intent of sexual gratification for the perpetrator. It is a lesser crime than rape or other sexual assault crimes on minors, which are regulated by
other penal code sections.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 15

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

Taxpayers Paid for 2,186 Needless Years in Custody
The 607 individuals in our sample spent 797,794 days—2,186 years—in prison and jail, all of which was
paid for by Californians. None of this time and cost resulted in sustained convictions.22 The vast majority
these wasted years (78%) were in state prison, with the remainder in county jail and less than one
percent in federal facilities.
The individuals subjected to the flawed prosecutions in this report were collectively incarcerated for
2,186 years, which was paid for by California taxpayers.

The taxpayer cost counts only a portion of the damage done by the errors catalogued here, as it cannot
quantify the effect on those who were wrongfully or illegally incarcerated and their families. The
individuals in this report were often incarcerated in the prime of their lives, in the time when they could
have been getting an education, starting a career, and building a family.23 More than half were 35 or younger
at the time of their conviction, and 21% were between 15 and 24.24 Their prosecutions are typically a matter
of public record, leaving these future job seekers vulnerable to employer Internet searches that disclose
the prosecution but not the dismissal. Their incomes are likely to decline after release from custody.25
They lost their right to vote while incarcerated in prison.26 Their children are stigmatized and more
likely to suffer long-term emotional and behavioral challenges.27 Indeed, a parent’s incarceration alone
increases the risk that his or her children will live in poverty or suffer household instability.28

FOOTNOTES
22	 The database includes 46 cases in which the serious crimes were dismissed but one or more lesser crimes remained. These cases were included only
when it could be established that the individual was incarcerated for at least one full year longer than he or she would have been, had the erroneous
conviction for the more serious crime never occurred. A small percentage of the 797,794 days may have been attributed to the lesser or minor conviction
that was not reversed.
23	 “While he was in prison, Carrillo missed out on a lot. His son was born. His father died. He says it hurts him to think that his father did not live to see
him free. Carrillo lost 20 years—good years, the years when most people go to college and find love and build careers,” Mary Harris and Colleen Williams,
“Freed after 20 years wrongly imprisoned, Franky Carillo hits the books,” NBC News, March 16, 2012,
http://usnews.nbcnews.com/_news/2012/03/16/10722742-freed-after-20-years-wrongly-imprisoned-franky-carrillo-hits-the-books?lite.
24	 Date of birth and date of conviction is known for 81% of the individuals in the database.
25	 The National Longitudinal Youth Survey indicates that male wage earnings decrease up to 30% after incarceration. Men who were previously incarcerated
contributed on average $1,400 less to their families with small children. See Christopher Wildeman, and Bruce Western, “Incarceration in Fragile
Families,” The Future of Children 20.2 (2010): 157-177; see also Steven Raphael and Michael A. Stoll, Do Prisons Make Us Safer? The Benefits and
Costs of the Prison Boom, (New York: Russell Sage Foundation, 2009). Family income declines by an average of $8,726 several years following a father’s
release (1997 dollars); see also Bruce Western, Punishment and Inequality in America (New York: Russell Sage Foundation, 2006). Earnings drop after
incarceration (2004 dollars).
26	 People convicted of felony offenses do regain their right to vote in California once they have completed the terms of their sentence including parole.
http://www.sfgov2.org/index.aspx?page=880. Federally, in March 2015, Senators Cardin and Conyers introduced the Democracy Restoration Act which
seeks to restore the voting rights in federal elections to all Americans who have been released from prison, ensure that probationers never lose their right
to vote in federal elections, and notify people of their right to vote upon leaving prison. As of May 2015 the Democracy Restoration Act had not been signed
into law. https://www.aclu.org/sites/default/files/field_document/dra_-_fact_sheet_114th_congress_3_2015_v2.pdf.
27	 Nancy G. LaVigne, Elizabeth Davies, and Diana Brazzell, “Broken Bonds: Understanding and Addressing the Needs of Children with Incarcerated Parents,”
Urban Institute, 2008, http://www.urban.org/UploadedPDF/411616_incarcerated_parents.pdf.
28	 Susan D. Phillips et al., “Disentangling the Risks: Parent Criminal Justice Involvement and Children’s Exposure to Family Risks,” Criminology and Public
Policy 5 (2006): 677–702, http://devepi.duhs.duke.edu/library/pdf/20970.pdf.
16 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

Such long-lasting effects may not always be quantifiable,29 but they are nonetheless a rallying cry for reform.
More than half of the individuals convicted and imprisoned in these dismissed cases were 35 or
younger when they were convicted; 21% were younger than 25.

Most of the Individuals Subjected to these Flawed Prosecutions
Did Not Plea Bargain
As in most jurisdictions, California overwhelmingly resolves criminal prosecutions through pleabargaining: in 2013 only two percent (2%) of criminal cases in California were resolved through a trial.30
In our sample, by contrast, seventy-four percent (74%) of the individuals did not plead guilty and were
instead convicted after a trial. This may be a result of the fact that these cases were problematic to
begin with: many of these cases had limited or unreliable evidence against the defendant, and in that
situation a defendant might rationally choose trial instead of a guilty plea.31 It is also a result of the near
impossibility of discovering error in a case that has been resolved by guilty plea. Additionally, many of
the errors discussed in this report are errors that occurred during trial. Such errors, of course, do not
occur in cases that do not go to trial.
Not only did the majority of these cases go to trial, but many of them went through multiple mistrials,
hung juries, and retrials. These defendants’ multiple trips through the system wasted 5,244 confirmed
days of jury time.32 Put differently, they required the equivalent of a full courtroom, with attorneys,
judges, juries and court staff, for over twenty continuous years.
The reversed and dismissed cases in this report used 5,244 days of jury time, equating to the dedication
of a full courtroom with attorneys, judges, juries, and staff for over twenty continuous years.

FOOTNOTES
29	 Although wrongful incarceration may not always be quantifiable, in at least one case in Massachusetts the government was ordered to pay over $100
million in damages to six defendants and their families for the defendants’ wrongful conviction and incarceration. In making the order Judge Nancy
Gertner said, “Now is the time to say and say without equivocation: this ‘cost’ to the liberty of four men, to our system of justice—is not remotely
acceptable.” Nancy Gertner, quoted in Robert Barnes and Paul Lewis, “FBI Must Pay $102 Million in Mob Case,” The Washington Post, July 27, 2007.
http://www.washingtonpost.com/wp-dyn/content/article/2007/07/27/AR2007072700282.html. One of the defendants in the case commented, “All the
money in the world wouldn’t bring back my 33 years.” Peter Limone, quoted in Scott Malone, “Judge awards 100 mln for unjust convictions,” Reuters,
July 26, 2007, http://www.reuters.com/article/2007/07/26/us-usa-fbi-murder-idUSN2643274020070726.
30	 Judicial Council of California, “2014 Court Statistics Report: Statewide Caseload Trends 2003-2004 through 2012-2013”, 2014, Figure 35, 72,
http://www.courts.ca.gov/documents/2014-Court-Statistics-Report.pdf.
31	 The vast majority of the 26% of cases in our database that were resolved by guilty plea follow the same pattern: The defendant is charged with a felony.
He files a motion alleging either that the evidence against him was obtained in violation of the Fourth Amendment, or that the charges against him were
filed too late and therefore violate the statute of limitations. The trial court denies the defendant’s motion, so the defendant pleads guilty. The defendant
then files an appeal, alleging that the trial court should have granted the motion. The appellate court agrees with the defendant, and orders that the
evidence should have been suppressed or that the charges were brought too late. The government then dismisses the case.
32	 These 5244 trial days were utilized by the 447 cases in which there were one or more trials, mistrials or hung juries.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 17

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

On March 18, 1993, Juan Carlos C. was an 18-year-old high school student on a school bus in San Joaquin
County. There were 10 to 15 other students, some of whom were opening the emergency window on the
bus. The driver pulled the bus over and called her supervisor. The supervisor responded in his work van,
boarded the bus, and ordered the students off the bus. When one of the students mentioned that the
supervisor had left his keys in the van, Mr. C. got in the van and began driving. The supervisor ran after
the van and grabbed onto the door as it was moving. He leaned in the open window and struggled for
control of the van. The car veered left and was struck by an 18-wheeler truck. The supervisor was killed.
In 1994 Mr. C., who had no criminal history, was charged and convicted of felony murder, robbery and
vehicle theft. He filed an appeal and in 1995, the court reversed the murder and robbery conviction for
insufficient evidence. In 1996, the prosecutor brought Mr. C. to trial a second time. This time the charge
was second-degree murder. Mr. C. was found guilty and sentenced to 15 years to life. He appealed and
in 1999 the court reversed the conviction for failure to instruct the jury on the lesser-included crime
of vehicular manslaughter. The prosecution was given the option of accepting a verdict on vehicular
manslaughter or retrying the case; the prosecution elected to retry the case. The prosecution tried
to bring Mr. C. to trial again in 2000, but the court dismissed the charges because they had not been
brought within the required time period. In 2001, the prosecution brought Mr. C. to trial for the third
time, this time for murder and hit and run. The jury could not reach a verdict and the court declared a
mistrial. In 2002, almost 10 years after the incident, the prosecution brought Mr. C. to trial for a fourth
time. He was convicted of second-degree murder and sentenced to 15 years to life.
In 2004, the court of appeals held that the prosecutor did not have any legal authority to proceed against
Mr. C. after the verdict in the first trial was reversed for insufficient evidence. Four trials, three appeals,
and over ten years later, Mr. C. was released from prison.

The fact that most of the unjust and illegal convictions in our sample came about after trial should
not be used to suggest that innocent people never plead guilty to crimes they did not commit,
or that errors do not occur in cases that resolve by plea bargain. Indeed, one comprehensive
review of two large group exonerations—the police scandals in Tulia, Texas and in the Rampart
division of the Los Angeles Police Department—documents the fact that the defendants in these
scandals only rarely went to trial.33 Instead, the vast majority of the Tulia and Rampart defendants
entered into plea bargains before they were exonerated. The fact that these defendants had not
actually committed the crimes to which they pled guilty would never have been known if the police
wrongdoing had not later become public knowledge.

FOOTNOTES
33	 Russell Covey. “Police Misconduct as a Cause of Wrongful Convictions” Washington University Law Review 90.4 (2013): 1133-1189.
http://openscholarship.wustl.edu/law_lawreview/vol90/iss4/2/
18 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

Other than the subsequently-discovered police wrongdoing, the hundreds of individuals in Tulia
and Los Angeles who pled guilty look no different than the thousands of individuals who plead
guilty every day in our nation’s courthouses. For all of them, it is next to impossible to establish
a wrongful conviction after a plea without new evidence or a fortuitous change in circumstances.
The guilty plea of Henry Samueli, for example, was dismissed only after his alleged co-conspirator
went to trial, a rare event in a state where only about 2% of cases proceed to trial. Mr. Samueli had
agreed to a deal that required him to plead guilty to making a false statement, and required that
he testify against his alleged co-conspirator, in order to avoid extensive prison time. During the
alleged co-conspirator’s trial, after hearing Mr. Samueli’s testimony, the trial judge dismissed Mr.
Samueli’s guilty plea and stated that the prosecutor had intimidated and improperly influenced Mr.
Samueli.34 “Needless to say, the government’s treatment of Mr. Samueli was shameful and contrary
to American values of decency and justice,’’ the judge said.35 “I have looked at the plea agreement.
I have listened to your testimony and you didn’t make a false material statement.”36
We do not know how many Californians plead guilty due to administrative convenience, the
desire to end incarceration, rational risk-balancing given the risk of a loss at trial, a need to work
and support their families, or a lack of faith in the system’s ability to accurately determine factual
innocence. The question of how many people plead guilty even when they were innocent is still
very much at issue.37
A robbery and carjacking took place in the early morning hours of May 23, 2005, near James Ochoa’s
home in Buena Park, California (Los Angeles County). The victims gave a description to police. One
officer thought that 19-year- old Mr. Ochoa matched the description and showed Mr. Ochoa’s picture
to the victims, who said that Mr. Ochoa looked like the perpetrator.38 Two hours later a bloodhound
was brought to the scene and given a swab from a baseball cap found in the car. The dog followed
the scent to Mr. Ochoa’s front door. Despite several family members stating that Mr. Ochoa had been
home all night, he was arrested and charged with armed robbery, carjacking, and street terrorism.
Prior to trial, DNA tests were done on the interior of the car and on objects found in the car. The DNA
did not match Mr. Ochoa’s, but the district attorney continued pursuing him. Mr. Ochoa was presented
with a plea offer but rejected it.

FOOTNOTES
34	 U.S. v. Ruehle, Reporter’s Transcript December 15, 2009, Case No. SACR 08-00139-CJC
http://clients.oakbridgeins.com/clients/blog/ruehletranscript.pdf
35	 http://www.nbclosangeles.com/news/Judge-Throws-Out-Chargess-Against-Broadcom-Founder-Nicholas-79357367.html
36	 http://www.ocregister.com/articles/samueli-223282-options-guilty.html
37	 “Courts work hard to equalize the incoming and outgoing cases—to resolve today’s crimes and infractions to make way for tomorrow’s…Courts use plea
bargains to help equalize inboxes and outboxes. And experts say the use of enticements to wheedle admissions of guilt out of the accused makes the plea
bargain process a lodestone for wrongful convictions.” David J. Krajicek, “America’s Guilty Mill,” The Crime Report, February 9, 2015,
http://www.thecrimereport.org/news/inside-criminal-justice/2015-02-americas-guilt-mill.
38	 Mr. Ochoa’s age is taken from the National Registry of Exonerations and was not independently verified with court records.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 19

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

After he rejected the plea, the judge indicated that Mr. Ochoa would be given the maximum life
sentence if he were found guilty. Mr. Ochoa’s attorney felt that the judge was pressuring him to get
Mr. Ochoa to take a plea offer.39 Mr. Ochoa’s jury trial began on December 7, 2005, but after three days
he decided, against the advice of his attorney, to plead guilty to second degree armed robbery for a
guaranteed sentence of two years state prison. Mr. Ochoa told the judge that he was innocent but that
he was scared at the possibility of facing a sentence of 25 years to life in prison, especially because he
had a two-year-old son. A year later, in October of 2006, another man confessed to the robbery and
carjacking. This man proved to be a DNA match to the DNA found on the items in the car. The district
attorney filed a petition for habeas corpus, which was granted. Mr. Ochoa was released the next day.
A month later, on November 22, 2006, the court granted Mr. Ochoa’s petition for a finding of factual
innocence. Mr. Ochoa subsequently settled a civil lawsuit for $550,000. The lawsuit was based in part
on improper police practices used to obtain the eyewitness ID, as well as improper reliance on the
dog as evidence. Mr. Ochoa also received state compensation of $31,700 in 2008, although the state
compensation board initially refused to grant him compensation because of the guilty plea.

The Original Sentences Imposed in these Flawed Cases Were
Disproportionately Severe
Had the errors described in this report not been uncovered and the cases dismissed, the 607 individuals
in our sample would have spent an enormous additional amount of time in custody. Nineteen percent
(19%) of the unjustly convicted individuals in this report were originally sentenced to imprisonment
for life or life without parole, and an additional 11% were originally sentenced to more than 10 years or
more than 20 years in prison (see Figure 3).40 All of these sentences were reversed and the convictions
dismissed—but if they had remained hidden, California taxpayers could have paid millions of dollars to
house these individuals for the remainder of their lives, including end of life medical costs that can run
into the millions of dollars.
19% of the individuals subjected to these faulty prosecutions were originally sentenced to life or life
without parole.

FOOTNOTES
39 	 Victims Compensation Government Claims Board Report, “In the Matter of the Claim of James Ochoa, Claim No. 565437”, February 5, 2008, 64,
http://www.vcgcb.ca.gov/docs/pc4900/PC-4900-Approved-Ochoa.pdf.
40	 Those who served the longest include Kash Register (33 years, Los Angeles County), James Shortt (27 years, Los Angeles County), Frank O’Connell
(27 years, Los Angeles County), Adam Miranda (27 years, Los Angeles County), Elmer Pratt (25 years, Los Angeles County), Thomas Goldstein (24 years,
Los Angeles County), Bruce Lisker (24 years, Los Angeles County), Willie Green (23 years, Los Angeles County), Kenneth Marsh (20 years, San Diego
County), Mark Sodersten (20 years, Tulare County), Maurice Caldwell (20 years, San Francisco County), Timothy Atkins (19 years, Los Angeles County),
John Stoll (19 years, Kern County), Francisco Carrillo (18 years, Los Angeles County), and Norma Croy (17 years, Siskyou County).

20 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

Figure 3: Original Sentence Imposed, Cases in this Report
.33% Death
27% 1–5 Years

3% Unknown
3% Life Without Parole
5% More Than 20 Years

20% Less Than a Year in Custody

N = 607

6% More Than 10 Years
7% Probation Without Custody

16% Life
13% 6–10 years

Moreover, the individuals in our sample were penalized more severely than those in the system at large,
which is likely a logical outcome of the fact that our sample contained more prosecutions for violent
crime than the system as a whole. More than two thirds of the cases (68%) in our sample resulted in the
defendant being sentenced to a term in prison, as compared to an annual California rate of about 20%
prison sentences imposed for felony convictions. Similarly, only 28% of the individuals in our sample
were sentenced to probation and/or less than a year in jail, while annually about 80% of all felony
convictions in California receive such sentences (see Figure 4).41
Figure 4: Sentences Imposed for Felony Convictions, Annual California Proportion v. Proportion in this Report
100%

Annual Sentences in California

80%

Sentences Imposed in this Report

60%
40%
20%
0%
Prison

Probation or Less Than a Year in Jail

FOOTNOTES
41	 Office of the California Attorney General, “Crime in California 2011,” Table 40, http://oag.ca.gov/sites/all/files/agweb/pdfs/cjsc/publications/candd/cd11/
cd11.pdf? and Office of the California Attorney General “Crime in California 1996,” Table 41 (1991 through 1996). http://ag.ca.gov/cjsc/publications/
candd/cd96/cd96obts.pdf. The year 2011 is used because Public Safety Realignment changed the ways in which defendants were sentenced to prison
and jail starting in October 2011. Prior to Realignment, the ratio between prison and non-prison sentences imposed in felony cases did not change
significantly for over 20 years; from 1996 to 2011 about 20% of sentences imposed annually in felony cases were prison sentences. In 2011, the exact
numbers were: 195,821 total felony convictions and 37,972 prison sentences (19.4%). In 1996 the exact numbers were: 197,309 felony convictions and
43,691 prison sentences (22.1%). In 1991, the exact numbers were: 195,727 felony conviction and 38763 prison sentences (19.8%). As with Figure 1,
the comparison is not exact because our dataset stretches over 25 years, while the Attorney General figures cover one calendar year.

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 21

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

These Faulty Convictions Were Not Resolved Quickly
Our criminal justice system does not acknowledge or repair its errors quickly, particularly in California.42
On average, the people in our sample fought for four and a half years before their cases were ultimately
dismissed.43 Fifty-eight of the individuals in the sample were imprisoned for more than ten years, and
eleven were in prison for more than twenty years, before their cases were ultimately reversed and
dismissed. Roughly one-quarter of the cases (24%) took more than five years from initial appearance
to final resolution, and one defendant, Kash Delano Register, fought for 34 years.
In April of 1979, an 18-year-old African-American teenager named Kash Register was arrested for the
murder of an elderly white man in Los Angeles County. At trial, Mr. Register presented five alibi witnesses
including family members, his girlfriend, and a job coach. The fingerprints found on the scene did not
match Mr. Register’s and no evidence specifically linking him to the crime was found in his home or on
his person. The case was based primarily on the eyewitness accounts of two people. One testified before
trial that she could not see the suspect very well and that she was not sure of her identification. At trial,
however, she said that she had no doubt that the suspect was Kash Register. The other eyewitness had an
elaborate story about chasing down the perpetrator but later stated that he did not remember any of the
events on the day in question. Following a nine-day trial before an all-white jury, Mr. Register was found
guilty of murder and sentenced to life in prison without the possibility of parole. The sentence was later
reduced to 27 years to life in prison. Mr. Register appeared before the parole board eleven times and
professed his innocence. His requests for parole were denied.
In 2012, Mr. Register filed a habeas petition alleging that the prosecution withheld evidence (including the
fact that one of the eyewitnesses was on probation at the time she testified), that eyewitnesses had lied
about what they saw, and that there were two additional eyewitnesses, one of whom could not identify
the shooter and the other of whom would have testified that Mr. Register was not the shooter. The court
held that the prosecutor had improperly withheld this exculpatory evidence. In light of the new evidence,
the court also found that the main witness against Mr. Register was not credible. The case was dismissed
in December 2013 and Mr. Register was released after serving 34 years in prison.44

Within these averages are large variances based on the type of prosecution: the median time from initial
arrest to reversal and release in flawed homicide prosecutions was almost nine years, while the median in
theft and drug cases was just about two years (see Figure 5).

FOOTNOTES
42	 Particularly when it comes to payment of compensation, other states have made resolution a priority in a way that California has not. New York, Ohio and
Texas have all worked to speed the process by which wrongfully convicted individuals are compensated.
43	 Date ranges in this section are from the date of arrest or first appearance in court, to the date of final dismissal. The average was 4.5 years and the
median was 2.7 years.
44	 Mr. Register’s claim for compensation had not been resolved as of the date of this report.
22 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

Figure 5: Median Case Resolution Length (Years), By Conviction Type
Larceny / Theft
Receiving / Possessing Stolen Property
Possession of Drugs
Possession of Weapons
Burglary
Other
Robbery
Failure to Register as a Sex Offender
Assault
Lewd Act on a Minor
Forgery / Fraud
Rape

N = 607

Homicide
0

1

2

3

4

5

6

7

8

9

Flawed Homicide Convictions Were the Most Expensive
Costs associated with the 607 cases in our sample were not distributed equally among the various
types of convictions. Faulty convictions with more serious charges, and particularly wrongful homicide
prosecutions, were the most expensive. Indeed, the average cost of the flawed homicide prosecutions
in our sample was more than twice the average cost of any other type of crime. Moreover, these failed
homicide prosecutions constituted more than half (52%) of the total cost calculated for all cases in the
sample. (See Figure 6)
The high cost for unjust homicide convictions is due in part to the fact that the individuals subjected to
these prosecutions spent longer terms in prison than others in the sample. In addition, they filed more
lawsuits and received more settlements than individuals prosecuted for other types of crime, all of which
contributed to the total cost.

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 23

SECTION 1 | THE CHARACTERISTICS OF INJUSTICE

Figure 6: Cost By Type of Crime
Crime

Sum of Total Cost

% of Total Cost

# of Cases

Average Cost Per Case

Homicide

$115,837,375

52%

92

$1,259,102

Lewd Act on a Minor

$27,387,054

12%

47

$582,703

Possession of Drugs

$19,987,150

9%

187

$106,883

Rape

$12,362,329

6%

21

$588,682

Assault

$11,213,758

5%

40

$280,344

Robbery

$9,242,437

4%

26

$355,478

Burglary / Theft / Stolen Property

$8,219,649

4%

74

$458,681

Other

$7,994,023

4%

57

$140,246

Possession of Weapons

$5,760,472

3%

43

$133,964

Failure to Register as Sex Offender

$3,597,483

2%

20

$179,874

Grand Total

$221,601,730

100%

607

$365,077

More than half of the cost associated with the reversed cases in this sample came from flawed
homicide prosecutions.

24 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

SECTION 2:

THE CAUSES
OF INJUSTICE

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 25

SECTION 2 | THE CAUSES OF INJUSTICE

This section examines the errors that caused the reversal of the convictions in our sample.
Four hundred and eighty three (483) jury trials, 26 mistrials, 16 hung juries, and 168 plea bargains
were required to resolve the 607 cases profiled in this report. Something went wrong in each of these
proceedings. They all contained at least one instance of misconduct, incompetence, or mistake that
invalidated the conviction.
The 607 defendants in this report burdened the California criminal justice system with 483 jury trials,
26 mistrials, 16 hung juries, and 168 plea bargains before their cases were eventually dismissed.

In total, this report catalogues 756 confirmed errors,45 each representing a finding by a California court
that the particular act of misconduct, incompetence or mistake justified reversal of the conviction.
We divided the errors into eight different categories; each category is discussed separately in this section
(see Figure 7).46 The eight categories are addressed in descending order of cost, from highest total cost
to lowest.

FOOTNOTES
45	 There were 693 reversed trials, mistrials, hung juries and plea bargains, but we coded 756 errors. The number of errors is larger than the number of
cases or the number of defendants because courts sometimes reverse a conviction for more than one reason. If the court specified that it was reversing
a conviction on two separate grounds, each ground justifying reversal was counted as a separate error.
46	 We did not assign a source to the types of error, because most errors have multiple sources. A Fourth Amendment search and seizure error, for example,
requires a mistake by the police regarding the applicable law, a mistake by the prosecution in determining that the evidence should be used, a potential
mistake by the defense in not adequately objecting, and a mistake by the judge in allowing the evidence to be used for the initial conviction.
47	 Fifty-nine miscellaneous errors make up the final 8%. Miscellaneous errors include retroactive change in the law subsequent to conviction and unknown.
The full list of coding categories for all errors can be found in Appendix F.
26 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

Figure 7: Eight Categories of Error47
Judicial Mistake
In Trial
22% of all errors
(164 of 756)

Unreliable or
Untruthful Official
Testimony

$32 million

3% of all errors
(24 of 756)

Failure of
Prosecutorial Discretion
15% of all errors
(117 of 756)
$19 million

$15 million

Ineffective
Assistance of Defense
Counsel

Prosecutorial
Misconduct
11% of all errors
(86 of 756)
$53 million

DEFENDENT

Fourth
Amendment Violations
19% of all errors
(146 of 756)
$12 million

Inadequate
Police Practices
Before Trial
3% of all errors
(35 of 756)
$21 million

11% of all errors
(81 of 756)
$27 million
Eyewitness
Misidentification
6% of all errors
(44 of 756)
$31 million

Category

Definition

Judicial Mistake In Trial

Jury instructions, erroneous admission/ exclusion of evidence, Sixth Amendment
(confrontation, impartial jury), jury misconduct, sentencing error

Failure of Prosecutorial Discretion

Insufficient evidence, defendant’s conduct doesn’t meet the legal definition of the
crime, statute of limitations, double jeopardy

Ineffective Assistance of Defense Counsel

Failure to investigate, failure to adequately represent

Eyewitness Misidentification

Eyewitness recanted, eyewitness unreliable or lying, mistaken eyewitness, police
eyewitness practices inadequate

Inadequate Police Practices Before Trial

New evidence found after trial, invalid confession or statement, Fifth Amendment
violation, inadequate police practices

Fourth Amendment Violations

Improper search and seizure

Prosecutorial Misconduct

Prosecutorial misconduct, Brady violations

Unreliable or Untruthful Official Testimony

Unreliable/untruthful police officer testimony, confidential informant unreliable
or lying

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 27

SECTION 2 | THE CAUSES OF INJUSTICE

Although all these errors appear throughout the sample, certain errors were responsible for a
disproportionate share of the overall cost. Prosecutorial misconduct, for example, accounted for 11% of
the errors but 24% of the cost, while Fourth Amendment (search and seizure) violations accounted for
19% of the errors but only 5% of the cost (see Figure 8).
Figure 8: Cost By Type of Crime

PROPORTIONATELY
MORE COST

PROPORTIONATELY
LESS COST

Type of Error

% of Error

% of Cost

Average Cost Per Error

Prosecutorial Misconduct

11%

24%

$617,513

Eyewitness Identification

6%

14%

$714,172

Ineffective Assistance	

11%

12%

$327,870

Official Testimony

3%

7%

$615,873

Police Practices

3%

10%

$620,832

Judicial Mistake

22%

14%

$194,962

Prosecutorial Discretion

15%

9%

$165,503

Fourth Amendment

19%

5%

$78,377

Prosecutorial misconduct, inadequate eyewitness identification, unreliable official testimony, and
improper pre-trial police practices appear to be more costly errors per incident than other types of
errors.

