Police Misconduct as a Cause of Wrongful Convictions Russell Covey Wash. L. Rev. 2013
Download original document:
This text is machine-read, and may contain errors. Check the original document to verify accuracy.
POLICE MISCONDUCT AS A CAUSE OF WRONGFUL CONVICTIONS RUSSELL COVEY ABSTRACT This study gathers data from two mass exonerations resulting from major police scandals, one involving the Rampart division of the L.A.P.D., and the other occurring in Tulia, Texas. To date, these cases have received little systematic attention by wrongful convictions scholars. Study of these cases, however, reveals important differences among subgroups of wrongful convictions. Whereas eyewitness misidentification, faulty forensic evidence, jailhouse informants, and false confessions have been identified as the main contributing factors leading to many wrongful convictions, the Rampart and Tulia exonerees were wrongfully convicted almost exclusively as a result of police perjury. In addition, unlike other exonerated persons, actually innocent individuals charged as a result of police wrongdoing in Rampart or Tulia only rarely contested their guilt at trial. As is the case in the justice system generally, the great majority pleaded guilty. Accordingly, these cases stand in sharp contrast to the conventional wrongful conviction story. Study of these groups of wrongful convictions sheds new light on the mechanisms that lead to the conviction of actually innocent individuals. I. INTRODUCTION Police misconduct causes wrongful convictions. Although that fact has long been known, little else occupies this corner of the wrongful convictions universe. When is police misconduct most likely to result in wrongful convictions? How do victims of police misconduct respond to false allegations of wrongdoing or to police lies about the circumstances surrounding an arrest or seizure? How often do victims of police misconduct contest false charges at trial? How often do they resolve charges through plea bargaining? While definitive answers to these questions must await further research, this study seeks to begin the Professor of Law, Georgia State University College of Law. Special thanks are owed to Eric Coffelt, who worked hard and traveled far to help me collect data for this study. I also wish to thank Brandon Garrett and Sam Gross for their excellent comments on earlier drafts, and to Max Compton for his research assistance. 1133 Electronic copy available at: http://ssrn.com/abstract=2349872 1134 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 inquiry. Specifically, this study attempts to improve our understanding of the intersection of police misconduct and criminal justice, and, more generally, to contribute to the ever-growing bank of knowledge about wrongful convictions. What we do know about wrongful convictions comes largely from studies of cases terminating in exonerations. These exoneration studies have produced a rich dataset from which several factors that contribute to wrongful convictions have been identified. While critically important, this dataset has significant limitations, chief of which is selection bias. The vast majority of the exonerations studied to date occurred in rape cases following DNA testing and murder cases often involving the death penalty. Such cases, comprising a tiny sliver of the criminal justice system workload, are relatively unrepresentative of the vast majority of felony convictions. As a result, and as researchers compiling these datasets acknowledge, the most closely analyzed data on wrongful convictions does not capture a representative sample of the probable distribution of wrongful convictions that occur.1 Drawing on new empirical data, this article adds to the wrongful convictions dataset by assessing another group of exonerees—those exonerated following revelations of systemic police and/or prosecutorial misconduct. Specifically, this Article examines two high-profile scandals that saw the wrongful conviction and later formal exoneration of large numbers of persons. To date, little attention has been paid to such exonerees. This, I argue, has affected perceptions of the scope and nature of the wrongful conviction problem. The profile of persons exonerated following revelations of major police misconduct varies dramatically from that of the typical capital murder or DNA exoneree. The defendants in the mass exoneration cases were convicted of different types of crimes, faced less severe punishments, and were far more likely to plead guilty than other exonerated defendants. Using extant data, earlier exoneration studies have found that the primary cause of the wrongful convictions in those studies is witness misidentification.2 Based on those findings, some 1. See Samuel R. Gross & Barbara O’Brien, Frequency and Predictors of False Conviction: Why We Know So Little, and New Data on Capital Cases 11–12 (Mich. Law Sch. Pub. Law Res. Paper No. 93, 2007), available at http://ssrn.com/abstract=996629. 2. See, e.g., D. Michael Risinger, Innocents Convicted: An Empirically Justified Factual Wrongful Conviction Rate, 97 J. CRIM. L. & CRIMINOLOGY 761, 773 n.19 (2007) (finding, based on analysis of sixty-two DNA exonerations as of 2000, that mistaken eyewitness identification is the most common factor); Samuel R. Gross et al., Exonerations in the United States 1989 through 2003, 95 J. CRIM. L. & CRIMINOLOGY 523, 542 (2005) (reporting based on study data that the “most common cause of wrongful convictions is eyewitness misidentification”). Edwin Borchard reached the same Electronic copy available at: http://ssrn.com/abstract=2349872 2013] POLICE MISCONDUCT 1135 commentators and reform-minded organizations have drawn the conclusion that witness misidentification is, in general, the leading cause of wrongful convictions.3 While the former claim was correct at the time those observations were made, the latter generalization likely was not warranted. The data we currently have is simply too limited to permit any accurate generalizations about how frequently wrongful convictions occur, or which contributing factors generate them, in specific or even rough proportion.4 Moreover, as one of the nation’s leading experts on exonerations, Professor Samuel Gross, has frequently emphasized, the primary causes of wrongful convictions are almost certainly crime-specific.5 That is, the factors that tend to cause wrongful convictions in rape cases are different from those that cause wrongful convictions in murder cases, and different from the causes of wrongful convictions in burglary cases, assault cases, and drug cases. The next generation of research must approach wrongful convictions in a more fine-grained manner. To that end, this study gathers data from two mass exonerations resulting from major police scandals. These exonerations are starkly different than most of the exonerations previously studied. In the “minerun” cases (if there is such a thing) resulting in individual exonerations, often as a result of DNA testing, several contributing factors, ranging from eyewitness misidentification to false confessions to faulty forensic evidence and testimony, have been identified. In contrast, wrongful convictions in the mass exoneration cases are tied together by a single dominant causal factor: police misconduct. Prior exoneration studies have not focused on this group of exonerees, nor, by and large, have they conclusion in his 1932 study of wrongful convictions. See EDWIN M. BORCHARD, CONVICTING THE INNOCENT: ERRORS OF CRIMINAL JUSTICE xiii (1932). 3. Drawing on its database of persons exonerated through DNA testing, the Innocence Project also claims that mistaken witness identifications are the leading cause of wrongful convictions. See Facts on Post-Conviction DNA Exonerations, INNOCENCE PROJECT, http://www.innocenceproject.org/ Content/351.php. See also Keith A. Findley & Michael S. Scott, The Multiple Dimensions of Tunnel Vision in Criminal Cases, 2006 WIS. L. REV. 291, 292 (2006) (describing “mistaken eyewitness identifications” as “the most frequent single cause of wrongful convictions”). 4. In any event, new data on wrongful convictions suggests that witness misidentification no longer can claim the top spot as a contributing cause of known exonerations. See Samuel R. Gross & Michael Shaffer, Exonerations in the United States, 1989–2012 40 (Mich. Law Sch. Pub. Law and Legal Theory Working Paper Series, Working Paper No. 277, 2012), available at http://ssrn.com/ abstract=2092195 (identifying “perjury or false accusation” as the leading known cause of wrongful convictions during time period studied). 5. See id. at 102. Electronic copy available at: http://ssrn.com/abstract=2349872 1136 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 incorporated data from these cases into the datasets.6 This Article does just that. The experiences of those wrongfully convicted as a result of police misconduct differ from other exonerees in interesting and potentially important ways. The exonerations resulting from the Rampart and Tulia police scandals raise the profile of police misconduct as a known cause of wrongful convictions. In addition, the tendency of exonerees in these cases to plead guilty rather than go to trial confirms what many have long suspected: that the problem of wrongful convictions is not limited to the small number of cases in which innocent defendants unsuccessfully contest their guilt in a jury trial. The article proceeds as follows: Part II describes the Rampart and Tulia cases in detail, and sets forth the circumstances leading to mass exonerations in those cases. Part III provides a brief description of data about wrongful convictions that has been generated in the literature to date. Part IV begins by describing the data used in this study. It then examines the Rampart and Tulia exonerations in more detail, identifying a subset of “actually innocent” exonerees in these cases that can be compared with exonerees in other studies. Part IV then undertakes a closer analysis of the circumstances and mechanisms that led to convictions of innocent Rampart and Tulia defendants. One of the most interesting contrasts between the mass exoneration cases studied here and other instances of wrongful conviction is that, as compared to other groups of exonerees, the great majority here pleaded guilty. Part V takes a closer look at this critically important phenomenon and offers some hypotheses as to why innocent defendants plead guilty in these cases at such a high rate. Disturbingly, the evidence suggests that the factors at work here— coercive penalties for contesting guilt at trial, coupled with few effective defense strategies and unsympathetic forums—may describe the prevailing conditions for a large number of, perhaps even most, criminal defendants. Part VI considers various types of police misconduct, documenting the prevalence of both “procedural perjury” and “substantive perjury,” and the fine line between them. Part VI then compares the experiences of actually innocent exonerees with those who the evidence suggests, though wrongfully convicted, were probably not actually innocent. Based on this comparative data, it finds evidence that innocence does dissuade some defendants from pleading guilty, but that any “innocence effect” has only a minor impact on guilty plea rates. Part VII briefly concludes. 6. One major exception is a new project attempting to compile a comprehensive catalogue of known exonerations, including mass exonerations, in a “National Registry of Exonerations.” See id. at 2. 2013] POLICE MISCONDUCT 1137 II. THE RAMPART AND TULIA EXONERATIONS On December 16, 1997, the L.A.P.D. arrested police officer David Mack on charges of stealing $722,000 from a Los Angeles area Bank of America. Three months later, the department fired two other officers, Brian Hewitt and Daniel Lujan, for severely beating a handcuffed prisoner in an interrogation room.7 The common thread was that all three officers were either former or current members of the Rampart CRASH, or Community Resources Against Street Hoodlums, unit.8 Rampart is an area covering 7.9 square miles to the northwest of downtown Los Angeles.9 It is the most densely populated portion of Los Angeles, with 36,000 people per square mile, and is widely known as a locus of gang activity. 10 At the same time, the Rampart CRASH unit had a reputation for operating in a largely autonomous fashion with little to no oversight.11 The arrest of Officer Mack and termination of Officers Hewitt and Lujan motivated L.A.P.D. Chief of Police Bernard C. Banks to form a special task force to investigate the Rampart CRASH unit.12 Then, on March 2, 1998, six-and-a-half pounds of cocaine disappeared from an evidence room in Los Angeles.13 Within a week, the special task force investigators honed in on Los Angeles police officer Raphael Perez, a member of Rampart CRASH, as the primary suspect. A year later, trial on the charge ended with a hung jury.14 Shortly thereafter, Perez cut a deal with prosecutors, agreeing to cooperate with a government investigation of police wrongdoing in the Rampart CRASH unit.15 Perez worked with investigators over the next year, divulging over 4,000 pages of interrogation transcripts.16 Perez’s testimony revealed police corruption on 7. BOARD OF INQUIRY INTO THE RAMPART AREA CORRUPTION INCIDENT, LOS ANGELES POLICE DEPARTMENT PUBLIC REPORT 2 (2000). 8. Id. at 1. 9. Id. at 55. 10. Id. 11. Id. at 69 (“Separate roll calls from the patrol division, a unique patch, jackets, an emphasis on narcotics enforcement, and an outward appearance of elitism were common CRASH traits that Rampart shared with other CRASH . . . units.”); see also id. at 77–78 (discussing the lack of oversight of the Rampart CRASH unit due to the physical separation of the unit from the rest of the Rampart division and other affirmative acts taken by the unit to isolate itself). 12. Id. at 1. 13. Peter J. Boyer, Bad Cops, THE NEW YORKER, May 21, 2001, at 67. 14. Id. at 69. 15. CRASH is the acronym given to L.A.P.D.’s specialized anti-gang units. Terry McDermott, Rafael Perez: The Road to Rampart, L.A. TIMES, Dec. 31, 2000, available at http://articles.latimes .com/2000/dec/31/news/mn-6831. 16. RAMPART RECONSIDERED: THE SEARCH FOR REAL REFORM SEVEN YEARS LATER 53 (2006). 1138 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 an unimagined scale, implicating police officers in wrongful killings, indiscriminate beatings and violence, theft, and drug dealing. Perez’s testimony also implicated dozens of police officers in systematic acts of dishonest law enforcement, exposing hundreds of instances in which evidence or contraband was planted on suspects, false statements were coerced or fabricated, and police officers offered perjured testimony in court. Perez’s confessions prompted the L.A.P.D. to re-name its investigative task force the “Rampart Task Force.”17 The Task Force was charged with corroborating Perez’s allegations of corruption within Rampart CRASH.18 What followed was, in the words of one independent commission, one of the “worst police scandals in American history.”19 Ultimately, the District Attorney was able to corroborate hundreds of Perez’s allegations and the L.A.P.D. entered into a consent decree with the U.S. Department of Justice, submitting to federal oversight of departmental operations.20 As a result of the scandal, more than three hundred prisoners filed writs of habeas corpus seeking to overturn allegedly tainted convictions, and approximately 156 felony convictions were dismissed or overturned as a result of “Rampart related” writs,21 110 of which were either initiated or unopposed by the District Attorney.22 The extent of wrongdoing by the L.A.P.D., however, remains a mystery to this day largely due to the overall ineffectiveness of the L.A.P.D.’s internal investigation of the police force.23 Although Officer Perez claimed that “ninety percent of the officers that work CRASH, and not just Rampart CRASH, falsify a lot of information” and “put cases on people,”24 no investigation or follow-up was ever undertaken to explore, or 17. Id. 18. Id. 19. REPORT OF THE RAMPART INDEPENDENT REVIEW PANEL 1 (2000), available at http://www. ci.la.ca.us/oig/rirprpt.pdf. 20. RAMPART RECONSIDERED: THE SEARCH FOR REAL REFORM SEVEN YEARS LATER, supra note 16, at 6–7, 37–38. 21. RAMPART RECONSIDERED: THE SEARCH FOR REAL REFORM SEVEN YEARS LATER App. C. Fourteen Rampart related writs remained pending at the time these statistics were compiled. Id. 22. Id. To the best of my knowledge, all of the writs initiated or unopposed by the D.A.’s office were granted. Courts were far more skeptical in writ cases initiated by defendants if the D.A. opposed the writ. Although relief was granted in approximately forty such cases, the large majority of contested writ applications were denied. 23. RAMPART RECONSIDERED: THE SEARCH FOR REAL REFORM SEVEN YEARS LATER, supra note 16, at 47 (concluding that basic facts regarding Rampart corruption scandal remain unknown seven years afterwards, including “[t]he extent of Rampart CRASH-like misconduct in the CRASH units of other divisions, other specialized units and LAPD policing generally”); see also id. at 54 (“The [L.A.P.D.] appeared to lack a clear and well-defined investigative approach and strategy and did not establish a plan for interagency coordination.”). 24. Id. at 53. 2013] POLICE MISCONDUCT 1139 even clarify, those allegations.25 In speaking with an investigative panel, some officers, who spoke anonymously out of fear of retribution, expressed concerns that the department did not genuinely seek to uncover the extent of the corruption.26 In fact, the L.A.P.D. failed to produce a promised “after-action” report which, according to the department, was going to include “the exact nature and disposition of each allegation.”27 Consequently, whatever the department may have uncovered about widespread corruption throughout the force remains outside the public domain.28 In the spring of 2003, while the Rampart story was winding down, news of another major police scandal broke, this time not out of a major metropolitan police force but instead in the tiny west Texas town of Tulia, located in Swisher County. The Tulia operation began as a roundup of suspected drug dealers in the summer of 1999, but transformed into what some described as a wholesale assault on the black residents of Tulia. The operation was spearheaded by a freelance agent named Tom Coleman. 29 Working undercover, Coleman claimed to have bought powder cocaine from more than 20% of the adult black residents of Tulia.30 In all, nearly fifty persons were convicted of selling drugs to Coleman, in most cases based solely on Coleman’s uncorroborated testimony.31 The first several Tulia defendants fought the drug charges at trial and were convicted and sentenced to draconian prison terms.32 After seeing the 25. Id. at 48, 54. 26. Id. at 65 n.95. 27. Id. at 62 (internal quotations omitted). 28. Id. at 63. 29. New York Times reporter Bob Herbert had covered the story even before Coleman’s credibility was shattered and the cases fell apart. See Bob Herbert, Tulia’s Shattered Lives, N.Y. TIMES (Aug. 5, 2002), http://www.nytimes.com/2002/08/05/opinion/tulia-s-shattered-lives.html?scp=7&sq= tulia&st=cse. 30. NATE BLAKESLEE, TULIA: RACE, COCAINE, AND CORRUPTION IN A SMALL TEXAS TOWN 5 (2005). 