The luck of the draw: Using random case assignment to investigate attorney ability

University of Chicago Law Review

Volumn 74, Issue 4, 2007, Pages 1145-1177

  Abrams, David S ^a,b,c,d   Yoon, Albert H ^c,e,f,g  

^a UNIVERSITY OF CHICAGO LAW SCHOOL   (United States)

^b HARVARD UNIVERSITY   (United States)

^c STANFORD UNIVERSITY   (United States)

^d MASSACHUSETTS INSTITUTE OF TECHNOLOGY   (United States)

^e NORTHWESTERN UNIVERSITY SCHOOL OF LAW   (United States)

^f YALE UNIVERSITY   (United States)

^g STANFORD LAW SCHOOL   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 38049141633     PISSN: 00419494     EISSN: None     Source Type: Journal    
DOI: 10.2307/20141859     Document Type: Article

References (98)

1. For example, consider the case of Attorney A, trained at a top law school, and Attorney B, trained at a lower-ranked school. Attorney A may have better objective measures of performance (that is, sentence length, acquittal rate) than Attorney B, but may actually be less effective. This can happen if Attorney A is able to choose defendants with stronger chances of acquittal, while Attorney B chooses defendants with weaker chances of acquittal. Even controlling for observed case characteristics will not fully eliminate this problem in evaluating lawyer performance, as there may be unobserved case differences.
2. For example, there are rankings such as the AmLaw 100, published annually by The American Lawyer, which rank the top 100 law firms based on their gross revenue. See, for example, The American Lawyer, 2006 AmLaw 100: Two More Billion Dollar Firms (May 2006), online at http://www.law.com/jsp/ article.jsp?id=1145803851920 (visited Sept 23, 2007).
3. See Emily Rubin, Ineffective Assistance of Counsel and Guilty Pleas: Toward a Paradigm of Informed Consent, 80 Va L Rev 1699, 1700 (1994) (noting that approximately 90 percent of criminal cases are disposed of in guilty pleas).
4. For example, in federal courts during the twelve-month period ending March 31, 2006, the number of civil cases filed was 244,068, while the number of criminal cases filed was 68,670. See Administrative Office of the U.S. Courts, Federal Judicial Caseload Statistic tables C and D (Office of Judges Programs Statistics Division, Mar 31, 2006), online at http://www.uscourts.gov/ caseload2006/contents.html (visited Sept 23, 2007). The National Center for State Courts reports that in 2003, the total numbers of state civil and criminal cases filed in general and limited courts were 13,318,182 and 10,143,783, respectively. See Court Statistics Project, State Court Caseload Statistics 2005, 137-38 table 7 (National Center for State Courts 2006), online at http://www.ncsconline.org/D_Research/csp/2004_Files/SCCSTabl5-8.pdf (visited Sept 23, 2007).
5. See Marc Galanter and Mia Cahill, Most Cases Settle: Judicial Promotion and Regulation of Settlements, 46 Stan L Rev 1339, 1340 (1994) (noting that two-thirds of civil cases are resolved without "definitive judicial ruling").
6. See U.S. Census Bureau, North American Industry Classification System (NAICS) Professional, Scientific, and Technical Services (Except Notaries), Estimated Revenue for Taxable Employer Firms: 1998 Through 2005 table 6.1, online at http://www.census.gov/svsd/www/services/sas/sas_data/54/2005_NAICS54. pdf (visited Sept 23, 2007).
7. See Office of Management and Budget, Department of Justice, The Budget for Fiscal Year 2006, online at http://www.whitehouse.gov/omb/budget/ fy2006/pdf/budget/justice.pdf (visited Sept 23, 2007) (listing actual amounts spent in 2004);
8. U.S. Department of Justice, FY 2003 Performance and Accountability Report: Consolidated Statements of Net Cost, online at http://www.usdoj.gov/ ag/annualreports/ar2003/pdf/afsnetcost.pdf (visited Sept 23, 2007) (describing the costs incurred by protecting "the Rights and Interests of the American People by Legal Representation, Enforcement of Federal Laws and Defense of U.S. Interests").
9. See American Bar Association, Law School Tuition: Average and Median 1985-2005, online at http://www.abanet.org/legaled/statistics/charts/ stats%20-%205.pdf (visited Sept 23, 2007) (stating that the average law school tuition in 2005 was $30,520).
10. See, for example, Michael Sauder and Ryon Lancaster, Do Rankings Matter? The Effects of U.S. News & World Report Rankings on the Admissions Process of Law Schools, 40 L & Socy Rev 105, 115 (2006) (arguing that small differences in school rankings have a significant influence on applicants' perceptions of schools).