Moreover, errors were not distributed evenly across the different types of prosecution (see Figure 9).
Figure 9: Most Common Error for Each Type of Prosecution
Type of Prosecution

% of
Prosecutions

Most Common Error

Possession of Drugs

31%

Fourth Amendment Violations

Homicide

15%

Prosecutorial Misconduct

Other

9%

Failure of Prosecutorial Discretion

Burglary / Theft / Stolen Property

8%

Failure of Prosecutorial Discretion / Judicial Mistake During Trial

Lewd Act on a Minor48

8%

Judicial Mistake During Trial

Assault

7%

Judicial Mistake During Trial

Possession of Weapons

7%

Fourth Amendment Violation

Forgery / Fraud / Embezzlement

4%

Judicial Mistake During Trial

Rape

4%

Inadequate Police Practices Before Trial / Problems With Eyewitness Testimony

Robbery

4%

IAC

Failure to Register as Sex Offender

3%

Failure of Prosecutorial Discretion

Prosecutorial misconduct was the most common error in flawed homicide prosecutions in
this sample.
28 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

Prosecutorial Misconduct
$53 million overall (24% of the total)
$617,513 average cost per error
11% of the errors

•	 Prosecutorial misconduct includes Brady 49 violations (failure to meet the constitutional
obligation to turn potentially exculpatory information over to the defense) and other
misconduct (e.g., overzealous closing argument, improper admission of evidence, violation
of judicial order, knowing admission of false testimony).
•	 Of the 86 errors in the prosecutorial misconduct category, 45 (52%) are Brady violations.
These violations of the prosecutorial obligation to share exculpatory information account for
$44 million of the total cost.
•	 The largest share of the prosecutorial misconduct (44%) occurred in homicide cases, followed by
prosecutions for lewd act on a minor (13%).
•	 Prosecutorial misconduct accounted for 11% of the errors but 24% of the total cost.
•	 Total sample cost associated with all prosecutorial misconduct was $53 million.

PROSECUTORIAL MISCONDUCT
Prosecutorial misconduct includes any act or failure to act by a prosecutor that violates a professional
code of ethics. Under Federal law, a prosecutor’s behavior violates the U.S. Constitution “if it is
so egregious that it infects the trial with such unfairness as to make the conviction a denial of due
process.”50 Under California state law, a prosecutor’s actions constitute prosecutorial misconduct if
they involve the use of deceptive or reprehensible methods to attempt to persuade either the court or
the jury.51 Not every act of misconduct will result in the reversal of a conviction, however, because the
defendant must also show that he was prejudiced or actually harmed by the misconduct.52
Misconduct might consist of inappropriate remarks in the courtroom, presenting false or misleading
evidence, tampering with evidence, or soliciting perjury or false evidence. It also might include the failure
to disclose exculpatory evidence (i.e., Brady violations); we address these cases separately below. At the
same time, misconduct can occur without any intent to create an unjust result. The federal Professional

FOOTNOTES
48	 Lewd Act on a Minor is the name given to California Penal Code Section 288(a), defined as an act upon a person under 14, on skin or through clothes, with
the intent of sexual gratification for the perpetrator. It is a lesser crime than rape or other sexual assault crimes on minors, which are regulated by other
penal code sections.
49	 Brady v. Maryland, 373 U.S. 83 (1963)
50	 People v. Harris, 47 Cal.3d 1047, 1084 (1989), citing Donnelly v. DeChristoforo, 416 U.S. 637, 642-643 (1974).
51	 People v. Hill, 17 Cal. 4th 800, 819 (1998); People v. Pitts, 223 Cal. App.3d 606,691 (1990); People v. Strickland, 11 Cal.3d 946, 955 (1974).
52 	 Chapman v. California, 386 U.S. 18, 22-23 (1967); see also People v. Crew, 21 Cal.4th 822, 839 (2003).
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 29

SECTION 2 | THE CAUSES OF INJUSTICE

Misconduct Review Unit (PMRU), for example, investigates allegations of misconduct by Department of
Justice attorneys and divides misconduct into four categories: 1) intentional professional misconduct;
2) reckless professional misconduct; 3) poor judgment; and 4) excusable mistake.53
In July of 2004 in Los Angeles County, police arrested Jaime A. after a 16-year-old boy alleged that 27
year old Mr. A. stole the boy’s bicycle. The main evidence against Mr. A. was the identification by the
victim. Despite a lack of physical evidence linking Mr. A. to the crime, he was found guilty by a jury
after a three-day trial. Because he had two prior strikes he was sentenced to 35 years to life in state
prison. Two years later, the court reversed the conviction based on prosecutorial misconduct because
the prosecutor had improperly vouched for the integrity of her office and stated her own personal
belief that Mr. A. was guilty in closing argument. Mr. A. was retried, and in his second trial the jury
deadlocked. The prosecutor sought to try him a third time, and this time the court declared a mistrial.
The prosecutor sought permission to try him a fourth time, but the court found that the evidence was
so weak that no reasonable jury could find Mr. A. guilty. The court finally dismissed the case over the
prosecutor’s objection. Mr. A. was in prison for over three years before his release.

Although our criminal justice system is an adversarial system, prosecutors are entrusted not with the
burden to win their case but rather with the burden to do justice. A failure to uphold this obligation is all
the more damaging given their unique role: “Prosecutorial misconduct fundamentally perverts the course
of justice and costs taxpayers millions of dollars in protracted litigation. It undermines our trust in the
reliability of the justice system and subverts the notion that we are a fair society.”54
Although the cases in this report stretch from 1989 to 2012, prosecutorial misconduct continues to
make headlines.55
In February 2015 in Riverside County, Johnny Baca was ordered to stand trial for a third time after both
a jailhouse informant and the district attorney who prosecuted Mr. Baca were found to have lied under
oath.56 Mr. Baca was accused of killing a friend’s father and the father’s partner.

FOOTNOTES
53	 U.S. Attorney General Eric Holder, Memorandum for Heads of Departments (January 14, 2011), http://www.documentcloud.org/documents/26309-memoon-professional-misconduct-review-unit.html. The Professional Misconduct Review Unit will review only findings of intentional or reckless professional
misconduct, but “findings of poor judgment and excusable mistake will continue to be referred to the relevant U.S. Attorneys for appropriate action.
“Attorney General Creates Professional Misconduct Review Unit, Appoints Kevin Ohlson Chief” U.S. Department of Justice, press release, January 18, 2011,
http://www.justice.gov/opa/pr/attorney-general-creates-professional-misconduct-review-unit-appoints-kevin-ohlson-chief.
54	 Kathleen M. Ridolfi and Maurice Possley, “Preventable Error: A Report on Prosecutorial Misconduct in California 1997-2009,” Northern California
Innocence Project, Santa Clara School of Law, October 2010, 4. http://digitalcommons.law.scu.edu/cgi/viewcontent.cgi?article=1001&context=ncippubs.
55	 This is particularly true in Orange County, which is embroiled in a misconduct scandal involving a “massive cover-up by both law enforcement and
prosecutors.” https://www.washingtonpost.com/news/the-watch/wp/2015/07/13/the-jaw-dropping-policeprosecutor-scandal-in-orange-county-calif/.
56	 Unless noted otherwise, case summaries in this report are based on court records. Mr. Baca’s summary is based on newspaper accounts instead of court
records and he is not included in our sample because his case was not final as of May 2015. See, Sarah Burge, “Riverside County: Prosecutor accused of
lying in murder trial,” The Press Enterprise, February 23, 2015, http://www.pe.com/articles/baca-760956-attorney-case.html?page=1. See also Maura
Dolan, “U.S. judges see ‘epidemic’ of prosecutorial misconduct in state,” Los Angeles Times, January 31, 2015,
http://www.latimes.com/local/politics/la-me-lying-prosecutors-20150201-story.html#page=1.
30 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

His original 1995 conviction was based in part on testimony from the jailhouse informant, who claimed
that the then-twenty-four-year-old Mr. Baca had confessed to him. The informant stated under oath
that he received no deal from the prosecution in return for his favorable testimony. The first conviction
was reversed for ineffective assistance of defense counsel. In Mr. Baca’s second trial, the district
attorney took the stand and testified under oath that the informant had received no deal in return for
his testimony. In fact, the informant had received a lower sentence in exchange for his testimony against
Mr. Baca. Mr. Baca appealed to the state appellate court and although the state court found that the
informant and the district attorney had both lied on the stand, the conviction was affirmed. Mr. Baca
then filed a federal Petition for Habeas Corpus, which was ultimately heard by the Ninth Circuit Court
of Appeals. The Ninth Circuit oral arguments took place on January 8, 2015. During oral argument, the
judges demanded to know why the informant and prosecutor had not been charged with perjury, and
expressed frustration about the fact that the prosecutor had never been referred to the State Bar for
investigation. One of the judges asked if Attorney General Kamala Harris was aware of the misconduct
and called on the district attorney to make the Attorney General aware of the prosecutor’s lying in the
next 48 hours.
The Attorney General then withdrew its opposition to Mr. Baca’s habeas petition, and on January
30th, 2015, the 9th Circuit granted the petition, vacating the conviction. The District Attorney’s Office,
however, decided shortly thereafter to proceed with a third trial against Mr. Baca. Mr. Baca was still
awaiting his third trial as of June 2015.

BRADY VIOLATIONS
In Brady v. Maryland, 373 U.S. 83 (1963), the United States Supreme Court articulated the prosecutor’s
Constitutional duty to turn over to the defense any material evidence that is favorable to the defendant.
The prosecutor’s duty extends to evidence that is held by law enforcement.
“A robust and rigorously enforced Brady rule is imperative because all the incentives prosecutors
confront encourage them not to discover or disclose exculpatory evidence. Due to the nature of a
Brady violation, it’s highly unlikely wrongdoing will ever come to light in the first place. This creates a
serious moral hazard for those prosecutors who are more interested in winning a conviction than serving
justice. In the rare event that the suppressed evidence does surface, the consequences usually leave the
prosecution no worse than had it complied with Brady from the outset. Professional discipline is rare,
and violations seldom give rise to liability for money damages.”57

FOOTNOTES
57	 U.S. v. Olsen, Case No. 10-36063, 11 (9th Cir. 2013) (Kozinski dissenting), http://cdn.ca9.uscourts.gov/datastore/opinions/2013/12/10/
10-36063%20web.pdf. In that same opinion, Justice Kozinski stated “There is an epidemic of Brady violations abroad in the land. Only judges can put
a stop to it.” US v. Olsen at 2.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 31

SECTION 2 | THE CAUSES OF INJUSTICE

On November 2, 1984, a mother was found beaten, burned and dead in her apartment in Tulare County.
Her two children were covered in soot but otherwise unharmed. The three-year-old girl stated that
“Mark did it,” identified Mark Dare in a lineup, and indicated that Mark Dare was her little brother’s
father. Mark Dare, who was her brother’s father, had an alibi. The little girl later changed her story and
said that there was a good Mark and a bad Mark and that the bad Mark was Mark Sodersten. Police later
found a fingerprint at the scene that belonged to the victim’s neighbor, Lester Williams. Mr. Williams
denied any knowledge of the murder in two interviews, but in his third interview he said he had been
there and that Mark Sodersten had fought with the victim and murdered her. Two other witnesses
placed Mark Sodersten with the victim, but there was no physical evidence that linked Mr. Sodersten to
the crime. At trial, the little girl’s father testified that he thought the government investigator switched
around the photos of Mark Dare and Mark Sodersten so many times that the girl got confused. Both
he and the victim’s mother testified that they had seen the victim fighting with Mark Dare, not Mark
Sodersten, days prior to the murder. In May of 1986, Mr. Sodersten, at age 28, was convicted by jury
of murder and sentenced to life without the possibility of parole. He unsuccessfully appealed. Mr.
Sodersten later learned that the prosecutors and the police had withheld a series of exculpatory tapes.
The tapes included recordings of interviews with the little girl, demonstrating her confusion between
the two Marks. Also on the tapes, Lester Williams repeatedly told police that he was high and did not
remember the night in question, until police threatened to charge him with murder and seek the death
penalty, after which he changed his story. Another tape revealed a conversation between Mr. Williams
and Mr. Sodersten, in which Mr. Williams told Mr. Sodersten that he lied to police when he implicated
Mr. Sodersten because he thought that Mr. Sodersten had implicated him in the murder. Mark Sodersten
challenged his conviction in light of these newly discovered tapes, but he died in prison at age 48
before the case was resolved. Despite Mark Sodersten’s death, the court held that the prosecutor had
committed a serious Brady violation and that it was reasonably probable that Mr. Sodersten would not
have been convicted had the prosecutor complied with his Constitutional obligations.

California is the only state in the nation that has failed to adopt some version of American Bar
Association Model Rule 3.8,58 which is broader than the Federal Brady standard because it requires
that prosecutors “make timely disclosure to the defense of all evidence or information known to the
prosecutor that tends to negate the guilt of the accused or mitigates the offense.”59 Unlike Brady, Rule 3.8
does not require prosecutors to make a determination as to the credibility of the evidence and therefore
leaves less discretion in the hands of prosecutors. Outside of the standard oath that all attorneys

FOOTNOTES
58	 Laurie Levenson and Barry Scheck, “California is overdue in adopting rule on exculpatory evidence,” Los Angeles Times, December 15, 2014,
http://www.latimes.com/opinion/op-ed/la-oe-1216-levenson-prosecutorial-misconduct-20141216-story.html.
59	 American Bar Association, Model Rules of Professional Conduct (Chicago: ABA Publishing, 2014), Model Rule 3.8(d),
http://www.americanbar.org/groups/professional_responsibility/publications/model_rules_of_professional_conduct/rule_3_8_special_responsibilities_
of_a_prosecutor.html. See American Bar Association, ABA Standards for Criminal Justice: Prosecution and Defense Function, 3d ed., 1993,
http://www.americanbar.org/content/dam/aba/publications/criminal_justice_standards/prosecution_defense_function.authcheckdam.pdf, (outlining the
special standards and duties for prosecutors).
32 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

practicing in California are required to take to uphold the U.S. and California State Constitutions,60
the California Bar Association provides very little guidance on the specific ethical duties of prosecutors.61
Although the vast majority of prosecutors comply with their Brady obligations, the failure to vigorously
and fully enforce Brady obligations has enormous ramifications: “When a public official behaves with
such casual disregard for his constitutional obligations and the rights of the accused, it erodes the public’s
trust in our justice system, and chips away at the foundational premises of the rule of law. When such
transgressions are acknowledged yet forgiven by the courts, we endorse and invite their repetition.”62
Two cases, 35 years apart, demonstrate how the same Brady violations that were taking place over forty
years ago continue today.
On July 28, 1972, 21-year-old, Elmer “Geronimo” Pratt was convicted of murder, robbery, and assault
with intent to commit murder in Los Angeles. He was sentenced to life in prison. Mr. Pratt was a member
of the Black Panther Party (BPP). He did not become a suspect until a letter indicating that he had
confessed to the murders was delivered to the police. Thereafter, two eyewitnesses in a photo lineup
identified Mr. Pratt and the police connected him to the gun and the getaway car. The defense presented
evidence that Mr. Pratt was at a BPP meeting in Oakland at the time of the murders, that the car at the
crime scene was often used by other BPP members, and that the gun was found inside a residence where
it was available to numerous people. Mr. Pratt was convicted after a thirty-one day jury trial.
The letter that incriminated Mr. Pratt was given to police by a man named Julius Butler, a former Los
Angeles Sheriff’s Deputy turned Black Panther. After the trial, it was discovered that Mr. Butler had
been an FBI informant. Information regarding Mr. Butler being an informant had been withheld from the
defense at trial. On the stand during trial, Mr. Butler had denied that he was an informant for the FBI.
Mr. Pratt filed an unsuccessful direct appeal and several unsuccessful habeas petitions. In 1997, after Mr.
Pratt had been incarcerated for 27 years, the conviction was finally overturned because the government
had violated its Brady obligation by failing to turn over evidence to the defense. Among other things, the
court found that Mr. Butler had been an FBI informant for at least three years prior to the trial, that Mr.
Butler had provided information to the FBI on at least 30 occasions, and that Mr. Butler had also been an
informant for the Los Angeles Police Department.

FOOTNOTES
60	 California Business and Professions Code Section 6068, “Duties of Attorneys,” https://www.calbarxap.com/applications/CalBar/PDFs/code_section_6068.
pdf. See also California Rules of Court, Rule 9.4, http://www.courts.ca.gov/cms/rules/index.cfm?title=nine&linkid=rule9_4. “In addition to the language
required by Business and Professions Code section 6067, the oath to be taken by every person on admission to practice law is to conclude with the
following: ‘As an officer of the court, I will strive to conduct myself at all times with dignity, courtesy and integrity.’” Judges also take a similar oath to
support and defend the Constitution of the United States and the Constitution of California. California Constitution, See Article 20, Sec. 3
61	 The California Bar Association has drafted some proposed rules using the relevant language from Model Rules 3.8, but they have yet to go before
the California Supreme Court. Supreme Court approval is required before the rules become binding. See The California Bar Association, Proposed
Rules of Professional Conduct, 2010, http://ethics.calbar.ca.gov/Portals/9/documents/CRRPC/RRC%20Final%20Docs/Proposed%20Rules%20of%20
Professional%20Conduct%20v.24%20%287-21-14%29.pdf.
62	 U.S. v. Olsen, (Judge Kozinski dissenting) at 15.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 33

SECTION 2 | THE CAUSES OF INJUSTICE

Mr. Pratt also presented evidence that the FBI knew that he was not in the LA area at the time of the
crime and was in fact at the BPP meeting in Oakland. The evidence included a declaration by a 25year FBI veteran who had heard another agent state that Mr. Pratt had been in Oakland at the time
of the murders. Evidence later surfaced that wiretaps installed in both the LA and Oakland Black
Panther hangouts confirmed Mr. Pratt’s presence in Oakland. The police had originally claimed that the
information from the wiretaps had been destroyed.63 None of this information had been turned over to
the defense prior to trial.

Ten years after Mr. Pratt’s release, in Sacramento County, another man was wrongfully convicted due to
the withholding of Brady information connected to an FBI informant.64 Eric McDavid had been arrested
on charges that he and two others conspired to build bombs to destroy the Sacramento Nimbus Dam,
a U.S. Forest Services Lab, and cellphone towers. No attacks took place. In 2007, at age 29, he was
convicted and sentenced to 20 years in prison.65
In November 2014, in response to a Freedom of Information Act request filed by Mr. McDavid,
the government provided a series of emails to Mr. McDavid’s defense attorneys. The emails related to
communications between Mr. McDavid and a key witness in the government’s case who had been an
FBI informant. In the emails, the informant promised a romantic relationship with Mr. McDavid once
a series of attacks against government targets were carried out. She urged Mr. McDavid to attack the
targets with promises of a sexual relationship. She also provided money, food, and shelter over an
18-month period while she was influencing Mr. McDavid to plan the attacks. The emails had been found
in files in the FBI’s Sacramento Office and the prosecutors claimed that they knew nothing of the letters
at the time of the trial.
Mr. McDavid’s conviction was reversed due to the Brady violation. He subsequently agreed to plead
guilty to a lesser charge of conspiring to attack a government facility in exchange for a five year prison
sentence. He also promised not to appeal or sue the government. He was released in January of 2015
because he had already served nine years.

FOOTNOTES
63	 Kate Coleman, “Elmer ‘Geronimo’ Pratt: The Untold Story of the Black Panther Leader, Dead at 63,” The New Republic, June 27, 2011,
http://www.newrepublic.com/article/politics/90735/black-panther-geronimo-pratt-murder-conviction-prison-huey-newton.
64	 Mr. McDavid is not in our database, and the errors and costs in his case are not reflected in our totals, because his exoneration came after
we had stopped collecting cases. Summary of the proceedings against Mr. McDavid are based on newspaper accounts and online sources.
See Danny Walsh and Sam Stanton, “Convicted ‘eco-terrorist’ freed amidst claims FBI hid evidence,” The Sacramento Bee, January 8, 2015,
http://www.sacbee.com/news/local/crime/article5641188.html; “Convicted ‘eco-terrorist’ freed amid evidence dispute,” USA Today, January
9, 2015, http://www.usatoday.com/story/news/nation/2015/01/09/eco-terrorist-freed/21528535/; Wikipedia, accessed March 26, 2015,
http://en.wikipedia.org/wiki/Eric_McDavid.
65	 Age calculated from http://www.sacbee.com/news/local/crime/article5641188.html.
34 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

The topic of prosecutorial immunity from liability is closely connected with Brady and prosecutorial
misconduct. Prosecutors are absolutely immune from civil lawsuits, meaning that they cannot be held
liable for their conduct as prosecutors even if they act knowingly and intentionally, and even if their
intentional acts result in the deliberate conviction of an innocent person. We are not aware of any other
profession—including police officers—that receives such immunity from intentional wrongdoing that
gravely injures another.
While prosecutors have strongly advocated to keep their immunity, the existence of absolute immunity
raises many concerns and the lack of accountability seems likely to contribute to the prevalence of
prosecutorial misconduct. It has been argued that “[t]he main problem with absolute immunity for
prosecutors is the incentives it creates. The problems with shielding a public servant in whom we grant
the enormous powers granted to prosecutors should be pretty self-evident. Now consider that nearly
every professional incentive (reelection, promotions, election to higher office, high-paying jobs at whiteshoe law firms) points prosecutors toward procuring as many convictions as possible, and that courts
and bar organizations are notoriously lax at sanctioning misconduct. You get a system that not only fails
to sanction bad behavior, but also often rewards it.”66
In 2014, the California legislature introduced California State Assembly Bill 885 (AB 885). The bill
addressed only prosecutorial misconduct that was (a) revealed during trial (a substantial minority of all
known Brady violations) and (b) determined by a judge to be intentional. Its remedy was measured:
the judge was permitted, but not required, to inform the jury of the prosecutor’s intentional misconduct,
and further could inform the jury that it was free to consider that fact in its assessment of the defendant’s
innocence or guilt. AB 885 passed both houses of the California legislature but was vetoed by Governor
Brown because “judges have an array of remedies at their disposal if a discovery violation comes to light
during trial.”67 The veto has been sharply criticized.68 While the need for prosecutors to have the freedom
to pursue their cases in good faith without fear of civil liability is plain, it is troubling that such a modest
and carefully tailored remedy for intentional acts of misconduct is not the law in California.69

FOOTNOTES
66	 Radley Balko, “7th Circuit Pokes a Hole In Prosecutorial Immunity,” The Washington Post, January 30, 2014,
http://www.washingtonpost.com/news/opinions/wp/2014/01/30/7th-circuit-pokes-a-hole-in-prosecutorial-immunity/.
67	 Text of the veto message can be found at: http://gov.ca.gov/docs/AB_885_Veto_Message.pdf.
68	 Radley Balko, “Jerry Brown Vetoes Bill Aimed At Holding Prosecutors More Accountable, The Washington Post, October 1, 2014,
http://www.washingtonpost.com/news/the-watch/wp/2014/10/01/jerry-brown-vetoes-bill-aimed-at-holding-prosecutors-more-accountable/.
69	 Texas is the first state in the nation that has sent a prosecutor to jail for his part in a wrongful conviction. See Mark Godsey, “For the First Time Ever a
Prosecutor Will Go to Jail for Wrongfully Convicting an Innocent Man,” The Huffington Post, November 8, 2013,
http://www.huffingtonpost.com/mark-godsey/for-the-first-time-ever-a_b_4221000.html.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 35

SECTION 2 | THE CAUSES OF INJUSTICE

Judicial Mistake During Trial
$32 million overall (14% of the total cost)
$194,962 average cost per error
22% of the errors

•	 Judicial mistakes during trial include jury misconduct, improper jury instructions,
erroneous admission or exclusion of evidence, and Sixth Amendment violations such
as witness confrontations, failure to seat an impartial jury, and denial of the right to
self-representation.
•	 Judicial mistakes were distributed evenly across types of crimes, with 16% in drug cases, 13% in
prosecutions for lewd act on a minor, 12% in homicide cases, and 11% in assault causes.
•	 Judicial mistakes accounted for 22% of the errors but only 14% of the total cost.
•	 The total cost associated with judicial mistakes was $32 million.

JURY MISCONDUCT
Jury misconduct includes any improper act by a juror, including the failure to be forthcoming during jury
selection or discussing the case with non-jury members during deliberations. Jury misconduct is one of
many errors that can hide other types of error, mistake and misconduct.70
On January 9, 1997, a man was killed at a party in Marin County. Police suspected that the victim
was shot in retaliation for an earlier altercation between the victim and another man about drugs.
The police questioned Darrell Hunter because he had been associated with the other man involved in
the altercation, but Mr. Hunter denied being at the party. Despite his denials, 23-year-old Mr. Hunter
was charged with murder, burglary, false imprisonment, assault with a deadly weapon, and possession of
a firearm. During the trial several witnesses identified Mr. Hunter as being at the party. He testified that
he was with the man involved in the earlier altercation on the night of the party, but that he did
not enter the party and remained outside. On February 2, 2000, Mr. Hunter was convicted of murder
and sentenced to life in prison without the possibility of parole. Shortly thereafter, Mr. Hunter’s
attorney learned that one of the jurors did not answer truthfully and withheld information when
questioned during jury selection. The court agreed that there had been juror misconduct and reversed
the conviction.

FOOTNOTES
70	 Judicial error, and the case of Mr. Hunter, illustrates why some of the numbers in this report may differ from the National Registry of Exonerations. Although the
actual ground for reversal of Mr. Hunter’s conviction was the juror misconduct, the NRE categorizes Mr. Hunter’s exoneration as eyewitness mistake. Although
there was eyewitness error, we have categorized the error as juror misconduct because that is the basis on which the court reversed the conviction.
36 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

Mr. Hunter was retried in January of 2008. He had a different attorney who presented some new
evidence, including the fact that witnesses said the shooter held the gun in his right hand but that Mr.
Hunter was left handed, and the fact that the gun found at the scene had someone else’s palm print on
it. In addition, counsel argued that police lineup procedures were improper because Mr. Hunter was
in prison clothes when others in the lineup were not. The jury returned a verdict of not guilty, and Mr.
Hunter was released on April 30, 2008, after 10 years in prison.