31. Id. 32. Id. at 6. Joe Moore was charged with making two drug deliveries: a single gram of crack cocaine and an eight ball of powder cocaine. With two prior drug felonies on his record, Moore faced a potential sentence, if convicted, of twenty-five to ninety-nine years in jail. Id. at 44. At trial, Coleman testified that he purchased the drugs at Moore’s house. Moore acknowledged that Coleman and another individual had come by his house, but contended that he chased Coleman away and that Coleman was lying about purchasing the drugs. Id. Apart from a small bag of cocaine that Coleman claimed to be the fruits of the sale, no evidence corroborated either of the alleged transactions. Id. In a contest of credibility pitting law-enforcement officer against convicted drug dealer, the jury banked on the officer. Moore was convicted and sentenced to ninety years. Id. at 59. The next two defendants, Chris Jackson and Jason Williams, were convicted based on virtually identical evidence. Id. at 82–83. Jackson received a twenty year sentence for allegedly selling an eight ball of cocaine. Id. at 83. Williams was convicted of multiple transactions and, based on the additional 1140 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 writing on the wall, however, most of the remaining defendants agreed to plead guilty. In all, forty-seven persons were charged and thirty-five were convicted. Of the twelve who were not convicted, several were placed on deferred adjudication.33 As these cases were tried, however, it became increasingly evident that Coleman’s testimony was not credible. Defense attorneys discovered that Coleman had been arrested on theft charges in a neighboring county and lied about it on his employment application to the task force.34 They also learned that Coleman had a history of employment problems, mental health problems, and significant unpaid debts.35 Worse still, it became increasingly evident that Coleman’s bosses in Tulia, as well as the prosecutor in the Tulia cases, knew of Coleman’s problems and lied about them under oath in the course of the Tulia trials.36 After the Texas Court of Criminal Appeals remanded four of the Tulia convictions for evidentiary hearings on claims that the prosecutor had failed to turn over material exculpatory evidence as required by Brady v. Maryland,37 hearings were conducted before a different trial judge. In the course of the hearings, it became clear that Coleman had perjured himself on numerous occasions during the Tulia trials, and that other law enforcement officials may have done so as well.38 Ultimately, the state agreed to a global settlement with defense attorneys in which it stipulated that Coleman was not a credible witness, vacated every conviction obtained as a result of the sting operation without seeking new trials against any of the defendants, and provided $250,000 to be divided among the defendants.39 In exchange, the defendants agreed not to sue the county.40 The state judge who presided at the hearing found “that Mr. allegation that the sales occurred in a drug-free zone, received a forty-five year sentence, notwithstanding that he was only nineteen years old and had no prior convictions. Id. Cash Love, one of the few white suspects arrested in the sting operation, was convicted of making eight sales, some in a drug-free zone, and was sentenced to 361 years. Id. at 92. Another defendant, Donnie Smith, was convicted of the first of several drug charges and sentenced to two years. Id. at 136. He pleaded guilty to the remaining charges and received a twelve and a half year sentence. Id. at 136–37. After Smith, two more defendants gambled on jury trials and lost. Freddie Brookins, Jr. was sentenced to twenty years. Id. at 157. Kareem Abdul Jabbar White received a sixty year sentence. Id. at 177. 33. Id. at 409. 34. Id. at 103–04. 35. Id. at 302–05. 36. Id. at 305–07, 388–89. 37. 373 U.S. 83 (1963). See BLAKESLEE, supra note 30, at 317. 38. BLAKESLEE, supra note 30, at 388–89. 39. Id. at 384, 386. 40. Id. at 385. 2013] POLICE MISCONDUCT 1141 Coleman had engaged in ‘blatant perjury’ and was ‘the most devious, nonresponsive law enforcement witness this court has witnessed in twenty-five years on the bench in Texas.’”41 Coleman was eventually tried and convicted of one count of perjury and sentenced to ten years probation.42 Although the settlement was contingent on approval by the Court of Criminal Appeals, when that approval was not immediately forthcoming, the Texas legislature passed a bill “specifically authorizing” the judge “to grant bond to the defendants.”43 Texas Governor Rick Perry then asked the Texas Board of Pardons and Paroles to review the cases. Pardons were granted to all thirty-five Tulia defendants convicted as a result of the sting operation.44 Two more individuals later were exonerated by courts.45 Rampart and Tulia together account for nearly two hundred cases of wrongful conviction and represent two large sets of exonerations stemming from police corruption scandals. But these are not the only major scandals that have recently beset law enforcement organizations in the United States, or even in Texas. In Hearne, Texas, numerous cases in 2001 were dismissed following revelations that a drug task force was systematically targeting black residents in an effort, allegedly, to drive them from the community.46 As in Tulia, the evidence against the defendants in these cases consisted solely of the uncorroborated assertions of a single, unreliable, police informant. Although most cases were dismissed prior to conviction, some innocent defendants pleaded guilty before the police wrongdoing was exposed.47 A year later, in the so-called “Dallas Sheetrock scandal,” at least thirty-nine criminal cases were dropped or dismissed after it was discovered that white powder allegedly recovered from criminal suspects and identified through field-tests as cocaine was actually ground up sheetrock packaged to look like cocaine. 41. Steve Barnes, Rogue Narcotics Agent in Texas Is Found Guilty of Perjury, N.Y. TIMES (Jan. 15, 2005), http://www.nytimes.com/2005/01/15/national/15tulia.html (last visited, Sept. 30, 2011). 42. BLAKESLEE, supra note 30, at 408. 43. Id. at 389. 44. A lawsuit brought against Coleman and the “26-county Panhandle Regional Narcotics Trafficking Task Force” by forty-five individuals caught up in the sting operation was settled in 2004 after the defendants in the lawsuit agreed to payment of six million dollars. See Barnes, supra note 41. 45. Janelle Stecklein, Judge reverses probation decision of Tulia brothers, AMARILLO GLOBENEWS (Feb. 5, 2011, 1:12 AM), http://amarillo.com/news/local-news/2011-02-05/decision-reversedtulians. 46. A.P., Texas: Suit Filed Over Drug Arrests, 11/2/02 N.Y. TIMES (Abstracts), 2002 WLNR 10823176. 47. See First Am. Compl. For Inj. Relief And Damages And Jury Trial at 5, Kelly v. Paschall, No. 02-A-02-CA-702 JN (W.D. Tex. Apr. 2, 2003). 1142 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 All of the victims in the scandal were blue-collar Mexican immigrants who spoke little or no English.48 Another Tulia-like scandal erupted more recently in St. Charles Parish, Louisiana, where seventy narcotics cases made by a single undercover officer were dismissed following revelations that the undercover officer had lied under oath in a criminal investigation. Before the scandal broke, at least twenty persons in cases made by the undercover officer had already pled guilty.49 An even larger Rampart-style corruption case has unfolded in Camden, New Jersey.50 Other incidents have also grabbed recent headlines.51 Gross and Shaffer have identified twelve separate incidents involving group exonerations based on police misconduct involving exonerations of at least 1,100 people.52 48. See Paul Duggan, ‘Sheetrock Scandal’ Hits Dallas Police: Cases Dropped, Officers Probed After Cocaine ‘Evidence’ Turns Out to Be Fake, WASH. POST (Jan. 18, 2002), http://www.washington post.com/ac2/wp-dyn/A229-2002Jan18?language=printer. 49. See Cop’s Alleged Past Ruins 70 St. Charles Parish Drug Cases, FOX EIGHT LIVE (Sept. 20, 2010, 2:40 PM), http://www.projectnola.com/the-news/news/42-fox-8/106713; Lori Lyons, St. Charles drug arrests in jeopardy after undercover cop discredited, authorities say, THE TIMES-PICAYUNE (Sept. 20, 2010, 8:15 PM), http://www.nola.com/crime/index.ssf/2010/09/st_charles_drug_arrests_in_jeo. html. To the best of my knowledge, however, and unlike Tulia, none of those guilty pleas have been overturned. 50. The full scope of the Camden scandal is still not clear. More than two hundred people have had convictions vacated or charges dismissed as a result of confessed misconduct implicating at least five officers in wrongdoing. The misconduct includes planting evidence on innocent persons and providing false testimony to convict them of crimes they did not commit. See Barbara Boyer, Two Former Camden Officers Face More Federal Charges, THE INQUIRER (Sept. 10, 2011), http://articles. philly.com/2011-09-09/news/30135226_1_original-indictment-amount-of-illegal-drugs-special-operati ons-unit. 51. In Tulsa, at least five Tulsa police officers have been charged with perjury and witness tampering. One defendant faced fifty-eight counts of wrongdoing. At least eleven people were released from prison as a result, with more cases under review. See Emory Bryan, Five Tulsa Police Officers Indicted in Corruption Probe, THE NEWS ON 6 (July 20, 2010, 9:17 PM), http://www.newson6.com/ story/12840428/five-tulsa-police-officers-indicted-in-corruption-probe?redirected=true. In Denver, one out of every seventeen police officers has been subject to administrative discipline for “‘departing from the truth’” or similar conduct in matters related to their official duties. That figure counts only those who have been formally sanctioned. It excludes those who are currently under investigation for similar violations, those who were investigated but insufficient proof of wrongdoing was presented to sustain a charge, and those whose misdeeds have not yet been detected. Christopher N. Osher, Denver cops’ credibility problems not always clear to defenders, juries, DENVER POST (July 10, 2011), http://www.denverpost.com/news/ci_18448755 (reporting that eighty-one officers still on the force out of 1,434 are on a list citing violators “in at least one of the following categories: departing from the truth, violating the law, making false reports, making misleading or inaccurate statements, committing a deceptive act, engaging in conduct prohibited by law, engaging in aggravated conduct prohibited by law, conspiring to commit conduct prohibited by law, soliciting or accepting a bribe, removing reports or records, destroying reports or records or altering information on official documents”). This list only includes the names of officers against whom violations have been formally substantiated. It does not include officers who are under investigation, or who were investigated but not cited. Id. 52. Gross & Shaffer, supra note 6, at 84. 2013] POLICE MISCONDUCT 1143 In short, Rampart and Tulia produced numerous exonerations and received a significant amount of national attention, but they are not unique. Revelations of large-scale police misconduct both preceded and post-dated them, suggesting that police misconduct leading to the wrongful conviction of innocent persons is a disturbingly common feature of the criminal justice system. III. WHAT WE KNOW ABOUT EXONEREES GENERALLY When it comes to wrongful convictions, very little hard data exists.53 After all, it is extraordinarily difficult to systematically identify erroneous convictions of innocent persons. Because the criminal justice process itself is assumed to provide the definitive test of guilt or innocence, there are few external Archimedean points from which its results may be tested. Our knowledge about innocent persons who are wrongfully convicted, therefore, is derived primarily from exonerations—that is, cases in which some government official, acting in an official capacity, has made a formal finding or declaration that a defendant is “not guilty of a crime for which he or she had previously been convicted.”54 Wrongful convictions have been the subject of academic inquiry since Edwin Borchard published his pathbreaking studies on the matter in the early part of the twentieth century.55 Other studies, including an influential article by Hugo Bedau and Michael Radelet, followed.56 Until quite recently, however, the leading study of criminal exonerations has been Gross, Jacoby, Matheson, Montgomery, and Patil’s analysis of exonerations occurring between 1989 and 2003 (“Gross Study”).57 That study has now been updated and greatly expanded in an examination of exonerations through 2012.58 Brandon Garrett has also made major contributions to the bank of knowledge of the exonerated through a series of articles, and a book (collectively, the “Garrett Study”) on DNA 53. See Robert Carl Schehr, The Criminal Cases Review Commission As A State Strategic Selection Mechanism, 42 AM. CRIM. L. REV. 1289, 1290–91 (2005) (describing the number of wrongfully convicted persons as a “dark number, a statistical unknown”). 54. Samuel R. Gross et al., Exonerations in the United States 1989 Through 2003, 95 J. CRIM. L. & CRIMINOLOGY 523, 524 (2005). 55. See Edwin M. Borchard, European Systems of State Indemnity for Errors of Criminal Justice, 3 J. AM. INST. CRIM. L. & CRIMINOLOGY 684 (1913); EDWIN BORCHARD, CONVICTING THE INNOCENT: ERRORS OF CRIMINAL JUSTICE (Yale Univ. Press 1932). 56. See Hugo Adam Bedau & Michael L. Radelet, Miscarriages of Justice in Potentially Capital Cases, 40 STAN. L. REV. 21 (1987). 57. Gross et al., supra note 54. 58. See Gross & Shaffer, supra note 6. 1144 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 exoneration cases.59 As the authors of both the Gross and Garrett Studies concede, the fact that an individual has been exonerated does not conclusively prove that the individual is actually innocent, although often the nature and circumstances of the evidence will leave little doubt. Nonetheless, formal exoneration is the best that our legal system is usually capable of doing, and thus provides the best indicator we have of instances in which an actually innocent person has been wrongfully convicted. These studies have identified some commonalities in cases resulting in exoneration. First, most exonerees were convicted of very serious crimes, typically resulting in sentences of death or long terms of imprisonment.60 Second, the vast majority of exonerees contested their guilt at trial. Only a tiny handful of exonerees, about 6%, pled guilty.61 This fact is particularly striking because the vast majority of criminal convictions, upwards of 90%, are obtained through guilty pleas. Third, although many types of procedural and evidentiary errors have been identified in cases of wrongful conviction, earlier studies consistently pointed to eyewitness misidentification as the leading cause of wrongful convictions, followed closely by faulty forensic evidence.62 Fourth, the studies suggest that persons of color are at far greater risk of false conviction than whites.63 The authors of these studies are quick to deny that the data is in any way representative of the wrongfully convicted more generally. As Gross and Shaffer observe in a more recent study, “[e]xonerations are unlikely, uncommon and unrepresentative of the mass of invisible false 59. Garrett’s main findings regarding the profile of innocent persons who were wrongfully convicted can be found in Brandon L. Garrett, Judging Innocence, 108 COLUM. L. REV. 55 (2008). Those findings were updated through 2010 in his book. See BRANDON L. GARRETT, CONVICTING THE INNOCENT (Harvard Univ. Press 2011). 60. Gross et al., supra note 54, at 535 (“With a handful of exceptions, everyone on our list of exonerees was sentenced to death or to a long term of imprisonment. Ninety-three percent were sentenced to ten years in prison or more; 77% were sentenced to at least twenty-five years. . . .”) (footnote omitted). 61. Id. at 536 (finding that 5.8% of exonerees (20/340) pleaded guilty); GARRETT, supra note 59, at 150 (reporting that 6%, or sixteen of 250 DNA exonerees, pleaded guilty). 62. Gross, supra note 2, at 542; GARRETT, supra note 59, at 48, 89 (reporting that eyewitness misidentifications factored in 76% of DNA exoneration cases, and faulty forensic testimony or evidence in 74%). Gross’s more recent study, which includes data from a broader source of exonerations, including the mass exonerations, finds that perjury or false accusation is the leading contributing factor to wrongful convictions overall, but that the prevalence of various contributing factors turns heavily on crime of conviction. See Gross & Shaffer, supra note 6, at 40. 63. Of the first 250 persons exonerated by DNA evidence, 62% were black, 30% were white, and 8% were Hispanic. Asians constituted less than 1% of the total. See GARRETT, supra note 59, at 5. Moreover, as Brandon Garrett has observed, although minorities are overrepresented in the prison population, their numbers among exonerees, or at least DNA exonerees, are even greater. Garrett, supra note 59, at 66. 2013] POLICE MISCONDUCT 1145 convictions.”64 Nonetheless, because their data is the best and most reliable that we have concerning wrongful convictions, there is an inevitable tendency to generalize or draw inferences about the characteristics, frequency, and causes of wrongful convictions from the data.65 As this Article will show, however, the mechanisms that produce exonerations in the police scandal cases differ substantially from those in other sorts of cases where exonerations have been most common. In trying to understand the role of police misconduct in causing wrongful convictions, then, it is imperative to deploy a narrow lens. It is impossible to know how frequently police misconduct of the type uncovered in Rampart and Tulia occurs, or how many wrongful convictions result from such misconduct. What happened in Rampart and Tulia appears to have involved widespread misconduct by police officers and prosecutors. In these cases, investigators discovered a culture of corruption that fostered official misconduct. Even if instances in which entire police departments, or at least entire units within a department, succumb to such cultural corruption are rare, the type of misconduct that led to the wrongful conviction of defendants in Rampart and Tulia could just as easily be perpetrated by smaller groups of corrupt officers, or even by officers acting on their own. Wrongful convictions resulting from occasional police misconduct involving only a single officer, or a relatively small group of corrupt police officers, scattered throughout the nation’s police departments, would be almost impossible to detect. And yet, the aggregate effect of such misconduct could easily generate a very large number of wrongful convictions. It is also possible that such cases may truly be rare. We simply do not have any way to know. In either case, the lack of attention paid to date to the mass exoneration cases has tended to reinforce some misconceptions about the causes and characteristics of the convictions of innocent persons. The vast majority of exonerations studied to date arose from murder and rape cases, in which defendants received typically severe sentences—often long prison terms or death sentences. The vast majority of these exonerations—some 94%— involved defendants who contested their guilt at trial and who were, as a result, able to pursue the full panoply of post-conviction remedies 64. Gross & Shaffer, supra note 6, at 9. 