Slipping away from justice: The effect of attorney skill on trial outcomes

Vanderbilt Law Review

Volumn 63, Issue 1, 2010, Pages 267-306

  Shinall, Jennifer Bennett ^a  

^a Philosophy in Law and Economics ^*  (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 78650631087     PISSN: 00422533     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Review

References (105)

1. David Margolick, Simpson Tells Why He Declined To Testify as Two Sides Rest Case, N.Y. TIMES, Sept. 23,1995, at A1.
2. Frank Rich, Journal: The L.A Shock Treatment, N.Y. TIMES, Oct. 4, 1995, at A21.
3. Seth Mydans, Not Guilty: The Lawyers: In the Joy of Victory, Defense Team is in Discord, N.Y. TIMES, Oct. 4, 1995, at A11.
4. David Margolick, Simpson's Lawyer Tells Jury That Evidence Doesn't Fit,'N.Y. TIMES, Sept. 28, 1995, at A1.
5. Scott Turow, Simpson Prosecutors Pay for Their Blunders, N.Y. TIMES, Oct. 4, 1995, at A21.
6. See Paul Dale Roberts et al, Letters to the Editor, Prosecutors Fumble Another Celebrity Case, USA TODAY, June 15, 2005, at A10 (criticizing the verdict and discussing the prosecutors' mistakes in the Michael Jackson case).
7. See Greg Risling, Blake Not Guilty of Wife's Murder: Prosecution 'Couldn't Put the Gun in his Hand,' Jury Foreman Says, CHI. SUN-TlMES, Mar. 17, 2005, at 3 (detailing the shortcomings of the prosecution in the Robert Blake case).
8. See Gary Cartwright, The Verdict: The Safe Money Had Robert Durst Getting Convicted of Murder in Seconds Flat. How on Earth Did He Get Off? Two Words: Dick DeGuerin, TEXAS MONTHLY, Feb. 2004, at 54, 56 (discussing the role of Dick DeGuerin in Robert Durst's acquittal).
9. Lawrence Buser, Dream Team Took Jury Into Nightmare, Emerged With Murder No Longer an Issue, COMMERCIAL APPEAL, Apr. 21, 2007, at A1.
10. Id.
11. Wife Guilty of Manslaughter in Minister's Killing, N.Y. TIMES, Apr. 20, 2007, at A17 (emphasis added).
12. See, e.g., Strickland v. Washington, 466 U.S. 668, 710 (1984) (Marshall, J., dissenting) (noting that the "difficulties of estimating prejudice after the fact are exacerbated by the possibility that evidence of injury to the defendant may be missing from the record precisely because of the incompetence of defense counsel");
13. McMann v. Richardson, 397 U.S. 759, 771 (1970) (holding that "if the right to counsel guaranteed by the Constitution is to serve its purpose, defendants cannot be left to the mercies of incompetent counsel, and that judges should strive to maintain proper standards of performance by attorneys who are representing defendants in criminal cases in their courts");
14. Powell v. Alabama, 287 U.S. 45, 68-69 (1932) (discussing the consequences of inadequate access to counsel).
15. For example, a defense attorney may practice in a location with highly empathetic juries, leading to systemically better results for this attorney than a similar attorney practicing in a less defense-friendly jurisdiction. Similarly-whether fair or not-the race or gender of a defense attorney could plausibly affect trial outcomes. Juries in areas of the country highly sensitized to race may take greater note of the race of all attorneys involved.
16. See David S. Abrams & Albert H. Yoon, The Luck of the Draw: Using Random Case Assignment to Investigate Attorney Ability, 74 U. CHI. L. REV. 1145, 1146-48 (discussing the problem of attorney matching in evaluating the effect of attorney skill).
17. See id. at 1146 (explaining that "the nonrandom pairing of attorney and client in most cases makes it difficult, if not impossible, to distinguish between attorney ability and case selection").
18. See id. at 1147 (noting that "case outcomes may reflect the matching process between clients and attorneys as much as the ability of the attorneys who represent their clients").