FAULTY JURY INSTRUCTIONS
Jury instruction errors vary, but examples include misstating the applicable law, failing to provide
information about lesser-included offenses, and instructing the jury in a manner that improperly sways
the jury’s opinion or decision-making process.
Lisa Pineda and her co-defendant were involved in an intimate abusive relationship in Los Angeles
County. In 2004, the co-defendant shot and killed someone in a gang-related shooting. A few days
earlier, the co-defendant had beaten Ms. Pineda because she had asked him to move out, and during
that altercation he had told her that he had a gun. On the day of the shooting, he made Ms. Pineda drive
him to the store even though she told him that she was not feeling well. He then insisted that she drive
him to a particular house. Fearing more abuse, she agreed. While in the car, he suddenly shot the victim,
who was outside the house. Ms. Pineda testified that she did not know he had a gun in the car and did
not know that he was planning to shoot anyone. The government later charged Ms. Pineda, age 26 and
the co-defendant with murder, and both were convicted.71 Ms. Pineda received a sentence of forty
years to life. The appeals court reversed her conviction because the trial judge had given incorrect jury
instructions, failing to instruct the jury on the lesser-included offense of involuntary manslaughter.
Ms. Pineda was retried. On retrial, after hearing evidence of the abuse that had not been presented
in the first trial, she was found not guilty of murder and the jury was unable to reach a decision on
involuntary manslaughter. A mistrial was declared on the involuntary manslaughter charge and the
government dismissed the case.

FOOTNOTES
71	 Ms. Pineda’s age was obtained from the National Registry of Exonerations.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 37

SECTION 2 | THE CAUSES OF INJUSTICE

EXCLUSION OR INCLUSION OF EVIDENCE
Juries in criminal trials can consider evidence only if it has been admitted by the court; the prosecution
and the defense must present their proposed evidence to the judge before trial so that the judge can
decide what will be admitted and what will be excluded. When the judge excludes or admits evidence
improperly, the result can be an unfair trial and an unjust conviction. A judge might, for example,
improperly refuse to admit evidence that impeaches a witness’ credibility. Or the judge might improperly
admit unsubstantiated scientific evidence (such as lie detector results or testimony derived from
hypnosis) or permit an expert witness to overstate the certainty of conclusions drawn from scientific
evidence. Although we have called this judicial error because the judge makes the ultimate decision
about what evidence is permitted, these errors can often also be attributed to inadequate representation
on the part of the defense attorney, or inadequate training and experience on the part of the prosecutor.
The judge, however, remains the final arbiter.
In 1997, a 14-year old girl told her drug counselor that her boyfriend had raped her nine months earlier
while at a resort with her family in Lake County. When the counselor informed her that they had to
report the rape to authorities, the girl changed her story and said that Brendan Loftus and his friend
had raped her. She described being tied to a bedpost by the friend and cut with a knife. Both Mr. Loftus,
age 23, and his friend were arrested and charged with rape.72 Mr. Loftus testified that he had been at the
resort and had seen his friend with the girl, but that he had been at the pool during the alleged attack.
Two witnesses testified that they saw him at the pool, and 15 character witnesses testified on Mr. Loftus’
behalf. Other evidence included the facts that there were no bedposts on the bed, that the victim had
given several conflicting reports, and that Mr. Loftus weighed140 pounds and would not have been able
to restrain the 200-pound victim.
The victim’s diary was made available to the defense. After reporting the rape, the victim had ripped the
pages from her diary that corresponded to the time of the alleged attack. She said she did so because
she was angry that her counselor had read them, but the counselor stated that she never read the diary.
In addition, eight months after the alleged rape, the victim had written a diary entry about having sex for
the first time with another man. The defense sought to have the diary introduced into evidence, but the
judge refused. Mr. Loftus was convicted of rape in 1997 and sentenced to five years in state prison.
After Mr. Loftus filed an appeal, the court held a hearing and heard testimony from the man referred to
in the diary entry. The man testified about his belief that the victim had been a virgin when he had sex
with her, eight months after the alleged rape. In 2000, the California Court of Appeal vacated Mr. Loftus’
conviction. The court held that, given the inconsistent testimony of the victim, the introduction of the
diary into evidence could have raised a reasonable doubt as to Mr. Loftus’ guilt or innocence. Mr. Loftus’
case was finally dismissed in 2000.

FOOTNOTES
72	 Mr. Loftus’ age was obtained from the National Registry of Exonerations.
38 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

Problems with Eyewitness Testimony
$31 million overall (14% of the total cost)
$714,172 average cost per error
6% of the errors

•	 Problems with eyewitness testimony include eyewitnesses who retract their testimony,
eyewitnesses who are found to be unreliable or lying, eyewitnesses who are mistaken in
their identifications, and police eyewitness identification practices that are improper
or inadequate and result in mistaken or false identifications.
•	 Eyewitness identification problems occurred most often when the individual was erroneously
convicted of homicide (41%), followed by cases in which the individual was erroneously
convicted of lewd act on a minor (25%) and rape (16%).
•	 Eyewitness identification accounted for 6% of the errors, but 14% of the cost.
•	 Total cost associated with eyewitness identification was $31 million.
Mistaken eyewitness identifications plague our criminal justice system, and yet eyewitness testimony
continues to be heavily relied upon by courts, prosecutors, and juries. In several of the cases
we reviewed, eyewitness testimony was the only evidence linking the individual to the crime.
Misidentifications are one of the most common types of error found in convictions of the innocent,73
and a growing body of research continues to document concerns and flaws.74
In 1979, Kevin Lee Green was a 21-year old corporal in the United States Marine Corps, living in Orange
County with his 20-year old pregnant wife. On the morning of September 30, 1979, Mr. Green drove to
Jack in the Box. When he arrived home he found his wife unconscious in their bedroom. She had been
raped, strangled, and hit in the head. Mr. Green called the police and an ambulance. His wife eventually
woke up but she had no memory of the event, had forgotten how to speak, and needed constant care.
The baby was stillborn. The police focused on Mr. Green as a suspect. Mr. Green and his wife had had
a tumultuous marriage and the police had been called to their home on several occasions. Although
Mr. Green reported to police that he had seen another man loitering in the apartment complex when
he left to get food that day, their attention remained on Mr. Green. In 1980 he was arrested and tried
for murder. At trial, despite her memory loss, his wife testified that the couple had quarreled about
having sex and that Mr. Green had beaten and raped her. The cashier at the Jack in the Box provided an
alibi for Mr. Green and the police report confirmed that the hamburgers were still warm when the first
responders arrived on the scene. Mr. Green was sentenced to 15 years to life for the murder of their unborn

FOOTNOTES
73	 “Eyewitness Misidentification,” The Innocence Project, accessed December 19, 2014,
http://www.innocenceproject.org/understand/Eyewitness-Misidentification.php.
74	 The National Academy of Science, “Identifying the Culprit: Assessing Eyewitness Identification,” (Washington D.C.: National Academies Press, 2014).
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 39

SECTION 2 | THE CAUSES OF INJUSTICE

child, assault with a deadly weapon, and the attempted murder of his wife. The only evidence against
him was his wife’s testimony. Mr. Green maintained his innocence throughout his entire time in prison.
He petitioned the California Supreme Court but the trial court decision was affirmed in 1982.
After the California DNA database was established, it was discovered that the DNA found at the scene
matched that of a serial killer. In 1996, the true perpetrator confessed to the wife’s attack as well as to five
other murders. The district attorney brought a petition for habeas corpus in the trial court based on the
newly discovered DNA evidence. Mr. Green was released after serving 16 years in prison. He was awarded
$620,000 by the State of California for his wrongful conviction in 1999.

In 2014, the National Academy of Science released a seminal report on eyewitness identification in
which the authors noted that “[T]he law enforcement community, while operating under considerable
pressure and resource constraints, is working to improve the accuracy of eyewitness identifications.
These efforts, however, have not been uniform and often fall short as a result of insufficient training,
the absence of standard operating procedures, and the continuing presence of actions and statements
at the crime scene and elsewhere that may intentionally or unintentionally influence eyewitness’
identifications.”75 The report identified a number of promising and research-based practices that have
been found to reduce inaccurate or erroneous identifications, often referred to as “false positive”
identifications, including:
•	 The use of blind procedures in lineups (i.e., administration of the photographs by an officer who
does not know who the suspect is, or whether the suspect is in the lineup);
•	 Ensuring that other people in the lineup resemble the suspect so that the suspect does not stand
out for any particular reason;
•	 Instructing the witness that the suspect may or may not be in the lineup and that the witness
should not look to the administrator for guidance;
•	 Telling the witness that the investigation will continue regardless of whether an identification
is made;
•	 Having the witness write a statement indicating the level of confidence in their identification
immediately following the lineup; and
•	 Recording the lineup to ensure that best practices are followed.76

FOOTNOTES
75	 The National Academy of Science, “Identifying the Culprit: Assessing Eyewitness Identification,” at 2. (Washington D.C.: National Academies Press, 2014).
76	 The National Academy of Science, “Identifying the Culprit”. See also “Eyewitness Identification,” The Innocence Project, accessed December 19, 2014,
http://www.innocenceproject.org/fix/Eyewitness-Identification.php.
40 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

Some counties in Northern California, such as Santa Clara County, are embracing these best practices,
as are cities of Los Gatos and El Cerrito.77 San Francisco County adopted the use of sequential lineups
and photo spreads,78 and Alameda County has been working to implement eyewitness best practices.78
In January of 1995, a man saw someone breaking into his truck in Santa Clara County. He shouted, and
the perpetrator left the truck and got into a nearby Trans Am. The victim said that the woman driving
the Trans Am had pointed a gun at him, and had then driven away in the Trans Am with the perpetrator.
The police were able to trace the license plate on the Trans Am. Later that day, Kenneth Foley’s wife
borrowed the Trans Am from the registered owner, and let Mr. Foley drive the car. Twenty-six-year-old
Mr. Foley was pulled over while driving the Trans Am. The witness identified Mr. Foley and the owner of
the Trans Am as the perpetrators in photo lineups.
Mr. Foley was tried and convicted by jury for his alleged involvement in the unsuccessful theft of a
car stereo. Two witnesses at Mr. Foley’s trial testified that they were, in fact, the actual perpetrators of
the crime and denied that Mr. Foley had any involvement. He was nonetheless convicted based almost
exclusively on the testimony of the single eyewitness. Throughout this time Mr. Foley maintained
his innocence.
Mr. Foley was sentenced to a “three strikes” sentence of 25-to-life in state prison. His appeals were
denied. In 2006, the District Attorney’s office re-opened its investigation and determined that there had
been no gun, and that there was strong possibility that Mr. Foley did not commit the crime. Mr. Foley’s
conviction was finally vacated in 2007, after the district attorney joined in a habeas petition brought by
the Northern California Innocence Project. By then Mr. Foley had been in prison for more than 11 years.
His claim for compensation from the state Victim Compensation and Government Claims Board was
subsequently denied because the Board found that he did not prove his innocence and did not prove
that he did not intentionally contribute to his arrest or conviction.

FOOTNOTES
77	 Police Chief’s Association of Santa Clara County, “Line-up Protocol for Law Enforcement, accessed March 26, 2014, http://www.ccfaj.org/documents/
reports/eyewitness/expert/Santa%20Clara%20County%20Eyewitness%20Identification%20Protocols.pdf. Also to be commended are Police Chief Greg
Suhr from San Francisco, Chief Scott Seaman from Los Gatos (an Executive Fellow and Past President of the California Police Chiefs’ Association) and
Chief Sylvia Moir from El Cerrito.
78	 Ari Burack, “New Police Protocols in San Francisco could help witnesses ID criminals,” The San Francisco Examiner, November 23, 2011,
http://www.sfexaminer.com/sanfrancisco/new-police-protocols-in-san-francisco-could-help-witnesses-id-criminals/Content?oid=2186742.
78	 Mistaken eyewitness identification played a role in 56 of the National Registry of Exoneration’s 154 California state cases; almost half of those 56 came
from Los Angeles County.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 41

SECTION 2 | THE CAUSES OF INJUSTICE

Ineffective Assistance of Counsel (IAC)
$27 million overall (12% of the total cost)
$327,870 average cost per error
11% of the errors

•	 IAC includes failure to adequately represent, failure to investigate, and failure to act
competently on the part of defense counsel.
•	 The largest share of the IAC was concentrated in homicide cases (27%), with another 19% in
drug cases.
•	 IAC accounted for 11% of the errors and 12% of the total cost.
•	 Total cost associated with IAC was $27 million.
An individual’s right to effective counsel is guaranteed by the Sixth Amendment of the Constitution.80
Situations in which a defendant does not receive such effective representation are referred to as
Ineffective Assistance of Counsel (IAC). Just as prosecutorial misconduct includes a variety of errors
attributed to prosecutors, IAC includes a variety of errors attributed to defense counsel. In order to
demonstrate IAC, the defendant must show that his or her attorney’s performance fell below an objective
standard of reasonableness and that there is a reasonable probability that, but for counsel’s deficient
performance, the result of the proceeding would have been different.81
In this report we have grouped all types of ineffective assistance together and have categorized them all
as error on the part of defense counsel. IAC can arise at almost any stage of an attorney’s representation,
and may include failing to adequately investigate a case before trial, not calling the appropriate witnesses
or experts to the stand during trial, or failing to object to the introduction of improper evidence during
trial, among other things.
Because claims of ineffective assistance can encompass almost anything that occurs during or before
trial, our classification has the impact of attributing certain errors to defense counsel when in reality they
may have been committed by more than one actor or part of the criminal justice system. For example,
some of the IAC cases in this sample were caused because the defense attorney did not pursue a motion
regarding a Fourth Amendment violation committed by the police. The initial mistake was made by the
police, who failed to comply with the Fourth Amendment. As a matter of the judicial record, however,
the proximate error justifying reversal was the ineffectiveness of defense counsel in failing to object to
the erroneous search, and therefore we have classified it as ineffective assistance of counsel.

FOOTNOTES
80	 Gideon v. Wainwright, 372, U.S. 335 (1963); Strickland v. Washington, 466 U.S. 668 (1984).
81	 Strickland v. Washington, at 669.

42 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

On July 8, 2002, in Los Angeles County, an African-American man opened fire on two men in South Los
Angeles, wounding one and killing the other. Two witnesses at a nearby liquor store identified twentyyear-old DeAndre Maurice Howard from a photographic lineup administered by the police. One of the
eyewitnesses later recanted her identification. Mr. Howard pleaded not guilty and went to trial in 2003.
He was found guilty of murder after a six-day trial, and sentenced to 75 years to life in state prison.
Two years later, the surviving victim of the shooting gave a sworn statement that Mr. Howard was not the
shooter, and stated further that had he been called at trial he would have testified that Mr. Howard was
not the shooter. Mr. Howard filed numerous appeals and habeas petitions, all of which were denied until
2011, when the court held that Mr. Howard was entitled to an evidentiary hearing on his claim that his
trial attorney was ineffective because he failed to interview or call the surviving victim as a witness.
After listening to testimony at the hearing, the court reversed Mr. Howard’s conviction and set the case
for a new trial. In 2013, after a nine-day jury trial, Mr. Howard was found not guilty on all charges and
released the same day. He spent 11 years in custody.

In 1990, twenty-eight-year-old Kelvin Wiley was arrested and charged with burglary and assault of
his former girlfriend in San Diego County. The former girlfriend testified that Mr. Wiley entered her
apartment in a rage, hit her repeatedly with a wrench and strangled her with a belt until she lost
consciousness. She also said that he attacked her because she tried to break off their relationship.
Mr. Wiley testified that he had been at his apartment at the time of the crime. His landlord testified that
he saw Mr. Wiley’s truck in front of his home and did not see Mr. Wiley enter or exit his apartment on
the day in question. Mr. Wiley testified that it was he who had broken off the relationship and since that
time she had been following him and harassing him. On the night before the attack he had declined a
dinner invitation from her and accidentally ran into her at a restaurant while he was having dinner with a
friend. He returned home to find an obscene note on his door. Based almost solely on the ex-girlfriend’s
testimony, Mr. Wiley was convicted of battery with serious bodily injury and sentenced to four years in
prison. He subsequently alleged ineffective assistance of counsel because his attorney failed to interview
several key witnesses, including neighbors who saw the ex-girlfriend leave her home in the morning
and who saw another man banging on her door shouting to be let in. The court found that Mr. Wiley’s
counsel had been ineffective, and the case was dismissed.

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 43

SECTION 2 | THE CAUSES OF INJUSTICE

Yvonne Eldridge was an experienced foster parent in Contra Costa County who was selected to
participate in a fragile infant care program for babies born with grave medical problems. When two
young infant girls under her care had multiple medical ailments requiring major medical interventions,
she was referred to Child Protective Services and eventually was prosecuted for child abuse.
Her trial took place in May of 1996. The babies at issue included one who was born seven weeks
premature to a drug-abusing mother, and a baby with a birth weight of only 1½ pounds. Both had
problems feeding and breathing, and issues with their bowel movements. Both had tubes placed into
their stomachs for feeding. One of the girls had 24 hospitalizations, 13 ER visits, and 200 outpatient visits.
Both of the babies ended up being removed from Ms. Eldridge’s care and their conditions improved
over time. The prosecution argued that Ms. Eldridge had Munchausen-by-proxy syndrome whereby she
would lie about the ailments of children in her care in order to get attention from doctors and medical
personnel. The defense was able to exclude mention of Munchausen-by-proxy but the prosecution was
still able to argue that Ms. Eldridge had manufactured her foster children’s illnesses and as a result the
babies endured unnecessary medical procedures. In 1996, at age 43, Ms. Eldridge was found guilty of
child abuse and was sentenced to three years in state prison.
Ms. Eldridge filed a motion for a new trial, arguing that her attorney had been ineffective. Her trial
attorney had failed to call witnesses who could have confirmed that the babies were genuinely ill,
and had failed to call a medical expert on Ms. Eldridge’s behalf although the prosecution had called
two medical experts. The defense attorney had consulted with a doctor on the case, but had given
the doctor only a limited portion of the medical records that had been prepared by the prosecution.
The doctor was never hired as an expert witness and was not provided adequate or complete medical
records to form a solid medical opinion. The doctor stated that after reviewing the complete medical
files, he would have testified that there were likely medical explanations other than abuse and that the
symptoms described were consistent with the medical histories of the two children. He further stated
that the issues suffered by the two girls were possibly the result of side effects of medications given
by the doctors and that there were several instances of medical mistake. Finally, Ms. Eldridge provided
evidence that one of the prosecution’s medical experts had a history of accusing women of Munchausenby-proxy, that he had made a pass at Ms. Eldridge but had been rejected, and that he had a history of
making sexual advances to patients and hospital staff. Ms. Eldridge’s motion was granted in 2002 and in
January 2003 the district attorney chose to dismiss the case rather than proceed with a new trial.

44 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

Failure of Prosecutorial Discretion
$19 million overall (9% of the total cost)
$165,503 average cost per error
15% of the errors

•	 Failures of prosecutorial discretion include court rulings that there was insufficient
evidence to support the conviction, a finding that the conviction could not be upheld
because the defendant’s conduct did not meet the definition of the crime, or a dismissal
because the prosecution violated the statute of limitations or double jeopardy.
•	 These errors were distributed across all types of convictions, with the largest concentrations in
drug cases (16%), prosecutions for failing to register as a sex offender (11%), fraud, forgery or
embezzlement (9%), and homicide (9%).
•	 These errors accounted for 15% of the errors but only 9% of the cost.
•	 Total cost associated with failures of discretion was $19 million.
From time to time, convictions are overturned because the conduct that has been proved does not fit
the definition of the crime charged or is determined by a court not to be criminal conduct. In addition,
convictions can be overturned when it is determined that the prosecution proceeded in violation of the
double jeopardy clause in the Constitution or brought the case after the end of the statute of limitations.
We have grouped these errors as “failures of prosecutorial discretion.” They have been combined with
cases in which a trial judge or an appellate court has ruled that, given the evidence presented at trial,
no reasonable trier of fact could have convicted the defendant beyond a reasonable doubt.82
We group these errors and use the term “failure of discretion” because all of these errors reflect a
prosecutor’s decision to proceed with a case when the evidence, facts, or law do not support the
prosecution. There is an obvious complexity that accompanies this sort of retrospective assessment of
error, as there may well be instances where the good faith pursuit of a criminal case that ends up being
unsuccessful is a reasoned and appropriate decision. At the same time, not only is it the obligation of the
prosecutor to review all the evidence and proceed only when the evidence warrants it, but it is an obligation
given specifically to the prosecutor in his or her sole discretion, and one with tremendous negative
downstream implications for those charged whether or not the individual is ultimately exonerated.
In addition, the selection of which charges to bring for conduct that a prosecutor believes in good faith to
be criminal can be a substantial and complex undertaking, and the decision regarding what to charge may
be based upon information known to the prosecutor or believed to be true that is either excluded from
evidence or subsequently found to be false. In such instances, the error would not be the levying of the

FOOTNOTES
82	 The United States Supreme Court has stated that “[t]he relevant question is whether, after viewing the evidence in the light most favorable to the
prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.” Jackson v. Virginia, 443 U.S. 307,
319 (1979). The California state standard is articulated in People v. Johnson, 26 Cal. 3d 557, 578 (1980). “[T]he court must review the whole record in the
light most favorable to the judgment below to determine whether it discloses substantial evidence, that is evidence which is reasonable, credible, and of
solid value such that a reasonable trier of fact could find the defendant guilty beyond a reasonable doubt.”
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 45

SECTION 2 | THE CAUSES OF INJUSTICE

initial charges, but the decision to proceed to trial with unavailable or compromised evidence. Likewise,
a court finding of insufficient evidence may indicate that the prosecutor misjudged or overestimated
the value and strength of the evidence against the defendant to an unreasonable extent, as subsequently
determined by an objective observer (the court). These errors as a whole reflect misjudgment or
overzealousness by the prosecution and are therefore grouped as “failure of discretion.”
In 1980 a Los Angles man was found dead in the basement of his fix-it shop. Next to his body were three
different size posts that were part of a turnstile unit typically found in grocery or retail stores. It was
believed that the murder weapon was the three-foot post. In 1985, five years after the murder, the police
matched some fingerprints on the alleged murder weapon to 25-year-old Melvin Mikes. Although no
other prints were found in the shop and there was no evidence that placed Mr. Mikes at the shop on the
day of the crime, Mr. Mikes was charged with first-degree murder and robbery. He pleaded not guilty and
went to trial.
The fingerprints found on the post were the only piece of evidence linking Mr. Mikes to the crime.
There were a total of 46 prints on the post, of which only sixteen were identifiable. Of those sixteen,
six belonged to Mr. Mikes. None of Mr. Mikes’ prints were found anywhere else in the fix-it shop,
including the area where jewelry boxes were found strewn about. There was an alibi witness who would
have testified that Mr. Mikes was elsewhere at the time of the murder, but the defense never called her
to the stand.
Mr. Mikes was convicted of first-degree murder and sentenced to 25 years to life. Six years later, in
1991, the court found that there was insufficient evidence to convict Mr. Mikes of murder since the
prints could have been made prior to the crime and there was no other evidence connecting him to the
murder. Mr. Mikes was released in 1992, after seven years of incarceration.

Prabhat Goyal, the Chief Financial Officer of Network Associates Inc., (NAI) formerly MacAfee,
was charged in federal court with one count of securities fraud, six counts of making material false
statements to auditors, and six counts of making false filings with the Securities and Exchange
Commission. The prosecution challenged the accounting method used by NAI under Mr. Goyal’s
supervision, and alleged that NAI was overstating its revenue.
Mr. Goyal was convicted at trial in 2007. On appeal, the court found no evidence that a crime had
been committed, and held that the government did not prove that any technical accounting violations
materially affected the revenue reported by the company. Further, the government did not prove that
Mr. Goyal knowingly made false statements. The case was overturned in 2010. Judge Kozinski, in a
concurring opinion, wrote, “This is just one of a string of recent cases in which courts have found that
federal prosecutors overreached by trying to stretch criminal law beyond its proper bounds.”83

FOOTNOTES
83	 U.S. v. Goyal, 629 F.3d 912 (2010), (Kozinski concurring)
46 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

Unreliable or Untruthful Official Testimony
$15 million overall (7% of the total cost)
$615,873 average cost per error
3% of the errors

•	 Unreliable or untruthful official testimony includes testimony from police officers,
confidential informants, and jailhouse informants.
•	 Official testimony problems occurred most often in homicide cases (50%), followed by drug
cases (46%).
•	 Unreliable or untruthful official testimony accounted for 3% of the errors, but 7% of the cost.
•	 Total cost associated with official testimony was $15 million.
Unreliable or untruthful testimony, whether by police officers or by government-sponsored confidential
informants, reflects a problem with the investigative arm of the prosecution. Although the sources of
error are different (officer testimony or informant testimony), the structural reasons behind the errors
are similar in that each may be the result of institutional pressure to obtain a conviction against a
particular defendant.84 Such errors are often multi-stakeholder errors—law enforcement has generated
the error during its investigation; the prosecutor has failed to identify the unreliability or lack of
truthfulness in his or her assessment of how to bring the case to trial; and the defense counsel, judge,
and jury are unaware of the lack of truthfulness or credibility and are thus unable to serve in their
intended “check and balance” roles.
In 1978, a man was shot and killed at a bathhouse in Los Angeles County. Several months after the
murder, a man contacted police and implicated Oscar Lee Morris in the murder. The man had known
Mr. Morris since childhood but they had recently had a falling out. The case was mistakenly closed in 1979
but in 1982, after being arrested for auto theft, the man again told police that Mr. Morris had committed
the murder. Mr. Morris was serving an unrelated eight-year prison sentence at the time. At trial the
witness testified that he drove Mr. Morris to the bathhouse and gave him the murder weapon. The
witness stated that Mr. Morris had told him that he had to kill a homosexual. The witness also testified
that he had received nothing in return for his testimony. After a five-day jury trial, thirty-seven-year-old
Mr. Morris was found guilty of both the murder and robbery. He was sentenced to death in 1983.
It was later discovered that the witness had received considerable leniency for testifying: His auto theft
charge had been reduced, his sentence for a parole violation had been terminated, and the prosecutor
had written two letters on his behalf asking a fellow prosecutor and the parole board to grant him leniency

FOOTNOTES
84	 In August of 2011, California Governor Jerry Brown signed SB 687, which requires that informant testimony be corroborated by other evidence. With the
passage of SB 687 defendants can no longer be convicted solely on the basis of informant testimony. See, “Unreliable Witness,” Editorial, Los Angeles
Times, August 11, 2011, http://articles.latimes.com/2011/aug/11/opinion/la-ed-informants-20110811.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 47

SECTION 2 | THE CAUSES OF INJUSTICE

in reward for his testimony against Mr. Morris. None of this information was made available to the
defense. Mr. Morris appealed his conviction and sentence on multiple grounds, including the withheld
information about the informant. The robbery conviction and death sentence were reversed, but
the murder conviction remained. The court acknowledged that the prosecutor violated Mr. Morris’
Constitutional rights when he withheld information regarding the leniency given to the witness,
but decided that the violation was harmless. The case was remanded for resentencing and Mr. Morris
was resentenced to life plus two years.
In 1997, the witness against Mr. Morris recanted his testimony and stated that he had fabricated the
entire story. Mr. Morris then filed a habeas petition and his conviction was reversed based on the
false testimony. The district attorney chose not to retry Mr. Morris and he was released in 2000 after
spending sixteen years in prison, six of them on death row.