65. For example, based on this data some have concluded that the rate of false convictions is much lower among guilty pleas than trials. See, e.g., Larry Laudan, Is It Finally Time to Put ‘Proof Beyond a Reasonable Doubt’ Out to Pasture? 15 (U. of Tex. Public Law & Legal Theory Working Paper Series, Paper No. 194, 2011), available at http://ssrn.com/abstract=1815321. 1146 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 available to them. These defendants make up an unrepresentative pool in a criminal justice system characterized by overwhelming rates of guilty pleas.66 As a result, while the exoneration data currently available tells us something about how innocent people can be convicted in rape and murder cases at trial, it tells us very little about how innocent people might agree to plead guilty in cases involving the more mundane criminal offenses that make up the bulk of criminal courts’ daily workloads. Little research to date illuminates this important corner of the wrongful convictions problem.67 The present study seeks to partially rectify that problem in two ways. First, the study examines, to the extent available data permits, the causes and characteristics of the wrongful convictions identified in the mass exonerations. Second, the study contrasts that data with the data gathered in earlier exoneration studies to challenge some common assumptions about wrongful convictions more generally. By examining data from exonerations which arose in settings very different from the typical DNAbased exonerations, my goal is to provide more nuance to our understanding of wrongful convictions, to debunk some suggested inferences from earlier data sets, and to identify new avenues for investigation and reform. IV. ACTUALLY INNOCENT RAMPART AND TULIA EXONEREES A. The Data The data used in this study comes from two well-publicized incidents of systematic police misconduct, the Rampart and Tulia scandals. Both incidents involved dozens of criminal defendants. More than 150 persons were exonerated as a result of Rampart, and thirty-seven as a result of Tulia.68 Unlike some other recent police misconduct scandals, 66. See, e.g., United States Sentencing Commission’s 2011 Annual Report, Chapter Five: Research 33 (2011), available at http://www.ussc.gov/Data_and_Statistics/Annual_Reports_and_ Sourcebooks/2011/2011_Annual_Report_Chap5.pdf (reporting that 96.9% of federal felony convictions were the result of guilty pleas). 67. See Jon B. Gould & Richard A. Leo, One Hundred Years Later: Wrongful Convictions After a Century of Research, 100 J. CRIM. L. & CRIMINOLOGY 825, 834–35 (2010) (observing that “research offers few glimpses into errors” affecting “‘lesser’” felonies, and certainly misdemeanors”). 68. Thirty-five persons were pardoned by Texas Governor Rick Perry. Adam Liptak, Texas Governor Pardons 35 Arrested in Tainted Sting, N.Y. TIMES (Aug. 23, 2003), http://www.ny times.com/2003/08/23/us/texas-governor-pardons-35-arrested-in-tainted-sting.html. Two others were exonerated by courts on petitions for habeas relief. See Ex parte Barrow, 2010 WL 2618851 (Tex. Crim. App., June 30, 2010); Janelle Stecklein, Judge reverses probation decision of Tulia brothers, 2013] POLICE MISCONDUCT 1147 investigations of alleged police misconduct did not reveal wrongdoing until tens of defendants in Tulia, and hundreds in Rampart, had seen their cases through to conviction. Accordingly, the Rampart and Tulia cases provide an extensive set of data regarding not only how police misconduct can lead to wrongful arrest or wrongful charges, but to wrongful convictions. The vast bulk of the data on the Rampart scandal relied on in this study was extracted from files obtained from the Los Angeles District Attorney’s Office. These files contained office memoranda tracking developments in the Rampart investigation and most helpfully, writs filed by the District Attorney and defense counsel seeking relief for wrongfully convicted defendants. This data was supplemented with other information gleaned from official reports, newspaper articles, and other articles on the scandal appearing in the popular press and in academic commentary. With respect to Tulia, I relied extensively on the facts and case descriptions compiled by Nate Blakeslee in his thorough and engaging account of the Tulia scandal.69 I have cross-checked Blakeslee’s data, to the extent possible, with other published reports about Tulia, and with data made available to me by attorneys involved in the Tulia cases. Of the two, the Rampart material provides the greatest insight into how police misconduct “on the ground” can trigger a disastrous chain of events for innocent persons directly resulting in criminal convictions. Because the writs filed on behalf of wrongly convicted Rampart defendants often included narrative accounts of the circumstances of arrest, the Rampart cases provide an illuminating glossary of the many ways that police misconduct can lead to wrongful convictions. Study of these cases in the aggregate provides a fairly detailed empirical picture of wrongful convictions resulting from dishonest policing. The data pertaining to the Tulia cases shows less variation in the factual circumstances surrounding the charges, primarily because of the relatively uniform way in which the Tulia convictions were generated: each Tulia defendant was convicted based almost exclusively on the uncorroborated testimony of a single corrupt undercover agent. However, the Tulia data permits useful observations about the adjudicative procedures in such cases, and deepens the data pool in this regard. AMARILLO GLOBE-NEWS (Feb. 5, 2011), http://amarillo.com/news/local-news/2011-02-05/decisionreversed-tulians. 69. NATE BLAKESLEE, TULIA: RACE, COCAINE, AND CORRUPTION IN A SMALL TEXAS TOWN 409–17 (2005). 1148 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 The data breaks down as follows. Although more than 150 exonerations resulted from the Rampart scandal, the District Attorney’s files contained case-specific data for only ninety-seven of those cases, and detailed case data as to eighty-seven of these. None of the case information was complete in any sense of the word.70 As a result, there is more information about some cases than others, depending on the extent of factual detail provided in the affidavits and writs for habeas corpus. Not all of the individuals whose convictions were reversed or vacated, however, were “actually innocent” of the crimes for which they were convicted. Many defendants obtaining relief in Rampart did so because of procedural misconduct on the part of the police, not because the police were without evidence of wrongdoing. For example, many Rampart defendants were exonerated when it became clear that the police officers who had arrested them lied about the circumstances leading to the discovery of contraband. Where evidence of this sort of police misconduct surfaced convictions were rightly reversed, but there is no reason to believe that these defendants were not in fact engaged in criminal conduct. Prior exoneration studies have focused on cases involving what often is referred to as “actual” or “factual” innocence.71 Actual innocence cases are those in which either the wrong person was convicted of a crime committed by another, or a person was convicted of a crime that did not actually occur.72 In the first Gross study, all 340 exonerees had been absolved through “an official act declaring a defendant not guilty of a crime for which he or she had previously been convicted”73 premised on “strong evidence of factual innocence” and no “unexplained physical evidence of the defendant’s guilt.”74 The first Gross Study excluded from its purview exonerations in cases where the evidence indicated that the exonerees had been “involved in the crimes for which they were convicted.”75 The Garrett Study similarly focused only on those who “did 70. The D.A.’s office has maintained files from the Rampart cases, but the information was not well-organized and consisted mostly of the legal pleadings prepared in habeas corpus proceedings. We did not have access to the investigative files in individual cases. Useful information that might further clarify the circumstances in these cases no doubt could be found in such files, if they exist. In any event, we did not have access to them. 71. See, e.g., D. Michael Risinger, Innocents Convicted: An Empirically Justified Factual Wrongful Conviction Rate, 97 J. CRIM. L. & CRIMINOLOGY 761 (2007). 72. Bousley v. United States, 523 U.S. 614, 615 (1998) (“Actual innocence means factual innocence, not mere legal insufficiency.”). 73. Gross, supra note 2, at 524. 74. Id. at 524 n.4. 75. Id. at 527. 2013] POLICE MISCONDUCT 1149 not commit the charged crime,”76 adopting the criteria used by the Innocence Project to identify its set of innocent exonerees.77 Accordingly, I subdivided the Rampart cases into three categories: the “actually innocent” group, the “maybe innocent” group, and the “not innocent” group. In making these divisions, I followed Gross and Garrett in defining an actually innocent exoneree as one who was not involved in the commission of the supposed crime, did not commit the charged crime, or who was convicted of a crime that never occurred. In assigning the Rampart exonerees to these various groups, I relied heavily on the factual findings presented to reviewing courts by the District Attorney’s Office in petitions for writs of habeas corpus filed by the state, conceding the wrongful conviction and seeking the release of the defendant. I also relied on the factual statements included in the declarations and affidavits filed by investigating agents accompanying the D.A.’s filings. Where habeas petitions were initiated by defendants, I relied on factual allegations made by the petitioners only where those allegations were conceded in the state’s response. I also only assumed the truth of factual allegations made by defendants and their witnesses if the D.A.’s Office affirmatively stated in its filings that prosecutors or investigators had discovered evidence corroborating those accounts. Based on my review of the files, misconduct unrelated to the factual guilt or innocence of the defendant seems to have been the primary basis for exonerations in forty-nine cases.78 In those cases, defendants were exonerated because police officers lied about probable cause, about where a search took place, or about whether the suspect consented to a search. In thirty-eight cases, however, the police misconduct plainly did implicate the guilt/innocence determination. In these cases, police planted drugs or guns on suspects, lied about observing defendants committing crimes, or coerced confessions from innocent individuals. Where convictions were reversed based on reliable evidence of such misconduct, they constitute “exonerations” in the fullest sense of the term and are consistent with the 76. Brandon L. Garrett, Claiming Innocence, 92 MINN. L. REV. 1629, 1645 (2008). 77. GARRETT, supra note 59, at 285–86 (explaining that the list of 250 DNA exonerations matches that maintained by the Innocence Project, which conservatively defines DNA exonerations by, for example, omitting “cases in which there has been no exoneration despite DNA evidence of innocence,” and only includes cases in which there is no doubt that the “convicts are actually innocent”). 78. In twenty-seven cases, defendants were exonerated for procedural misconduct by the police, but there remained evidence of criminal culpability on the part of the defendants. Exonerees in these cases were coded as “not innocent.” In another twenty-two cases, the guilt or innocence of the exonerees was not clear from the record. These cases were coded as “maybe innocent.” For a further discussion, see infra Part VI. 1150 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 criteria used by others, including Gross and Garrett, who have studied the wrongful convictions of innocent persons. This study thus focuses primarily on those thirty-eight cases. Tulia supplies another thirty-seven cases. Thirty-five Tulia exonerations resulted from pardons recommended by Texas Governor Rick Perry. Two defendants were granted writs of habeas corpus vacating their convictions. Several other defendants were also wrongly charged or convicted in the scandal but were never pardoned. One such defendant was a minor who was sent to boot camp and who had already completed his sentence when Governor Perry recommended the Tulia pardons. Other defendants negotiated deals for deferred adjudication. Because convictions were never formally entered in those cases, pardons were not deemed necessary. Although these defendants too were wrongfully convicted, they were not included in the study because no official exonerations were ever granted in their cases.79 There are those who continue to assert that at least some of the Tulia defendants were, in fact, guilty. Indeed, about a half-dozen of the Tulia defendants admitted that they helped undercover agent Coleman purchase crack cocaine.80 None of the defendants, however, ever admitted involvement in the sale of powder cocaine, and it was the powder cocaine charges that provided the basis for the most serious sentences imposed on the Tulia defendants. Differentiating among Tulia defendants is made more difficult because the Tulia defendants were pardoned en masse, based on the fact that the cases were uniformly predicated on the word of a proven liar, and thus no formal individual findings of innocence were ever made. However, what evidence we do have points strongly toward innocence of virtually all of the Tulia defendants. First, in a sting resulting in the arrest of forty-six individuals, where most arrests occurred in the early morning hours, by surprise, at the suspects’ homes, not a single suspect was caught in possession of cocaine or crack.81 Second, none of the alleged drug transactions were recorded on audio or video tape. Indeed, there was virtually no corroborating evidence presented to implicate any of the defendants in the charged crimes. Third, the charges were relatively implausible by their nature. The defendants were drawn 79. At least one person wrongfully charged in the Tulia drug sting, Etta Kelly, did not receive a pardon because she pleaded guilty in exchange for deferred adjudication and thus a conviction was never actually entered in her case. See John Reynolds, Pardons Urged in Drug Cases, LUBBOCK AVALANCHE-J. (July 31, 2003), http://lubbockonline.com/stories/073103/reg_073103064.shtml. 80. BLAKESLEE, supra note 30, at 296. 81. See Herbert, supra note 29. 2013] POLICE MISCONDUCT 1151 from Tulia’s poorest classes where the drug of choice was crack cocaine. Yet all of the alleged transactions involved small amounts of powder cocaine. These facts strongly undermine the credibility of the charges and support the theory that Tom Coleman, the investigating undercover officer, may have been scamming the Drug Task Force for money by claiming to have engaged in fake sales and then logging into evidence powder cocaine that he himself severely diluted.82 In short, while it is possible that one or two of the Tulia defendants were in fact guilty, substantial evidence demonstrates that the vast majority of the Tulia defendants were innocent of any criminal wrongdoing, and the exonerations granted them by the Texas governor based on this evidence is sufficient to bring all of the Tulia defendants within the category of the “actually innocent.” B. Rampart and Tulia Exoneree Demographics The vast majority of those wrongfully convicted in Rampart and Tulia were persons of color. Although the data available for this study did not specify race or ethnicity of the Rampart defendants, an informal review of the surnames of the defendants strongly suggests that most, if not all, of the Rampart exonerees were of Hispanic origin. That conclusion fits with the population of the Rampart area, which is heavily Hispanic,83 and the demographics of the Rampart area street gangs that the CRASH unit at the center of the Rampart scandal policed. More precise data exists with respect to the Tulia defendants, the overwhelming majority of whom were persons of color. Of the thirty-five Tulia defendants who received pardons, thirty-one were black, two were Hispanic, and two were white. The average age of the Tulia exonerees was 29.8 years. There was insufficient data to determine the average age of the Rampart exonerees.84 With respect to gender, as is true in criminal law generally, the great 82. Coleman was subsequently convicted of perjury for false testimony given during hearings of Tulia defendants. See Steve Barnes, Rogue Narcotics Agent in Texas Is Found Guilty of Perjury, N.Y. TIMES (Jan. 15, 2005), http://www.nytimes.com/2005/01/15/national/15tulia.html?ref=tomcoleman. 83. One 2010 study of the area reported that 63% of Rampart’s residents are Latino, 26% are Asian-American, and 5% are African-American. The dominant ethnic groups in Rampart are Mexicans, Salvadorans, and Guatemalans. THE URBAN INSTITUTE, EVALUATION OF THE LOS ANGELES GANG REDUCTION AND YOUTH DEVELOPMENT PROGRAM: RAMPART ZONE PROFILE 2 (2010), available at http://www.urban.org/uploadedpdf/412274-Rampart.pdf. The two most violent gangs operating in Rampart are 18th Street and Mara Salvatrucha, or MS-13. Both gangs recruit heavily, if not exclusively, from Latinos. Id. at 3. 84. See Summary of Data (on file with author). 1152 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 majority of the exonerees from the Rampart and Tulia scandals were male. Of ninety-seven Rampart exonerees, approximately 90% were male.85 Interestingly, among the Rampart exonerees determined in this study to be “actually innocent,” all were male.86 Although there is no definitive answer as to why there were no females among the actually innocent Rampart exonerees, it is possible to speculate. The strong gender disparity might reflect an array of factors, including the fact that street gangs are predominantly, if not exclusively, male organizations; that the CRASH unit’s specific mission was to target gang activity in Rampart; and that females tend less often to be involved in the types of activities—streetlevel drug dealing and armed conflict—around which most of the false allegations arose. It might also be the case that false uncorroborated allegations of wrongdoing by males are more plausible than similar allegations against females, and thus dishonest police trying to lie credibly are more likely to make such allegations against male suspects. A somewhat larger percentage—19% (7/37)—of exonerated defendants in Tulia were female. An even larger percentage—24% (11/47)—of the total Tulia defendants were female.87 The somewhat smaller percentage of women who received pardons reflects the fact that more women had their cases dismissed prior to prosecution, negotiated a deferred prosecution, or otherwise avoided receiving the type of lengthy prison sentence for which a pardon was needed. C. Offenses of Conviction The types of crimes leading to wrongful convictions in the mass exoneration cases are strikingly different from those leading to exonerations in other cases. Whereas most known exonerees typically have been convicted of rape or murder, the vast majority of the exonerees in the police scandal cases were convicted of relatively low-level drug crimes. All thirty-seven of the Tulia exonerees were convicted of drug crimes,88 while nearly half of the actually innocent Rampart exonerees (18/38) were convicted of drug crimes.89 In addition, an almost equal number of actually innocent Rampart exonerees (16) were convicted of 85. Id. 86. Id. 87. See BLAKESLEE, supra note 30, at 409–17 (providing list of defendants). 