19. Here, and throughout the paper, the "attorney skill effect" simply refers to the effect of attorney skill on trial outcomes. Attorney skill is defined in Part I as the qualities that an attorney brings to the courtroom independent of his case's strength, such as rhetorical abilities, tactical strategies, and knowledge of the law.
20. E.g., Abrams & Yoon, supra note 14, at 1147-49 (investigating whether systematic differences in defense attorney ability exist and their effects on trial outcome);
21. Floyd Feeney & Patrick G. Jackson, Public Defenders, Assigned Counsel, Retained Counsel: Does the Type of Criminal Defense Counsel Matter?, 22 RUTGERS L.J. 361, 407-10 (1991) (exploring the differences between different types of defense attorneys);
22. Daniel Linz et al, Attorney Communication and Impression Making in the Courtroom: Views From Off the Bench, 10 L. & HUM. BEHAV. 281, 281-301 (1986) (designing an experiment to test the effect of prosecution and defense attorney skill on trial outcomes).
23. U.S. CONST, amend. VI.
24. Powell v. Alabama, 287 U.S. 45, 71-73 (1932).
25. Gideon v. Wainwright, 372 U.S. 335, 343-45 (1963).
26. 466 U.S. 648, 659 (1984).
27. See, e.g., McMann v. Richardson, 397 U.S. 759, 771 (1970) (arguing that the Sixth Amendment right to counsel includes a right to competent counsel);
28. Powell, 287 U.S. at 68-69 (discussing the consequences of inadequate access to counsel).
29. Powell, 287 U.S. at 68-69.
30. Id. at 69.
31. Id.
32. See REGINALD HERBER SMITH, JUSTICE AND THE POOR 123 (1919) (finding that retained counsel had higher acquittal rates than public defenders, but public defenders had better probation outcomes).
33. See, e.g., PETER W. GREENWOOD ET AL., PROSECUTION OF ADULT FELONY DEFENDANTS IN Los ANGELES COUNTY: A POLICY PERSPECTIVE 52-56 (1973), available at http://www.rand.org/pubs/reports/R1127/ (finding that assigned counsel had better acquittal rates, but public defenders had better sentencing outcomes);
34. DALLIN H. OAKS & WARREN LEHMAN, A CRIMINAL JUSTICE SYSTEM AND THE INDIGENT: A STUDY OF CHICAGO AND COOK COUNTY 162-63 (1968) (finding that private counsel and publicly appointed counsel had similar trial outcomes);
35. LEE SILVERSTEIN, DEFENSE OF THE POOR IN CRIMINAL CASES IN AMERICAN STATE COURTS 20-29 (1965) (finding retained counsel had better trial outcomes than appointed counsel).
36. Feeney & Jackson, supra note 18, at 407.
37. Id. at 409 (
38. quoting Dallin H. Oaks & Warren Lehman, Lawyers for the Poor, in THE SCALES OF JUSTICE 91-93 (Abraham S. Blumberg ed., 1970)).
39. Id.
40. Linz et al, supra note 18, at 281-302.
41. Id. at 284.
42. Id. at 286.
43. Id. at 286-87.
44. Id.
45. Id. at 296.
46. Id. at 297.
47. Id. at 299.
48. Id. at 300-01 (suggesting that future studies should gather ratings on attorney performance from judges and juries).
49. Richard T. Boylan, Salaries, Turnover, and Performance in the Federal Criminal Justice System, 47 J.L. & ECON. 75, 75-92 (2004).
50. Richard T. Boylan, What do Prosecutors Maximize? Evidence from the Careers of U.S. Attorneys, 7 AM. LAW & ECON. REV. 379, 379-402 (2005).
51. Id. at 396.
52. 18 U.S.C §3006A(a)(3) (2006).
53. Radha Iyengar, An Analysis of Attorney Performance in the Federal Indigent Defense System 2-5 (Nat'l Bureau of Econ. Research, Working Paper No. 13187, 2007), available at http://www.nber.org/papers/w13187.