In 1979 a man was killed in Long Beach, Los Angeles County. One of the witnesses was shown a picture
of Thomas Lee Goldstein and identified him as the potential shooter but noted that he was not certain.
Mr. Goldstein did not match the witness’s description of the perpetrator. Despite the lack of evidence,
29 year-old Mr. Goldstein was arrested and placed with a jailhouse informant named Eddy Fink, who later
told police that Mr. Goldstein had confessed. There was no physical evidence that tied Mr. Goldstein
to the crime. Mr. Goldstein went to trial. The informant testified against Mr. Goldstein, and testified
that he (the informant) received no benefit for his testimony and that he had never received benefits
from cooperating with the police. In 1980 a jury found Mr. Goldstein guilty of the murder and he was
sentenced to 27 years to life in state prison. Mr. Goldstein filed numerous unsuccessful appeals.
In 1990 a Grand Jury investigation found that the Los Angeles District Attorney’s Office had regularly
presented false testimony by jailhouse informants from 1979 to 1990. In 1998 Mr. Goldstein filed a federal
habeas petition claiming that the prosecution withheld Brady evidence by failing to reveal that the jailhouse
informant received benefits in exchange for his testimony. Mr. Goldstein also argued that there was Brady
evidence related to the eyewitness identification of Mr. Goldstein, because the police had coached the
witness by pointing out Mr. Goldstein in the photo spread during the identification. Mr. Goldstein also
alleged that his attorney had been ineffective.
In 2000, the eyewitness recanted his testimony. He stated that he did not initially recognize anyone
in the lineup but police coached him into pointing out Mr. Goldstein. In 2002 the court agreed with
Mr. Goldstein and ordered him released or retried. The district attorney initially filed new charges and
Mr. Goldstein remained in custody, but the district attorney eventually decided to dismiss the charges.
Mr. Goldstein was released after serving 24 years in prison.
Mr. Goldstein filed a civil suit against the city and county, the police department, the district attorney
and his chief deputy. The California Supreme Court ruled that the district attorney and his chief deputy
had absolute immunity and could not be sued, but Mr. Goldstein settled with the city of Long Beach for
almost $8 million.

48 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

Fourth Amendment Violations
$12 million overall (5% of the total cost)
$78,377 average cost per error
19% of the errors

•	 Fourth Amendment violations occur when the police violate the defendant’s Constitutional right
to be free from unreasonable search or seizure, such as when the police search without
a warrant.
•	 Fourth Amendment violations occurred overwhelmingly in drug cases (73%), with a smaller
number in weapons cases (16%).
•	 Fourth Amendment violations accounted for 19% of the errors in our sample, but only 5% of the
total cost.
•	 The total cost associated with Fourth Amendment violations was $12 million.
The Fourth Amendment of the United States Constitution protects individuals against unreasonable
searches and seizures. When the police violate a person’s Fourth Amendment rights, the court prohibits
the government from using the illegally obtained evidence against that person even if the person
committed the crime. This often results in dismissal of the case. Dismissals due to Fourth Amendment
violations have been called “technicalities,”85 but framing the debate in this manner focuses on the costs
of the remedy (the exclusion of illegally obtained evidence) as opposed to its benefits (the protection
of each American’s right of privacy).86 Moreover, “if a rule is even subconsciously viewed as merely a
technicality, the courts will far more easily let it bend to countervailing concerns and will defer to other
legal actors’ judgments about whether the rule has been met or requires an exception. To avoid that
result in the area of search and seizure law, a substantive vision of the Fourth Amendment’s value to our
republic must replace the near-sighted view of mere technicalities.”87
Whether one values Constitutional enforcement or bemoans the suppression of evidence in these
situations, a more productive conversation centers on error prevention. The question is not how we
should or should not enforce Constitutional rights, but rather how we can ensure that police are properly
trained and supervised so that future violations do not occur. Doing so will actually aid the purposes
of law enforcement and ensure that individuals in possession of illegal materials are appropriately held
accountable for their acts, and that valuable police and prosecutorial resources are not spent on cases
tainted from the outset.

FOOTNOTES
85	 Andrew E. Taslitz, “Respect and the Fourth Amendment,” 94 Journal of Criminal. Law. & Criminology 15 (Fall 2003): 28-29.
86	 Andrew E. Taslitz, “Respect and the Fourth Amendment,” 94 Journal of Criminal. Law. & Criminology 15 (Fall 2003): 23-24.
87	 Andrew E. Taslitz, “Respect and the Fourth Amendment,” 94 Journal of Criminal. Law. & Criminology 15 (Fall 2003): 29.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 49

SECTION 2 | THE CAUSES OF INJUSTICE

Heriberto F. was stopped by police in Los Angeles County during the day on May 3rd, 2006, when officers
saw him riding his bicycle on the sidewalk and in the street against the flow of traffic. The officers asked
him to get off his bicycle and began a pat down search, a permissible activity only if there is reasonable
suspicion to believe that an individual is armed, dangerous, or committing a crime. After the pat down,
the officers asked Mr. F. if he had any weapons or needles; 29 year-old Mr. F. volunteered that he had
a knife in his jacket pocket that was draped over the handlebars of the bike. The officers arrested Mr.
F., then performed a full search which uncovered several small baggies of methamphetamine and drug
related paraphernalia. Mr. F. told the officers that he carried the knife for protection and that he sold
some of the drugs to his friend. Mr. F. was charged with possession of methamphetamine for sale.
He pleaded no contest and was sentenced to 180 days in jail and three years’ probation.
Mr. F. appealed and the court reversed his conviction. The court found that the officers violated Mr. F.’s
Fourth Amendment rights because they lacked reasonable suspicion to conduct the initial pat down
search. There was nothing about Mr. F.’s actions or behavior that gave rise to a reasonable suspicion that
he was armed or dangerous, and he should not have been searched simply for riding his bicycle on the
sidewalk. On June 18, 2007, the trial court dismissed the case after the prosecution announced that it
was unable to proceed because the evidence was suppressed.

50 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

Ruse Reports
Ruse reports are fake reports—that is, false reports generated by law enforcement—that are
sometimes used in police investigations to obtain confessions from suspects. In this legally
sanctioned procedure, police will create a report demonstrating that the person being questioned
has been implicated in the crime, then show the report to the suspect in an attempt to trick
the suspect into confessing his or her participation in the crime. Ruse reports might state that
fingerprints were found, that DNA was a match, or that other scientific findings point to the
suspect. Although ruse reports are permissible, their ability to increase the potential for error,
mistake, and false confessions is obvious.
In the case of fifty-four year old Michael Kerkeles in Santa Clara County, a ruse report caused serious
harm when the investigator gave a ruse report to the prosecuting attorney along with the rest of a
case file. A 22-year-old developmentally delayed woman had accused Mr. Kerkeles of rape. The woman
stated that the assault took place on a pink and peach colored blanket but investigators were unable to
obtain a coherent timeline from her. Her mother indicated that she was mentally only seven years old.
Prior to trial, the prosecution held a preliminary hearing where the judge was called upon to determine
whether there was sufficient evidence to proceed against Mr. Kerkeles. The prosecution lost two
preliminary hearings because the court found that there was insufficient evidence to hold Mr. Kerkeles
for trial.
At the third preliminary hearing, the district attorney introduced the ruse report into evidence.
The ruse report indicated that there was semen on the blanket where the alleged assault took place.
In fact there was no semen on the blanket, only some blood. After the third preliminary hearing, the
defense attorney noticed that he had two different crime lab reports and learned that the analyst
whose name was on one of the reports did not in fact exist. Faced with the falsified report, the district
attorney initially sought to enter into a plea deal with Mr. Kerkeles. Mr. Kerkeles refused and the
case was eventually dismissed prior to trial. Mr. Kerkeles filed a civil law suit and, in 2013, he won a
settlement of $150,000 plus legal fees. Mr. Kerkeles filed a petition of factual innocence but the District
Attorney opposed the petition and the petition was denied.88

FOOTNOTES
88	 Information on this case was gathered from the following sources: Kerkeles v. City of San Jose, 199 Cal. App. 4th 1001 (2011); Leslie Griffy,
“Sex case hinged on phony lab report,” The San Jose Mercury News, December 16, 2007, http://www.mercurynews.com/politics/ci_7736235?nclick_
check=1.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 51

SECTION 2 | THE CAUSES OF INJUSTICE

Inadequate Police Practices Before Trial
$21 million overall (10% of the total cost)
$620,832 average cost per error
3% of the errors

•	 Inadequate police practices before trial include obtaining confessions or statements in
violation of the Fifth Amendment, cases in which new evidence was discovered after
trial, and other inadequate or improper police practices.
•	 Pretrial inadequate police practices accounted for 3% of the errors, but 10% of the cost.
•	 Total cost associated with inadequate police practices was $21 million.

FIFTH AMENDMENT VIOLATIONS
The Fifth Amendment to the United States Constitution protects defendants from self-incrimination,
among other things. A violation of the Fifth Amendment is often referred to as a Miranda violation from
the Supreme Court case, Miranda v. Arizona 384 U.S. 436 (1966), which held that statements made by a
defendant in custody can be used against the defendant only if the police have previously warned the
defendant that he or she has a right to remain silent, that anything he or she says might be used against
him or her in court, and that he or she has a right to an attorney. Fifth Amendment violations include the
failure to give timely Miranda warnings, as well as the improper use of coercive interrogation tactics to
obtain a confession.
On December 25, 2007, in Los Angeles County, a security guard saw two people rummaging in debris
inside the perimeter of a fence near a partially demolished building. The security guard called the police.
An officer arrived and saw, but then lost sight of, Arturo M. Officers later found Mr. M. in a nearby pool
hall, handcuffed him, and told him that he was not under arrest. They then asked him what he had been
doing at the building. After further questioning, Mr. M. stated that he went into the building to get
some wire.
Mr. M., age 40, was found guilty by a jury of burglary and sentenced to 16 months in state prison.
On appeal, Mr. M. argued that the trial court erred in admitting his statements at trial because his
questioning constituted a custodial interrogation and therefore he should not have been questioned
without being advised of his Fifth Amendment Miranda rights. The court found that a reasonable person
in Mr. M.’s position would have concluded that he was not free to leave after being placed in handcuffs.
The appellate court reversed the conviction based on the erroneous admission of Mr. M.’s statements in
violation of his Miranda rights. On July 31, 2009, the trial court dismissed the case.

52 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

Convictions based upon a false confession are often reversed not because of a Fifth Amendment violation,
but rather because of other fortuitous errors that bring the false confession to light. Without other evidence
demonstrating that the confession is false or the conviction is illegal, a defendant’s contention that his
confession was false will generally not be sufficient to reverse his conviction. As with plea bargains,
therefore, it is impossible to know how often defendants confess to crimes they did not commit.
This is troubling given the fact that juries continue to give great weight to confessions, despite a growing
awareness that false confessions happen across a wide variety of cases.
In 1992, police in Los Angeles County found the dead bodies of four women believed to be prostitutes.
In January of 1993, police interviewed David Allen Jones about the murders. Mr. Jones, a thirty-two-yearold developmentally disabled man with an IQ of 62, confessed to the murders under police questioning and
was charged with four murders and a rape. There were no eyewitnesses, and Mr. Jones’ blood type did not
match the biological material found at the scene. Nonetheless, in 1995 a jury convicted him of one count
of murder, one count of rape and two counts of the lesser-included offense of voluntary manslaughter
(he was acquitted on one of the murder counts). He was sentenced to 36 years to life in prison. After his
conviction, more murders of women occurred in a similar area of Los Angeles. DNA testing was done on
two of the four rape kits in Mr. Jones’ case and the DNA matched another man who was a serial murderer
already in prison.
In 2004, the district attorney filed a habeas petition on Mr. Jones’ behalf, on innocence grounds.
The court granted the habeas petition with respect to the murder and voluntary manslaughter charges,
but could not set aside the rape charge through a habeas petition because Mr. Jones had fully served
his sentence. The court later vacated the rape conviction based on a Brady violation because the
prosecution had failed to disclose the rape victim’s rap sheet. In 2005, the trial court finally dismissed the
case entirely, freeing Mr. Jones after 19 years in prison. In 2006, the City of Los Angeles settled with Mr.
Jones for $720,000. The state Victims Compensation Government Claims Board granted Mr. Jones’ claim
and awarded him an additional $74,600.

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 53

SECTION 2 | THE CAUSES OF INJUSTICE

IMPROPER AND INADEQUATE POLICE PRACTICES
Improper and inadequate police practices cover a variety of investigative errors, including failing to find
evidence, ignoring evidence, or failing to follow leads for other suspects.
On March 10, 1983, in Los Angeles County, Bruce Elliot Lisker went to his parents’ home, looked through
a window, and saw his mother lying on the floor, injured. He broke into the house and called 911.
She was transported to the hospital and died there from the injuries. A police detective suspected Mr.
Lisker of the crime and arrested him at the scene. Although Mr. Lisker was 17, he was tried as an adult.
His first trial began in November 1984. In the middle of trial he pleaded guilty to murder as part of deal
to receive a juvenile disposition, which would have had him released from custody when he turned 26.
The court refused to accept the deal, and Mr. Lisker was tried as an adult in a second trial in October
1985. The evidence at trial included a bloody footprint that allegedly matched blood on Mr. Lisker’s
clothing, the fact that Mr. Lisker had a tumultuous relationship with his parents, and the allegation that
Mr. Lisker would not have been able to see in the house through the window. The detective did not
obtain an expert to examine shoe prints, did not test for fingerprints, did not obtain the telephone bill
to verify phone calls, and did not preserve the bloody entry rug. The prosecution alleged the motive
to be a robbery because Mr. Lisker’s father had given his wife $120 in cash the night before in their
son’s presence, and this cash was said to be missing from her wallet. In addition, a jailhouse informant
testified that Mr. Lisker had confessed to the murder.
During the investigation, Mr. Lisker’s father had mentioned to the detective that one of Mr. Lisker’s
friends had unexpectedly come by their home the day before the murder looking to earn some money
by doing some chores. The friend had been in and out of foster homes, mental institutions, and juvenile
hall. The detective looked into him but abandoned him as a suspect, and the court refused to allow the
defense to admit any evidence about the friend at trial.
Mr. Lisker was found guilty of second degree murder and sentenced to 16 years to life. He filed an appeal
and two state court habeas petitions, all of which were denied. In 2004, Mr. Lisker filed a federal habeas
petition, and in 2005 the court held an evidentiary hearing. The lack of proper investigation was critical
to the hearing. After a crime scene reconstruction, it was clear that Mr. Lisker could have seen inside
the window and could have seen the body, contrary to the detective’s trial testimony. A comparison of
bloody footprints from the crime scene with the shoes Mr. Lisker wore that day demonstrated that the
footprints were not Mr. Lisker’s. The court also found that the blood on Mr. Lisker was as consistent
with innocence as it was with guilt. In addition, the court heard from two officers who had responded
to the scene. One stated that he probably stepped in blood and left a footprint. The other stated that it
was possible that he had stepped in blood and left a footprint. Also at issue in the hearing was the fact
that the detective dismissed the friend as a suspect even though the friend had lied about where he was
at the time of the murder. The detective also did not discover that the friend had a significant criminal
record. The court also found that the allegedly missing money had been in Mr. Lisker’s mother’s wallet
and had never been missing. Finally, the court found that the informant was unreliable.

54 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

The court reversed the conviction due to ineffective assistance of defense counsel, a due process
violation from the use of false evidence produced by the detective, and cumulative error giving rise to a
due process violation. Mr. Lisker was released after spending 26 years in prison and the court dismissed
the case in September 2009.

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 55

SECTION 2 | THE CAUSES OF INJUSTICE

DNA
DNA testing has been a valuable tool in exonerating the wrongfully convicted,89 although common
expectations about both its availability and applicability are sometimes exaggerated. Indeed, in our
report DNA appears as a factor in only 14 of the 607 cases.90 Of those fourteen, eleven of the defendants
alleged that they were innocent, and nine of the fourteen were declared innocent by the court. Only eight
of the fourteen received a civil settlement, and only five of those eight were also awarded compensation
from the state Victim Compensation Government Claims Board (VCGCB). Two people received
compensation from the VCGCB but no civil settlement.
DNA is not itself a cause of the wrongful conviction, because the DNA is not usually in evidence at the
trial that originally results in conviction. (In fact, had the DNA evidence been introduced, the likelihood
of an acquittal in these cases was presumably substantial.) Rather, the errors causing the wrongful
conviction often include mistaken eyewitness identification, a failure by defense counsel to adequately
represent his or her client, a failure by the police to investigate, a judicial mistake in excluding evidence,
or the improper withholding of evidence by the prosecution. Subsequent to the conviction, the DNA
provides objective information that assists the courts in recognizing that an error occurred.
In the DNA cases in our sample, the DNA was either unavailable or untested at the time of the original
trial. In most of our cases the actual error that caused the unjust conviction was an eyewitness
misidentification. This is consistent with Innocence Project data, which has found that eyewitness
misidentification played a role in 72% of the convictions that were subsequently overturned as a result
of DNA testing.91

FOOTNOTES
89	 There have been 321 DNA exonerations in the United States. “DNA Exonerees Case Profiles,” Innocence Project, accessed December 19, 2014,
http://www.innocenceproject.org/know/.
90	 The Innocence Project does not include three people who are included in our Report.
91	 “Eyewitness Misidentification,” The Innocence Project, accessed December 19, 2014,
http://www.innocenceproject.org/understand/Eyewitness-Misidentification.php.
56 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 2 | THE CAUSES OF INJUSTICE

On January 10, 1991, a woman was discovered murdered in the guest bedroom of her home in Santa
Clara County. She had been bound and stabbed, and duct tape had been placed across her mouth.
Investigators were able to obtain fingerprints from the crime scene and matched latent fingerprints
from the duct tape to a drug dealer. The drug dealer was arrested on March 7, 2001. He initially denied
involvement in the murder and then confessed and implicated thirty-year-old Quedillis Ricardo Walker as
his co-participant. The dealer claimed that he and Mr. Walker had accompanied the victim to her house,
where two armed white men jumped out of a closet and assisted Mr. Walker in the killing. The dealer
claimed that he assisted Mr. Walker in the crime because Mr. Walker had threatened to kill him. During a
polygraph test, the dealer changed his story about the two white men being involved and said it was just
he and Mr. Walker. Mr. Walker denied any involvement and stated that he was with another woman at a
motel on the night of the crime. That woman denied being with Mr. Walker that night.
Mr. Walker and the dealer began trial together in August of 1991. Three weeks into the trial the dealer
struck a plea deal by testifying against Mr. Walker. Another witness testified that Mr. Walker was a violent
person and that he had previously threatened her with a knife and a gun. The woman from the motel
testified that she was with Mr. Walker for a three-day period including the day of the crime, but that she
had earlier lied because she was married and was afraid her husband would find out that she had been
with Mr. Walker. The only evidence linking Mr. Walker to the crime was his co-defendant’s testimony.
The jury convicted Mr. Walker of first-degree murder and on April 3, 1992, he was sentenced to 26 years
to life in prison. Mr. Walker filed several petitions for habeas corpus alleging a Brady violation due to the
prosecutor’s failure to disclose the plea agreement with his co-defendant, and ineffective assistance of
counsel. All were denied. Mr. Walker and his attorneys subsequently learned that, after the trial, five people
had come forward to say that another man, not Mr. Walker, had accompanied his co-defendant to the
victim’s home the night of the murder. They also learned that the woman who testified against Mr. Walker
had made a deal with the prosecutor to reduce her drug charges in an unrelated matter in exchange for
her testimony against Mr. Walker, but the prosecutor had not provided that information to Mr. Walker’s
attorneys at the time. And, new DNA evidence from a cigarette butt at the scene linked the cigarette butt
to a different person. In June 2003, Mr. Walker filed a new petition for habeas corpus based on newly
discovered evidence pointing to factual innocence. On June 16, 2003, the district attorney conceded that
Mr. Walker was factually innocent, that the prosecution’s key witnesses at trial had provided perjured
testimony, and that the prosecution failed to inform the defense that the prosecution had promised a
benefit to the witness in exchange for her testimony. The court granted the habeas corpus petition and
made a judicial finding of factual innocence. Mr. Walker was released from prison after serving 12 years.
He was awarded $409,500 from the California Victim Compensation and Government Claims Board and
received a $2,058,356 civil settlement.

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 57

SECTION 2 | THE CAUSES OF INJUSTICE

Forensic Errors
Although not categorized as a separate error in this report, forensic and laboratory errors can seriously
compromise the integrity of a conviction. Sometimes these errors become known when science changes
over time, as in the cases of George Souliotes and Cameron Todd Willingham. Mr. Souliotes was
convicted of first-degree murder and arson, and subsequently sentenced to life without parole, after
a 1997 apartment fire in a building he owned in California. Sixteen years later, in 2013, he was released
when new evidence called into doubt the science used to argue that the fires were in fact arson.
Mr. Willingham was similarly convicted of arson and murder after the deaths of his children in a
house fire in 1991, in Texas. As with Mr. Souliotes, Mr. Willingham’s conviction was based on scientific
methodology and expert testimony that has now been largely discredited. Unlike Mr. Souliotes,
Mr. Willingham was not able to fully refute the evidence in court, and he was executed in 2004.
Many now believe that the State of Texas executed an innocent man.92
Forensic errors can also arise through faulty lab analysis or bad actors, as in the recent scandals involving
the FBI’s microscopic hair comparison unit and the San Francisco drug lab. In San Francisco, a crime
lab technician admitted to stealing cocaine from the lab, causing the dismissal of hundreds of cases.93
With the FBI, it was recently disclosed that the unit overstated forensic matches in favor of the
prosecution in over 95% of 268 trials examined. Thirty-two of those defendants were sentenced to death,
and 14 of the 32 died in prison or have already been executed. The analysis included five California cases,
three of which were death penalty cases. More cases are being investigated.94
The next section discusses the types of costs that gave rise to the cost calculations discussed earlier:
incarceration, trials and attorneys and courts, claims against the state, and lawsuits
and settlements.

FOOTNOTES
92	 “A conviction up in smoke,” Mora Dolan, Los Angeles Times, May 31, 2010, http://articles.latimes.com/2010/may/31/local/
la-me-arson-20100531; Trial by Fire, The New Yorker, Sept. 7, 2009, http://www.newyorker.com/magazine/2009/09/07/trial-by-fire; Report of the Texas
Forensic Science Commission, Willingham/Willis Investigation, April 15, 2011, http://www.fsc.state.tx.us/documents/FINAL.pdf.
93	 “Former SF Lab Technician Sentenced for Cocaine Possession,” CBS SF Bay Area, July 19, 2013,
http://sanfrancisco.cbslocal.com/2013/07/19/former-sf-crime-lab-technician-sentenced-for-cocaine-possession/.
94	 “FBI admits flaws in hair analysis over decades,” Spencer S. Hsu, The Washington Post, April 18, 2015, http://www.washingtonpost.com/local/crime/fbioverstated-forensic-hair-matches-in-nearly-all-criminal-trials-for-decades/2015/04/18/39c8d8c6-e515-11e4-b510-962fcfabc310_story.html.
58 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 3 | THE COSTS OF INJUSTICE

SECTION 3:

THE COSTS
OF INJUSTICE

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 59

SECTION 3 | THE COSTS OF INJUSTICE

This section explains the bases for the cost calculations in this report, separating them into costs of
incarceration, costs for court proceedings, costs for state compensation, and costs for lawsuits and
settlements. It ends with an analysis of the potentially substantial costs that we did not include in
our calculations.
For the 607 failed prosecutions profiled in this report, taxpayers paid over $220 million.94 This money—
which represents only the direct costs we can definitively show—reflects the prosecution, defense,
incarceration and compensation of the 607 people in our sample, including:
•	 $80 million in incarceration costs for time spent in jail and prison;
•	 $68 million for trials and appeals, including prosecutors, defense attorneys, judges, trial courts
and appellate courts;
•	 $5 million awarded by the California Victims Compensation and Government Claims Board
(VCGCB) as compensation for wrongful imprisonment; and
•	 $68 million to settle lawsuits brought by these individuals against California counties and cities.

Figure 10: Failed Prosecutions Profiled in this Report
$5 million Compensation for
Wrongful Imprisonment
$80 million Incarceration Costs
$68 million Trials and Appeals

N = 607

$68 million Lawsuit Settlements

FOOTNOTES
94	 $221,601,730, adjusted for inflation to 2013 dollars.
60 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 3 | THE COSTS OF INJUSTICE

At its most basic level, this equates to over $360,000 for each of the 607 people in our sample, although
the actual estimated costs arising from each faulty conviction range from $4,000 to over $10 million
per person.
For perspective, if California and its cities and counties had $220 million today, the funds could purchase:
•	 A year of K-12 instruction for almost 40,000 children;96
•	 117 million school lunches;97
•	 Testing of 214,000 rape kits;98
•	 2,467 additional California Highway Patrol Officers;99
•	 2,300 additional police officers; or
•	 1,836 additional firefighters.100
The cost numbers in this report reflect only those costs we could locate and confirm; they are
undoubtedly an incomplete picture. Moreover, we have not calculated additional costs such as lost
wages, which would require both the knowledge of each individual’s employment prior to their
conviction and a model to calculate the negative impact on future earnings of the conviction. We have
also not attempted to calculate the less quantifiable costs imposed on these individuals, including the
impact on their families and the impact on their mental health. Although perhaps less quantifiable,
these costs are substantial and ongoing, and should not be ignored.