88. Drug crimes here refer to any narcotics offense, including possession, transportion, and trafficking, of illegal narcotics, usually cocaine, crack cocaine, and heroin. 89. See supra note 84. 2013] POLICE MISCONDUCT 1153 gun possession offenses.90 Of these, defendants were most frequently charged as felons in possession of firearms, although California law provides a variety of unlawful gun possession offenses, making it unlawful for minors, gang members, probationers, and parolees to possess firearms as well.91 Actually innocent exonerees include persons convicted of each of these various offenses. A few were convicted of both drug and gun offenses. A larger number were initially charged with drug and gun offenses, but, through plea bargaining, negotiated a conviction on only one offense. Four of the actually innocent defendants also were convicted of assaulting police officers. Often, those convictions were enhanced with false allegations that the assailant used a gun or another deadly weapon to commit the assault. One Rampart exoneree was convicted of the offense of “giving false information to a police officer.”92 What these offenses of conviction primarily have in common is that they are all easily manufactured by the arresting officers. Drugs and guns are easily planted and, once “found,” constitute completed offenses. To the extent there were alleged victims in any of these cases, the victims uniformly were police officers. In none of the cases was there a need to obtain any corroborating evidence or eyewitness testimony from persons other than police officers. As a result, it was easy for police to falsely charge suspects with commission of these crimes, and extremely difficult for defendants to defend against them. Most of the actually innocent Rampart exonerees received relatively light sentences. Several of the exonerees were sentenced only to terms of probation. Most were sentenced to short prison terms ranging from 6 months to a few years. A few of the actually innocent defendants, however, received quite severe sentences. The median sentence of the actually innocent Rampart exonerees was two years. The average sentence was a little more than three years, and the disparity between the median and average sentence reflects the small number of severe sentences that were imposed on a few of the defendants. The most severe, and in many ways the most egregious, Rampartrelated sentence was imposed on Javier Ovando. In the case that probably did the most to trigger the Rampart scandal, police officers shot Ovando, then nineteen years old and a member of the 18th Street gang, four times 90. Id. 91. Id. 92. Pet. for Writ of Habeas Corpus at 2, In re James Edward Thomas, No. JJ04795 (Cal. Super. Ct. May 22, 2000). 1154 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 in the neck and the chest.93 In official reports of the incident submitted by police officers Rafael Perez and Nino Durden, the officers claimed that Ovando had broken into a vacant apartment where Perez and Durden were conducting surveillance. Ovando allegedly was armed with a semiautomatic rifle and a “military-style ‘banana clip.’”94 The officers said, and later testified under oath, that they shot Ovando after he refused to comply with their order to put the weapon down.95 As a result of the shooting Ovando was left paralyzed from the waist down.96 Despite the severe injuries he suffered and the fact that he had no prior felony convictions, Ovando was charged with two counts of assault with a firearm on a police officer and one count of exhibiting a firearm in the presence of a police officer.97 Firearm use enhancements were also alleged. A jury convicted Ovando essentially as charged, and the court sentenced him to twenty-three years and four months in state prison.98 The allegations regarding Ovando’s conduct, however, were pure fiction. Officer Perez subsequently admitted in a deposition that Ovando was unarmed at the time of the shooting, that the shooting was unprovoked, and that he and Officer Durden had planted a gun on Ovando to cover up that fact.99 According to Perez, the gun planted on Ovando had been obtained during a “gang sweep” a few days prior to the incident,100 and the serial number had been filed off so that the officers could use the weapon as a “throwaway.”101 Perez further stated that the gun was wiped clean of prints by the officers before it was placed next to the injured man.102 Thus, Ovando had the double misfortune of being shot and paralyzed, and then convicted of a serious crime he did not commit. Apart from the Ovando case, the longest Rampart sentences were imposed in drug cases. Russell Newman was sentenced to twelve years, 93. For a news analysis discussing details of the case, see Lou Cannon, One Bad Cop, N.Y. TIMES, Oct. 1, 2000, at SM32. 94. Pet. for Writ of Habeas Corpus, Ex. A at 2, Decl. of Richard A. Rosenthal, In re Javier Francisco Ovando, No. BA139642 (Cal. Super. Ct. Sept. 16, 1999). 95. See Cannon, supra note 93. 96. Id. 97. Id. 98. Rosenthal Decl., supra note 94, at 2. 99. Id. at 3. 100. Id., Ex. B at 1–2. The reasons for the shooting remain unclear. Perez testified that Durden shot Ovando during the course of an argument, and that he shot Ovando “reflexively.” Others have speculated that after Durden shot Ovando, Perez “tried to finish off Ovando so that there would be no witness to the crime.” Cannon, supra note 93, at SM37. Ovando himself has little recollection of the shooting, and “still has no idea why the officers shot him.” Id. 101. Cannon, supra note 93, at *5. 102. Id. 2013] POLICE MISCONDUCT 1155 Esaw Booker to nine years, and Walter Rivas to seven years, all for allegedly dealing cocaine.103 In comparison, the Tulia cases resulted in substantially harsher sentences. All of the Tulia defendants were convicted of selling small quantities of powder cocaine. Despite the small quantities, many of the defendants received draconian sentences, often as a result of prior felony convictions. William Cash and Joe Moore both received prison sentences in excess of ninety years. Kareem Abdul Jabbar White was sentenced to sixty years. Jason Jerome Williams and Kizzie White received sentences of forty-five and twenty-five years, respectively. The average sentence of the exonerated Tulia defendants was 157.8 months or a little over thirteen years. The median sentence, however, was thirty-six months, which was well below the average sentence primarily because of the large number of defendants who were sentenced to extended terms of probation in lieu of prison.104 Two defendants, Mandis and Landis Barrow, spent ten years in jail as a result of a probation revocation before a Texas court granted their habeas writ and ordered them released.105 D. Causes of Wrongful Conviction While the leading identified cause of wrongful convictions in past studies of exonerations is witness misidentification, a very different dynamic is at work in the police misconduct cases. Police misconduct generally, and perjury in particular, was the primary cause of wrongful convictions in every Rampart and Tulia case resulting in exonerations. Witness misidentifications played virtually no role in any of the cases. Police misconduct in these scandals took many forms. Police officers filed false police reports detailing observations of criminal conduct the defendants never engaged in, or describing circumstances that if true would have established criminal conduct. In most of the cases, police either physically planted drugs or weapons on the defendants and then lied about how they found the contraband, or simply misstated that they had 103. See Resp. to Pet. for Writ of Habeas Corpus at 2, In re Russell Newman, No. BA050279-01 (Cal. Super. Ct. Apr. 27, 2000); Resp. to Federal Pet. for Writ of Habeas Corpus at 2, In re Esaw Booker, No. BA057055-02 (Cal. Super. Ct. Feb. 15, 2000); Resp. to Pet. for Writ of Habeas Corpus at 2, In re Walter Rivas, No. BA165829 (Cal. Super. Ct. Feb. 15, 2000). 104. For purposes of calculating average and median figures, sentences of probation were treated as the equivalent of zero jail time. Because a later violation can result in a probationer serving the entire term in jail, the numbers understate actual punishment. 105. Janelle Stecklein, Judge reverses probation decision of Tulia brothers, AMARILLO GLOBENEWS (Feb. 5, 2011, 1:12 AM), http://amarillo.com/news/local-news/2011-02-05/decision-reversedtulians. 1156 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 found drugs or weapons when they had not. Police officers then testified to these same false facts at preliminary hearings and at trial in those rare cases that did not end in guilty pleas. For example, Emmanuel Chavez was arrested and ultimately convicted of possession of a firearm by a minor.106 In the arrest report, Officer Perez stated that “he and his partners observed” Emmanuel Chavez pass “a sawed-off shotgun” to another minor named Sergio Salcido.107 According to later evidence gathered by investigators, however, police never saw either minor handle a gun. Instead, Chavez and Salcido were stopped because police knew them to be members of a “tagging crew.”108 As Perez frisked Salcido, a gun dropped down Salcido’s pant leg and struck the pavement. The officers then made up a story that allowed them to charge Chavez as well as Salcido for possession of the gun.109 Similar police misconduct led to the wrongful conviction of Diego Barrios. Barrios and several others were socializing in the parking lot of a Jack-In-the-Box fast food restaurant when a “police car drove up to the group and shined its high-beam lights on the group.”110 The officers ordered everyone in the group to kneel down. They then searched and questioned each person. Four persons, including a juvenile by the name of Raymond C., were placed into a police car and taken to the police station. Unknown to the police, Raymond C. had a handgun in his possession at the time which he deposited, during the ride, behind the back seat of the squad car.111 Police discovered the gun after searching the car at the station and demanded to know who had dropped the gun. Raymond C. admitted the gun was his, but according to Barrios, “the officers said they did not ‘want’ a juvenile,” and instead “‘put the gun’ on Barrios.”112 Barrios pled guilty to a charge of unlawful gun possession. On a different occasion, police approached another group of youths in a parking lot. After police recovered a handgun from underneath a parked car, they arrested one of the youths and brought him to the station where they asked him, among other things, who owned the gun. When he failed 106. Pet. for Writ of Habeas Corpus at 2, In re Emmanuel Chavez, No. FJ10183 (Cal. Super. Ct. Feb. 23, 2000). 107. Id., Ex. A at 1. 108. Id. 109. Id. 110. Return and Answer to Pet. for Writ of Habeas Corpus, at 9–11, Decl. of Michael Gannon, In re Diego Barrios, No. BA126209 (Cal. Super. Ct. July 13, 2000) (recounting statements made by several witnesses). 111. Id. 112. Id. at 11. 2013] POLICE MISCONDUCT 1157 to provide an answer, the interrogating officer “told him that he was going to jail for the gun and rubbed it against Lobos’ fingers.”113 Lobos pled guilty at his arraignment to a charge of unlawful possession of a firearm by a felon.114 In these cases, contraband discovered by police in the possession of one person, or an unknown person, was attributed to others in order to permit an arrest to be made or to facilitate additional arrests. In other cases, police planted guns or drugs obtained elsewhere on suspects, or simply claimed that they found guns or drugs on suspects who in fact were not in possession of them.115 This is precisely what officers did in the Ovando case, where, after shooting Ovando, the officers planted a weapon on him picked up elsewhere to falsely implicate him in criminal conduct and cover up their own misdeeds.116 Other examples include the case of Ivan Oliver, who was charged with unlawful possession of a gun after police raided a party at which he was present. In Oliver’s case, police searched the residence where the party was held and located several guns. One officer then, investigators concluded, “arbitrarily decided who would be arrested for possessing them,” while other “CRASH officers created scenarios accounting for the recovery of each gun and . . . wrote the arrest report accordingly.”117 In several cases, defendants did not even know what offenses they were alleged to have committed until long after being arrested. One defendant, who was charged with narcotics possession, stated that he “did not find out why he was being arrested until he got to the jail and asked a jailer to tell him what his ‘pink slip’ indicated.”118 113. Pet. for Writ of Habeas Corpus, Ex. B, Decl. of Brian Tyndall, In re Allan Manrique Lobos, No. BA 131378 (Cal. Super. Ct. Jan. 31, 2000). 114. Pet. for Writ of Habeas Corpus, In re Allan Manrique Lobos, No. BA 131378 (Cal. Super. Ct. Jan. 31, 2000). 115. A gun was planted on Jose Armando Lara, for instance, by Officer Durden, and the weapon was booked into evidence only after Durden obliterated the serial number. See Pet. for Writ of Habeas Corpus, Ex. A at 1, Decl. of Richard A. Rosenthal, In re Jose Armando Lara, No. BA145000 (Cal. Super. Ct. Jan. 25, 2000). 116. See id. at 2. 117. Pet. for Writ of Habeas Corpus, Ex. A at 1, Decl. of Laura Laesecke, In re Ivan Oliver, No. BA135752 (Cal. Super. Ct. Feb. 11, 2000). 118. Pet. for Writ of Habeas Corpus, Ex. B at 1–2, Decl. of Brian Tyndall, In re Carlos Guevara, No. BA165829 (Cal. Super. Ct. Feb. 15, 2000). Another Rampart exoneree, Rene Matlong, similarly stated that “he had no idea he had been arrested for possession of a gun until he was given his paperwork several hours later at the jail division.” Pet. for Writ of Habeas Corpus, Ex. B, Decl. of Brian Tyndall, In re Rene Barela Matlong, No. BA131589 (Cal. Super. Ct. Apr. 14, 2000). Two other Rampart exonerees made almost identical allegations. Pet. for Writ of Habeas Corpus, Decl. of Brian Tyndall, In re Juan Torrecillas, No. BA143145 (Cal. Super. Ct. Apr. 7, 2000) (reporting that Torrecillas told investigators that “he had no idea he had been arrested for possession of cocaine” until 1158 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 Although present in every Rampart case, police perjury was not the sole cause of the false convictions. In some cases, police coerced an individual to make false statements inculpating the defendant. George Alfaro, for instance, was arrested for violating a gang injunction based on such evidence.119 According to police officers, Alfaro and two other suspects were arrested after police recovered a baggie of rock cocaine at the scene.120 The officers claimed that one of the suspects admitted that he possessed the drugs for sale. Rampart investigators, however, concluded that the drugs were planted at the scene, and the officers coerced the admission.121 As a result of the incident, Alfaro’s probation was revoked and Alfaro was sentenced to two years in state prison.122 In other cases, police simply falsely reported incriminating statements made by others. This happened in the case of Gregorio Lopez.123 Lopez and another man named Omar Alonso were arrested after police claimed they saw Alonso in possession of a magnetic key holder containing cocaine and Lopez attempt to discard a similar item. According to the arrest report, police searched Lopez and recovered a gun from his waistband. In fact, investigators found, the drugs said to belong to Lopez were planted, and the gun was found in his car rather than on his person. The prosecution’s case was also bolstered by inculpatory statements allegedly made by Alonso. No such statements, it turns out, were ever made, nor did the officers administer Miranda warnings as they claimed to have done.124 informed of charges at jail); Answer to Pet. for Writ of Habeas Corpus at 9, Decl. of Barbara Moulton, In re Alex Umana, No. BA144035 (Cal. Super. Ct. June 22, 2000). These accounts of being framed match up with Officer Perez’s own statements about how police typically went about framing suspects. “When we talk about planting or putting a case on someone, for some reason, some investigators or some attorneys have thought that we actually—I go into the car, take the three bindles [of drugs], lay it next to him and go, ‘See, that’s what you dropped.’ It’s not the way it works. I take them into custody, put them in the car, do whatever. And you know, when it’s time to book evidence, we go and get the evidence from our car. There’s no need to go and, you know, lay it on the ground next to them or put fingerprints on the baggies or anything like that. That’s just not the way it works. We get the evidence when it’s needed and move forward.” McDermott, supra note 15 (quoting Rafael Perez at LAPD Board of Rights hearing on June 3, 2000). 119. See Pet. for Writ of Habeas Corpus at 2, In re George Kenneth Alfaro, No. BA159394 (Cal. Super. Ct. Jan. 31, 2000). 120. Id. at 6. 121. Id. at 2. 122. Id. at 2. 123. Lopez is an alias. His real name is Leonel Ramos Estrada. Because court papers refer to Estrada as Lopez, I use that name here. 124. Pet. for Writ of Habeas Corpus, Decl. of Brian Tyndall, In re Omar Ramos Alonso and Leonel Ramos Estrada, No. BA148402 (Cal. Super. Ct. Apr. 10, 2000). 