11. The median starting salaries for lawyers from the top ten most selective law schools was 26 percent higher than the median starting salary for lawyers from the next ten most selective schools. See NALP Foundation for Law Career Research and Education and American Bar Foundation, After the J.D.: First Results of a National Study of Legal Careers, 44 table 5.2 (2004), online at http://www.nalpfoundation.org/webmodules/articles/articlefiles/87- After_JD_2004_web.pdf (visited Sept 23, 2007).
12. See Richard H. Sander, The Racial Paradox of the Corporate Law Firm, 84 NC L Rev 1756, 1819 (2006) (arguing that affirmative action programs at both law schools and law firms, designed to benefit black law school graduates, actually serve to their detriment).
13. See also Richard H. Sander, A Systemic Analysis of Affirmative Action in American Law Schools, 57 Stan L Rev 367 (2004) (analyzing legal education and law school admissions policies and concluding that racial preferences result in higher attrition rates and lower bar passage rates). For a critical discussion of Sander's claims about affirmative action,
14. see David L. Chambers, et al, The Real Impact of Eliminating Affirmative Action in American Law Schools: An Empirical Critique of Richard Sander's Study, 57 Stan L Rev 1855, 1857 (2005) ("The conclusions in Systemic Analysis rest on a series of statistical errors, oversights, and implausible assumptions.").
15. See also Daniel E. Ho, Why Affirmative Action Does Not Cause Black Students to Fail the Bar, 114 Yale L J 1997, 1997 (2005) (explaining that Sander "misapplies basic principles of causal inference, which enjoy virtually universal acceptance in the scientific community");
16. David B. Wilkins, A Systematic Response to Systemic Disadvantage: A Response to Sander, 57 Stan L Rev 1915, 1919 (2005) (arguing that the evidence "does not come anywhere close to proving" that black lawyers would be better off going to lower-ranked law schools);
17. Michele L. Dauber, The Big Muddy, 57 Stan L Rev 1899, 1907-08 (2005) (arguing that Sander "had an inadequate empirical basis for making" his claims and that he failed to disclose the data he used to the editors of the Stanford Law Review or the academic community);
18. Kevin R. Johnson and Angela Onwauchi-Willig, Cry Me a River: The Limits of "A Systemic Analysis of Affirmative Action in American Law Schools," 7 Afr Am L & Policy Rep 1, 4 (2005) (noting the many factors Sander fails to take account of in Systemic Analysis);
19. Ian Ayres and Richard Brooks, Does Affirmative Action Reduce the Number of Black Lawyers?, 57 Stan L Rev 1807, 1809 (2004) (arguing that "even within [Sander's] framework, there is not persuasive evidence indicating that affirmative action is responsible for lowering the numbers of black attorneys").
20. See Marianne Bertrand and Antoinette Schoar, Managing with Style: The Effect of Managers on Corporate Policy, 118 Q J Econ 1169, 1204 (2003) (finding that "[t]he realization of all investment, financing, and other organization variables appear to systematically depend on the specific executives in charge").
21. See Jonah Rockoff, The Impact of Individual Teachers on Student Achievement: Evidence from Panel Data, 94 Am Econ Rev 247, 247-48 (2004) (showing that teacher quality and experience have demonstrable effects on students' test scores).
22. The findings about attorney characteristics are impacted by potential selection in the hiring process based on these characteristics. This is discussed further in Part VI.
23. The Sixth Amendment of the Constitution states: In all criminal prosecutions, the accused shall enjoy the right to a speedy and public trial, by an impartial jury of the State and district wherein the crime shall have been committed, which district shall have been previously ascertained by law, and to be informed of the nature and cause of the accusation; to be confronted with the witnesses against him; to have compulsory process for obtaining witnesses in his favor, and to have the Assistance of Counsel for his defense. (emphasis added). The Supreme Court has since held that the right to counsel extends to defendants in state court. See Gideon v Wainwright, 372 US 335, 342 (1963) (holding that the right to counsel is fundamental and was therefore incorporated into the Due Process Clause of the Fourteenth Amendment).