FOOTNOTES
96	 Average current spending for instruction per pupil in California by Elementary and Secondary School Systems is $5,595.78 (adjusted for 2013).
Mark Dixon, “Public Education Finances 2012,” United States Census Bureau, 22 May 2014, Table 8,
http://www2.census.gov/govs/school/12f33pub.pdf.
97	 Average cost per school lunch is $1.89. California Department of Education, “Food Programs,” March 2014, School Nutrition Program (SNP)
2012-2013 County Profile, Table Lunch 12-13, http://www.cde.ca.gov/ds/sh/sn/#annual.
98	 Average cost for processing a rape kit is $1,035.65 (adjusted for 2013). Susan Price, “Rape Kits Testing, Backlogs, and Model Statutes,”Office of
Legislative Research, September 1, 2011, http://www.cga.ct.gov/2011/rpt/2011-R-0260.htm.
99	 California Highway Patrol employee’s average wage was $89,825 in 2013. California State Controller’s Office, “Government Compensation in California:
Highway Patrol 2013,” July 14, 2014, http://www.publicpay.ca.gov/Reports/State/StateEntity.aspx?fiscalyear=2013&entityid=3743.
100	 California police officers made, on average, $96,400, including overtime, incentive pay and retirement payouts during 2012, according to a Sacramento
Bee analysis of new data from the state controller’s office. Firefighters and engineers earned, on average, $120,700. Average pay for police captains
across the state was $164,800; for fire captains, it was $148,200. Phillip Reese, “See What California Cities Pay Police, Firefighter,” The Sacramento Bee,
October 20, 2014. http://www.sacbee.com/site-services/databases/article2573210.html.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 61

SECTION 3 | THE COSTS OF INJUSTICE

Incarceration, Trials, and Appeals: $148 million
The 607 individuals in our sample spent 797,794 days (2,186 years) in prison and jail, at a cost of $80
million dollars.101 In addition, we estimate an additional $68 million for the cost of prosecution, defense,
trials, and appeals. These costs are incomplete and they are at best estimates—but when estimates rise
to $148 million, policymakers and advocates should take action.102
The $80 million attributed to incarceration reflects the actual days in custody for each individual,
calculated using court records. The number of days per year was multiplied by the cost per day,
as reflected in county and state budget documents for the year in which the person was incarcerated.103
Totals were then adjusted for inflation to 2013.
The $68 million attributed to prosecution, defense, trials and appeals reflects court resources such as the
judge, clerks, bailiffs, and other judicial personnel, a prosecutor and, usually, a public defender or countypaid attorney for the defense. Our methodology for these costs is fully set out in Appendix C. In brief:
•	 For court costs associated with trials, mistrials, hung juries and plea bargains, we used the
California Administrative Office of the Courts’ 2002 Judicial Workload Assessment104 to ascertain
the average number of court minutes by type of case (homicide, property, drug, and other)
in the trial courts, both for cases that went to trial and for cases that resolved by plea prior to
trial. We used the same study to determine the average cost of a court day and a court minute.
The number of minutes multiplied by the cost per minute formed our estimate for trial court
costs, differentiated by type of crime and whether the case went to trial or plead guilty. For cost
purposes, trials, mistrials and hung juries were all considered to be trial cases despite the fact that
the jury does not reach a verdict in mistrials and hung juries.

FOOTNOTES
101	 In many of our cases the exact date on which the person entered custody and the exact date on which they left custody were unknown. To be conservative,
if the dates were unknown we used as a starting date the first court date where the defendant appeared in custody (although it is almost certain that he
or she was in custody for at least a day prior to the court appearance and potentially much longer). For an ending date, we used the last entry in the court
file where the defendant was in custody; this was often the date on which the court ordered dismissal of the charges (although it is almost certain that
the defendant was in custody for some number of days after the court issued its order). On rare occasions we did not have a release date, and we knew
that the release date occurred prior to the final court date. This was the only situation in which we would fill in the missing data. In those rare situations,
we calculated the number of days left to serve after the date on which the sentence was imposed. This may over-count the actual time in custody for these
defendants, as some of them will have been released prior to the calculated date due to good time credits.
102	 If these 797,794 days in prison and jail had all been imposed in 2013, the custody time alone would have cost taxpayers $120 million. The estimate of
$120 million is based on 173,065 county jail days times $114, plus 619,240 state prison days times $161.14. The prison calculation is based on CDCR’s
average cost per person per day in 2013, as determined by CDCR in its budget. The $114 a day estimate for County Jail is taken from the Public Policy
Institute of California, “Just the Facts: California’s County Jails,” (June 2013), http://www.ppic.org/content/pubs/jtf/JTF_CountyJailsJTF.pdf.
103	 For state prison cost, we used the average annual cost per inmate published by the California Department of Corrections and Rehabilitation (CDCR).
CDCR calculates this cost by dividing total annual cost by number of inmates. Some have argued that marginal cost of an additional inmate is lower than
CDCR’s estimate because so many incarceration costs are set and those costs may not change with the addition or subtraction of one person. We adopt
the methodology and calculations used by CDCR. The methodology by which we determined daily County
Jail cost is set out in Appendix B.
104	 Brian Ostrom et al “California Judicial Workload Assessment,”National Center for State Courts, May 2002,
http://cdm16501.contentdm.oclc.org/cdm/ref/collection/ctadmin/id/1503.
62 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 3 | THE COSTS OF INJUSTICE

•	 For attorney costs associated with trials, mistrials, hung juries and plea bargains, we used state
budget documents to calculate the cost of a fully loaded public defender hour and the cost of
a fully loaded district attorney hour. We then used a 2003 study by the San Francisco Office
of the Controller to ascertain the number of out-of-court minutes spent by public defenders
on the different types of cases, and assumed the same number of out-of-court minutes for
district attorneys. Finally, we multiplied the number of out-of-court minutes by the hourly cost
to estimate an average cost for the out-of-court work of the public defender and the district
attorney, by type of crime.
•	 For attorney costs associated with appeals, we used a three-year per-case average cost for courtappointed appellate counsel in California, obtained from the California Administrative Office of
the Courts. We assumed that same cost to estimate the prosecution side.
•	 We were unable to develop a methodology by which to estimate the appellate court costs for
appeals and post-conviction challenges, so those costs are unaccounted for.
•	 All costs were adjusted for inflation to 2013 dollars.
Developing a full methodology to ascertain these costs was beyond the scope of this report, particularly
because these types of estimations have not been attempted previously. As with any pioneer effort,
they will undoubtedly improve over time and with scrutiny. That said, we have attempted to keep our
estimates conservative at every level and we believe that we are substantially underestimating actual
costs. For example, we based our estimate for the cost of appellate counsel on three years of court data
demonstrating the amounts paid to court-appointed appellate counsel between 2010 and 2013. Over
those three years, court-appointed appellate counsel handled over 15,000 appeals in California; the
statewide average attorney cost per case was $9,171.19 and we applied that average to every case in our
sample. Court records make clear, however, that actual numbers are often far greater. In one of our cases,
for example, the true cost of defense counsel alone was over one million dollars.105 Due to limitations
in the data accessible on these cases, no reliable case-specific calculation exists for appellate counsel
and we have thus used the more conservative average per-case estimate across the board, despite the
virtual certainty of a substantial underestimation of the actual costs. We have also excluded the often
substantial costs of expert witnesses, as no data was available from which the costs could be estimated.106

FOOTNOTES
105	 Mr. Atkins’ counsel was not court-appointed, but he sought attorneys’ fees in the lawsuit that followed his wrongful conviction, and the court ordered that
the county pay Mr. Atkins’ attorneys $1,368,834 in fees and costs. Atkins v. Miller, Order Granting Plaintiff’s Motion For Attorneys’ Fees and Costs, No. CV
01-01574 DDP, C.D. Cal., August 27, 2007.
106	 In one recent case from Idaho, the expert witness charged $6500 a day.
http://www.kboi2.com/news/local/Murder-trial-costs-tallied-after-baby-sitters-acquittal-306982321.html?mobile=y
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 63

SECTION 3 | THE COSTS OF INJUSTICE

California Victims Compensation and Government Claims
Board (VCGCB): $5 million
Fifty-eight of the 607 individuals in our sample filed claims for compensation with the California Victims
Compensation and Government Claims Board (VCGCB).107 Of the 58, only fourteen (24%) were granted
compensation despite the fact that many of those who sought compensation were unquestionably
wrongfully convicted. Eight of the 58 claims were still pending as of February 2015, and the remaining 36
claims (62%) were denied. The money awarded to compensate those fourteen people for their wrongful
convictions and incarcerations totaled over just $5 million.
Although only 14 people in our sample received compensation from the state entity charged with
providing compensation for the wrongfully convicted, 56 people in our sample successfully filed lawsuits
against the counties and cities that led the wrongful prosecutions. (These lawsuits and settlement are
discussed in the next section.). The remarkably low compensation rate is likely due to the substantial
procedural hurdles raised by the VCGCB prior to 2013. Prior to 2013 these hurdles, combined with the
discretionary power of the VCGCB hearing officers, often resulted in the denial of compensation even
in cases with judicial rulings of factual innocence and clear wrongful incarceration.108 For example, in
four out of the 36 cases that were denied compensation in our sample, a different court had earlier ruled
that the applicants were factually innocent, only to have the VCGCB reach a different result. In another
nine of the denied cases, the VCGCB refused compensation because the hearing officer found that the
applicant had somehow contributed to his arrest or conviction, despite the fact that the conviction had
been overturned and invalidated by the court in the criminal proceedings.
California law allows anyone erroneously convicted of a felony to file a claim for government
compensation if he or she was incarcerated in a state prison as a result of that erroneous conviction.109
The VCGCB, originally created in 1965, handles those claims. To apply for compensation, a person must
submit an application and provide information about his or her conviction and dates of imprisonment.110
He or she must provide facts that show that the crime charged was either not committed at all or was
not committed by the applicant. Each applicant must also provide a pecuniary injury statement that
demonstrates financial loss as a result of the erroneous conviction and imprisonment.111

FOOTNOTES
107	 VCGCB compensation is only available for an individual incarcerated in a state prison, not for exonerated individuals who were prosecuted and
incarcerated by the federal government.
108	 Although difficulties in compensation are not unique to California, the state’s compensation system falls behind states like Texas, that have sought to
improve both the speed and the quality of their compensation programs.
http://www.latimes.com/nation/la-na-exoneree-national-20141220-story.html#page=3; http://spreadsheets.latimes.com/texas-compensation-payments/.
109	 California Penal Code Section 4900 et seq.
110	 “Erroneously Convicted Person Claim Form,” California Victims Compensation and Government Claims Board, accessed December 19, 2014,
http://www.vcgcb.ca.gov/docs/forms/claims/PC4900ClaimForm.pdf.
111	 Erroneously Convicted Person Claim Form.
64 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 3 | THE COSTS OF INJUSTICE

California Penal Code Section 4900 originally capped the amount of any claim at $10,000 per person.
As a practical matter, this acted as a substantial disincentive to file a claim since the maximum reward
was less than the likely legal costs needed to pursue the claim. In 2000, Governor Gray Davis signed a
bill amending the compensation scheme to allow compensation of $100 per day for incarceration in
state prison. Even after the 2000 amendment, however, substantial hurdles remained. Claimants were
subject to a two-year statute of limitations after release to file a claim for wrongful incarceration, and
the law limited compensation to only time spent in state prison, leaving aside the often substantial time
spent in county jail awaiting the resolution of the case. For example, if a wrongfully convicted person
spent two years in county jail awaiting the resolution of his case, and two years in state prison before the
wrongful conviction was overturned, he could be compensated only for the two years in state prison.
The VCGCB also had complete discretion to decide whether someone was eligible for compensation,
regardless of any earlier court ruling. This meant that VCGCB hearing officers were permitted to ignore
the original triers of fact (court and jury), disregard court findings of factual innocence, and make their
own credibility determinations. The law also allowed the VCGCB to deny compensation if it found that
the claimant had contributed to his or her arrest or conviction in any way.112
On July 26, 1978, nineteen-year-old Longino Acero was convicted in Santa Clara County of committing
a lewd act in public with an adult woman, a violation of California penal code section 647(a). A clerical
error was then made in his file and his crime was listed as penal code 647a (child molestation) instead of
647(a) (lewd act in public). The crime for which Mr. Acero was convicted does not require sex offender
registration, but the crime that was incorrectly recorded does. In 1978, Mr. Acero was incorrectly
told that he was required to register as a sex offender. He registered in 1978, but in 1994, 2001, 2002,
and 2003 he did not register and, as a result, he was charged with the crime of failing to register and
imprisoned for 465 days. He was repeatedly told by his public defenders and by government officials
that he was required to register, and on that basis he pled no contest to the felony of failing to register.
Finally, in 2005, he received a letter from the Santa Clara Police Department informing him that he was
not required to register as a sex offender. On March 28, 2006, with the support of the district attorney,
the court made a finding of factual innocence regarding the crime of failure to register. Mr. Acero filed
a claim before the VCGCB to receive compensation for the 465 days he spent in prison. The VCGCB
denied compensation because it found that Mr. Acero had contributed to his incarceration by pleading
no contest to the crime. Mr. Acero argued that he believed he had to register because his attorneys and
the court told him so, and therefore he had pleaded no contest. The Board denied him any compensation
and stated that it did not find his testimony credible, despite the fact that the Police Department and
the district attorney acknowledged telling him that he had to register, and despite the fact that the court
subsequently found him factually innocent of the crime.

FOOTNOTES
112	 The VCGCB generally used this provision to exclude anyone who had pled guilty, even if the guilty plea had been vacated and the conviction overturned.
A plea of guilty was considered a way in which the defendant contributed to his conviction and incarceration.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 65

SECTION 3 | THE COSTS OF INJUSTICE

On October 13, 2013, Governor Jerry Brown signed California Senate Bill 618, which attempted to address
many of these concerns.113 Under the new law, applicants can be compensated for time spent wrongfully
incarcerated in county jail as well as time in state prison. Factual findings and credibility determinations
made by a different court are now binding on the VCGCB, and a finding that new evidence points
unerringly towards innocence will likewise be binding on the VCGCB. In such cases the VCGCB will
be required to recommend compensation without holding its own separate hearing. The new law also
removes the requirement that applicants prove that they did not intentionally contribute to their arrest
or conviction.114 This means that people who pled guilty at some point in their case are not disqualified
from receiving compensation. The new law also places some parameters on the Board to ensure that
decisions are made in a reasonable time frame. Governor Brown continued on the path toward reform in
2015, when he signed SB 635 and raised the compensation rate for the wrongfully convicted to $140 a day.
The cases in our sample were all decided under the old law, although eight individuals from our sample
had pending claims before the VCGCB at the time of this report’s writing. The impact of the new law
remains unclear, though its passing sends a hopeful signal that fair compensation will be provided to
those individuals that the State of California has wrongfully convicted and incarcerated.

Settlements: $68 million
Cities and counties in California spent an additional $68 million (adjusted for inflation) to settle
lawsuits brought by the individuals in our sample as a result of their unjust and illegal convictions, and
many of the individuals in our sample have unresolved lawsuits that are expected to settle for many
millions of dollars.
Jeffrey Deskovic, who was awarded over $13 million after 16 years in prison in New York for a rape and
murder that he did not commit, said “I would be willing to not only give the money back, I’d be willing
to go into debt for that amount of money, maybe even double it, to have had my years back and had a
normal life.”115 This sentiment was mirrored by Obie Anthony, who was wrongfully convicted of murder
in Los Angeles County and who received an $8.3 million settlement. “The money will never make up for
it,” said Anthony, who was only 19 when he was convicted.116

FOOTNOTES
113	 California State Senate Bill 618. Approved by Governor, October 13, 2013.
http://leginfo.legislature.ca.gov/faces/billNavClient.xhtml?bill_id=201320140SB618.
114	 As of June 2015, however, the application forms available on the VCGCB website continue to require the applicant to affirm that he or she did not
contribute to his arrest or conviction. “Erroneously Convicted Person Claim Form,”
http://www.vcgcb.ca.gov/docs/forms/claims/PC4900ClaimForm.pdf.
115	 Jeffrey Deskovic, quoted in Saski De Melker, “Four wrongfully convicted men, four very different outcomes,” PBS NewsHour, November 9, 2014,
http://www.pbs.org/newshour/updates/life-exoneration/.
116	 http://abc7.com/news/la-to-pay-$83-million-settlement-to-wrongly-convicted-man/656569/
66 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 3 | THE COSTS OF INJUSTICE

Of the 607 cases examined, 56 led to lawsuits and settlements. Certain types of convictions were
more likely to lead to lawsuits and settlements. Faulty homicide convictions, for example, represent only
15% of the cases in our sample but account for 41% of the settlements. In contrast, drug cases represent
30% of our sample, but only 16% of the settlements went to those individuals (see Figure 11).117
The reasons for this disproportionality are unknown, but may be related to the types of error seen in the
faulty homicide convictions. It may also be because the true perpetrator can sometimes be found after
a faulty homicide conviction, thus eliminating doubt about whether the conviction was wrongful and
strengthening the claim of the falsely convicted individual.
Figure 11: Distribution of Settlements
Failure to Register as a Sex Offender
Forgery / Fraud / Embezzlement
Robbery
Rape
Assault
Possession of Weapons
Lewd Act on a Minor
Property Crime
Other
Homicide
Possession of Drugs
0

5%

10%

% of Database

15%

N = 607

20%

25%

30%

35%

% of Settlements

40%

45%

50%

N = 56

FOOTNOTES
117	 These numbers could change if, for example, a large number of defendants had their convictions overturned due to a lab scandal.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 67

SECTION 3 | THE COSTS OF INJUSTICE

Costs and Cases Not Captured in this Report
Our methodology excluded an unknown—and potentially enormous—number of errors and resulting
costs because we included only felonies, and then only those felonies in which (a) the defendant was
convicted of the felony, (b) the conviction was reversed, and (c) the case was ultimately dismissed or
the defendant was acquitted on retrial.118 This section identifies some of the areas that were beyond the
scope of this report, including:
•	 Cases that never reached conviction;
•	 Errors and reversals that were followed by a guilty plea;
•	 Errors that were found to be legally harmless;
•	 Misdemeanors; and
•	 Juveniles.
Each of these areas represents an additional potential well of injustice, with costs for taxpayers, victims,
and the individuals caught up in the failures of our criminal justice system.119

CASES THAT NEVER REACHED CONVICTION
Each year, tens of thousands of people are arrested and prosecuted, but released before trial or acquitted
at trial. Californians pay millions of dollars to incarcerate these individuals, none of whom are ever
convicted of a crime.
Almost 15% of California’s annual felony prosecutions are dismissed before trial or result in acquittal.
Put differently, roughly 37,000 Californians each year are caught up in the criminal justice system but

FOOTNOTES
118	 As noted earlier, we included a small number of cases in which the defendant took a plea deal on remand, but only if the defendant had already been in
custody for at least a year longer than the amount of time negotiated in the deal and only if the plea was to a lesser crime. For example, one defendant in
our database was in custody for almost ten years for serious felony convictions relating to a shooting, and after those convictions were dismissed he took
a deal for a one-year misdemeanor sentence. There are 46 such cases.
119	 This report does not specifically focus on the death penalty, but the presence of so many different errors highlights the problematic nature of having a
system that imposes death sentences. University of Michigan Law Professor Samuel Gross believes that over four percent of people sentenced to death
in the United States are probably innocent, while nationwide 117 people have been exonerated after being sentenced to death. (Michael McLaughlin,
“Shocking Numbers of People Sentenced to Death Study Finds,” The Huffington Post, April 28, 2014, http://www.huffingtonpost.com/2014/04/28/
innocent-death-penalty-study_n_5228854.html.) Governor Wolf of Pennsylvania cited these numbers when announcing a moratorium on the death
penalty in February 2015. (“Governor Tom Wolf Announces a Moratorium on the Death Penalty in Pennsylvania,” Pennsylvania Office of the Governor,
press release, February 13, 2015, http://www.prnewswire.com/news-releases/governor-tom-wolf-announces-a-moratorium-on-the-death-penalty-inpennsylvania-300035913.html.) A 2011 report found that the death penalty cost California over $4 billion dollars since 1978 and had resulted in only
thirteen individuals being executed, suggesting a substantial cost for little to no potential deterrent effect. (Arthur L. Alarcon & Paula M. Mitchell,
Executing the Will of Voters? A Roadmap to Mend or End the California Legislature’s Multi-Billion Dollar Death Penalty Debacle, Loyola of Los Angeles Law
Review 44 (2011): S41-S224, http://digitalcommons.lmu.edu/cgi/viewcontent.cgi?article=1030&context=llr.) A follow up study found that maintaining
the death penalty in California through 2050 will cost taxpayers $5.4 to $7.7 billion more than it would cost to impose life without parole. (Arthur L
Alcaron & Paula M. Mitchell, Costs of Capital Punishment in California: Will Voters Choose Reform in November?, Loyola of Los Angeles Law Review 46
(2012): S1-S34, http://digitalcommons.lmu.edu/cgi/viewcontent.cgi?article=2803&context=llr.)
68 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 3 | THE COSTS OF INJUSTICE

never convicted of anything.120 It is not known how much time each of these individuals spent in
custody prior to the dismissal or acquittal, but even if each of them spent only five days in custody—
a number that seems quite low—these individuals would cost the state over $20 million dollars annually
for incarceration alone. It is more likely that many of them spent far more than five days in jail before
their cases were dismissed or they were acquitted, since between 2006 and 2012 the average length of
stay in a California jail for individuals who were not sentenced was about 17 days.121 If each of these
37,000 individuals is spending 17 days in custody each year, the total cost to California taxpayers for their
incarceration is over $70 million a year.122
Each year, approximately 37,000 Californians are arrested and prosecuted, but released before any
conviction. If each of them is spending only 17 days in custody, Californians are paying $70 million a
year for their incarceration.

In 2005, twenty-four-year-old Edmond Ovasapyan was arrested for murder in Los Angeles County.
The mother of the victim said that Mr. Ovasapyan was not one of the perpetrators but that he did look
like one of them. After Mr. Ovasapyan had been in jail for 8 months, detectives collected DNA evidence
that led to another suspect and Mr. Ovasapyan was released. In 2008, Mr. Ovasapyan filed a civil rights
lawsuit against the city. The jury found that the officers unlawfully arrested and caused the malicious
prosecution of Mr. Ovasapyan. It also found that the officers lacked probable cause to arrest, misled the
district attorney, and withheld exculpatory information. In 2009 the jury awarded Mr. Ovasapyan $1.31
million dollars.123

In August 2005, Michael Walker was arrested for robbery of a Los Angeles County convenience store.
He was detained for 27 months because two police detectives withheld evidence of Mr. Walker’s
innocence. The charges were dismissed when the deception came to light, and a judge later found Mr.
Walker to be innocent. A jury subsequently awarded him $106,000 and the award was upheld by the
United States Supreme Court in May 2015.124

FOOTNOTES
120	 California Attorney General, “Crime in California 2013,” Tables 38 A and 38B. The six-year average of dismissal and acquittals from 2008 through 2013
was 36,657 a year. An average of $114 a day is used for jail costs. Public Policy Institute of California, “Just the Facts: California’s County Jails,”
June 2013, http://www.ppic.org/main/publication_show.asp?i=1061.
121	 Board of State and Community Corrections, Jail Dashboard Average Length of Stay,
https://public.tableausoftware.com/profile/kstevens#!/vizhome/ACJROctober2013/About.
122	 Calculated based on PPIC’s 2013 average cost of $114 a day for county jail in California.
http://www.ppic.org/main/publication_show.asp?i=1061.
123	 Melanie Hicken, “Glendale Will Pay Man Who Was Wrongly Detained In Murder Probe,” Los Angeles Times, January 18, 2011,
http://articles.latimes.com/2011/jan/18/local/la-me-glendale-settlement-20110118; http://justicedenied.org/issue/issue_42/1.3_Million_jd42.pdf
124	 “Supreme Court permits LAPD to be sued for concealing evidence,” David G. Savage, Los Angeles Times, May 18, 2015,
http://www.latimes.com/nation/la-na-supreme-court-lapd-suit-20150518-story.html.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 69

SECTION 3 | THE COSTS OF INJUSTICE

In an attempt to gather further information about systemic criminal justice costs for cases that do not
include conviction and reversal, we submitted requests under the California Public Records Act to each
of the 58 counties in California and to the thirteen largest cities in the state, seeking information on all
payments made to any individual who alleged wrongful arrest, prosecution, or conviction based on events
occurring between January 1, 1989 and 2013.125 We received information on 1,916 settlement or payments, the
vast majority of which were made between 2000 and 2012.126 Of the 1,916 settlements disclosed as a result
of our Public Records Act request, ninety-six were already in our sample set. However, 1,820 of them were
cases that we had not identified, potentially because those 1,820 cases never reached conviction. According
to the documents provided by the counties and cities, these 1,820 cases cost taxpayers an additional
$135 million for civil settlement and legal fees paid to settle claims of wrongful arrest, incarceration or
prosecution.127 (Responses to the PRA request are more fully detailed in Appendix G.)

GUILTY PLEAS ON REMAND
Of the over 1400 cases that we examined for potential inclusion, 645 were excluded because the
defendant pled guilty on remand and stayed in custody as part of the plea deal. In 177 of those cases, we
were able to locate information about the original sentence and the sentence imposed as part of the plea
bargain after reversal. The facts of these 177 cases indicate potential injustices worthy of further study.128
In 68 of the 177 cases that we excluded because the defendant pled guilty on remand, the defendant had
originally been sentenced to life in prison. In 40 of those 68 (60%), after the conviction was reversed
and the defendant entered into a plea bargain, the defendant no longer had a life sentence. This includes
seven cases where the original sentence was life without the possibility of parole. In one example, the
original sentence of life without possibility of parole was reduced to just eleven years after the reversal.
In another, a sentence of 88 years to life was reduced to 20 years, and in another a sentence of 100 years
to life was reduced to six years.
In 82 of the 177 cases that were excluded because the defendant pled guilty on remand, the defendant
was originally sentenced to prison or jail for a term less than life. The average reduction in those 82 cases
FOOTNOTES
125	 The methodology by which the responses were reviewed and our estimate reached is described in Appendix G. The PRA request sought: “(1) All documents
that indicate or authorize payment by [respondent] to any individual who alleged wrongful arrest, wrongful prosecution, and/or wrongful conviction based
on events occurring any time between January 1, 1989 and the present day. This request includes document in the possession of [respondent] that reflect
the full name(s) of each person receiving such compensation; the alleged factual basis for the claim; the identifying court case number and jurisdiction,
if a court case was associated with the request for compensation; copies of any court order(s) requiring that the County provide compensation; and
minutes of any meeting at which one or more employees, agents or representatives of the County approved payment of the financial compensation
referenced above. (2) All documents that reflect payment by [respondent] to outside legal counsel or for other professional services engaged to defend
[respondent] against claims raised in paragraph (1); and/or (3) All documents that display payment of monetary compensation to any individual between
January 1, 1989 and the present, in response to the individual’s allegation of wrongful arrest, imprisonment, or prosecution or related police misconduct.”
126	 As with our unjust conviction review, our inquiry into wrongful arrests and prosecutions sought information dating back to 1989. As a practical matter,
we received almost no information on settlements paid prior to 2000, and in most cases the response was even more limited due to the record retention
and analysis capabilities of the responding counties and cities.
127	 The total for all 1,916 was $217 million.
128	 There are undoubtedly far more than 177 cases in which a conviction was overturned and the defendant plead guilty on remand, but this data was not
easily accessible without accessing each file and reviewing it by hand, which would have been cost prohibitive. The examples presented here are merely
a sample of the potential injustices that a more complete study would likely uncover.
70 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 3 | THE COSTS OF INJUSTICE

was just under five years per case, but the total sum of years reduced in those 82 cases was 3,925 years.
In one example, the defendant was originally sentenced to 36 years and the sentence was reduced to six
years. In another, an 18-year sentence was reduced to four years.
In 2006 in Riverside County, a woman told officers that her bicycle had been taken; officers found the
bicycle sitting on the lawn of a nearby house where two men were sitting in a car. Carl Lee Mallet was
one of the men in the car. The woman identified him as the bicycle thief from at least 60 feet away with
a flashlight shining on him. Later, she was shown a picture of Mr. Mallet and said he did not look like the
thief. Then, she denied that he was the thief. At trial she testified that she had wrongly identified Mr. Mallett
because the lighting was poor and she was far away; she also stated that it was her ex-boyfriend who had
stolen the bike and that she was afraid of him because he had threatened to kill her. The jury was unable
to reach a decision, but returned to deliberations after the judge said “[Y]ou’ve got to understand that it
would be a real shame for the County of Riverside and you to pay for another trial just because you did
not put in enough effort.”129 When the jurors returned they indicated that they would like more time, but
because of scheduling difficulties the Judge excused two jurors and inserted two alternates. Only fifteen
minutes after the two alternates were sworn, the jury found Mr. Mallett guilty of second degree burglary.
The court sentenced him to 12 years in state prison. In 2009, the appellate court reversed the conviction
because of judicial misconduct, and in 2010 Mr. Mallett was released from custody. The district attorney
decided to refile the charges but offered Mr. Mallett a sentence of four years, which meant he would not
have to return to custody if he plead guilty. Mr. Mallett took the deal.