2013] POLICE MISCONDUCT 1159 Several of the Rampart exonerees falsely confessed, or were reported to have done so. Clinton Harris, for example, was convicted of possession of a firearm by a felon after police reported that they had observed Harris wearing a gun in the waistband of his pants, and that he had admitted that the gun was his, saying “[d]amn, I knew I shouldn’t have bought the gun. . . .”125 In fact, Harris never made any such statement. At the time of arrest Harris was seated on the couch of a friend’s apartment. Police officers entered the apartment without consent and found a gun on a table. According to Officer Perez, they decided to attribute the gun to Harris “because he was an ex-con.”126 Delbert Carrillo was arrested after police officers allegedly “‘noticed a large bulge in his front shirt pocket.’”127 The arresting officers explained in the police report: Knowing defendant to be on active parole and having a criminal history, we asked him what he had in his pocket (to ensure that it was not a weapon or narcotics). The defendant’s expression went from that of being calm to nervous, and he hesitantly reached into his pocket and removed a clear plastic baggy containing approximately nine white paper bindles, the type routinely used to package rock cocaine, and stated, “its [sic] rocks.” [W]e recovered the bag and found it to contain nine paper bindles, each one, containing approximately ten off-white wafers resembling rock cocaine.128 After Carrillo was arrested, police obtained a signed statement reading, with original misspellings, as follows: I DelBert Carrillo contacted officer Cohan and BRehm to discouse a matter at the time I had cocane in my posseion. and Because I new them I thought It would not Be a proBlem. Officers then overed it in my Shirt pocket. DEC. I make this statement freely.129 Carrillo was charged with possession for sale of cocaine base and ultimately pleaded guilty to an amended complaint that charged him with 125. Pet. for Writ of Habeas Corpus, Decl. of Barbara Moulton, In re Clinton Harris, No. BA140224 (Cal. Super. Ct. 2000) (alterations in original). 126. Id., Ex. A, Decl. of Natasha S. Cooper, at 1. Harris allegedly admitted that he was in possession of a gun at the time of the arrest. 127. Pet. for Writ of Habeas Corpus at 3, In re Delbert Carrillo, No. BA169722 (Cal. Super. Ct. Sept. 8, 2000) (internal quotation marks omitted). 128. Id. 129. Id. at Ex. C (statement form attached to police report). 1160 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 possession for sale of a controlled substance. He was sentenced to the statutory minimum term of two years. Carrillo later alleged that the drugs were planted and police officers coerced him into signing the statement by threatening to file additional charges against him if he refused.130 His conviction was vacated after the state discovered evidence corroborating Carrillo’s account of the incident.131 In short, then, the primary “cause” of false convictions in the Rampart and Tulia scandals was police perjury, some form of which was present in 100% of the cases. Innocent defendants who won exonerations primarily had been convicted in the first instance on the basis of the false reports and false testimony of corrupt police officers. That same police misconduct, however, was also responsible for the generation of other types of false evidence, including false witness statements and false confessions that supported the police officers’ false reports and perjurious testimony in court. After police perjury, the most common “causes” of false convictions were the false confessions generated through police misconduct. False confessions were present in about 13% of the Rampart cases. Interestingly, that figure is consistent with findings by Gross and Garrett on the approximate frequency of false confessions in wrongful conviction cases.132 While a substantial amount of commentary has focused on the problem of false confessions, and commentators have probed how innocent defendants might be induced to confess to crimes they did not commit, very little discussion exists regarding the problem of entirely fabricated confessions. Yet, as the Rampart cases show, some false confessions “occur” simply because police lie about what suspects actually said. When the mass exoneration data is added to the existing data regarding the causes of wrongful convictions, there is ample room to doubt the claim that witness misidentification is the leading cause of false convictions. Indeed, when the Rampart and Tulia cases are combined with the data gathered by Gross in his first study (which intentionally excluded these cases), perjury dislodges witness misidentification as the most prevalent cause of known wrongful convictions during the time period covered in 130. Answer to Pet. for Writ of Habeas Corpus at 2, In re Delbert Carrillo, No. BA169722 (Cal. Super. Ct. Nov. 9, 2000). 131. See id. at 3. 132. Gross found false confessions in 15% of the 340 exonerations examined in his first study. Gross et al., supra note 54, at 544. Gross’ more comprehensive second study also found false confessions in 15% of the 873 cases. See Gross & Shaffer, supra note 4, at 40. Garrett found false confessions in 16% of the 250 DNA exoneration cases he studied. GARRETT, supra note 59, at 18. 2013] 1161 POLICE MISCONDUCT the study. In the combined data set, perjury is a factor in 221 of the 415 exonerations of innocent defendants occurring between 1989–2003. Witness misidentifications are a close second, factoring into 219 cases during this same period. Gross’ more comprehensive study, which includes mass exonerations in the data set, confirms that perjury and false accusation, and not witness misidentification, is known to be the leading factor contributing to wrongful convictions.133 TABLE 1: CAUSES OF FALSE CONVICTIONS FOR EXONERATIONS Eyewitness Mis-i.d. Reported Perjury False Confession Gross Study (340) 64% (219/340) Mass Exon’s (75) 0% (0/75) Combined Total 53% (219/415) 43% (146/340) 100% (75/75) 53% (221/415) 15% (51/340) 13% (5/38)134 15% (56/378) While there of course is no way to know how generalizable these numbers are, the data does suggest that efforts to reform the criminal justice system in order to prevent wrongful convictions should include greater focus on the prevention of police misconduct. During the last decade, a major effort has been made to improve the reliability of lineup identification procedures. The revised data set suggests that those concerned with decreasing the incidence of wrongful convictions should devote similar attention to enhancing the integrity and reliability of police officer statements and testimony. E. Method of Conviction Perhaps the most striking insight to be drawn from the mass exoneration data concerns the high rate of guilty pleas seen in these cases, which provides strong evidence that the wrongful conviction problem extends to defendants who plead guilty as well defendants who contest guilt at trial. Earlier studies of exonerations found only a negligible number of innocents who were exonerated after pleading guilty. In Gross’ 133. Gross & Shaffer, supra note 4, at 40. 134. Excludes Tulia data because information about those investigations was not sufficient to make a determination. 1162 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 first study of 340 exonerations, only twenty of the exonerees, or approximately 6%, pled guilty.135 The vast majority, about 94%, were convicted after a trial. Garrett’s data tell the same story. In Garrett’s study of 250 DNA exonerations, sixteen, or 6%, pled guilty. The rest were convicted at trial. This data has been interpreted by some to mean that innocent people generally do not plead guilty, or if they do, they do so only under extraordinary circumstances.136 The accuracy of guilty pleas is a major determinant of the scope of the problem of wrongful convictions. After all, the vast majority of criminal convictions, upwards of 90%, are a result of guilty pleas.137 If innocent people do not plead guilty but rather insist on going to trial, then the upper estimate of wrongful convictions is bounded by the small proportion of persons overall who are convicted at trial. In other words, even if 100% of defendants who were convicted at trial were actually innocent, the wrongful conviction “rate” would still be only about 5%, since approximately 95% of all defendants plead guilty. If, on the other hand, the fact that a defendant pleads guilty provides no guarantee that the defendant is not actually innocent, then the potential magnitude of the wrongful conviction problem is many times greater. Even if the rate of false guilty pleas is low, the far-greater size of the guilty plea pool ensures that it adds up to a quantitatively large problem. 1. Evidence that the Innocent Do Plead Guilty138 It has long been apparent that the innocent do, on occasion, plead guilty.139 The more important question, however, is how often false guilty pleas occur, and how false guilty plea rates compare with false trial 135. Gross, supra note 2, at 536. 136. See, e.g., Ronald J. Allen and Larry Laudan, Deadly Dilemmas, 41 TEX. TECH L. REV. 65, 71 (2009) (concluding that false guilty plea rate is much lower than false conviction rate at trial based on evaluation of Garrett data). 137. See Sourcebook of criminal justice statistics online (2006), http://www.albany.edu/source book/pdf/t5462006.pdf. 138. See Pet. for Writ of Habeas Corpus at 11, Decl. of Brian Tyndall, In re Gerald Peters, No. BA131401 (Cal. Super. Ct. May 9, 2000) (reporting that “Peters plead[ed] guilty to the charges on the advice of his attorney because he believed he would face a stiffer penalty if he chose to fight the charges in a trial and lost”); Scott Glover & Matt Lait, 10 More Rampart Cases Voided, LA TIMES (Jan. 26, 2000), http://www.streetgangs.com/topics/rampart/012600more10.html (“Davalos, 41, an upholstery worker who served 91 days in jail. He said he only agreed to a plea bargain because he was threatened with eight to 16 years in prison.”). 139. For a list of sources discussing the problem of innocent persons pleading guilty, see Gabriel J. Chin & Richard W. Holmes, Jr., Effective Assistance of Counsel and the Consequences of Guilty Pleas, 87 CORNELL L. REV. 697, 740 n.305 (2002). 2013] POLICE MISCONDUCT 1163 conviction rates. If there ever was any real doubt that false guilty pleas can occur in large numbers, the Rampart and Tulia data put those doubts to rest, indicating that at least in some types of cases, innocent defendants are far more likely to be convicted through a guilty plea than at trial. In the Rampart cases not involving alleged probation violations, twenty-five of thirty-two exonerees pled guilty. In Tulia the numbers were about the same: twenty-seven of thirty-four. Overall, fifty-two of the exonerees, or 81%, were convicted through guilty pleas, and twelve, or 19%, were convicted after trial. Those numbers represent a far more typical distribution of guilty pleas and trial convictions than was seen in the Gross and Garrett data, and provide strong reason to believe, notwithstanding prior exoneration studies showing a low incidence of guilty pleas among exonerees, that the problem of wrongful convictions is not contained to those who contest their guilt at trial.140 Indeed, the mass exoneration cases make clear that, at least with respect to the types of charges at issue in the Rampart and Tulia cases, the method of conviction makes very little difference to the reliability of the conviction. In Rampart and Tulia, wrongful convictions resulted from guilty pleas and trials alike, and as is true in the criminal justice system generally, guilty pleas accounted for the majority of the convictions. TABLE 2: METHOD OF CONVICTION OF EXONERATED DEFENDANTS Guilty plea Trial conviction Gross Study Mass Exon’s 6% (20/340) 94% (320/340) 81% (52/64) 19% (12/64) Combined Total 18% (72/404) 82% (332/404) As Table 2 suggests, in comparison with other exonerees, the Rampart and Tulia exonerees pled guilty at much higher rates. The percentage of convictions obtained through guilty pleas, however, still falls short of the average. What is the significance of these numbers? On one hand, a trial rate approaching 20% in low level drug cases might seem remarkable. These are typically simple cases to prosecute and the vast majority of such cases undoubtedly would normally be resolved through guilty pleas. On the other hand, the evidence in many of these cases, especially the Tulia 140. See, e.g., Ronald J. Allen & Larry Laudan, supra note 136, at 71 (calculating wrongful conviction rate by guilty plea based on Garrett study data showing that 9/200 exonerees pled guilty). Michael Risinger critiques that reasoning in a responsive essay. See D. Michael Risinger, Tragic Consequences of Deadly Dilemmas: A Response to Allen and Laudan, 40 SETON HALL L. REV. 991, 995 (2010). 1164 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 cases, was extremely weak. These were cases built on the testimony of a single undercover cop, with no electronic recording of the transactions or corroborating evidence, and in some cases in the face of significant alibi defenses. Given this, it is difficult to tell whether a 20% trial rate is high or low. Regardless, the fact that so many mass exoneration cases were resolved by guilty pleas should erode any perception that actually innocent defendants almost uniformly refused to plead guilty. Of course, it is possible that the Rampart and Tulia cases, rather than the exoneration cases studied by others, represent the outlier. There are several reasons, however, to believe that with respect to the frequency of false guilty pleas, the Rampart and Tulia cases provide a more typical distribution, and that wrongful convictions, like convictions generally, are usually the end product of a guilty plea rather than a trial verdict. First, as both Gross and Garrett acknowledge, cases resulting in exonerations are the beneficiaries of a phenomenally rare confluence of events that are simply not present in typical cases.141 For a DNA exoneration to occur, for example, the crime must have been one involving biological evidence, where that evidence is dispositive of the defendant’s guilt. That evidence must have been gathered but not tested, or not tested properly, preserved for years or decades, and located in quantities and in sufficient quality to permit testing, and defendants must have preserved the means to launch a legal challenge against their conviction once the evidence is discovered. The preconditions for exoneration after a trial conviction are only rarely satisfied; rarer still will they exist where the defendant pleads guilty. In guilty plea cases, the state is less likely to preserve evidence for later testing, and because no trial record exists, even where such evidence was preserved, it is difficult to assess the significance of exculpatory test results. Defendants who plead guilty typically waive their rights to appeal and to post-conviction review.142 As a result, innocent people who plead guilty almost invariably lack a viable procedural mechanism to prove their innocence in a post-conviction proceeding, at least absent the type of extraordinary circumstances that occurred in Rampart and Tulia. To make matters worse, many statutes governing access to post-conviction DNA 141. See, e.g., Gross & Shaffer, supra note 6, at 4–5 (explaining how exonerations tend to be the product of “blind luck” or “improbable chains of happenstance”). 142. See Nancy J. King, Appeal Waivers And The Future Of Sentencing Policy, 55 DUKE L.J. 209, 209 (2011) (finding in her empirical study that defendants who pleaded guilty waived their right to appeal in nearly two-thirds of cases). 2013] POLICE MISCONDUCT 1165 testing specifically preclude defendants who plead guilty from obtaining testing.143 There is good reason, moreover, to view the mass exoneration cases as far more typical of garden-variety wrongful convictions than the cases included in the earlier Gross and Garrett studies, a very large percentage of which (100% in the Garrett study) involved post-conviction DNA testing. In the Rampart and Tulia cases, most defendants were convicted of drug or gun crimes, which are far more common than the rape, murder, or rapemurder convictions making up the vast majority of the earlier studied exoneration cases. Although some sentences were draconian, especially in Tulia, most sentences were relatively modest in severity, as are most felony sentences imposed on typical felony convicts. As noted above, most of the Rampart exonerees received relatively light sentences, with the average sentence being approximately three years, and the median sentence less than two years. These figures are consistent with national averages for state felons.144 In contrast, exonerees in the first Gross study had almost uniformly received harsh sentences for the most serious crimes. This was especially true among the non-DNA exonerations in the data pool, of which 85% (166/196) were serving sentences for murder or manslaughter, and 22% among all of the exonerees (74/340) were sentenced to death.145 Moreover, the exonerations in Rampart and Tulia were largely the product of happenstance. The Rampart exonerations in particular involved run-of-the-mill drug and gun cases that never would have received even passing interest from the outside world had it not been for the cooperation deal struck by Rafael Perez. Unlike typical DNA exonerations, the exonerations in Rampart came about without the intervention of Innocence Projects or big-firm pro-bono advocacy. There were few trial transcripts, physical evidence, or other compelling evidence from which a defendant’s actual innocence could be determined.146 The Tulia exonerations did benefit from substantial pro-bono advocacy, but one suspects that none of the events leading to the uncovering of misconduct in Tulia would have been uncovered had the extent of the misconduct not been as sweeping, the sentences not as draconian, and the racial component not as overt as it 143. See, e.g., FLA. STAT. ANN. § 925.11(1)(a)(2) (West 2006). But see TEX. CODE CRIM. PROC. ANN. art. 64.03(b) (West 2007). 144. See supra note 137. 145. Gross, supra note 2, at 531. 146. This, in part, was a necessary byproduct of a set of convictions obtained largely through plea bargains. 1166 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 was. Setting aside the extraordinary manner in which the police misconduct was discovered, the kinds of convictions at issue in Rampart and Tulia were far more typical, substantively and procedurally, than those that have eventuated in DNA exonerations. V. EXPLAINING WRONGFUL PLEAS In addition to providing an empirical basis for the claim that innocent people plead guilty, the mass exoneration cases vividly illustrate how and why actually innocent defendants plead guilty. In general, there appear to have been three main factors driving innocent Rampart and Tulia defendants to plead guilty: an outsized trial penalty, a lack of viable strategies to contest the charges, and presumptively or actually unsympathetic forums. Each is considered briefly below. A. New Data on the Trial Penalty Without a doubt, the overwhelming reason that innocent Rampart and Tulia exonerees pleaded guilty to crimes they did not commit was that they feared that they would do much worse at trial if they did not plead guilty. Typical are the sentiments expressed by one innocent Rampart exoneree who on advice of his attorney pleaded guilty in exchange for a three-year term of probation, believing that “he would face a stiffer penalty if he chose to fight the charges in a trial and lost.”147 That exoneree likely was not wrong. The existence of a trial penalty has been long acknowledged, albeit bemoaned by many.148 It is an institutionalized feature of contemporary criminal justice. Nonetheless, the coercive impact of the trial penalty is unmistakable, and is plainly evident in the Rampart and Tulia cases. Tulia provides an extreme example of the coercive impact of the trial penalty. Of the thirty-seven innocent Tulia exonerees, seven went to trial and were convicted, twenty-seven pleaded guilty, one did both, and two 147. In re Gerald Peters, supra note 138. Peters also alleged that he was physically abused by officers in an interview, but that “he never made a complaint regarding this incident because he felt ‘it would do no good.’” Id. Similarly, two months after pleading guilty, Ruben Rojas had second thoughts and wrote a letter to the judge who had sentenced him. In the handwritten letter, Rojas explained that “I was informed that I was facing 25 years to life by my defense counsel and that there was no way I could have won my case because I was up against a police officer.” He added: “I never did what I was charged for. . . I’m not guilty.” Matt Lait, Another Inmate Set to Be Freed in Police Probe, L.A. TIMES, Nov. 17, 1999, at A1. 