24. See Strickland v Washington, 466 US 668, 687 (1984) (holding that "the proper standard for attorney performance is that of reasonably effective assistance").
25. See, for example, Wheat v United States, 486 US 153, 165 (1988) (holding that "the Sixth Amendment provides protection for a criminal defendant's choice of counsel").
26. For discussions about the effectiveness of counsel, see William S. Geimer, A Decade of Strickland's Tin Horn: Doctrinal and Practical Undermining of the Right to Counsel, 4 Wm & Mary Bill Rts J 91, 95 (1995) (arguing that Supreme Court decisions have undermined indigent defendants' right to counsel, and proposing a minimal standard of effectiveness for capital defense counsel);
27. Barbara A. Babcock, Fair Play: Evidence Favorable to an Accused and Effective Assistance of Counsel, 34 Stan L Rev 1133, 1135-36 (1982) (using an analogy to fair play in sports to examine what the adversary system requires of its players).
28. See also Gabriel J. Chin and Richard W. Holmes, Effective Assistance of Counsel and the Consequences of Guilty Pleas, 87 Cornell L Rev 697, 703 (2002) (arguing that failing to share information regarding "collateral consequences" of a guilty plea with a client should constitute ineffective assistance).
29. For example, Georgia and Tennessee courts have held that state ethics rules do not define standards of civil liability. See Davis v Findley, 262 Ga 612, 422 SE2d 859, 861 (1992) (holding that a cause of action for malpractice could not rely solely on a violation of the Code of Professional Conduct); Lazy Seven Coal Sales, Inc v Stone & Hinds, PC, 813 SW2d 400, 404 (Tenn 1991) (holding that the Code of Professional Responsibility itself prohibits its use as a standard of civil liability). Other states, such as Michigan and California, have held that state ethics rules are probative for determining standards of care. See Lipton v Boesky, 110 Mich App 589, 313 NW2d 163, 167 (1981) (holding that "a violation of the Code is rebuttable evidence of malpractice"); Mirabito v Liccardo, 5 Cal Rptr 2d 571, 573 (Ct App 1992) ("The attorney's violation of those rules [of Professional Conduct] established his negligence.").
30. See generally, for example, Richard T. Boylan and Cheryl X. Long, Salaries, Plea Rates, and the Career Objectives of Federal Prosecutors, 48 J L & Econ 627 (2005) (analyzing salaries and experience levels of federal prosecutors and, among other things, connections between experience and case outcomes);
31. Talia Roitberg Harmon and William S. Lofquist, Too Late for Luck: A Comparison of Post-Furman Exonerations and Executions of the Innocent, 51 Crime & Delinquency 498 (2005) (finding, through logistic regression, evidence that attorney ability was critical to the outcome of capital cases);
32. Joyce S. Sterling, Retained Counsel versus the Public Defender: The Impact of Type of Counsel on Charge Bargaining, in William F. McDonald, ed, The Defense Counsel 167 (Sage 1983) (finding that "defendants with privately retained attorneys did not consistently obtain more lenient outcomes");
33. Robert V. Stover and Dennis R. Eckart, A Systematic Comparison of Public Defenders and Private Attorneys, 3 Am J Crim L 265 (1975) (finding comparable performance between public defenders and private attorneys).
34. But see Morton Gitelman, The Relative Performance of Appointed and Retained Counsel in Arkansas Felony Cases - An Empirical Study, 24 Ark L Rev 442, 450 (1971) (finding that although the performance of individual attorneys did not vary depending on whether they were appointed or retained, indigents with appointed counsel "plead guilty more frequently, receive fewer dismissals," and receive heavier sentences than defendants with retained counsel).
35. In a recent series of investigative articles, the Las Vegas Review-Journal examined criminal defense representation by court-appointed private attorneys in Clark County, Nevada (the court appoints these attorneys when there is a conflict within the CCPD). The journal found that the system was expensive and inefficient, strongly suggesting that defendants fare significantly worse when represented by court-appointed private attorneys than by CCPD attorneys. See Alan Maimon, Court Officials Review Indigent Defense, Las Vegas Rev-J 1A (Mar 27, 2007);