On September 24, 1997 in Los Angeles County, seven-year-old Eric O. was abducted from his elementary
school and molested. He told police that the perpetrator was an African-American man wearing a red,
white, and blue shirt. He described the home where he had been taken as an empty red and white
house, and his brother also said the man had been African-American. The police went to a red and white
house where they found Myron Howard, a forty-two year old man who said he was homeless and that
he had arrived at the house a few minutes earlier. Eric identified Mr. Howard, but his brother did not.
Mr. Howard’s defense attorney said that neither of the boys could identify anyone from the lineup, and
that it was only after some words were exchanged in Spanish between the detective and the boys that
eventually Eric said that one person looked familiar. After a nine-day jury trial, the jury found Mr. Howard
guilty and in 1999 the court sentenced him to 85 years to life. Six years later, in 2003, DNA testing
excluded Mr. Howard as a source of the DNA. The judge vacated Mr. Howard’s conviction, but ordered
that Mr. Howard remain in custody so that he could be tried again on the same charges despite the DNA
evidence. In November of 2004, after spending more than seven years in prison, Mr. Howard accepted
an offer by the prosecutor for a sentence of six years in prison and immediate release to his family. Mr.
Howard says that he pleaded guilty because he was told he could get out of prison if he did so.

FOOTNOTES
129	 People v. Mallett, 2009 WL 1879251, 2 (Cal. App. 4 Dist.) (unpublished). Mr. Mallett’s age at the time of arrest or conviction could not be determined.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 71

SECTION 3 | THE COSTS OF INJUSTICE

It is worth questioning a system where the same case can be worth life in prison without the possibility
of parole at the original sentencing, only to be decreased to 11 years in a plea deal after the original
conviction is overturned.130 The “black-box” of plea-bargaining, however, makes analyzing this issue and
constructing a thoughtful policy response a nearly impossible task.131

HARMLESS ERROR
The legal doctrine of “harmless error” allows appellate courts to reject a defendant’s appeal even
when error is present, on the basis that the error would not have changed the outcome of the case.132
To some degree our system needs the harmless error doctrine because “[i]f every error was ‘reversible
error,’ an extraordinary percentage of trial judgments would be reversed” and the system could not
function.133 At the same time, however, the harmless error doctrine allows error and misconduct to fester
unnoticed. Judges “send a message through our criminal justice system each time we reverse or remand
a conviction on the grounds that the police or prosecutors have violated a defendant’s individual rights.
Upon receiving such a message, the criminal justice process corrects itself accordingly. Thus, when we
shrink from our duty to overturn convictions in individual cases, we accomplish nothing less than
a subversion of the rules that we have devised to protect our shared values.”134
Criminal defendants in California file about 5,000 appeals a year, and approximately 95% of those
appeals are denied.135 Some of those denials are because the defendant did not demonstrate that an error
occurred. Others, however, are denied despite an acknowledgement that the error occurred, because the
error was deemed to be harmless. And while that error may be legally “harmless,” it still has an economic

FOOTNOTES
130	 Concerns about the great discretion of prosecutors, particularly as it relates to plea bargains, are not new. See The Economist, “The Kings of the
Courtroom.” Oct. 4, 2014, http://www.economist.com/news/united-states/21621799-how-prosecutors-came-dominate-criminal-justice-system-kingscourtroom; Bibas, Stephanos, “Plea Bargaining Outside the Shadow of Trial” (2004). Faculty Scholarship. Paper 924,
http://scholarship.law.upenn.edu/faculty_scholarship/924.
131	 The case of Weldon Angelos from Utah exemplifies some of the potential problems in our plea bargaining system: “The ability to procure a sentence that’s
far longer than necessary is a well-used item in some prosecutors’ tool kits. A case in point is the 55-year sentence a judge was forced to impose on
a man named Weldon Angelos, charged with marijuana trafficking and weapon possession. The government had offered him a 16-year sentence if he
pleaded guilty. When he refused, the prosecutor charged him in such a way that had he been convicted of every charge, he would have been sentenced to
105 years in prison. [. . .] Clearly, the government believed that the goals of punishment would have been fulfilled by a 16-year sentence. The additional
89 years it sought? Punishment for going to trial, plain and simple.” Mary Price ”Wanting A Trial By Jury Is Not A Crime. So Why Do We Treat It Like One?,”
Forbes Magazine, September 24, 2014,
http://www.forbes.com/sites/realspin/2014/09/24/wanting-a-trial-by-jury-is-not-a-crime-so-why-do-we-treat-it-like-one/.
132	 People v. Watson, 46 Cal. 2d 818 (1956), Chapman v. California, 386 U.S. 18 (1967). In California, a conviction can be upheld notwithstanding a
substantive or procedural error if a judge determines that “it is reasonably probable that a result more favorable to the appealing party would have been
reached in the absence of the error.” People v. Watson, at 836. This is a much broader standard than that required in federal cases, where a judge can
reverse the conviction only if he or she determines that the error was harmless beyond a reasonable doubt. Chapman v. California at 24. The impact of
both of these doctrines is the same, however: many defendants who have been subjected to illegal proceedings are not able to reverse their convictions
because the error is found to be legally “harmless.”
133	 Harry T. Edwards, A Conversation with Judge Harry T. Edwards, Washington University Journal of Law & Policy 16 (2004): 61, 71
134	 Harry T. Edwards, To Err is Human, But Not Always Harmless: When Should Legal Error Be Tolerated?, N.Y.U. L. Rev. 70 (1995): 1167, 1198-99.
135	 Judicial Council of California, “2013 Court Statistics Report: Statewide Caseload Trends 2003-2004 through 2011-2012,,” 69,
http://www.courts.ca.gov/documents/2013-Court-Statistics-Report.pdf.
72 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 3 | THE COSTS OF INJUSTICE

cost in time and money since, had the mistake or misconduct not existed, there would have been no
grounds for appeal, saving appellate court and attorney resources for more substantive matters.
There is no way to determine how many wrongful and faulty convictions are upheld under the harmless
error doctrine. One potential proxy is “Preventable Error: Prosecutorial Misconduct in California, 19972009” published in 2010 by the Veritas Initiative at Santa Clara University School of Law. The report
examined 707 substantiated findings of prosecutorial misconduct in California. In 548 (77%) of those
cases, the court determined that the misconduct was “harmless.” Only 159 of the 707 were deemed to
have overcome the harmless error standard.136 In other words, almost 80% of the cases of misconduct
were hidden by the harmless error doctrine. It is unknown whether this rate applies to other types of
error, but the matter remains nonetheless worthy of further study.

MISDEMEANORS
In general, about 80% of all criminal filings are misdemeanors, while only about 20% are felonies.137
The maximum sentence for a misdemeanor is one year in jail, and many misdemeanors are relatively minor
crimes which are resolved by probation or other non-custodial alternative sentences. Given the potentially
enormous pool of misdemeanors, and the difficulties in locating cases that are reversed and dismissed after
conviction, we could not include misdemeanors in our methodology. Yet, a faulty misdemeanor conviction
can often impact an individual’s life and clog our courts just as much as a faulty felony conviction.
The potential universe of injustice within misdemeanor convictions is worth examining.

JUVENILES
By law, juvenile court files are not available to the public, preventing their inclusion in our analysis.
There is no reason to believe, however, that the errors discussed in this report do not apply to juvenile
prosecutions in all the same ways that they apply to those of adults. If anything, inflicting these injustices
on juveniles should give rise to a louder call for reform. These failed prosecutions and faulty convictions
have the potential for an even greater negative impact on the defendant’s life, since these individuals are
torn from their families and communities even before they have reached adulthood.
A small window into these potential cases exists when juveniles are prosecuted as adults, because
the files then become public. Eight of the defendants in our sample were arrested and incarcerated as
juveniles, but tried as adults.

FOOTNOTES
136	 Kathleen M. Ridolfi and Maurice Possley, “Preventable Error: A Report on Prosecutorial Misconduct in California 1997-2009,” Northern California
Innocence Project, Santa Clara School of Law, October 2010, 3, http://digitalcommons.law.scu.edu/cgi/viewcontent. cgi?article=1001&context=ncippubs.
137	 These totals do not include infractions, which are far greater than either misdemeanors or felonies but which do not implicate potential jail or prison time.
Judicial Council of California, “2013 Court Statistics Report:” http://www.courts.ca.gov/documents/2013-Court-Statistics-Report.pdf.

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 73

SECTION 3 | THE COSTS OF INJUSTICE

In July of 2002 in Los Angeles County, Brian Banks was accused of raping his classmate. Mr. Banks was
a 16-year-old rising football star; he and the girl had had a consensual sexual encounter not involving
intercourse before she accused him of rape. In 2003, at the age of 17, he was charged as an adult.
The only evidence against him was the preliminary hearing testimony of the girl. Her testimony was
inconsistent with her prior report to police, no sperm was found on her body and no male DNA was
detected on her clothing. Faced with the prospect of a harsh sentence of 41 years to life in prison, in
2003 Mr. Banks decided to plead no contest in return for a 6 year sentence and the requirement that he
register as a sex offender for the rest of his life.
In 2006, Mr. Banks filed a writ of habeas corpus alleging inadequate legal counsel and alleging that
there was no evidence of rape. The court denied his petition, and Mr. Banks was released on parole
after serving five years in state prison. In 2011, the girl contacted Mr. Banks and recanted her story,
acknowledging that she lied because she did not want her mother to find out that she was sexually active.
She had been reticent to come forward sooner because she had sued the Long Beach Unified School
District alleging inadequate security and received a $1.5 million settlement. Armed with her recantation,
Mr. Banks filed another habeas corpus petition. The case was dismissed in May of 2012 without any
objection from the district attorney.
In 2015, Mr. Banks received $142,200 from the VCGCB. He had not yet settled his lawsuit as of the time of
this report.

74 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 4 | A NOTE ON GROUP EXONERATIONS

SECTION 4:

A NOTE ON GROUP
EXONERATIONS

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 75

SECTION 4 | A NOTE ON GROUP EXONERATIONS

The 607 unjust convictions profiled in this report thus far reflect individual, unrelated instances of errors
in the criminal justice system. There are other situations in which multiple individuals are arrested,
charged, or convicted through a systematic and related set of facts; 85 cases fitting that criteria are
included in our database. We refer to these interconnected injustices as “group exonerations.”
We are aware of several group exonerations that have occurred in California, involving allegations of
systematic police corruption and crimes against a substantial number of people whose arrests and
convictions were ultimately dismissed. The most prominent of these was the 2002 Rampart police
scandal in Los Angeles, in which a group of Los Angeles police officers admitted to falsely arresting or
accusing hundreds of mostly Latino residents. Two hundred and twenty-eight individuals received civil
settlements for the corrupt misconduct of the Rampart officers, while a report released by the City of Los
Angeles pointed to 156 felony convictions and 15 misdemeanor convictions overturned.138 Los Angeles
paid more than $78 million in verdicts and settlements related to the scandal, and total costs related to
the scandal and its review have ranged from $125 million139 to upwards of $1 billion dollars.140 The scandal
caused the Los Angeles Public Defender to review upwards of 8000 cases for error, a herculean task
required above and beyond their regular caseload.141 And, in an ironic twist, the settlement costs included
three $5 million payouts to officers of the LAPD who were wrongly enmeshed in the scandal and lost
their jobs, showing that substantial errors can occur even in the investigation of other errors.
Only slightly less prominent was the Riders scandal in Oakland, which led the City of Oakland to settle
over 119 cases of police misconduct for $10.9 million. While Rampart and Riders received the most
publicity, other group exonerations have happened in California amidst lesser fanfare. For example,
the District Attorney in Sacramento was forced to dismiss 79 criminal cases in 2010, most of them DUI
convictions, after it was discovered that the arresting officer had falsified his reports, offered perjured
testimony, and committed other errors.142 And in 2010, the San Francisco Police Department was forced
to temporarily close its crime lab after it was revealed that a technician had been taking drugs captured
by police and supplied to the lab for evidentiary testing. As a result, the San Francisco District Attorney
was forced to review over one thousand drug convictions, ultimately dismissing 701 cases.143

FOOTNOTES
138	 “Rampart Reconsidered: The Search for Real Reform Seven Years Later,” Appendix C, Rampart CRASH Scandal Scorecard, p. 1
139	 Young, Rick, “Frontline: The Outcome of the Rampart Scandal Investigations,” available online at
http://www.pbs.org/wgbh/pages/frontline/shows/lapd/later/outcome.html.
140	 Kaplan, Paul J., “Looking Through the Gaps: A Critical Approach to the LAPD’s Rampart Scandal,” Social Justice, Vol. 36, No. 1 (115), at 61 (citing
Pomerantz, Dorothy, “Final Cost of Rampart: $1 Billion,” Los Angeles Business Journal, February 28, 2000.
141	 Young, Rick, above at n. 138.
142	 “Sacramento County DA Cites Cop Misconduct in Dismissing 79 Cases,” Kim Minugh, Sacramento Bee, available at http://blogs.sacbee.com/crime/
archives/2010/09/sacramento-coun-34.html; “Sacramento Police Officer Brandon Mullock charged with 34 Counts of Perjury, Falsifying Police Reports,”
Kim Minugh, Sacramento Bee, republished at http://exceptionalcrimes.com/2011/05/19/officer-brandon-mullock/, accessed April 2, 2015.
143	 “Mistrial Declared for Woman At Center of SFPD Crime Lab Scandal,” ABC7 News, October 3, 2012, available online at http://exceptionalcrimes.com/
2011/05/19/officer-brandon-mullock/; Marina Neil, “Judge Asks DA to Drop Hundreds of Drug Cases, Says Gov’t Didn’t Check for Witness Issues,”
ABA Journal, April 16, 2010, available at http://exceptionalcrimes.com/2011/05/19/officer-brandon-mullock/, accessed April 2, 2015; communication
with the San Francisco Office of the Public Defender October 1, 2015.
76 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 4 | A NOTE ON GROUP EXONERATIONS

Although we collected group exoneration cases as we were building our database, we have not included
information on these individuals in the calculations made up to this point in the report.144 This is in
keeping with other observers of unjust outcomes in criminal cases. Moreover, including a large number
of similarly-situated police misconduct errors in our data and analysis would likely skew the remaining
data, as the characteristics of group exonerations, including Rampart, Riders, and the Tulia police scandal
in Texas, tend to be different than those of the 607 individual exonerations.145 For example:
Figure 12: Comparison of 85 Rampart Cases with 607 Cases in Report
85 Rampart Cases

607 Individual Cases

Average time from initial appearance to dismissal

3 years

4.5 years

% of violent crimes

7

30

% of drug crimes

67

31

% of cases with paid civil settlement

75

9

Settlement as % of total cost

89

33

In addition, with the Rampart scandal in particular it proved impossible to obtain accurate data. While
it has been published that as many as 228 people have received civil settlements from Los Angeles in
compensation for the violation of their rights related to the Rampart scandal, and there were additional
convictions overturned for individuals who did not sue the City of Los Angeles, we were ultimately able
to identify and confirm only 85 reversed convictions that met our criteria for inclusion in this report. In
attempting to get a precise count, we reviewed: (1) Russell Covey’s law review article, which provided a
list of Rampart exonerations and settlement payments based on the author’s primary review of Rampart
documents provided by the Los Angeles District Attorney’s Office;146 (2) a series of articles published
in The Los Angeles Times by Matt Lait and others, who observed court proceedings related to Rampart
and published roughly contemporaneous lists of individuals whose arrests or convictions were set aside;
and (3) information provided in response to a Public Records Request Act in which we sought amounts
paid in settlement of claims related to wrongful arrest, prosecution or conviction. There was variation
across all sources both in the names of those receiving payments and in the amount of payments that
they received. Across all sources of data, we identified 228 names and settlements totaling $79.8 million.
Once we went to the courts, however, we could locate only 85 court files corresponding to the names
and cases, with civil settlements totaling $54 million. The reasons for this are unknown. Deficient recordkeeping by the Los Angeles Courts played a role, as the County was unable to locate other requested case

FOOTNOTES
144	 We have included in our group of 607 several individuals whose convictions in Kern County on charges stemming from a supposed child sex abuse scandal
were ultimately reversed as individual convictions, rather than as a mass exoneration. For a detailed assessment of child sex cases in Kern County and its
effects, see Edward Hume, “Mean Justice: A Town’s Terror, A Prosecutor’s Power, A Betrayal of Innocence,” Simon & Schuster, 2012.
145	 Russell Covey, Police Misconduct as a Cause of Wrongful Convictions, Georgia State Law Review, 26 (2013).
146	 Russell Covey, Police Misconduct as a Cause of Wrongful Convictions, Georgia State Law Review, 26 (2013).
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 77

SECTION 4 | A NOTE ON GROUP EXONERATIONS

files in addition to the Rampart cases. In addition, some of the 228 settlements were paid to juveniles,
and those files are not available to the public. Other settlements were paid to people who were harmed
by the Rampart unit but never convicted of a crime due to the unit’s malfeasance, so they would not have
had a file in the criminal courts.
Ultimately, it is not the precise dollar amount paid or the number of cases confirmed, but rather
the inability to identify the costs and cases that is the problem. It is clear that the costs of the police
misconduct involved are huge, and that even a small number of corrupt police can have a massive direct
and indirect economic impact on the criminal justice system. It is equally clear, and perhaps equally
troubling, that after multiple internal and external investigations, we cannot quantify those costs with
any certainty. Without information systems that can link criminal court records to civil settlements and
records from specific compensation funds, it is impossible to accurately measure the costs of one of the
largest police scandals in the history of American law enforcement. Without such measurements, it is
difficult to prioritize thoughtful reforms that will prevent the next Rampart scandal. And if a scandal of
the magnitude of Rampart does not inspire such capabilities, it is difficult to envision what scandal will
be required to spur change.

78 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 5 | CONCLUSION: NEXT STEPS AND RECOMMENDATIONS

SECTION 5:

CONCLUSION – NEXT STEPS
AND RECOMMENDATIONS

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 79

SECTION 5 | CONCLUSION: NEXT STEPS AND RECOMMENDATIONS

This report covers a number of errors and flaws in California’s criminal justice system, each of which
could give rise to its own set of recommendations. We do not attempt to comprehensively evaluate all
possible reforms here. Instead, we call upon legislators, advocates, academics, and policy makers to
look to existing recommendations, knowing that no single reform will by itself solve everything but that
positive change can be generated with each individual reform. In particular, we look to the 2006 report
issued by the California Commission on the Fair Administration of Justice.147 The Commission, which
was formed of prosecutors, law enforcement, defense attorneys, judges, and policymakers, targeted six
different topics associated with wrongful convictions:
•	 Mistaken eyewitness identification;
•	 False confessions;
•	 Perjured informant testimony;
•	 Inaccurate scientific evidence;
•	 Prosecutorial and defense lawyer misconduct; and
•	 Inadequate funding for defense services.148
The Commission issued a series of unanimous recommendations in each area, some of which are
highlighted below. California embarked upon follow-up outreach and meetings in 2007,149 but substantive
practical reform has been stymied. To date, many of the recommended reforms have not been
implemented statewide.
The Commission’s report provides California a rare opportunity to build upon existing consensus.
Reaching agreement in any single area of criminal justice is an unusual occurrence—reaching agreement
in multiple areas, and coming to the point of specific, unanimous recommendations, is virtually unheard
of. We should not waste this past investment: California should not squander the opportunity to build
upon the work done by the Commission.
In this final section, we focus on four potential areas of reform. Prosecutorial misconduct is highlighted
because it was the most common error in flawed homicide prosecutions, which were the most expensive
errors in our sample. Mistaken eyewitness identification and judicial mistake were also sources of high
cost. Lastly, search and seizure (Fourth Amendment) errors affected the largest number of people,
although costs were lower. These areas are addressed below.

FOOTNOTES
147	 California Commission on the Fair Administration of Justice, California Commission on the Fair Administration of Justice Final Report, 2006,
http://www.ccfaj.org/documents/CCFAJFinalReport.pdf.
148	 California Commission on the Fair Administration of Justice, at 6.
149	 California Commission on the Fair Administration of Justice, Minutes of Public Hearing, July 11, 2007,
http://www.ccfaj.org/documents/reports/prosecutorial/expert/Public%20Hearing%20Meeting%20Minutes.pdf.
80 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 5 | CONCLUSION: NEXT STEPS AND RECOMMENDATIONS

PROSECUTORIAL MISCONDUCT
Prosecutorial misconduct, and more specifically the failure of prosecutors to disclose potentially
exculpatory material as required under Brady v. Maryland, has become a substantial focus of reform.
Ample avenues for change—some minor, some expansive—have been suggested. These include:
•	 Creating statewide, standard practices for open file discovery, so that defense counsel have access
to the prosecutor’s full file while appropriately protecting information that could endanger a
potential witness, based on the theory that truth is best served and Brady violations best avoided
by sharing information;150
•	 Allowing civil litigation against district attorney offices that engage in misconduct, potentially by
allowing lawsuits against offices that have multiple intentional or reckless Brady violations within
an agreed-upon period of time;151
•	 Reducing absolute prosecutorial immunity, to provide accountability for prosecutors who
intentionally manipulate or withhold evidence in order to secure a conviction they know may
be false;152
•	 Increasing investigation and disciplinary activity from the State Bar, including mandatory
reporting of certain types of prosecutorial misconduct to the Bar for potential disciplinary
proceedings;153
•	 Creating a Commission on Prosecutorial Misconduct;154

FOOTNOTES
150	 In 2014 Texas adopted an open file policy in the Michael Morton Act, named after a man who was wrongfully convicted in the murder of his wife and spent
25 years in prison. Prior to the law, individual prosecutors made their own decisions about whether a particular document in the file should be considered
exculpatory. See Terry Langford, “Costs and Questions as TX Implements New Discovery Law, The Texas Tribune, May 29, 2014,
http://www.texastribune.org/2014/05/29/michael-morton-act-driving-evidence-costs-das/.
151	 The Hyde Amendment (Pub.L. 105-119, § 617, Nov. 26, 1997, 111 Stat. 2519, codified as a note following 18 U.S.C. § 3006A) allows individuals
prosecuted in federal court to recover their attorneys’ fees and costs if the federal prosecutors’ position was “vexatious, frivolous, or in bad faith.”
The Hyde Amendment does not apply to prosecutions that originated in state court, and it helps only those defendants who paid for their own counsel.
In addressing a way to avoid future cases of prosecutorial misconduct, retired Associate Justice of the Supreme Court of the United States John Paul
Stevens recommends borrowing from tort law and the rule of respondeat superior. There, employers are required to pay damages for torts committed by
their employees during the normal course of business. Justice Stevens suggests a similar model for prosecutors.
See John Paul Stevens, Letter to the Editor, The New York Times, February 18, 2015,
http://mobile.nytimes.com/2015/02/18/opinion/prosecutors-misconduct.html?referrer=&_r=0.
152	 Edwin Chemerinsky, “Head In the Sand Over Prosecutorial Misconduct,” The National Law Journal, April 25, 2011, available online at Lexis/Nexis.
153	 Kathleen M. Ridolfi and Maurice Possley, “Preventable Error: A Report on Prosecutorial Misconduct 1997-2009, Northern California Innocence Project,
Santa Clara School of Law, October 2010, at 4. http://digitalcommons.law.scu.edu/cgi/viewcontent.cgi?article=1001&context=ncippubs. See also
Commission on the Fair Administration of Justice, at 79-80.
154	 Law Professor H Mitchell Caldwell advocates creating a Commission on prosecutorial misconduct fashioned on the existing California Judicial Conduct
Commission. The Commission would have the authority to investigate, conduct hearings, and issue sanctions for violations. See Douglas Kmiec,
“Prosecutorial Misconduct and ISSI Recruiting: The Hidden Linkage,” The Huffington Post, February 20, 2015, http://www.huffingtonpost.com/douglaskmiec/prosecutorial-misconduct-_1_b_6704524.html. Many in New York have also been advocating for such a commission.
http://observer.com/2015/06/supporters-rally-for-state-commission-on-prosecutorial-conduct-bill/.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 81

SECTION 5 | CONCLUSION: NEXT STEPS AND RECOMMENDATIONS

•	 Requiring district attorney offices to formulate and disseminate written policies governing Brady
compliance, providing for systematic collection, tracking and disclosure of the materials as soon
as relevance is known and prior to any guilty plea;155 and
•	 Address overzealous prosecutions by requiring counties to pay for every inmate they incarcerate
in state prison.156
These suggestions are not particularly new, and most of them are not particularly radical. California took
a significant step in the right direction when Governor Brown signed AB 1328 in October, 2015. The new
law will require judges to report certain deliberate and intentional Brady violations to the State Bar, and
in rare instances will allow judges to disqualify entire offices if there is a systemic pattern and practice of
violations. It is an encouraging first step, and one that should be watched as the law is implemented.
Many of these suggestions also allow for systems change that goes beyond blaming an individual
prosecutor or office. Although appropriate accountability for intentional acts is important, it is only
a partial solution. We must recognize the impact of high caseloads, insufficient resources for proper
investigation and prosecution, poorly trained and compensated defense attorneys, an investigator culture
that may not value evidentiary disclosure, and implicit bias that can negatively affective outcomes.157
By addressing all of these pressures, along with individual repercussions for intentional bad acts, we can
reduce misconduct and build towards a better system.