148. See, e.g., Candace McCoy, Plea Bargaining as Coercion: The Trial Penalty and Plea Bargaining Reform, 50 CRIM. L.Q. 67 (2005). 2013] POLICE MISCONDUCT 1167 others had their probation revoked.149 The first defendant to go to trial, Joe Moore, was convicted and sentenced to ninety years in prison for allegedly dealing 4.5 grams of cocaine.150 Moore had been offered an opportunity to plead guilty in exchange for a twenty-five-year sentence (the minimum available given the charges and Moore’s prior record), but he declined.151 Six more defendants stood trial, and were convictedand sentenced to prison terms ranging from 20 to 361 years.152 In light of this precedent, and with cases substantively indistinguishable in terms of the nature of the charges and the strength of the evidence, the remaining defendants all chose to plead guilty.153 Although the sentences imposed on those who pleaded guilty in Tulia were often quite harsh, the harshness of their sentences paled in comparison to those who were convicted at trial. On average, Tulia defendants who pleaded guilty were sentenced to approximately four years in prison.154 The Tulia defendants who contested their guilt at trial received an average sentence of 615.4 months, or 51.3 years.155 Trial sentences at Tulia, in other words, were nearly thirteen times harsher than sentences imposed following guilty pleas. The trial penalty evident in Tulia might be attributed, at least in part, to an apparently intentional prosecutorial strategy to frighten defendants into foregoing trial. Such an express strategy was made easier in small-town Tulia, where word of harsh sentences quickly spread among Tulia’s small defense bar and the defendants themselves. These dynamics were noticeably absent in Rampart. Unlike Tulia, there is no indication that prosecutors were aware of the defects in the cases they brought against innocent defendants. Indeed, after the scandal broke, the Los Angeles District Attorney’s Office took affirmative steps to investigate the scope of wrongdoing and to vacate convictions resulting from police misconduct.156 In terms of size and population, the L.A. justice system also obviously dwarfs Tulia’s. There is far less reason to believe 149. Donald Wayne Smith was charged with seven drug trafficking offenses, and the prosecutor elected to try the cases separately. See BLAKESLEE, supra note 69, at 117. After Smith was convicted in the first case and sentenced to two years in prison, he accepted a plea offer to resolve the remaining charges in exchange for a 12.5 sentence to run concurrently with his other conviction. Id. at 136–37. 150. Id. at 59. 151. Id. at 48. 152. Cash Love was sentenced to 361 years by the trial court. Id. at 92. 153. See id. at 160–61. 154. See Summary of Data (on file with author). 155. Id. In calculating this figure, I counted Cash Love’s sentence as 99, rather than 361, years. I also omitted Smith’s case. Smith’s two-year sentence was based on the least serious of only one of seven charges. 156. See generally CONSTANCE L. RICE ET AL., RAMPART RECONSIDERED: THE SEARCH FOR REAL REFORM SEVEN YEARS LATER (2006). 1168 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 that prosecutors sought to send any messages to specific classes of defendants by seeking harsh trial sentences. Any implicit threat inherent in the harsher trial sentences would seem to be endemic to the justice system in general. Nonetheless, the observable trial penalty in the Rampart cases, though not on the same order as the average trial penalty in Tulia, was still quite large. On average, actually innocent Rampart defendants who were convicted at trial were sentenced to 101.25 months, or nearly 8.5 years.157 Actually innocent Rampart defendants who pleaded guilty were sentenced to an average term of 18.5 months, or just over 1.5 years.158 Defendants who contested their cases at trial, in other words, received sentences on average more than five times harsher than those who agreed to plead guilty.159 The trial penalty for the larger sample of all Rampart exonerees, including those who did not appear to be actually innocent, was even bigger. For this group, the average plea sentence was 20.3 months.160 The average trial sentence was 136.3 months.161 Trial sentences were therefore on average 6.7 times longer than plea sentences, with no apparent qualitative differences among the types of crimes charged or the criminal history of the defendants.162 The longest sentence imposed on any Rampart exoneree was a term of fifty-four years to life, later reduced on appeal to twenty-nine years to life, 157. See Summary of Data (on file with author). 158. Id. 159. Arguably, one could object that these numbers are skewed by the inclusion of the Ovando case—by far the harshest sentence imposed in any of the Rampart cases. There are several reasons, however, to include that case in calculating the numbers. First, Ovando was not the only innocent Rampart exoneree to be charged with a crime of violence. Jose Perez was charged with assaulting a peace officer with a firearm, the same crime charged against Ovando, and pleaded guilty. Perez received a sentence of three-years probation. Raul Munoz and Cesar Natividad were also (falsely) charged with assaulting a peace officer with a deadly weapon—in their case—allegedly attempting to run over a police officer with a car. Both settled the cases by guilty plea. Munoz was sentenced to three-years prison, and Natividad was sentenced to a three year term of probation. None of these defendants suffered the kinds of serious injuries that Ovando did. Ovando’s trial sentence was thus about eight times harsher than the harshest plea sentence for comparable conduct, and if one treats a term of probation as equivalent to about one-third of a prison term, his trial sentence exceeded the average plea sentences in the three cases by 1400%. In Ovando’s case, prosecutors offered Ovando a thirteen-year deal to resolve the case, but Ovando’s lawyer thought the offer was “way too severe” and turned it down. Lou Cannon, One Bad Cop, N.Y. TIMES MAG., Oct. 1, 2000, available at http://www.nytimes.com/2000/10/01/magazine/one-bad-cop.html. 160. See supra note 159. 161. Id. 162. In calculating these figures, there were sixty-one plea sentences and twelve trial sentences included in the data. Sentencing data was unavailable in some of the cases. In other cases, defendants were never sentenced because they failed to appear, and a bench warrant issued. Terms of probation were not counted as punishment. Some defendants were deported as a result of their convictions. Deportation was also not counted as punishment in calculating average sentences. 2013] POLICE MISCONDUCT 1169 for Lorenzo Nava. Like most of the other Rampart exonerees, Nava was convicted of drug and gun offenses and contested the charges at trial.163 After conviction, Nava received an initial fifty-four-year sentence under California’s three-strikes law.164 Nava’s case, however, can be compared to Joseph Jones, another Rampart defendant, to show that the long trial sentence imposed on Nava was not simply a function of the three-strikes law or other factors unique to his case. Like Nava, Jones was charged with multiple drug counts and was potentially subject to prosecution under the three-strikes law. According to Detective Chris Barling, who interviewed Jones in the Rampart investigation, “Jones believed that he was facing a life term,” and notwithstanding his contention that he was innocent, decided to plead guilty on the advice of counsel.165 Pursuant to the plea, Jones was sentenced to a prison term of eight years.166 The disparity in sentence outcome between Nava and Jones is roughly consistent with the average trial penalty evident in the Rampart cases, amounting to at least a seven-fold penalty increase based on Nava’s initial trial sentence, and a four-fold penalty increase based on Nava’s reduced sentence on appeal.167 This data provides further evidence that the real trial penalty could be far larger than estimated in some studies.168 With trial sentences ranging anywhere from four to thirteen times longer than plea sentences, the costs of contesting a typical felony charge are prohibitive. Few defendants can afford to run the risk. The experience of those wrongly convicted in the Rampart and Tulia scandals demonstrates that the coercive power of the 163. See Summary of Data (on file with author). 164. Id. 165. Pet. for Writ of Habeas Corpus, Ex. B at 1, Decl. of Chris Barling, In re Joseph Jones, No. BA154853 (Cal. Super. Ct. Nov. 10, 1999). 166. Pet. for Writ of Habeas Corpus at 2, In re Joseph Jones, No. BA154853 (Cal. Super. Ct. Nov. 10, 1999). 167. These numbers represent minimums because they ignore the upper end of the sentencing range (life) and are based on the minimum sentence that Nava was required to serve. 168. Hans Zeisel estimated that trial sentences were 42% more severe than guilty plea sentences. See Hans Zeisel, The Disposition of Felony Arrests, 1981 AM. B. FOUND. RES. J. 407, 444–49 (1981). Other scholars have reached varying estimates. See, e.g., Nancy J. King et al., When Process Affects Punishment: Differences in Sentences After Guilty Plea, Bench Trial, and Jury Trial in Five Guidelines States, 105 COLUM. L. REV. 959 (2005) (finding that trial penalties varied from state to state, and offense to offense, ranging from 13% to 461%); Jeffery T. Ulmer & Mindy S. Bradley, Variation in Trial Penalties Among Serious Violent Offenses, 44 CRIMINOLOGY 631, 652 (2006) (finding that defendants convicted at trial received sentences 57% longer than defendants who pleaded guilty). The Federal Sentencing Guidelines provide a 35% discount for defendants who “accept responsibility.” Because of the variety of charge bargains that typically accompany federal plea bargains, actual discounts are almost always far larger than 35%. See Russell Covey, Reconsidering the Relationship Between Cognitive Psychology and Plea Bargaining, 91 MARQ. L. REV. 213, 224–30 (2007). 1170 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 trial penalty causes innocent defendants as well as guilty ones to plead guilty. B. Lack of Effective Trial Strategies for Falsely Accused Defendants Time and again, actually innocent defendants asked by investigators to explain why they pled guilty repeated a common mantra: it was their word against that of the police, and who were the prosecutors, judges, or jurors going to believe?169 While this dynamic is present in most cases, it is especially likely to have an effect where defendants have reason to believe they will not be treated fairly. Most of the Rampart exonerees were gang members, some with criminal histories. They were likely correct in believing that few middle-class jurors would give credence to their claims of police misconduct. In Tulia, racial dynamics clearly affected the calculations of the black defendants, who assumed (correctly, given the trial outcomes) that their protestations of innocence would be ignored. In these cases, innocent defendants often had little except their own word to prove their innocence, and their word was demonstrably not enough. In part because of the nature of the cases, and in part because of their lack of resources, the defendants were typically unable to amass any credible exonerating evidence. Given that the police already had demonstrated a willingness to testify falsely,170 many defendants realized that a successful trial defense was unlikely and simply decided to cut their losses. C. Unsympathetic Forums A third reason so many innocent Tulia and Rampart defendants pled guilty, even in cases where the evidence was flimsy, was an undoubtedly accurate perception that the system itself was not constituted in a way likely to give them much chance of prevailing. As one blue ribbon panel observed after investigating the Rampart scandal, the Los Angeles County criminal justice system is characterized by “assembly-line” justice.171 Many actors are complicit in pressuring innocent defendants to plead 169. See Pet. for Writ of Habeas Corpus, Decl. of Michael J. Hansen, In re William Zepeda and Argelia Diaz, No. BA156980 (Cal. Super. Ct. Mar. 17, 2000) (reporting that Zepeda “decided to plead guilty to the charge after he realized it was just his word against the officers”). 170. For example, Rafael Zambrano, who was charged with violating probation for unlawful gun possession after police planted a gun on him, claimed that he “decided to plead guilty to the charge after Officer Rafael Perez testified at his preliminary hearing.” See Pet. for Writ of Habeas Corpus, Ex. B, Decl. of Brian Tyndall, In re Rafael Zambrano, No. BA138148 (Cal. Super. Ct. Feb. 11, 2000). 171. RAMPART RECONSIDERED, supra note 16, at 49. 2013] POLICE MISCONDUCT 1171 guilty to crimes they did not commit, including prosecutors who “pressure defendants to accept plea deals in extremely short time frames,” “overworked public defenders” who counsel their clients to accept those pleas, and judges who are quick to impose “draconian” sentences on those who “drain judicial resources by demanding a trial.”172 Moreover, prosecutors are reluctant to doubt the credibility of the police officers with whom they work daily, and judges “are unwilling or unable to pursue their own suspicions of police perjury or misconduct.”173 As a result, a falsely accused defendant has little reason to believe that he will fare well by going to trial, and has great reason to believe that he will be much worse off by refusing to take a plea and cut his losses. Ample evidence also suggests that judges often are biased toward the prosecution.174 A large part of the bench is populated by former prosecutors. These former prosecutors often have difficulty shedding their former roles. Regardless of background, judges often form relationships with prosecutors who appear regularly in their courtrooms, and many think of themselves as part of a “law-enforcement” team. In addition, electoral politics drive many judges to more pro-prosecution positions. Some judges even campaign overtly on being “tough on crime” or “hard on criminals.”175 Actually innocent defendants tried before such judges likely are often led to believe, probably correctly, that they will not get the benefit of the doubt should they go to trial. Arguably, pro-prosecution judges played an especially prominent part in many of the cases in which actually innocent defendants were convicted at trial. The two judges presiding over the Tulia prosecutions initially barred defense lawyers from impeaching undercover agent Tom Coleman’s character. After defense counsel discovered that Coleman had 172. Id. 173. Id. 174. See, e.g., Susan D. Rozelle, Daubert, Schmaubert: Criminal Defendants and the Short End of the Science Stick, 43 TULSA L. REV. 597, 606 (2007) (arguing that judges admit dubious forensic science far more often on behalf of prosecutors than defendants); Rodney J. Uphoff, On Misjudging and Its Implications for Criminal Defendants, Their Lawyers and the Criminal Justice System, 7 NEV. L.J. 521, 529 (2007) (noting based on personal observation that “a significant number of judges with prior prosecutorial experience bring a decidedly pro-prosecution attitude to the bench, and that attitude invariably influences their decisionmaking”); Keith Swisher, Pro-Prosecution Judges: “Tough on Crime,” Soft on Strategy, Ripe for Disqualification, 52 ARIZ. L. REV. 317 (2010) (arguing that elected judges are biased against defendants and “tough on crime” prosecutors should recuse themselves in criminal cases under ethics rules). 175. Swisher, supra note 174, at 328–29 (quoting numerous expressions of overt bias in judges electoral campaigns, including one pledge from a Texas Court of Appeals judge who asserted “I’m a prosecution-oriented person” who “see[s] legal issues from the perspective of the state instead of the perspective of the defense” (internal quotation marks omitted)). 1172 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 been charged with theft in a neighboring county during the same time period in which the undercover operation was conducted, the judges still barred the defense from putting any of that evidence before a jury, thereby effectively precluding defendants from presenting their theory of the case.176 In the Rampart trial of Javier Ovando, the presiding judge made a similar ruling against the defense, precluding Ovando’s attorney from attempting to impeach Officers Perez and Durden with discrepancies between contemporaneous statements given by them to investigators and a new account of events proffered at trial.177 Decisions by trial judges to preclude defendants from introducing evidence calling into question the honesty and integrity of the police or challenging substantial inconsistencies in the prosecution case were a major part of the Rampart and Tulia stories, and a clear contributing cause to many of the wrongful convictions that occurred. D. Innocence as a Minor Factor in Plea Bargaining Although the empirical evidence from the mass exoneration cases leaves no doubt that innocent defendants plead guilty, the question remains: does innocence have a measurable impact on whether a defendant will hold out for trial? Anecdotally speaking, we know that some innocent defendants turn down favorable plea bargains because of innocence. One Rampart exoneree to do so was Alex Umana. Umana was returning from a barbecue with his daughter and her mother when police stopped him and placed him among a group of four to six people who had been detained by police in the lobby of an apartment building. Although Umana was not in possession of any drugs at the time, he was nevertheless charged with possession of cocaine. Prosecutors offered a plea bargain for a probationary sentence, but Umana rejected the offer “because he was innocent and wanted to fight the charges.”178 Umana was convicted at trial and sentenced to five years in state prison.179 Several Tulia defendants also refused to plead guilty to drug charges because they were innocent. Take the case of Freddie Brookins, Jr., for example, who was accused of selling an eight-ball of powder cocaine to Coleman. Before trial, the prosecutor 176. See BLAKESLEE, supra note 30, at 100. 177. See Tamar Toister, Rampart Hasn’t Changed How Criminal Courts Do Business, L.A. TIMES, Sept. 21, 2000, available at 2000 WLNR 8421802. 