36. Alan Maimon, Conflicted Justice, Las Vegas Rev-J 1J (Mar 25, 2007);
37. Alan Maimon, Probe Finds Uneven Justice, Las Vegas Rev-J 1A (Mar 5, 2007).
38. Within the academic scholarship, see, for example, Radha Iyengar, Not Getting Their Due Process: An Evaluation of Federal Indigent Defense 3-4 (Princeton University working paper 2006), online at http://law.bepress.com/cgi/ viewcontent.cgi?article=1681&context=alea (visited Sept 23, 2007) (finding that criminal defendants receive lower sentences when represented by federal public defenders than by court-appointed private attorneys);
39. Pauline Houlden and Steven Balkin, Costs and Quality of Indigent Defense: Ad Hoc vs. Coordinated Assignment of the Private Bar within a Mixed System, 10 Just Sys J 159, 170 (1985) (finding that the method of assigning attorneys to criminal cases did not affect their outcomes); Pauline Houlden and Steven Balkin, Quality and Cost Comparisons of Private Bar Indigent Defense Systems: Contract vs. Ordered Assigned Counsel, 76 J Crim L & Criminol 176, 199 (1985) (determining that there was little difference in performance of private and public attorneys); Stuart S. Nagel, Effects of Alternative Types of Counsel on Criminal Procedure Treatment, 48 Ind L J 404, 424 (1973) (finding that hired attorneys provide some benefits to their clients over public defenders, but they also are "more likely to appear later in the case, . . . to consume more time while the defendant is in jail pending trial, and [are] more likely to have clients who receive longer prison terms if the defendant is imprisoned").
40. See, for example, Timothy R. Johnson, Paul J. Wahlbeck, and James F. Spriggs II, The Influence of Oral Arguments on the U.S. Supreme Court, 100 Amer Polit Sci Rev 99, 108-09 (2006) (discussing how Justice Harry Blackmun's grades of the oral argument performance of attorneys appearing before the Court correlate with the Court's decisionmaking);
41. Susan Brodie Haire, Stefanie A. Lindquist, and Roger Hartley, Attorney Expertise, Litigant Success, and Judicial Decisionmaking in the U.S Courts of Appeals, 33 L & Soc Rev 667, 684 (2006) (finding that litigation experience positively correlates with court decisions);
42. Kevin T. McGuire, Explaining Executive Success in the U.S. Supreme Court, 51 Polit Rsrch Q 505, 522 (1998) (same);
43. Kevin T. McGuire, Repeat Players in the Supreme Court: The Role of Experienced Lawyers in Litigation Success, 57 J Polit 187, 187-88 (1995) (same).
44. See, for example, Central Railroad & Banking Co v Pettus, 113 US 116, 128 (1885) (reducing the attorney's fees from 10 percent of the award in the case to 5 percent).
45. See, for example, Lindy Bros Builders, Inc v American Radiator & Standard Sanitary Corp, 487 F2d 161, 167 (3d Cir 1973) ("A logical beginning in valuing an attorney's services is to fix a reasonable hourly rate for his time."). For a background discussion, see Justin Lamb, Comment, The Lodestar Process of Determining Attorney's Fees: Guiding Light or Black Hole?, 27 J Legal Prof 203, 214 (2003) (describing the development of the lodestar process and arguing that "the lode-star [process] has clearly . . . created more problems than it has solved").
46. See Johnson v Georgia Highway Express, Inc, 488 F2d 714, 717-19 (5th Cir 1974) (outlining factors such as attorney effort, the difficulty of the case, opportunity cost, norms of compensation within the relevant legal community, attorney ability and seniority, and awarded fees in similar cases).
47. See Hensley v Eckerhart, 461 US 424, 433-34 (1983).
48. See In re Fine Paper Antitrust Litigation, 751 F2d 562, 583 (3d Cir 1984).
49. See Court Awarded Attorney Fees: Report of the Third Circuit Task Force, 108 FRD 237, 242 (1985).
50. See id.
51. See Judith Chevalier and Glenn Ellison, Are Some Mutual Fund Managers Better Than Others? Cross-Sectional Patterns in Behavior and Performance, 54 J Fin 875, 877 (1999).
52. See Bertrand and Schoar, 118 Q J Econ at 1172-73 (cited in note 12).
53. Walter Nicholson, Microeconomic Theory: Basic Principles and Extensions 505-07 (Dryden 3d ed 1985) (applying the standard competitive model to minimum wage laws and predicting that increasing wages above the competitive market equilibrium will result in unemployment).
54. Collectively these amenities and disamenities are known as compensating differentials.