54. Id. at 28-30.
55. Abrams & Yoon, supra note 14, at 1176-77.
56. Id. at 1161-64.
57. Id. at 1149.
58. Id.
59. See id. at 1167-70 (discussing the effect that these attorney characteristics had on the duration and probability of incarceration).
60. Id. at 1173-75.
61. Id.
62. Id at 1175.
63. For a concise definition of attorney skill as the term is used in this paper, see supra Part I.
64. Sample-selection bias arises "when the availability of the data is influenced by a selection process that is related to the value of the dependent variable. This selection process can introduce correlation between the error term and the regressor, which leads to bias . . . ." JAMES H. STOCK & MARK W. WATSON, INTRODUCTION TO ECONOMETRICS 250 (2003). The classic example of sample selection bias occurs in a regression of wages on determinants like education, years of experience, and geographic location. Only people who have a job have wages, and the same factors that determine how much money a person makes also determine whether that person gets a job.
65. Id. at 251.
66. External validity is a term commonly used in econometrics that indicates whether the study is robust enough to extrapolate its results to other situations.
67. See id. at 243-45 (discussing when an empirical study is subject to threats of external validity).
68. PAULA L. HANNAFORD-AGOR ET AL., INTER-UNIVERSITY CONSORTIUM FOR POLITICAL AND SOCIAL RESEARCH, EVALUATION OF HUNG JURIES IN BRONX COUNTY, NEW YORK, LOS ANGELES COUNTY, CALIFORNIA, MARICOPA COUNTY, ARIZONA, AND WASHINGTON, DC, 2000-2001, http://www.icpsr.umich.edu/cocoon/ICPSR/STUDY/03689.xml [hereinafter EVALUATION OF HUNG JURIES].
69. Abrams & Yoon, supra note 14, at 1149.
70. NAT'L CRIMINAL JUSTICE REFERENCE SERV., EMPIRICAL STUDY OF FREQUENCY OF OCCURRENCE, CAUSES, EFFECTS, AND AMOUNT OF TIME CONSUMED BY HUNG JURIES: FINAL REPORT (1975), available at http://www.ncjrs.gov/App/Publications/abstract.aspx? ID=19026. This study was originally published in 1975, and so is somewhat old. However, according to the researchers who compiled EVALUATION OF HUNG JURIES, supra note 58, most current statistics on hung juries come from studies in the 1960s and 1970s, further motivating their project.
71. Recently, a prosecutor's decision whether to pursue a retrial has received a lot of publicity after the second hung jury in a Miami terrorism case. Six men are accused of planning to bomb the Sears Tower in Chicago and the FBI offices in Miami, yet two different juries have been unable to reach a verdict. Curt Anderson, Jury Hits Stalemate in Terrorism Retrial, USA TODAY, Apr. 11, 2008, available at http://www.usatoday.com/news/nation/2008-04- 113235714827-x.htm.
72. Individually, measures like the number of prosecution witnesses could be challenged as imperfect controls for the strength of the prosecution's case; for example, the prosecution could plausibly base their entire case on the testimony of one eyewitness. But together, these variables demonstrate the relative matchup of resources and evidence on both sides.
73. The fixed-effect variables allow each type of crime to have its own intercept; this is useful if, for example, defendants in murder cases are systematically more likely to be convicted than defendants in theft cases. See WILLIAM H. GREENE, ECONOMETRIC ANALYSIS 193-200 (6th ed. 2008) (giving a technically rigorous discussion of fixed effects regression);
74. STOCK & WATSON, supra note 56, at 278-83 (giving a basic overview of fixed effects regression).
75. Because only eleven defendants in the sample were convicted but not sentenced to prison, these defendants are not included in the sentencing analysis.
76. Abrams & Yoon, supra note 14, at 1173-74.
77. The results of a linear probability analysis are easier to interpret than the results of a probit analysis because the coefficients can be multiplied by one hundred and interpreted as percentage effects.
78. STOCK & WATSON, supra note 56, at 299-302 (giving a basic overview of the linear probability model and how to interpret it).
79. But see GREENE, supra note 63, at 772-73 (discussing problems with the linear probability model despite its ease of interpretation).
