Police misconduct as a cause of wrongful convictions.

Author:Covey, Russell
Position:Continuation of IV. Actually Innocent Rampart and Tulia Exonerees E. Method of Conviction through VII Conclusion, with appendix and footnotes, p. 1162-1189
  1. Evidence that the Innocent Do Plead Guilty (138)

    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 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.

    As Table 2 suggests, in comparison with other exonerees, the Rampart and Tulia exonerees pied 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 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 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 rape-murder 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 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.


      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.

      1. 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 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 convicted and 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...

To continue reading