EYEWITNESS IDENTIFICATION
As with prosecutorial misconduct, there has been a great deal of research, evaluation, and discussion
around reducing inaccurate eyewitness identifications. As a result, there is a considerable body of
existing recommendations surrounding erroneous eyewitness identifications. Moreover, eyewitness
identification is another area that can be viewed systematically:
Yes, the eyewitness does have to choose the wrong man from the photo array, but before
that, law enforcement officers have to decide to put him into the array, design the format of
the array and choreograph its display. Forensic evidence at the crime scene could have been
overlooked or—even if properly collected and then tested in the lab—distorted during the
courtroom presentation. Cell phone, mass transit card data, or other alibi information could
have been ignored. Tunnel vision—augmented by clearance rate and caseload pressures
from above—may have overwhelmed the investigators and the prosecutors. Poorly funded

FOOTNOTES
155	 Commission on the Fair Administration of Justice, 90-91.
156	 W. David Ball, “Defunding State Prisons,” Criminal Law Digital Commons, Bulletin 50.5 (2014): 1060-1090, http://digitalcommons.law.scu.edu/cgi/
viewcontent.cgi?article=1606&context=facpubs. Ball proposes that one way to curb overzealous prosecution is to make counties pay for the men
and women whom they send to state prison. Currently, the costs of state prison incarceration are borne by the state, so counties face no financial
repercussions from sending larger numbers of individuals to prison rather than county jail. Ball argues that changing the funding structure to force
counties to pay the costs of sending an individual to prison would improve accountability and transparency.
157	 James M. Doyle, “Learning From Error in the Criminal Justice System: Sentinel Event Reviews,” 8-9, in U.S. Department of Justice, National Institute of
Justice, “Mending Justice: Sentinel Event Reviews,” September 2014, https://ncjrs.gov/pdffiles1/nij/247141.pdf.

82 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 5 | CONCLUSION: NEXT STEPS AND RECOMMENDATIONS

or untrained defense counsel may have failed to investigate alternative explanations or to
execute effective cross-examination. The witness erred; the police erred; the technicians erred;
the prosecutors erred; the defense erred; the judge and the jury erred; and the appellate court
erred, too. 158
The National Academy of Sciences recently evaluated the research on eyewitness identification and
recommended a series of focused and specific practices. Some of these recommendations target the
actual moment of (mis)identification, while others seek to reduce error by viewing the identification as
one moment within the larger frame of potential error. These best practices include:159
•	 Train all law enforcement officers on scientifically established practices that reduce the likelihood
of “false positive” eyewitness identifications;
•	 Implement double-blind and photo array procedures;
•	 Use standard witness instructions;
•	 Document levels of witness confidence;
•	 Videotape the identification process;
•	 Conduct pretrial judicial inquiry;
•	 Make juries aware of prior identifications;
•	 Use scientific framework expert testimony; and
•	 Use jury instructions as an alternate means to convey information.
Other states have already begun to focus reform on eyewitness identification—Massachusetts, for
example, recently created a standing committee on eyewitness identification.160 New Jersey has set new
standards for how judges instruct juries in the credibility of eyewitness identifications,161 and Oregon
has done similarly.162 Although a few jurisdictions in California (e.g., Santa Clara County, Placer County,
the city of El Cerrito) have been leading the way, for the most part California has not been implementing
these best practices despite the fact that the 2006 Commission Report recommended similar reforms.
Given the enormous potential for error, and given the clear and well-researched recommendations set
out by the Commission and the National Academy of Sciences, it is time for California to act.

FOOTNOTES
158	 James M. Doyle, Learning From Error in the Criminal Justice System, 4.
159	 National Research Council, The National Academy of Sciences (2014), “Identifying the Culprit: Assessing Eyewitness Identification
http://www.nap.edu/catalog/18891/identifying-the-culprit-assessing-eyewitness-identification.
160	 In 2015, Massachusetts created a Standing Committee on Eyewitness Identification which provides guidance to courts on the evolving law and science
of eyewitness ID and designs seminars and training sessions on new eyewitness ID evidence procedures and protocols. See “Supreme Judicial Court
Announces New Standing Committee on Eyewitness Identification,” Massachusetts Court System, press release, January 12, 2015,
http://www.mass.gov/courts/news-pubs/sjc/sjc-announces-new-standing-committee-on-eyewitness-identification.html.
161	 http://www.nytimes.com/2011/08/25/nyregion/in-new-jersey-rules-changed-on-witness-ids.html?_r=1
162	 http://www.oregonlive.com/pacific-northwest-news/index.ssf/2012/11/oregon_supreme_court_ruling_de.html

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 83

SECTION 5 | CONCLUSION: NEXT STEPS AND RECOMMENDATIONS

JUDICIAL MISTAKE
Judicial mistakes during trial account for almost one quarter of the errors in this report, particularly
improper jury instructions and erroneous admission or exclusion of evidence. Additional training
for judges is a partial solution. But, as with misconduct and eyewitness identification, what is termed
“judicial mistake” in this report is more accurately viewed as a systemic problem. In our adversarial
system, we depend on competent and well-trained public defenders and district attorneys, both of whom
should be granted sufficient time to understand the issues and adequately brief the court. Our system
relies on each side speaking up to ensure that the judge has adequate information and that a fair trial is
had. A judge may act inappropriately, or even commit misconduct, but he or she is aided by a prosecutor
who missed or silently endorsed the mistake, as well as a defense attorney who did not or could not
respond adequately.163
Rather than address the specific jury instructions and evidentiary mistakes made by judges, this report
recommends the adoption of holistic checks and balances that reduce error more systematically. Those
checks and balances should ensure well-trained prosecutors and defense attorneys, with caseloads that
allow for proper representation, and resources that allow for adequate investigation and representation
on both sides of the aisle. The 2006 Commission Report, for example, makes explicit recommendations
regarding qualifications and compensation for court-appointed defense counsel who are retained by
contract with the court.164 It also calls for increased funding and oversight for defender services.165
On the government side, it is critical that the system encourage and implement a culture of doing justice,
rather than a culture of victory.

FOURTH AMENDMENT VIOLATIONS
Officer violations of the Constitution’s Fourth Amendment protections against search and seizure were
the largest category of individual error in our sample, although they were resolved relatively quickly and
at relatively low monetary cost. But these errors nonetheless impact the lives of those who are subject to
the illegal searches and seizures, as well as victims of crime who may not receive closure when the case
against the defendant is dismissed as a result of the Constitutional violation.
Additional training is again a useful recommendation; officers who know and understand the Fourth
Amendment and its application by the courts in specific circumstances are less likely to violate the law
when they search or arrest a defendant. Here again, however, viewing the issues systemically provides
a helpful framework. The search and seizure errors in our sample belong not only to the officer who
conducted the search, but to his or her supervisor, to the prosecutor in the case who failed to detect the
improper search during the investigation phase, and to the defense counsel who failed to adequately
identify and object to the violation prior to or during trial. Moreover, many jurisdictions lack an

FOOTNOTES
163	 Judicial misconduct and error raise the issue of whether judges should be elected, since the electorate often has little way to accurately asses the
qualifications of a particular candidate. www.mercurynews.com/opinion/ci_27778493/mercury-news-editorial-another-flawed-judge-begs-question.
164	 Commission on the Fair Administration of Justice, 91 – 100.
165	 Commission on the Fair Administration of Justice, 91 – 100.
84 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 5 | CONCLUSION: NEXT STEPS AND RECOMMENDATIONS

automatic feedback mechanism, meaning that officers do not always know the repercussions of their
errors and therefore cannot correct them:
“The police operate a ‘production stage,’ during which they make the cases, often with the participation
of the prosecutors. Then the prosecutors, together with defense lawyers and judges, conduct an
“inspection stage” that culminates in an adversary trial, at which the law enforcement team is required
to account for the work it has produced. It is axiomatic in all industries that end-of-process inspection
schemes, although they are necessary components of quality-control systems, are poor routes to
achieving overall system quality.”166
Achieving system quality in this context means reducing search and seizure errors. Although those
errors were caught in the cases studied here, they were not caught until the end of the process, after the
defendant had been convicted. A more efficient and effective system should strive to reduce or eliminate
error before it happens, rather than catching the errors after the fact. Requiring that individual officers
be informed of the court’s ruling on a search or seizure in which they participated, and requiring that the
court’s ruling be disseminated to the rest of the office, is a first step.

CONVICTIONS INTEGRITY UNITS
Several prosecutors’ offices across the United States are establishing Convictions Integrity Units to
investigate possible wrongful convictions and work toward error prevention.167 Although the Dallas unit
has been operating continuously since 2007,168 in fact one of the first units in the country started in
Santa Clara in 2002.169 In addition to Dallas, the trend has been adopted in Houston, Manhattan, Denver,
Baltimore, Philadelphia, Brooklyn, New York, Cleveland, New Orleans, and Portland, among others.170
In September of 2014, the United States Attorney’s Office in Washington D.C. established the first federal
Convictions Integrity Unit.170 Outside of district attorney offices, North Carolina established an Innocence
Commission in 2006, tasked with providing an independent truth-seeking forum for post-conviction claims
of innocence.172 Most recently, in 2015, the District Attorney for Los Angeles, the County responsible for
the largest share of convictions in California, announced the formation of a Convictions Integrity Unit.173

FOOTNOTES
166	 James M. Doyle, “Learning From Error in the Criminal Justice System, 8.
167	 Center for Prosecutor Integrity, “Conviction Integrity Units: Vanguard of Criminal Justice Reform,” 2014,
http://www.prosecutorintegrity.org/wp-content/uploads/2014/12/Conviction-Integrity-Units.pdf.
168	 Molly Hennessy-Fiske. “Dallas County district attorney a hero to the wrongfully convicted,” The Los Angeles Times, May 8, 2012,
http://articles.latimes.com/2012/may/08/nation/la-na-dallas-district-attorney-20120509.
169	 The Santa Clara unit operated from 2002 to 2007, then was relatively inactive from 2007 to 2010. It was revitalized and formally identified as a
Convictions Integrity Unit in 2010 by District Attorney Jeff Rosen.
170	 http://www.prosecutorintegrity.org/wp-content/uploads/2014/12/Conviction-Integrity-Units.pdf
171	 Spencer S. Hsu, “D.C. prosecutors create unit to find wrongful convictions,” The Washington Post, September 11, 2014,
http://www.washingtonpost.com/local/crime/dc-prosecutors-create-unit-to-find-wrongful-convictions/2014/09/11/91a3722c-39da-11e4-bdfbde4104544a37_story.html.
172	 The North Caroline Innocence Commission, accessed March 4, 2014, http://www.innocencecommission-nc.gov/.
173	 http://www.latimes.com/local/lanow/la-me-ln-conviction-review-unit-20150422-story.html
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 85

SECTION 5 | CONCLUSION: NEXT STEPS AND RECOMMENDATIONS

Recommendations for Convictions Integrity Units tend to cover many of the subjects already addressed
in this section, including Brady disclosures, prosecutorial misconduct, and police practices. As in other
areas, best practices for these units often include rewarding prosecutors for their commitment to justice
instead of their number of convictions, establishing open file discovery, working in partnership with
police officers and investigators, and using best practices in eyewitness identification.174 The fact that
these recommendations overlap with many of the recommendations made in other contexts, including
the 2006 Commission Report, highlights the need for reform.

CONCLUSION
This report focuses a spotlight on 607 wrongful and invalid convictions that burdened Californians with
483 jury trials, 26 mistrials, 16 hung juries, 168 plea bargains, over 700 appeals and habeas petitions, more
than 50 lawsuits and settlements and 2,186 years in prison and jail, for a taxpayer cost of $221 million.
California can do better.
The urgency of our call to action is grounded not only in the costs documented here, but also in the
unknown costs omitted by our methodology. We have surely not captured all of the flawed felony
convictions that could have been included in this report, nor have we quantified the often immense
emotional harm done to those subjected to these flawed proceedings and their families. We have
not included the untold costs that result from wrongful juvenile convictions, wrongful misdemeanor
convictions, and prosecutions that do not result in conviction, among other things. The costs
documented in this report are thus only a window onto the myriad ways in which our system fails
citizens, victims and defendants.
At most, examining 607 cases as we have done here is the start of a conversation, and the suggested
next steps and recommendations in this concluding section are only the first step. There are still many
unanswered questions: how many more people are not being served by our criminal justice system?
What are the true costs? Why does California seem to fall behind other states? Who will take the lead on
solving these problems?
The answer to at least the final question is that it will take all of us. All Californians should care about our
criminal justice system, and we all must make it a priority. The professionals who administer California’s
criminal justice system surely share our goal of a system without unjust arrests, and with convictions
that are fair and free from error, mistake and misconduct. Our communities, our legislators, our citizens,
and our criminal justice professionals themselves—police, prosecutors, defense counsel, judges, and
juries—overwhelmingly believe in and seek to serve in a fair and just criminal justice system. Achieving
this standard requires both an understanding about how and where we are not succeeding and the
willingness to openly discuss areas for improvement. These discussions should include all participants
across agency and adversarial lines, so that we can prioritize and promote improvements to the system.
The real cases highlighted in this report can be the starting point for those conversations.

FOOTNOTES
174	 Establishing Convictions Integrity Units in Prosecutors’ Offices, A Report of the Center on the Administration of Criminal Law’s Conviction Integrity Project, 9,
accessed March 30, 2015, http://www.law.nyu.edu/sites/default/files/upload_documents/Establishing_Conviction_Integrity_Programs_FinalReport_ecm_
pro_073583.pdf.
86 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

SECTION 6:

APPENDICES

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 87

SECTION 6 | APPENDICES

APPENDIX A – County Comparisons
It is not possible to generalize about individual counties based on the cases in this report. While there
are variances among the reversals and wrongful convictions from county to county, limitations on data
availability preclude any certainty regarding whether one county is “better” or “worse” than another in
this regard. One county may appear to have more wrongful convictions than another, but that may be
because, for example, an investigative journalist in that county published information about the reversed
and dismissed cases, or because there is a local Innocence Project. Likewise, a particular county’s lack
of sample cases in this report may simply be a result of poor record-keeping and barriers to access.
That said, the comparison below between the number of cases in our sample, and the number of felony
jury trials in each county, may provide a starting point for a conversation about the need for additional
data and review. This information is not intended to be a report card for any county, but rather to
demonstrate the need for additional data collection and transparency.175 (See Figure 13)
Figure 13: Felony Jury Trials and Sample Cases, by County

County

Total Number of
Felony Jury Trials
2002–2012176

Average # of
Felony Jury
Trials Per Year
2002–2012177

Total Cases
in Sample
1989–2012178

Total Cases in
Sample
2002–2012179

Sample cases
as % of Felony
Jury Trials
2002–2012180

Amador

20

1.8

3

3

15.0%

Alameda

1,113

101.2

7

4

0.4%

Alpine

3

0.3

0

0

0.0%

Butte

301

27.4

6

5

1.7%

Calaveras

40

3.6

0

0

0.0%

Colusa

38

3.5

0

0

0.0%

FOOTNOTES
175	 The ten counties with the highest number of felony convictions in 2013 include: Los Angeles, 43,545; Riverside, 12,734; Orange, 11,529; San Diego,
11,828; San Bernardino, 15,021; Sacramento, 6,509; Kern, 6,380; Fresno, 5922; Santa Clara, 5,743; and San Joaquin, 4,216. Judicial Council of California,
“2014 Court Statistics Report,” Table 8b, http://www.courts.ca.gov/documents/2014-Court-Statistics-Report.pdf.
176	 Judicial Council of California Administrative Office of the Courts, Court Statistics Reports, Table 8a. Data maintained by fiscal year 2002-03 through
2012-13. http://www.courts.ca.gov/12941.htm#id7495.
177	 Based on an eleven year average, 2002-03 through 2012-13, inclusive.
178	 These totals do not include 20 cases in the database from the California-based federal courts. Numbers in each column are based on the calendar year of
the date of reversal.
179	 These totals do not include 20 cases in the database from the California-based federal courts. Numbers in each column are based on the calendar year of
the date of reversal.
180	 Database cases from 2002-2012, inclusive, as a percentage of each county’s total jury trials 2002-2012, inclusive. Note that our sample includes cases
where the defendant was initially convicted by guilty plea, as well as cases where the defendant was initially convicted after trial. All cases in the
database are felony cases.
88 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

Contra Costa

1,349

122.6

18

15

1.1%

Del Norte

78

7.1

2

2

2.6%

El Dorado

335

30.5

0

0

0.0%

Fresno

1,122

102

3

2

0.2%

Glenn

66

6

0

0

0.0%

Humboldt

299

27.2

6

5

1.7%

Imperial

125

11.4

6

6

4.8%

Inyo

77

7

1

1

1.3%

Kern

2,017

183.4

65

40

2.0%

Kings

371

33.7

3

3

0.8%

Lake

228

20.7

5

1

0.4%

Lassen

87

7.9

0

0

0.0%

Los Angeles

22,391

2035.5

165

122

0.5%

Madera

270

24.5

3

3

1.1%

Marin

230

20.9

3

2

0.9%

Mariposa

12

1.1

0

0

0.0%

Mendocino

190

17.3

2

2

1.1%

Merced

301

27.4

2

1

0.3%

Modoc

23

2.1

0

0

0.0%

Mono

5

0.5

0

0

0.0%

Monterey

426

38.7

4

4

0.9%

Napa

168

15.3

3

2

1.2%

Nevada

59

5.4

1

1

1.7%

Orange

1,975

179.5

46

39

2.0%

Placer

256

23.3

1

1

0.4%

Plumas

15

1.4

1

1

6.7%

Riverside

5,552

504.7

14

13

0.2%

Sacramento

4,122

374.7

17

15

0.4%

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 89

SECTION 6 | APPENDICES

San Benito

50

4.5

2

2

4.0%

San Bernardino

2,521

229.2

16

13

0.5%

San Diego

2,991

271.9

38

26

0.9%

San Francisco

2,447

222.5

11

10

0.4%

San Joaquin

818

74.4

8

7

0.9%

San Luis Obispo

142

12.9

0

0

0.0%

San Mateo

569

51.7

19

16

2.8%

Santa Barbara

311

28.3

2

1

0.3%

Santa Clara

1,684

153.1

42

36

2.1%

Santa Cruz

186

16.9

4

3

1.6%

Shasta

844

76.7

8

8

0.9%

Sierra

11

1.1

0

0

0.0%

Siskiyou

77

7

4

3

3.9%

Solano

966

87.8

14

9

0.9%

Sonoma

486

44.2

11

10

2.1%

Stanislaus

942

85.6

4

3

0.3%

Sutter

180

16.4

1

1

0.6%

Tehama

96

8.7

5

5

5.2%

Trinity

18

1.6

0

0

0.0%

Tulare

728

66.2

5

5

0.7%

Tuolumne

195

17.7

2

2

1.0%

Ventura

948

86.2

0

0

0.0%

Yolo

912

82.9

3

3

0.3%

Yuba

137

12.5

1

1

0.7%

90 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

APPENDIX B – Methodology for Cost of County Jail
This appendix details the methodology and data sources employed to calculate the cost of detention in
California’s county facilities.181 Using data from the State Controller’s Office and the Board of State and
Community Corrections, we estimated per day, per capita detention costs for 58 counties in California
for each year from 1996 to 2011.

METHODOLOGY FOR ESTIMATING PER CAPITA DETENTION COSTS
To estimate the per capita cost of detaining an individual, we obtained annual detention expenditures
for each county and annual average daily detention population data. Using these data, we calculated an
average cost per inmate figure using the following formula:
Average Annual Cost Per Inmate

=

Total Annual Spending on Detention ($)
Annual Average Daily Population (ADP)

To calculate a daily per inmate cost, the annual per capita cost figure above is divided by 365 days:
Average Daily Cost Per Inmate

=

Average Annual Cost Per Inmate ($)
365 days

This calculation yielded an average cost per day, per inmate for each county in California, by year.

DATA SOURCES
County Detention Expenditures Data
The most comprehensive resource on county-by-county detention expenditures is the Counties Annual
Report published each year by the California State Controller’s Office (SCO). This publication reports all
county revenues and expenditures each year, including spending on adult detention. Data from 1991 to 2011
were obtained from the SCO. To calculate per capita daily detention costs, only operating expenditures
were included. County spending on juvenile detention has been excluded from these calculations.
FOOTNOTES
181	 The cost of detaining an inmate in a county detention facility may be broken down into two broad categories—operating costs and capital costs.
Only operating costs are included in our calculations.
	- Operating costs include the cost of personnel, services, and facility maintenance. These costs are the most relevant cost for estimating a per
capita cost figure, as they can vary based on the number of inmates in detention.
	- Capital costs include land purchases and construction costs. These may vary dramatically year-to-year and are not necessarily directly related to
the expense of detaining each additional (or, marginal) inmate.
	

Within the operational cost category, there is a further distinction between fixed costs and variable costs. By and large, facility maintenance costs are
mostly fixed, as are personnel costs. These fixed costs are a stepped-function: the addition of one additional inmate is unlikely to alter personnel and
maintenance costs. However, an increase of 300 inmates may require the hiring of additional custody staff and the expansion of a unit. Ordinary variable
costs—expenses than can change in the short run—include costs of food, supplies, clothing and medical care. Again, these are the costs likely to
change with the addition of one inmate (and are most relevant to an estimate of per capita costs). A full analysis of fixed and variable costs is beyond the
scope of this project.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 91

SECTION 6 | APPENDICES

Limitations
The SCO county detention data are high-level, aggregating all adult detention costs including jail, residential
rehabilitation centers, honor farms, as well as other non-secure, non-jail facilities, such as halfway houses,
temporary holding facilities, and work release centers. There is no way to isolate jail costs from these data.
As such, the per day costs calculated here may overstate the actual costs of detaining an inmate only in
a county jail facility.
Inflation Adjustment
Expenditure data from 1991 to 2011 were adjusted to 2013 dollars using the Bureau of Labor Statistic’s
Consumer Price Index for All Urban Consumers. The formula for adjusting for inflation used is as follows:
Cost in 2013$

=

(Nominal Value)

x

2013 CPI Index Value
Original Year CPI Index Value

County Jail Population Data
The denominator of the per capita cost calculation is the average daily population (ADP) of a facility
or county. In California, the Board of State and Community Corrections (BSCC) collects and maintains
ADP figures for each facility in each county across the State through the Jail Profile Survey (JPS). The JPS,
which has been conducted on both a monthly and quarterly basis since late 1995, covers Type II, III and
IV facilities, and excludes temporary holding facilities and juvenile detention facilities (including juvenile
halls and camps).
Facility types covered by the JPS:
•	 “Type II facility” means a local detention facility used for the detention of persons pending
arraignment, during trial, and upon a sentence of commitment.
•	 “Type III facility” means a local detention facility used only for the detention of convicted and
sentenced persons.
•	 “Type IV facility” means a local detention facility or portion thereof designated for the housing
of inmates eligible under Penal Code Section 1208 for work/education furlough and/or other
programs involving inmate access into the community.
Annual Average Daily Population Calculations
In order to estimate a yearly average daily population (ADP), monthly ADP figures were obtained from
the JPS survey data for all 58 counties for the years 1996-2012. An average annual ADP was calculated for
each county by summing each monthly ADP figure within a given year and dividing by 12 months. When
these numbers were calculated, the 2012 JPS data were available from January-June only, so the 2012
figures reflected a 6-month average, rather than an annual average.

92 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

Notes about the JPS data
For the majority of counties, detention facilities are included under the jurisdiction of the Sheriff
or County Corrections Department. However, the JPS data also include detention data for several
jurisdictions which were excluded from the calculations:
•	 Oakland Police Department (2000-2005) – Oakland City Jail, closed in 2005. Inmates were
transferred to Glenn E. Dyer or Santa Rita facilities.
•	 San Diego Probation Department (1996-2001) – Includes Barrett Honor Camp/Barrett Juvenile
(which also housed juveniles, beginning in 1997) and Camp West Fork (minimum security honor
camp, closed in 2001).
•	 San Diego Work Furlough (all years) – Work Furlough residential facility, operated by
Correctional Alternatives, Inc.
•	 Santa Clara Probation Department (1996-2007) – Mountain View Work Furlough program;
residential facility, operated by the Probation Department.
•	 Scapular House (1996-2002) – Los Angeles County’s Work Furlough Program, developed as part
of Community Based Alternatives to Custody Program, managed by the Probation Department.
•	 Ventura Work Furlough (all years) – Minimum security residential facility, operated by the
Ventura Probation Department.

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 93

SECTION 6 | APPENDICES

APPENDIX C – Methodology for Trial, Appeal,
and Attorney Costs
182

This Appendix describes the data sources and methods used to calculate court cost estimates for felony
cases defined by broad offense category, including whether the case involves a trial. We draw data from
several sources, including a judicial work load assessment, court costs from drug court cost evaluations,
and estimates of the hourly costs of public defenders and district attorney accounting for all supporting
personnel. For the purposes of comparison, we present alongside our estimates for California, courtcost estimates produced by the Washington State Institute of Public Policy (WSIPP) for the state of
Washington. The WSIPP estimates are used in their benefit-cost analysis of policy interventions that
impact crime rates, and in turn, resources consumed in processing criminal cases.
Our estimates are based on time use and costs per hour figures for courts, district attorneys, and public
defenders. We first discuss our time input estimates. We then discuss our estimates of hourly costs.
Finally, we pull these estimates together to generate total costs per case type.

TIME INPUT ESTIMATES
We begin by presenting estimates of the court time for criminal cases from the 2002 California Judicial
Workload Assessment conducted by the National Center for State Courts. Table 1 presents estimates
of the proportion of cases involving pre-trial proceedings, trials, and post-trial proceedings and with
estimates of average time in minutes of each stage in the case processing flow. The table also presents
estimates of average time taking into account the event occurrence rates for pre-trial proceedings, trials,
and post-trial proceedings, as well as average time with and without trials. For our cost estimates we
focus on the figures in the five columns pertaining to felony crimes. We use these estimates for several
purposes. First, in estimating district court costs, we use the time estimates in conjunction with an
hourly court cost estimate (to be discussed shortly) to estimate court time costs. In addition, we use the
court time disparity by whether a case involves a trial in conjunction with time use estimates for public
defenders to arrive at time use figures for public defenders and district attorneys.
Table 2 presents data from the San Francisco Public Defender’s Office Caseload/Workload analysis on
average time in hours for public defenders by type of felony offense. We use these data along with the
data in Table 1 to estimate time for case by whether the case involves a trial. We assume that differences
in prep time for cases involving trials and cases with no trials is proportional to the differences in
court time list in Table 1. To be specific, let R be the ratio of court time for a case involving a trial to a
comparable case (in terms of offense) not involving trial. Let X be the prep time for the public defender
when there is no trial. Let p be the proportion of cases of this type that go to trial, and let A be the
average amount of prep time for these cases. The average prep time can be expressed as A = pRX + (1-p)
X. Solving for X gives X= A/(pR +1 – p). With a solution for X (prep time for cases not involving a trial)

FOOTNOTES
182	 The authors of this report are grateful to Steven Raphael, Ph.D. Professor of Public Policy at the Goldman School of Public Policy, University of California
Berkeley, for developing the methodology set out in this section. A complete peer-reviewed methodology for trial, attorney and court costs was beyond the
scope of this report. We have attempted here only to roughly estimate these costs, using available information. Professor Raphael’s gracious contribution
to this report is deeply appreciated.
94 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

prep time for cases involving a trial require multiplication by the ratio R. We use the figures in the first
row of Table 2 for the parameter A, we use data on trial incidence from Table 1 to estimate p and 1-p,
and estimate R from the data in Table 1 by taking the ratio of average court time with trial to average
court without trial. The implied solutions for average time per case with and without a trial are presented
in the second and third rows of Table 2.
We do not have separate estimates of time use per case type for district attorneys. Hence, we assume that
prep time for district attorneys and public defenders are the same.