178. Answer to Pet. for Writ of Habeas Corpus at 9, Decl. of Barbara Moulton, In re Alex Umana, No. BA144035 (Cal. Super. Ct. June 22, 2000). 179. Pet. for Writ of Habeas Corpus at 2, In re Alex Umana, No. BA144035 (Cal. Super. Ct. Apr. 6, 2000). 2013] POLICE MISCONDUCT 1173 offered Brookins a plea of five years.180 The maximum sentence for the offense was twenty years. Brookins discussed the offer with his father, and the following colloquy reportedly occurred: “Did you do it,” his father asked him. “No, I didn’t,” Brookins replied. “Well then,” his father responded, “don’t take the deal.”181 Against the advice of counsel, Brookins declined the plea offer. He went to trial, was convicted, and was sentenced to the maximum term of twenty years.182 Although we know that some defendants decline plea offers because of innocence, it is also possible that other innocent defendants plead guilty at equal or higher rates to avoid draconian trial penalties. The question therefore remains: does innocence materially alter guilty plea rates? The Rampart data sheds some additional light. For the purposes of this analysis, I identified three groups of Rampart cases resulting in exonerations. The first group, discussed above, consisted of those who were actually innocent of the crimes of conviction. There were thirty-eight such defendants in the data set. Of the remaining forty-nine, twenty-seven were identified as clearly “not actually innocent.” This group consisted of defendants who in fact were in possession of contraband or who admitted that they were engaged in criminal conduct at the time of arrest, but whose convictions were reversed based on procedural violations. The remaining group of twenty-two consisted of defendants whose guilt or innocence remains unclear given the record evidence. I identify this group as the “may be innocent” group. Although the numbers are small, they are large enough to permit some tentative comparisons. With respect to plea rates, the data shows that innocence does appear to make some difference. Twenty-five actually innocent Rampart exonerees pleaded guilty, while seven were convicted at trial.183 Actually innocent exonerees thus pleaded guilty at a rate of 77%. In comparison, twenty-two of those who were not actually innocent pled guilty while three were convicted at trial. In other words, 88% of those who were not innocent pleaded guilty. Finally, of the remaining group of “may be innocents,” seventeen pled guilty while two were convicted at trial, providing an 89% guilty plea rate.184 180. BLAKESLEE, supra note 30, at 138, 148. 181. Id. at 148 (internal quotation marks omitted). 182. Id. at 157. 183. The other seven had their probation revoked, or were minors who were adjudicated delinquent. 184. Of the rest, two admitted probation violations and 1 had his probation revoked. 1174 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 It thus appears from the data that actual innocence does induce some defendants to refuse a guilty plea and hold out for trial, but that the incentive has only a marginal effect, leading the innocent to contest their cases at trial at an approximately 10% greater rate than those who are actually guilty. Nonetheless, the data underscore that the vast majority of the actually innocent resolve false charges against them by pleading guilty. Very few held out for trial, and, as the numbers above documenting the size of the trial penalty demonstrate, those who did and lost paid a heavy price for that decision.185 VI. THE WRONGFULLY CONVICTED VS. THE ACTUALLY INNOCENT: DOES THE DISTINCTION MATTER? In the national dialogue about wrongful convictions, definitions of terms like “innocence,” “exonerated,” and “wrongfully convicted” have been contested. Has a person, convicted on the basis of unconstitutionallyobtained evidence, been “wrongfully convicted”? The answer, technically, is yes, but commentators typically use terms like “legal innocence” to describe defendants whose convictions resulted from significant procedural error but who are not factually innocent, or at least cannot establish their factual innocence.186 Legally innocent defendants were “wrongfully convicted,” but typically are treated as occupying a lesser status in the wrongful conviction debates than those who are “factually” or “actually” innocent. The term “actually innocent” has tended to be reserved for those who succeed in establishing not only that their conviction was legally flawed, but that they did not engage in any significant criminal wrongdoing. Accordingly, both Gross and Garrett limited their datasets to those defendants who were both formally exonerated by official act declaring the defendant not guilty of the crime of conviction and who were “actually innocent.” By “actually innocent,” Gross and Garrett mean that the exoneration was based on evidence that the defendants “had no role in the crimes for which they were originally convicted.”187 Defendants who were 185. Of course, whether trial was a good or bad decision for the average innocent defendant falsely charged by corrupt Rampart officers is impossible to determine without information regarding acquittals and dismissals of such defendants, which is unavailable. The data does show that those who gamble on trial and lose fare far worse than those who plead guilty. 186. See Margaret Raymond, The Problem with Innocence, 49 CLEV. ST. L. REV. 449, 456 (2001) (distinguishing between legal and factual innocence); Emily Hughes, Innocence Unmodified, 89 N.C. L. REV. 1083 (2011) (critiquing the distinction). 187. Gross et al., supra note 2, at 524. 2013] POLICE MISCONDUCT 1175 not guilty of the convicted offense, but who were guilty of committing some lesser crime based on the same conduct, are not considered actually innocent and are typically excluded from any count of exonerations.188 All of the individuals who were included in the Gross and Garrett studies were thus persons whose convictions were formally vacated, either through pardon or court order, and who were able to produce strong evidence not only that the convictions in their cases were unreliable, but that they were affirmatively innocent of wrongdoing. Other commentators have also urged the importance of distinguishing between actually innocent and procedurally innocent defendants, because convictions of actually innocent people represent far more serious breakdowns in the truth-seeking function of the criminal process.189 Still, police misconduct remains troubling even where the victim of that misconduct is engaged in unlawful behavior. Such misconduct undermines the effectiveness of constitutional rules established to protect the bodily integrity, privacy, and autonomy of citizens from incursion by the state. When police evade these rules by lying about their conduct, they undermine those mechanisms and weaken the protections safeguarding the innocent and the guilty alike. A. Wrongful Convictions Resulting from Unconstitutional Police Conduct The primary aim of this article has been to use the Rampart and Tulia exonerations as a means to understand how police misconduct causes wrongful convictions. Accordingly, until now the article has focused on the Rampart and Tulia cases that meet the actual innocence criteria used by other researchers in studying known wrongful convictions. As discussed above, thirty-eight Rampart exonerees and thirty-seven Tulia 188. See id. at 524 n.4. Garrett’s study employed similar criteria, generally adopting the same screening mechanism—affirmative proof of innocence—used by the Innocence Project to identify potential clients. See GARRETT, CONVICTING THE INNOCENT, supra note 59, at 285–86 (explaining that list of 250 DNA exonerations does not include “cases in which there has been no exoneration despite DNA evidence of innocence” and only includes cases in which there is no doubt that the “convicts are actually innocent”). 189. See Jon B. Gould & Richard A. Leo, One Hundred Years Later: Wrongful Convictions After a Century of Research, 100 J. CRIM. L. & CRIMINOLOGY 825, 833 (2010). At least one commentator has taken issue with the narrowness of the definition. In a forthcoming paper, Keith Findley argues that the criteria used to define the “actually innocent” is too narrow, at least where proof of innocence rather than absence of proof of guilt is demanded. Findley thus contends that all persons whose convictions are formally vacated based on evidence of innocence should be considered innocent. See Keith A. Findley, Defining Innocence, 74 ALB. L. REV. 1157 (2011). 1176 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 exonerees meet these criteria.190 A list of all of the innocent Rampart and Tulia exonerees appears in the Appendix to this article. Forty-nine persons—all Rampart defendants—were excluded from the dataset notwithstanding the fact that their convictions were reversed or vacated by a court with the affirmative consent of the Los Angeles District Attorney on the basis of police misconduct. These defendants did not meet the actual innocence criteria because there either remained affirmative evidence of criminal wrongdoing or insufficient evidence of innocence in the records available for this study. Under the Gross criteria, these individuals were not “exonerated” in the relevant sense. There is no doubt, however, that they were wrongfully convicted. Their cases illustrate various ways in which police officers circumvent constitutional protections and then lie about their conduct in order to convict criminal defendants. I refer to this type of police misconduct as “procedural perjury.” B. Types of Police “Procedural Perjury” Procedural perjury occurs when police lie about the circumstances of an encounter in order to ensure that evidence obtained during the encounter is not excluded or excludable. Procedural perjury is a common enough problem that a word—“testilying”—has been coined to describe the phenomenon.191 In one survey, insiders in the criminal justice system estimated that police perjury occurs in 20% to 50% of all Fourth Amendment suppression hearings.192 Seventy-six percent of police officers also believed that police misrepresented the facts relevant to probable cause determinations.193 In general, procedural perjury arises in three main guises: lies about consent, lies about probable cause, and lies about compliance with other constitutional rules of criminal procedure, most commonly, the rules governing custodial interrogation set forth in Miranda v. Arizona.194 As I use the terms, procedural perjury differs from 190. See infra Part IV.A. One Tulia defendant, Jonathan Loftin, was a minor at the time of his wrongful conviction and did not receive a pardon because he had already served out his camp sentence. Although his case is indistinguishable from the other Tulia cases in every other respect, I have not included him in the dataset. 191. See generally I. Bennett Capers, Crime, Legitimacy, and Testilying, 83 IND. L.J. 835 (2008). 192. See Christopher Slobogin, Testilying: Police Perjury and What to Do About It, 67 U. COLO. L. REV. 1037, 1041 (1996); see also Myron W. Orfield, Jr., Deterrence, Perjury, and the Heater Factor: An Exclusionary Rule in the Chicago Criminal Courts, 63 U. COLO. L. REV. 75, 107 (1992). 193. See Myron W. Orfield, Jr., The Exclusionary Rule and Deterrence: An Empirical Study of Chicago Narcotics Officers, 54 U. CHI. L. REV. 1016, 1050 (1987). 194. 384 U.S. 436 (1966). 2013] POLICE MISCONDUCT 1177 substantive perjury in that when police commit procedural perjury, they lie to circumvent procedural rules that otherwise would prevent them from prosecuting apparently guilty suspects. Procedural perjury is a form of whitewashing that is intended to facilitate what most police officers likely perceive to be the most essential aspect of their jobs: to punish those who are believed to be committing or to have committed crimes. Substantive perjury, in contrast, occurs when police lie to incriminate innocent persons. From the perspective of the criminal justice system’s guilt/innocence sorting mechanism, substantive perjury is a far more destructive practice than procedural perjury, although both forms of perjury undermine the integrity of the criminal justice system and diminish the credibility and the legitimacy of the police. A New York state commission headed by Judge Milton Mollen issued a report in 1994 documenting the “commonplace” types of procedural perjury routinely committed by New York police officers in their day-today duties, which included lying about observing unlawful conduct or incriminating facts to justify a search and seizure, lying about where contraband was found to cover-up plainly unconstitutional conduct, and lying about compliance with various rules of constitutional criminal procedure.195 The commission found that perjury was a particular problem in drugs and weapons cases, a finding that is consistent with the pattern of police misconduct evident in the Rampart scandal.196 Indeed, all of the types of “testilying” identified by the Mollen Commission are on vivid display in the Rampart cases. 195. For example, when officers unlawfully stop and search a vehicle because they believe it contains drugs or guns, officers will falsely claim in police reports and under oath that the car ran a red light (or committed some other traffic violation) and that they subsequently saw contraband in the car in plain view. To conceal an unlawful search of an individual who officers believe is carrying drugs or a gun, they will falsely assert that they saw a bulge in a person’s pocket or saw drugs and money changing hands. To justify unlawfully entering an apartment where officers believe narcotics or cash can be found, they pretend to have information from an unidentified civilian informant or claim they saw the drugs in plain view after responding to the premises on a radio run. To arrest people they suspect are guilty of dealing drugs, they falsely assert that the defendants had drugs in their possession when, in fact, the drugs were found elsewhere where the officers had no lawful right to be. See Capers, supra note 191, at 869 (quoting COMM’N TO INVESTIGATE ALLEGATIONS OF POLICE CORRUPTION AND THE ANTI-CORRUPTION PROCEDURES OF THE POLICE DEP’T, CITY OF NEW YORK, COMM’N REPORT 36 (1994) (Milton Mollen, Chair)). 196. See Capers, supra note 191. 1178 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 1. Lies about Consent One common form of procedural perjury on display in the Rampart cases is a false claim that a suspect “consented” to a search. Consent is simple to manufacture. Police need only claim that a suspect orally consented to a search to establish the existence of consent, although police in some Rampart cases went further and either forged a suspect’s signature on a written consent form or tricked or coerced a suspect into signing such a form. In several cases, police falsely claimed to have obtained consent to justify a forcible warrantless entry into a home. For example, police stated that they had received information from unnamed sources that a woman named Laura Villatora was storing cocaine and marijuana in her apartment. According to the arrest report, police sought and Villatora consented to a search of her apartment that turned up approximately seven pounds of marijuana.197 Rampart investigators later concluded, however, that Villatora, who was home with her daughter when the police arrived, did not consent to a search. When it appeared that police efforts to force open the door would cause serious damage to it, Villatora’s teenage daughter, Laura Reyes, opened the door. Officers then “grabbed Reyes by the hair and arm and dragged her to the living room” where they “demanded to know where the drugs and money were located.”198 They then began searching the apartment, eventually finding both drugs and money.199 One of the officers later testified to the false version of events at the preliminary hearing.200 As a result, Villatora pleaded guilty to a charge of unlawfully possessing marijuana with intent to sell and was sentenced to two years in prison.201 Similarly, in several cases police falsely claimed to have obtained consent to search a vehicle. For example, in the course of accosting Villatora and her daughter, police officers also learned that Villatora’s marijuana supplier was a person named Porfirio Acosta. Police then induced Villatora to arrange for Acosta to deliver drugs to her home. When Acosta arrived, police detained Acosta and searched his car without 197. See Pet. for Writ of Habeas Corpus at 5, In re Laura Villatora, No. BA153152 (Cal. Super. Ct. Jan. 31, 2000). 198. Id. at 10. 199. Pet. for Writ of Habeas Corpus at 13, Decl. of Michael Gannon, In re Porfirio Acosta, No. BA153198 (Cal. Super. Ct. June 7, 2000) (detailing interviews with Laura Villatora and her daughter, Laura Reyes). 200. In re Laura Villatora, supra note 200, Ex. B (Decl. of Brian Tyndall). 201. Id. 2013] POLICE MISCONDUCT 1179 obtaining consent. Officers also searched his home. The officers then falsely stated in the arrest report that Acosta had consented to the searches. They even manufactured a false consent to search form.202 Acosta was also charged with possession of marijuana for purposes of sale. He pleaded guilty and was sentenced to three years’ probation and 120 days in county jail.203 Another Rampart case involved Charles Harris, who was arrested after police allegedly recovered 203 grams of rock cocaine and a handgun in a supposedly consensual search of his vehicle.204 Investigators later concluded that, in fact, Harris never consented to the search.205 Likewise, police claimed to have obtained consent to search Juan Rojo’s car when, according to witnesses the D.A.’s office concluded to be credible, “Rojo was taken out of his car at gunpoint.”206 Officers testified falsely about these events at the preliminary hearing and suppression hearing, at which one officer testified that he merely “asked Rojo to step out of his vehicle so that he could speak with him.”207 The officers also falsely claimed that during this encounter Rojo consented to a search of his residence. Gricelda Orellana was in the residence when police turned up seeking to search it. According to Orellana, she “tried to lock the door” to prevent the police from entering.208 Notwithstanding those efforts, the officers entered and found cocaine in her bedroom. Orellana said that the cocaine belonged to “some guy,” but told police officers that “Rojo was innocent.”209 Nonetheless, Orellana and Rojo both eventually pleaded guilty to one count of possession for sale of cocaine and each served two years in prison.210 2. Lies about Probable Cause Perhaps the most common sort of lies told by the police are those used to establish probable cause for searches and seizures that otherwise are 202. In re Porfirio Acosta, supra note 202, at 6–7. It appears that the officers not only lied about obtaining consent from Acosta, but also either forged his signature on a consent to search form or coerced or tricked him into signing it. Id. 203. Id. at 2. 204. See Pet. for Writ of Habeas Corpus at 5, In re Charles Edward Harris, No. BA157278 (Cal. Super. Ct. Mar. 6, 2000). 205. Id. 206. Pet. for Writ of Habeas Corpus at 6, In re Juan Cerna Rojo and Gricelda Orellana, No. BA156027 (Cal. Super. Ct. May 22, 2000). 