80. See app. tbl.I.
81. See supra note 66 (explaining how to interpret the results of a linear probability analysis).
82. See, e.g., Cartwright, supra note 8, at 1 (describing the media's reaction to the verdict in the murder trial of millionaire Robert Durst);
83. Risling, supra note 7, at 3 (describing the media's reaction to the verdict in the murder trial of Robert Blake);
84. Roberts et al, supra note 6, at A10 (describing the media's reaction to the verdict in the child molestation trial of Michael Jackson).
85. Abrams & Yoon, supra note 14, at 1173-75.
86. 22A C.J.S. Criminal Law §941 (2008).
87. 1 WHARTON'S CRIMINAL EVIDENCE §2:9 (15th ed. 2008).
88. See, e.g., Daniel Givelber, Meaningless Acquittals, Meaningful Convictions: Do We Reliably Acquit the Innocent?, 49 RUTGERS L. REV. 1317, 1394 (1997) ("Advantages which may enhance the case of the guilty defendant such as the right to silence and to exclude relevant inculpatory evidence, work no benefit for the innocent. Instead, those advantages justify the prosecution's withholding from the accused and the factfinder evidence which might undermine the prosecution's case.");
89. Fred C. Zacharias & Bruce A. Green, The Duty to Avoid Wrongful Convictions: A Thought Experiment in the Regulation of Prosecutors, 89 B.U. L. REV. 1, 57 (2009) (arguing that "the problem of false convictions implicates aspects of prosecutors' professional conduct that might plausibly be regulated by the legal ethics codes"). These variations in our criminal justice system led legal scholar John Merryman to joke that "if he were innocent, he would prefer to be tried by a civil law court, but that if he were guilty, he would prefer to be tried by a common law court."
90. JOHN HENRY MERRYMAN, THE CIVIL LAW TRADITION: AN INTRODUCTION TO THE LEGAL SYSTEMS OF WESTERN EUROPE AND LATIN AMERICA 132 (2d ed. 1985).
91. Aaron M. Clemens & Hale R. Stancil, Unhandcuffing Justice: Proposals to Return Rationality to Criminal Sentencing, FLA. B. J., Feb. 2009, at 54.
92. Id.
93. See Boylan, supra note 41, at 76 (finding that lower salaries lead to higher turnover, and higher turnover leads to lower output).
94. See Jeff E. Biddle & Daniel S. Hamermesh, Beauty, Productivity, and Discrimination: Lawyers'Looks and Lucre, 16 J. LAB. ECON. 172, 193 (1998) (finding a raw earnings difference of 158 percent between attorneys in the private and public sectors that did not disappear after controlling for productivity).
95. Id. at 195.
96. These suggestions are strengthened by Boylan's result that lower salaries for U.S. Attorneys lead to higher turnover and lower job performance. Boylan, supra note 41, at 75.
97. These surveys could be similar to the ones used to create EVALUATION OF HUNG JURIES. See supra note 58 (describing the data collection process).
98. Other scholars have also argued for prosecutor training programs. E.g., Zacharias & Green, supra note 73, at 19 (arguing that a training program "encourages prosecutors to temper overzealousness and consider how their behavior can produce untoward results").
99. In case the court-appointed attorney variable is highly correlated with the defense attorney skill measure, all regression specifications were re-estimated without the courtappointed attorney dummy variable. The results did not change.
100. American Civil Liberties Union, Criminal Justice: Indigent Defense, http://www.aclu.org/crimjustice/indigent/index.html (last visited Oct. 30, 2009).
101. Iyengar, supra note 45, at 2.
102. Since this data set does not contain information on defendants who did not go to trial, we cannot test this possibility here. This possibility is worth mentioning here, however, because the results of the present study do not preclude possible inequitable treatment before trial.
103. Margolick, supra note 1, at A1.
104. Turow, supra note 5, at A21 (emphasis added).
105. Frank Rich, Judge Ito's All-Star Vaudeville, N.Y. TIMES, Oct. 2, 1994, at E17.