HOURLY COST ESTIMATES
We estimate the average cost of court time using estimates from the California Drug Court study.
The Phase 1 evaluation of the Administrative Office of the Court includes estimates of hourly costs for
the Superior Court, district attorney, and public defender from three separate county drug courts.
These figures are presented in Table 3. While there is general agreement for Courts 1 and 2, average
hourly costs are much lower for Court 3. We believe that Court 3 is an outlier and thus base our estimate
of an hour of court time on the average of the values for Courts 1 and 2.
In our cost estimates, we use the Superior Court figures from Table 3 only. To estimate hourly costs
figures for prosecutors and public defenders, we rely on more conservative calculations based on total
annual operating expenditures in the state for public defenders and district attorneys and total attorney
staffing levels. Table 4 presents these tabulations using data for fiscal year 2010-2011 and inflating the
ultimate dollar figure per hour to 2013 dollars. Note, the tabulations in table 4 amortize all operating
expenditures (inclusive of non-attorney staff) to attorney hours. The hourly figures in Table 4 are
considerably lower than the hourly costs estimates from the drug courts study in Table 3. We use these
lower hourly costs estimates in our ultimate cost tabulations.

GENERATING THE FINAL COST ESTIMATES
Table 5 multiplies to time use figures and the hourly costs figures to estimate average court costs, public
defense costs, and prosecution costs by broad offense type. The table presents separate cost totals for
cases involving trials and cases not involving trials. Not surprisingly, costs are considerably higher when
a trial is involved.
There are few extant comparable estimates of court processing costs for felony convictions. The
Washington State Institute of Public Policy (WSIPP) has produced comparable estimates for the state
of Washington, though using a decidedly different methodological approach and for somewhat different
offense categories.183 Moreover, WSIPP does not produce separate estimates for cases distinguished
by whether they involve a trial. These differences aside, we present these estimate in Figure 1 to assess
whether we are in the same ballpark. The figures are generally consistent with our estimates.

FOOTNOTES
183	 WSIPP estimates a time series regression of total criminal justice expenditures against changes in the number of convictions by type. The coefficients on
the changes in convictions serve as estimates of the marginal cost of a criminal conviction. Table 1 comes from the coefficients in the regression model
presented in Washington State Institute of Public Policy, Benefit Cost Technical Manual, Page 35 exhibit 13. The complete report can be found on the
WSIPP webpage location, visited August 8, 2014, http://www.wsipp.wa.gov/TechnicalManual/WsippBenefitCostTechnicalManual.pdf.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 95

SECTION 6 | APPENDICES

Table 1: Estimates of Court Time for Criminal Cases from the 2002 California Judicial Workload Assessment184
Homicide

Felony against
person

Property

Drug

Other felony Class A & C

Class B & D

Infractions

EVENT OCCURRENCE RATE
Pre-trial

100%

100%

100%

100%

100%

100%

81%

27%

Trial

50%

6%

3%

2%

3%

3%

0.5%

4%

Post-trial

20%

20%

20%

50%

20%

10%

10%

0.5%

TIME IN MINUTES
Pre-trial

713

159

74

82

153

18

3

2

Trial

3,000

1,829

684

902

1,440

720

360

13

Post-trial

186

70

45

76

97

30

3

0.52

104

138

216

43

5

1

74

82

153

18

2.43

1

803

1,060

1,690

768

366

15.5

AVERAGE TIME WITH GIVEN OCCURRENCE RATES
Total

2,250

283

AVERAGE TIME FOR CASES WITHOUT TRIAL
Total

713

159

AVERAGE TIME FOR CASES THAT GO TO TRIAL
Total

3,899

2,058

FOOTNOTES
184	 Data come from Table 6-15 (page 110) from National Center for State Courts (2002), California Judicial Workload Assessment Final Report.
96 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

Table 2: Times Per Case for Public Defender and (by Assumption) District Attorney
Homicide

Felony against person

Property

Drug

Other

Average

250

40

10

10

10

No Trial

93.68

27.34

8.53

8.95

8.53

Trial

406.31

243.01

57.56

60.21

57.63

Average hours taken from the San Francisco Public Defender’s Office Caseload/Workload Analysis
(2013), Page 4, which in turn cites a report by the San Francisco Controller entitled: Public Defender
Project: Final Felony and Misdemeanor Caseload Standards (Controller’s Office, May 21, 2003) for the
hours estimates used above. Using the trial incidence from Table 1, we assume that differences in prep
time for cases involving trials and cases with no trials for public defenders and district attorneys is
proportional to the differences in court time list in Table 1. To be specific, let R be the ratio of court time
for a case involving a trial to a comparable case (in terms of offense) not involving trial. Let X be the
prep time for the public defender when there is no trial. Let p be the proportion of cases of this type that
go to trial, and let A be the average amount of prep time for these cases. The average prep time can be
expressed as A = pRX + (1-p)X. Solving for X gives X= A/(pR +1 – p). With a solution for X (prep time for
cases not involving a trial) prep time for cases involving a trial require multiplication by the ratio R.
Table 3: Estimated Costs Per Session Hour in Drug Courts for Three Courts Evaluated in Phase 1 of the California
Drug Court Cost Assessment (2001 Dollars)185
Court 1

Court 2

Court 3

Superior Court

$1,043.64

$1,048.80

$146.02

District Attorney

$355.14

$199.40

$227.70

Public Defender

$315

$631.07

$110.95

FOOTNOTES
185	 Figures come from Appendix F of Administrative Office of the Courts Judicial Council of California (2003), California Drug Courts: A Methodology for
Determining Costs and Avoided Costs, Phase 1: Building the Methodology.

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 97

SECTION 6 | APPENDICES

Table 4: Tabulating Per Hour Expenditures for District Attorneys and Public Defenders from Operating Expenditures186
Attorney Total 2010

Operating Expenditures,
FY10-11

Expenditures per hour
assuming 2000 hours
per year

Expenditures inflated to
2013.

Public Defenders

3,990

1,352,982,518

169

176

District Attorneys

2,496

730,869,200

146

152

Table 5: Estimates of Court, Prosecution, and Public Defense Costs for Criminal Trials Based on Time Use Estimates
and Cost Estimates Presented in Tables 1 through 3 (costs expressed in 2013 dollars)187
Homicide

Felony against person

Property

Drug

Other

Court

$51,644

$6,489

$2,376

$3,168

$4,948

Public Def.

$38,000

$6,080

$1,520

$1,520

$1,520

Prosecution

$44,000

$7,040

$1,760

$1,760

$1,760

Total

$133,644

$19,609

$5,656

$6,448

$8,228

Court

$16,364

$3,649

$1,698

$1,882

$3,511

Public Def.

$14,240

$4,110

$1,296

$1,364

$1,296

Prosecution

$16,489

$4,759

$1,501

$1,580

$1,501

Total

$47,093

$12,519

$4,496

$4,826

$6,308

Court

$89,485

$47,233

$18,430

$24,328

$38,787

Public Def.

$61,760

$36,939

$8,748

$9,152

$8,761

Prosecution

$71,511

$42,771

$10,130

$10,598

$1,501

Total

$222,757

$126,943

$37,307

$44,078

$57,693

AVERAGE COSTS

WITHOUT TRIAL

WITH TRIAL

FOOTNOTES
186	 Personnel totals come from the California Office of the Attorney General “Table 9: Criminal Justice Full Time Personnel by Type of Agency and Full Time
Classification,” http://oag.ca.gov/crime/cjsc-stats/2010/table9 accessed on June 13, 2014. Total expenditures come from the State Controller’s Office
Division of Accounting and Reporting Local Government Reporting System. Totals in the table represent the sums for each variable across all 58 counties.
187	 Hourly rates and time per case for public defenders and prosecutors come from the figures in Tables 2 and 4. Hourly court expenditures are based on the
average hourly figures for courts 1 and 2 presented in Table 3 from the Drug Court Evaluation study. All hourly figures are converted into 2013 dollars.
98 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

Figure 1: Estimated Marginal Court and Prosecution Costs Per Conviction for Washington State in 2013 Dollars
from the Washington State Institute of Public Policy
$180,000
$160,000
$140,000
$120,000
$100,000
$80,000
$60,000
$40,000
$20,000
$0
Homicide

Sexual Assault

Robbery

Assault

Non-violent

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 99

SECTION 6 | APPENDICES

APPENDIX D – Lack of Access to County Court Records
in California
Data collection for this report was severely hampered by a lack of access to public court records in
California, as well as inconsistency between counties regarding what could be provided.
Only a few counties in California have online records systems that can be accessed by the public.
At the time of this report’s case collection process, for example, only Riverside and Napa Counties had
complete online records systems that were searchable and that contained pdfs of court documents.
Orange County maintained an online database detailing basic case information only, and San Bernardino
County was the only Southern California county to provide free access to their online dockets, although
not all dockets were available online. Yolo County has begun digitizing its more recent case files and
other counties have made some progress at using technology to make records more available to the
public.188 Five of the counties visited, two in northern California (Napa and Yolo), and three in southern
California (Los Angeles, San Diego and Orange), had files available through an imaged database that
could be accessed through onsite kiosks.
In virtually every other county, accessing the public files required an in-person visit to the courthouse
and an in-person written request for the file, followed by a delay before the file could be located.
In Southern California counties the delay was generally a few days to a week or two, but in many
Northern California counties, particularly the smaller ones, the delay was up to three weeks. Moreover,
six out of the thirty-five Northern California counties that were visited required a fee from $10-15 to
produce each case file or volume; the others did not charge a fee.
San Joaquin County has an online form that can be used to request court files, which avoids multiple
visits to the courthouse.189 San Diego is also building to an electronic court file system. Although not all
courthouses have implemented the system, collecting records from the downtown San Diego Courthouse
was a seamless experience. Physical files and records were delivered within minutes of the request,
imaged files could be printed on site at a kiosk, and files generally were complete and in order.

FOOTNOTES
188	 Other counties that leveraged technology to their advantage by making (somewhat) detailed dockets available via internet or other electronic court record
system included San Bernardino, San Diego, Fresno, Solano, Kern, Butte, Shasta, Humboldt, Marin, San Mateo, Sonoma and Sacramento. Although Los
Angeles provided access to imaged case files from computer terminals in the County Hall of Records, a substantial number of the imaged case files were
missing or could not be located, and those that were available often had missing documents. Out of the 550 requested case files in Northern California,
only 12 were missing.
189	 In Los Angeles County, out of the thirteen different courthouses, only one permitted a file request over the phone. The others either required an in-person
request or did not answer the phone. Only two of the courthouses had a separate records counter. At the other courthouses, requests had to be made to
clerks who also handle general matters.
100 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

Counties had inconsistent policies regarding which documents were inappropriate for public view.
One court clerk prohibited the viewing of the Abstract of Judgment on the ground that it was
confidential; all others made the Abstract available. Many courthouses also initially prohibited the use
of a handheld scanner, including courthouses in Los Angeles County where the presiding judge has
expressly permitted such scanners. One court supervisor in a Southern California county stated that
using a scanner was akin to “stealing” the $.50 per page copying fee from the court, and refused to allow
it. A small minority of Northern California counties prohibited the scanner.
More troubling inconsistencies arose regarding confidentiality. In some counties, court clerks took steps
to omit or blot out the names of witnesses before making the public file available. In others, however,
case files were inappropriately released with un-redacted juror names, photos of child victims, and
witness information.

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 101

SECTION 6 | APPENDICES

APPENDIX E – Coding: All Errors Identified in Cases Reviewed
for this Report
Note: For most of the analysis in the main body of this report, we consolidated our 41 specific reversal
ground types into 8 broader categories. The categories and their component ground types are as follows:

Error Definition

Number of
Confirmed Errors

Number of Alleged
Errors the Court
Declined to Decide

Number of Alleged
Errors the Court
Rejected

Grand Total

756

119

107

4th Amendment Violations

146

2

0

4th Amendment search/seizure

146

2

1

Failure of Prosecutorial Discretion

117

6

0

Defendant’s conduct doesn’t meet legal definition of the crime

16

1

1

Double Jeopardy

8

1

1

Insufficient evidence

82

4

12

Statute of limitation

11

0

0

IAC

81

18

0

4th Amendment search/seizure

5

0

0

5th Amendment confession/Miranda violation

2

0

0

Defendant’s conduct doesn’t meet legal definition of the crime

2

0

0

DNA

1

0

0

Erroneous admission or exclusion of evidence

8

3

1

Eyewitness recanted his/her testimony

2

0

0

Eyewitness was unreliable or was lying

1

0

0

Failure to adequately represent

39

11

8

Improper jury instructions

3

2

0

Insufficient evidence

4

0

0

Lab error/individual or group mishandling of physical evidence

1

0

0

Mistaken eyewitness ID (honest mistake)

3

0

0

Other grounds

5

1

1

Police eyewitness ID practices were inadequate

2

0

0

Prosecutorial Misconduct

2

1

0

Ineffective defense counsel

1

0

0

102 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

Inadequate Police Practices Before Trial

35

4

0

5th Amendment confession/Miranda violation

8

1

5

New Evidence

24

2

0

Other police practices were inadequate

3

1

0

Judicial Mistake During Trial

164

45

0

6th Amendment

21

3

7

Erroneous admission or exclusion of evidence

50

13

13

Improper jury instructions

76

20

11

Juror Misconduct

16

2

5

Sentencing Error

1

7

2

Other

59

28

0

Unspecified error190

2

5

0

Other

37

23

27

Retroactive change or clarification in the law subsequent
to conviction

16

0

0

Unknown

4

0

0

Problems With Eyewitness Identification

44

3

0

Eyewitness recanted his/her testimony

18

0

0

Eyewitness was unreliable or was lying

16

2

1

Mistaken eyewitness ID (honest mistake)

5

0

0

Police eyewitness ID practices were inadequate

5

1

0

Prosecutorial Misconduct

86

13

0

Brady violation

45

5

5

Prosecutorial Misconduct

41

8

5

Unreliable or Untruthful Official Testimony

24

0

0

Officer testimony was unreliable or untruthful

17

0

0

Snitch/confidential informant was unreliable or was lying

7

0

1

FOOTNOTES
190	 In the final stages of quality control, it was discovered that the grounds for reversal (i.e., the type of error) in two cases had been entered with incorrect
or unclear coding. The total database cost attributed to these two individuals for incarceration, trials, appeals, prosecution, defense, and settlement is
reflected in this report. However, for those calculations which are broken down by type of error, the cost of these two individuals has not been included.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 103

SECTION 6 | APPENDICES

APPENDIX F – Total Cost By Type of Error, All Errors
This report calculates the cost of the catalogued errors, according to the eight error categories used
throughout the report. This appendix assigns cost to all of the errors coded for in the sample, without
combining them into the eight categories.
Error Type

Sum of Total Cost191

Grand Total

$221,601,730.19

Brady violation

$44,221,687.04

New Evidence

$19,132,785.76

Eyewitness was unreliable or was lying

$17,294,773.28

IAC – Failure to adequately represent

$15,401,844.95

Improper jury instructions

$12,568,521.16

Fourth Amendment search/seizure

$11,969,856.66

Insufficient evidence

$11,911,741.43

Erroneous admission or exclusion of evidence

$11,354,014.66

Officer testimony was unreliable or untruthful

$9,361,344.13

Prosecutorial Misconduct

$8,884,509.41

Eyewitness recanted his/her testimony

$8,597,709.29

Snitch/confidential informant was unreliable or was lying

$5,419,612.85

IAC – Mistaken eyewitness ID (honest mistake)

$5,323,264.65

Juror Misconduct

$4,410,436.79

Double Jeopardy

$3,709,087.82

Police eyewitness ID practices were inadequate

$3,707,046.08

Sixth Amendment

$3,617,975.30

Retroactive change or clarification in the law subsequent to conviction

$2,656,725.16

Statute of limitations

$1,905,946.97

Defendant’s conduct doesn’t meet legal definition of the crime

$1,837,075.92

Mistaken eyewitness ID (honest mistake)

$1,824,067.00

Other police practices were inadequate

$1,379,813.41

Fifth Amendment confession/Miranda violation

$ 1,216,520.69

IAC – Erroneous admission or exclusion of evidence

$1,130,789.12

FOOTNOTES
191	 The associated cost is the total calculated cost of incarceration, attorneys, trial courts, VCGCB payments, and settlements, adjusted for inflation to 2013 dollars.
104 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

Unknown

$1,083,230.63

IAC – Eyewitness was unreliable or was lying

$791,511.67

IAC – Fourth Amendment search/seizure

$650,064.50

IAC – Failure to adequately represent

$15,401,844.95

IAC – Lab error/individual or group mishandling of physical evidence

$592,318.66

IAC – Other grounds

$563,613.61

Ineffective defense counsel

$412,384.56

IAC – Police eyewitness ID practices were inadequate

$356,203.40

IAC – Fifth Amendment confession/Miranda violation

$255,058.30

IAC – Eyewitness recanted his/her testimony

$248,701.10

IAC – DNA

$225,466.23

IAC – Improper jury instructions

$219,905.10

IAC – Prosecutorial Misconduct

$171,924.44

IAC – Defendant’s conduct doesn’t meet legal definition of the crime

$119,843.35

Unspecified error192

$113,740.95

IAC – Insufficient evidence

$ 94,581.85

Sixth Amendment

$3,617,975.30

Sentencing Error

$22,837.38

FOOTNOTES
192	 In the final stages of quality control, it was discovered that the grounds for reversal (i.e., the type of error) in two cases had been entered with incorrect
or unclear coding. The total database cost attributed to these two individuals for incarceration, trials, appeals, prosecution, defense, and settlement is
reflected in this report. However, for those calculations which are broken down by type of error, the cost of these two individuals has not been included.
The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 105

SECTION 6 | APPENDICES

APPENDIX G – Public Records Act Request for Settlements
In April, 2013, we sent a request under the California Public Records Act, codified as California
Government Code §§ 6250 et seq., seeking information about payments made by each county and 13
large cities in California related to claims of malicious or inappropriate arrest, charging, or prosecution.
The request was intended to include payments made to settle one or more claims, as well as payments
ordered at the conclusion of litigation. Although many and perhaps most such settlements include
language expressly denying liability or inappropriate activity on the part of the county or city, the fact
of payment on a claim, even where liability is strictly disclaimed in writing, suggests a potential for
inappropriate action by law enforcement or the prosecution that should be subject to examination.
A limitation of our dataset is that it is dependent on the coding and recordkeeping of either the cities
or counties that responded to our Public Records Act request, or their third party risk management
providers. For example, many of the respondents to our request had a definition of “wrongful arrest”
that did not differentiate between cases of inaccurate arrest (i.e., “I was arrested for a crime I did not
commit”) and cases involving the use of excessive force during an arrest (i.e., “I am not contesting
the validity of my arrest, but I should not have been arrested in the manner that I was arrested.”)
Our request sought information related to settlements based on claims of inaccurate arrest, prosecution,
or conviction, but not based on excessive force, assault and battery, or other similar “manner of arrest”
cases. In some instances we were able to exclude “manner of arrest” claims based on other information
provided by the respondent, but in practice, this was as much an art as a science, in part based on
the complexities of specific cases and in part based on the lack of data and lack of standardization
of formats and supporting information among the counties and cities. Thus, our numbers are not a
precisely accurate accounting of costs arising from inaccurate arrests or prosecutions, but rather reflect
a general estimate.
The costs of these wrongful arrests and prosecutions were further divided, where possible, into the costs
of direct settlement and the costs arising from outside counsel or others outside the county or city paid
to help defend against the lawsuit. Many counties did not provide amounts paid to outside legal counsel,
either because they did not use outside counsel in the disputes or because data management limitations
prevented them from tracking or reporting such external costs.

106 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

Location

Population

Number
of Cases

$$ Paid Out

$$ in
Settlements

$$ in Outside
Legal Fees

Alameda County

1,510,000

17

$2,409,000

$1,405,000

$1,004,000

Alpine County

1,200

1

$0

$0

$0

Amador County

38,100

0

$0

$0

$0

City of Anaheim

336,000

11

$99,000

$99,000

$0

City of Bakersfield

347,000

11

$43,000

$43,000

$0

Butte County

220,000

4

$25,000

$25,000

$0

Calaveras County

45,600

2

$31,000

$31,000

$0

City of Chula Vista

244,000

8

$85,000

$20,000

$65,000

Colusa County

21,400

0

$0

$0

$0

Contra Costa County

1,050,000

20

$176,000

$176,000

$0

Del Norte County

28,600

0

$0

$0

$0

El Dorado County

182,000

21

$570,000

$352,000

$218,000

Fresno City

509,000

0

$0

$0

$0

Fresno County

930,000

22

$3,380,000

$2,339,000

$1,041,000

Notes

No documents responsive
to request

No documents responsive
to request

No documents responsive
to request

“Does not keep records in a
format that would enable [it]
to to provide the documents
requested without locating
and examining every individual
claim and lawsuit filed since
January 1, 1989.”

“Does not possess and
does not retain documents
requested and therefore cannot
provide the documents.”

Glen County

28,000

0

$0

$0

$0

Humboldt County

135,000

9

$152,000

$73,000

$79,000

Imperial County

175,000

0

$0

$0

$0

Failed to respond to Public
Records Request

Inyo County

19,000

0

$0

$0

$0

No documents responsive
to request

Kern County

840,000

75

$12,855,000

$12,120,000

$735,000

Kings County

153,000

0

$0

$0

$0

Lake County

65,000

2

$277,000

$174,000

$103,000

Lassen County

35,000

0

$0

$0

$0

“Does not maintain the files
in a manner which would
allow electronic access to the
documents requested.”

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 107

SECTION 6 | APPENDICES

City of Long Beach

462,000

21

$8,472,000

$8,472,000

$0

Los Angeles City

3,793,000

370

$93,264,000

$93,264,000

$0

Los Angeles County

10,018,000

99

$29,477,000

$23,372,000

$6,105,000

Madera County

151,000

3

$580,000

$81,000

$499,000

Marin County

252,000

3

$1,066,000

$886,000

$180,000

Mariposa County

18,000

0

$0

$0

$0

Mendocino County

88,000

4

$65,000

$62,000

$3,000

Merced County

256,000

23

$445,000

$383,000

$62,000

Modoc County

10,000

0

$0

$0

$0

Failed to respond to Public
Records Request

Mono County

14,000

20

$0

$0

$0

No documents responsive
to request

Monterey County

416,000

6

$8,000

$8,000

$0

No documents responsive
to request

Napa County

136,000

9

$367,000

$289,000

$78,000

Nevada County

99,000

3

$2,000

$2,000

$0

City of Oakland

391,000

364

$49,466,000

$49,466,000

$0

Orange County

3,010,000

49

$2,163,000

$1,546,000

$617,000

Placer County

348,000

14

$357,000

$189,000

$168,000

Plumas County

20,000

0

$0

$0

$0

No documents responsive
to request

City of Riverside

304,000

0

$0

$0

$0

Failed to respond to Public
Records Request

Riverside County

2,190,000

67

$5,328,000

$5,328,000

$0

Sacramento City

2,527

39

$482,000

$482,000

$0

Sacramento County

1,419,000

193

$7,747,000

$4,441,000

$3,306,000

San Benito County

55,000

12

$667,000

$502,000

$165,000

City of San Bernardino

210,000

15

$3,079,000

$2,376,000

$703,000

San Bernardino County

2,035,000

126

$5,325,000

$5,325,000

$0

San Diego City

1,323,000

51

$3,386,000

$3,386,000

$0

San Diego County

3,095,000

83

$7,004,000

$7,004,000

$0

City and County of
San Francisco

805,000

97

$12,643,000

$12,643,000

$0

Includes Rampart Scandal –
200 people, $78,192,000

No documents responsive
to request

Includes Rider Scandal –
119 people, $10,900,000

108 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

SECTION 6 | APPENDICES

San Joaquin County

685,000

8

$477,000

$375,000

$102,000

City of San Jose

946,000

58

$1,454,000

$1,454,000

$0

San Luis Obispo County

270,000

23

$925,000

$549,000

$376,000

San Mateo County

718,000

22

$838,000

$838,000

$0

Santa Barbara County

424,000

29

$1,421,000

$1,405,000

$16,000

Santa Clara County

1,782,000

102

$6,262,000

$4,440,000

$1,822,000

Santa Cruz County

262,000

2

$$65,000

$65,000

$0

Shasta County

177,000

23

$0

$0

$0

No documents responsive
to request

Sierra County

3,000

0

$0

$0

$0

No documents responsive
to request

Siskiyou County

45,000

20

$0

$0

$0

Failed to respond to Public
Records Request

Solano County

413,000

19

$843,000

$530,000

$313,000

Sonoma County

484,000

17

$80,000

$80,000

$0

Stanislaus County

514,000

1

$7,000

$7,000

$0

City of Stockton

292,000

45

$1,500,000

$1,500,000

$0

Sutter County

95,000

1

$118,000

$11,000

$107,000

Tehama County

63,000

13

$704,000

$324,000

$380,000

Trinity County

14,000

1

$3,000

$2,000

$1,000

Tulare County

442,000

7

$110,000

$103,000

$7,000

Tuolumne County

55,000

3

$134,000

$134,000

$0

Ventura County

823,000

146

$1,595,000

$420,000

$1,175,000

Yolo County

201,000

0

$0

$0

$0

Failed to respond to Public
Records Request

Yuba County

72,000

12

$0

$0

$0

No documents responsive
to request

Total

--

2,351

$268,031,000

$248,601,000

$19,430,000

Includes Rider Scandal –
119 people, $10,900,000

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 109

NOTES

110 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System

NOTES

The Chief Justice Earl Warren Institute on Law and Social Policy, Berkeley School of Law | 111

NOTES

112 | CRIMINAL (IN)JUSTICE: A Cost Analysis of Wrongful Convictions, Errors, and Failed Prosecutions in California’s Criminal Justice System