207. Id. 208. Id. (Tyndall Decl.). 209. Id. 210. Id. at 2. 1180 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 constitutionally unjustifiable.211 Numerous Rampart cases involved such misconduct. So-called “dropsy” cases are one well-known form of “testilying.”212 In a dropsy case, police claim that suspects in possession of drugs or guns “drop” the contraband before any Fourth Amendment seizure takes place.213 Since the contraband has been “abandoned” and is in “plain view,” the evidence is admissible. That police often resort to this type of perjury has been apparent for decades. Researchers observed a surge in dropsy cases shortly after Mapp v. Ohio was decided,214 in which the Supreme Court extended the exclusionary rule to the states.215 Dropsy cases were well-represented in the Rampart scandal. In one case, officers discovered a bag of cocaine after they had seized two suspects and then conducted a warrantless forty-five minute search of an apartment building. The two men were arrested and charged as a result. In their arrest report, the officers falsely claimed to have seen one of the men hold and then drop the bag of cocaine.216 In another case, a police officer saw a suspect suspiciously stuffing an object under his car seat. After searching the vehicle, the officer discovered cocaine. He then stated in the arrest report, and testified at the probation revocation hearing, that he had observed the suspect drop the cocaine on the ground.217 This was later determined to have been false.218 In a third case, police had information that the narcotics were located in one of the rooms in a hotel.219 After entering the room without a search warrant or consent, police found cocaine and the two defendants inside. Instead of these facts, the arrest report stated that police encountered the defendants in the hallway and saw them drop canisters of rock cocaine to the ground.220 In a fourth case, police seized a suspect and then, apparently after searching him and 211. See Jennifer Hunt & Peter K. Manning, The Social Context of Police Lying, 14 SYMBOLIC INTERACTION 51, 56 (1991). 212. See Capers, supra note 191, at 868. 213. Id. 214. Id. 215. See, e.g., Sarah Barlow, Patterns of Arrests for Misdemeanor Narcotics Possession: Manhattan Police Practices 1960–62, 4 CRIM. L. BULL. 549, 549–50 (1968). 216. See Pet. for Writ of Habeas Corpus at 10, Decl. of Olivia Rosales, In re Aristide Vanegas and Rodolfo Arevalo, No. BA146324 (Cal. Super. Ct. Mar. 28, 2000). One of the men admitted that they were delivering drugs to an apartment in the building at the time. Id. at 12 (Decl. of Brian Tyndall). 217. Resp. to Pet. for Writ of Habeas Corpus at 3, In re Oscar Lafarga, No. GA024373 (Cal. Super. Ct. Nov. 19, 1999). 218. Id. 219. Pet. for Writ of Habeas Corpus, Decl. of Natasha S. Cooper, In re Carlos Martinez Pena and Manuel Espinoza Ferrera, No. BA145491 (Cal. Super. Ct. Apr. 28, 2000). 220. Id. 2013] POLICE MISCONDUCT 1181 finding no contraband, searched the area where the seizure occurred. The officers located a firearm. Instead of reporting these facts, the officers falsely claimed to have actually seen the suspect discarding the weapon.221 The suspect, Michael Williams, was charged with possession of a firearm by a felon, contested the charge at trial, and was convicted. Williams was sentenced to serve twenty-five years to life.222 Arguably more egregious than dropsy cases are cases where police falsely claim to have seen the suspect actually engaged in criminal conduct. For instance, Rampart officers detained and searched two individuals without probable cause.223 After finding cocaine, the officers stated in the arrest report that they had seen the suspect “engaging in the sale of narcotics prior to her arrest” in order “to establish the necessary probable cause for the detention and search.”224 Reports involving false claims of direct observation of criminal conduct represent a potentially more serious form of procedural perjury because the false statement not only insulates the recovery of the contraband from suppression but provides direct affirmative (albeit false) evidence of the suspect’s guilt, increasing the chances that an innocent person will be convicted. This risk is apparent in the case of Edward Villanueva, who was convicted of possession of a firearm by a felon.225 According to the arrest report, Officer Perez was manning an observation point when he personally observed Villanueva with the firearms. Those statements turned out to be false. Officer Perez later admitted that he stopped and searched Villanueva based only on a report from a surveillance helicopter team who claimed to have observed Villanueva with the guns.226 Officer Perez decided to report the facts differently out of an apparent concern that the true facts left some room to doubt whether probable cause existed for the search, or indeed, whether Villanueva ever actually possessed the guns. Some cases combine dropsy testimony and false claims of observed criminal activity. This occurred in the prosecution of Octavio Fernandez.227 Officers searched Fernandez and discovered drugs. To justify the search, police falsely 221. Pet. for Writ of Habeas Corpus, Ex. A, Decl. of Richard A. Rosenthal, In re Michael Williams, No. BA128788 (Cal. Super. Ct. Jan. 31, 2000). 222. Id. at 2. 223. Id. 224. Pet. for Writ of Habeas Corpus, Ex. A at 1–2, Decl. of Richard A. Rosenthal, In re Sonia Castro, No. BA128771 (Cal. Super. Ct. Jan. 31, 2000). 225. Pet. for Writ of Habeas Corpus at 2, In re Edward Yumol Villanueva, No. BA135887 (Cal. Super. Ct. Apr. 28, 2000). 226. Id. Decl. of Natasha S. Cooper at 8. 227. Pet. for Writ of Habea[s] Corpus at 5, In re Octavio Fernandez, No. BA136807 (Cal. Super. Ct. Apr. 10, 2000). 1182 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 claimed both that they witnessed Fernandez selling drugs and that he dropped them prior to being seized.228 Another common type of procedural perjury involves misstatements regarding the location in which contraband was found. In one case, officers searched a suspect’s car and found a pouch of heroin-filled balloons above the car’s rearview mirror.229 Worried that they lacked probable cause for the automobile search, the officers falsely reported that the balloons had been found in one of the suspect’s socks.230 In another case, officers discovered marijuana during a search of a suspect’s residence and reported instead that it was found either on his person or in his car.231 In a third case, officers falsely stated they recovered a gun near the front door of the defendant’s apartment and ammunition in his pocket. These facts formed the basis for charging the defendant as a felon in possession of a gun, for which he served four years in state prison. Credible evidence later revealed that police had found the gun inside the defendant’s apartment, under a bed, in a search of doubtful constitutional validity. The gun, moreover, likely belonged to someone else.232 3. Lies about Miranda Compliance While police engage in procedural perjury most frequently to avoid Fourth Amendment suppression concerns, police also commit perjury to evade other constitutional rules. This is especially true with respect to compliance with the Miranda rules. Again, the willingness of police to lie about their compliance with Miranda in order to ensure that incriminating admissions or confessions made by suspects under interrogation are admissible has been noted by other scholars. While conducting observational studies of police interrogations, Professor Richard Uviller 228. Id. 229. Pet. for Writ of Habeas Corpus at 4, In re Blanca Sahagun and Carlos Carranza, No. BA147177 (Cal. Super. Ct. Jan. 24, 2000). 230. Id. at Ex. A, Decl. of Laura Laesecke. One or both suspects may have been innocent. Carranza’s account of the incident confirmed what Perez testified to, which is that Carranza never possessed heroin. However, Carranza admitted that “he helped Sahagun arrange the sale of heroin on the day of the incident,” which would seem to make him an accomplice. Sahagun, however, denied the charges, alleging that “the entire arrest report was fabricated because she refused to be an informant for Perez.” Sahagun claimed “that an officer at Rampart station displayed a clear baggie containing brightly colored objects and told Carranza in her presence that he would ‘put this in [Carranza’s] shoes.’” Id. at Ex. B (Decl. of Brian Tyndall (alterations in original)). 231. See Answer to Pet. for Writ of Habeas Corpus, Decl. of Vesna Maras, In re Gene Serrano, No. BA150628 (Cal. Super. Ct., June 9, 2000). 232. See Pet. for Writ of Habeas Corpus, Decl. of Natasha S. Cooper, In re Salvador Luis Arias, No. BA090914 (Cal. Super. Ct. May 12, 2000). 2013] POLICE MISCONDUCT 1183 noticed that officers often “advance slightly the moment at which the Miranda warnings were recited to satisfy the courts’ insistence that they precede the very first question in a course of interrogation.”233 This type of shading of the truth occurred while police were under the known observation of an outsider. When police have no reason to believe they are being observed, even more egregious deceptions are sometimes attempted. For instance, many of the Rampart cases included false statements by police that they had complied with Miranda. In the Charles Harris case mentioned above, not only did police falsely claim that Harris consented to a car search, they also falsely claimed to have advised him of his Miranda rights.234 These false claims of Miranda compliance are often only the necessary precedents to further false claims about incriminating statements falsely attributed to the suspect. This too was true in the Harris case, where not only did police lie about Mirandizing Harris, they also attributed incriminating admissions to Harris that he never made,235 casting his actual guilt into doubt. The same pattern appears in the arrest and prosecution of Carlos Romero. In that case, police conducted a warrantless and nonconsensual search of a residence. During the search, police threatened to arrest Romero’s sister, who in response “identified her brother as a narcotics dealer and directed them to a stash of cocaine.”236 The officers then arrested Romero. They sanitized the search and seizure by falsely claiming to have been directed to the stash of drugs by Romero himself after they had advised him of his Miranda rights.237 C. The Blurred Line Between Procedural and Substantive Perjury, and Other Forms Of Police Corruption In some cases, it is impossible to determine whether perjury was committed merely to secure a shortcut to conviction of a guilty suspect or instead to convict an innocent man. Such was the case with Julian Lopez Hernandez. According to the arrest report, Hernandez was arrested after police found eleven balloons of heroin during a consensual search of his 233. H. RICHARD UVILLER, TEMPERED ZEAL: A COLUMBIA LAW PROFESSOR’S YEAR ON THE STREETS WITH THE NEW YORK CITY POLICE 116 (Contemporary Books 1988) (cited and discussed in Slobogin, supra note 192, at 1043 (speculating that “lying about events in the interrogation room may be routine”)). 234. See In re Charles Edward Harris, supra note 205, Decl. of Laura Laesecke at 1. 235. See id. 236. Pet. for Writ of Habeas Corpus, Ex. A at 1–2, Decl. of Richard A. Rosenthal, In re Carlos Romero, No. BA154270 (Cal. Super. Ct. Nov. 10, 1999). 237. Id. 1184 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 apartment. After his arrest, police claimed that they Mirandized Hernandez, questioned him, and obtained incriminating statements. Hernandez pleaded guilty and was permitted to take advantage of diversion.238 During the subsequent Rampart investigation, Officer Perez admitted under oath that Hernandez did not consent to the search and that the officers never read Hernandez his Miranda rights prior to questioning him.239 Perez thus admitted to serious procedural flaws in the search and arrest of Hernandez, but appears to have maintained a belief in Hernandez’s substantive guilt. Hernandez, however, asserted an entirely different story. “[H]e claimed that the officers escorted him to an unfamiliar apartment building and used a set of keys owned by a man known as ‘Gerardo’ to open the apartment door. Inside, they found numerous colored balloons on a windowsill in the living room.” Hernandez “denied living at the location or knowing anything about the narcotics,” but stated that he “pled guilty to avoid going to prison.”240 Either way, Hernandez was wrongfully convicted, but whether he was actually innocent is impossible to determine on the scant available record. Procedural perjury also goes hand-in-glove with other forms of police corruption. William Zepeda and Argelia Diaz were convicted of possession of cocaine for purposes of sale.241 Both served two year sentences after agreeing to plea bargains.242 In making the arrest, the officers lied about seeing Zepeda and Diaz selling drugs, and falsely claimed to have obtained consent to search their apartment. While there, the officers “stole a large sum of money from their apartment.”243 They included these false facts in the arrest report, and then repeated the lies at the preliminary hearing. All of the misconduct occurring in the prosecution of Charles Harris was accompanied, similarly, by the unreported appropriation of $6,000 from Harris’s residence and the theft of at least $500 by the officers.244 In the Romero case, Officer Durden 238. See Pet. for Writ of Habeas Corpus at 2, In re Julian Lopez Hernandez, No. BA160279 (Cal. Super. Ct. Jan. 31, 2000). 239. Id., Decl. of Laura Laesecke. 240. Id., Decl. of Brian Tyndall. 241. Pet. For Writ of Habeas Corpus at 2, In re William Zepeda and Argelia Diaz, No. BA156980 (Cal. Super. Ct., Mar. 17, 2000). 242. Id. 243. Id., Decl. of Olivia Rosales. 244. In re Charles Harris, supra note 193, at Decl. of Laura Laesecke. According to Rafael Perez, the $6000 along with three guns seized from Harris’s house were given to Harris’s sister in exchange for information about drug dealers. Id. 2013] POLICE MISCONDUCT 1185 reportedly stole several pieces of jewelry from a suspect’s residence, as well as $1,500 in cash.245 All of the false convictions in the Tulia cases, similarly, may have been a collateral consequence of the efforts of a corrupt undercover officer to steal drug “buy” money from the Drug Task Force.246 At least some people believe that Officer Coleman pocketed the buy money, lied about buying powder cocaine from the Tulia defendants, and then evidenced his purported buys by turning in small amounts of white powder that he mixed himself, each containing only enough cocaine to trigger a positive reading on a lab test.247 In these cases, police lies about compliance with constitutional rules were concomitant with, or in service to, other acts of corruption. VII. CONCLUSION Police misconduct, when it occurs, is a major source of wrongful convictions. The profile of those most at risk of such wrongful convictions likely differs in some respects from that of other wrongfully convicted persons. The offenses are generally less serious, and the sentences less severe, than those involved in the DNA exoneration cases. These cases involve drugs and guns, assaults on police officers, charges of disturbing the peace, resisting arrest, or other allegedly violent or aggressive conduct directed at the police. Hundreds of thousands, perhaps millions, of people have been convicted of such crimes. There is simply no way to know how many persons convicted of such offenses were actually innocent, but both Rampart and Tulia provide stark evidence that police misconduct can, and does, result in wrongful convictions. Comparison of the mass exoneration data with prior exoneration studies suggests that two important adjustments to the empirical picture of wrongful convictions may be in order. Although earlier studies of wrongful convictions found only a small number of cases involving guilty pleas, in the mass exoneration cases, guilty pleas provided the main procedural vehicle to criminal conviction. In more than 80% of the combined Rampart and Tulia cases, innocent defendants pleaded guilty. While innocence did seem to provide a marginal incentive to some defendants to reject guilty pleas, actually innocent Rampart exonerees held 245. See Romero Pet., supra note 236, Ex. A at 2 (Decl. of Richard A. Rosenthal). 246. See BLAKESLEE, supra note 30, at 88. 247. Id. at 88–89. 1186 WASHINGTON UNIVERSITY LAW REVIEW [VOL. 90:1133 out for trials only slightly more frequently than their guilty counterparts.248 The Rampart and Tulia exoneration data thus provides strong reason to suspect that guilty pleas are not insulated from the risk of wrongful convictions. Consideration of this data should also raise the profile of perjury among the causes of wrongful conviction. Although eyewitness misidentification has received a substantial amount of attention as one of the main identified contributing factors in wrongful convictions, the mass exoneration cases make clear that the “causes” of wrongful convictions vary significantly by crime. These exonerations show that police misconduct is a potentially significant cause of wrongful convictions in its own right. Procedural reforms that reduce the incidence of police misconduct, therefore, should be high on the list of priorities among those working to reduce wrongful convictions. 248. See infra Part V.D. 2013] POLICE MISCONDUCT APPENDIX “Actually Innocent” Rampart Exonerees 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. Alfaro, George Kenneth Bailey, Samuel Joseph Barrios, Diego Booker, Esaw Candido, Roberto Carrillo, Delbert Chavez, Emmanuel Davalos, Octavio Gonzalez Escobar, Edgar Estrada, Leonel Ramos (aka Gregorio Ramos Lopez) Flores, Luis Manuel Gomez, Alfredo Guardado, Manuel Guevara, Carlos Harris, Clinton Hernandez, Miguel Lara, Jose Armando Lobos, Allan Manrique Madrid, Jose Hugo Matlong, Rene Barela (aka Rene Mationg) Montes, Roy Munoz, Raul Alfredo Natividad, Cesar aka Danny Banuelos Newman, Russell Oliver, Ivan Ordonez, Felipe Enriquez Ovando, Javier Francisco Perez, Jose Peters, Gerald Rivas, Walter Rodriguez, Raul Rojas, Ruben Tapia, Daniel Thomas, James Torrecillas, Juan Umana, Alex Wesley, Mohammed Wayman Zambrano, Rafael 1187 1188 WASHINGTON UNIVERSITY LAW REVIEW Tulia Exonerees 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. Allen, Dennis Mitchell Barrow, James Ray Barrow, Landis Barrow, Leroy Barrow, Mandis Charles Benard, Troy Brookins, Freddie Wesley Cooper, Marlyn Joyce Ervin, Aremnu Jerrod Fowler, Michael Fry, Jason Paul Fry, Vickie Hall, Willie B. Henderson, Cleveland Joe Henry. Mandrell L. Jackson, Christopher Eugene Kelly, Denise Kelly Sr., Eliga Klein, Calvin Kent Love, William Cash Marshall, Joseph Corey Mata, Laura Ann McCray, Vincent Dwight Moore, Joe Welton Olivarez, Daniel G. Powell, Kenneth Ray Robinson, Benny Lee Shelton, Finaye Smith, Donald Wayne Smith, Yolanda Yvonne Strickland, Romona Lynn Towery, Timothy Wayne White, Kareem Abdul Jabbar White, Kizzie Rashawn [VOL. 90:1133 2013] POLICE MISCONDUCT 35. Williams, Alberta Stell 36. Williams, Jason Jerome 37. Williams, Michelle 1189