55. The regulation occurs in two ways. Most states require that one (a) attend law school and (b) pass the state bar examination.
56. This is just one example of an important labor market that does not satisfy neoclassical assumptions. Others include the market for physicians, professional athletes, and accountants.
57. For example, if Firm A sues Firm B for $100 million and they settle for $30 million, is this a victory for Firm A or Firm B?
58. For example, the conviction rate for federal criminal defendants charged with felonies between 1989 and 2002 was over 80 percent. See Andrew D. Leipold, Why Are Federal Judges So Acquittal Prone? 83 Wash U L Rev 151, 180 table F (2005) (relying on statistics provided by the Federal Judicial Center).
59. The conviction rate for state criminal defendants between 1990 and 2000 was approximately 75 percent. See Erica J. Hashimoto, Defending the Right to Self Representation: An Empirical Look at the Pro Se Felony Defendant, 85 NC L Rev 423, 448 table 1 (relying on statistics provided by the National Archive of Criminal Justice Data).
60. We discuss this in detail below.
61. These assumptions include homoskedastic and normally distributed errors.
62. While this nonparametric approach has distinct advantages over a parametric approach, its use is less common because it requires more extensive programming, and is not built in to most statistical software packages, such as Stata and SAS. For our analysis, we wrote a nonparametric test.
63. See generally David Card, The Causal Effect of Education on Earnings, in David Card and Orley Ashenfelter, eds, 3 Handbook of Labor Economics 1801 (North Holland 1999) (citing studies finding that "better-educated individuals earn higher wages");
64. Joshua Angrist and Alan Krueger, Does Compulsory School Attendance Affect Schooling and Earnings?, 106 Q J Econ 979 (1991).
65. Stacy Dale and Alan Krueger provide evidence to suggest that students' post-university earnings are dictated more by the talent universities attract than by the education they provide. See Stacy Berg Dale and Alan B. Krueger, Estimating the Payoff to Attending a More Selective College: An Application of Selection on Observables and Unobservables, 117 Q J Econ 1491, 1494 (2002) (arguing that students maximize future income by finding the right "fit" in a school instead of attending the most selective school that accepts them).
66. See U.S. Department of Justice, Compendium of Federal Justice Statistics 2004 2 (GPO 2006), online at http://www.ojp.gov/bjs/pub/pdf/ cfjs0404.pdf (visited Sept 23, 2007) ("The proportion of convicted defendants who pleaded guilty increased from 87 percent during 1990 to 96 percent during 2004.").
67. This figure is as of 2005. See U.S. Census Bureau, Population Division, 100 Largest Counties (July 1, 2005), online at http://www.census.gov/ popest/counties/CO-EST2005-08.html (visited Sept 23, 2007).
68. See U.S. Census Bureau, Counties Ranked by Percent Change in Population, 1990-2000 (Apr 2, 2001), online at http://www.census.gov/ population/cen2000/phc-t4/tab04.pdf (visited Sept 23, 2007).
69. See U.S. Census Bureau, State & County QuickFacts, USA (Mar 23, 2007), online at http://quickfacts.census.gov/qfd/states/00000.html (visited Sept 23, 2007).
70. Calculation taken by the U.S. Department of Justice, Federal Bureau of Investigation, Crime in the United States 2004 table 5 (listing crime rates for 2004 for all regions and states in the United States), online at http://www.fbi.gov/ucr/cius_04/offenses_reported/offense_tabulations/table_05. html (visited Sept 23, 2007).
71. See id Section II: Offenses Reported at 68, 74-75 table 4 (listing crime rates for regions and states), 89, 105 table 6 (listing crime rates for 2004 by metropolitan area), online at http://www.fbi.gov/ucr/cius_04/documents/ CIUS_2004_Section2.pdf (visited Sept 23, 2007).
72. See Carri Geer Thevenot, Settlement Ends Ex-Inmate's Saga, Las Vegas Rev-J 1A (June 30, 2004).
73. See Miranda v Clark County, 279 F3d 1102, 1105 (9th Cir 2002).
74. Thevenot, Settlement Ends Ex-Inmate's Saga, Las Vegas Rev-J at 1A (cited in note 49).
75. See National Legal Aid and Defender Association, Evaluation of the Public Defender Office: Clark County, Nevada 57 (Mar 2003), online at http://www.nlada.org/DMS/Documents/1051974411.43/ Executive%20Summary%2C%20Table%20of%20Contents%20and%20Report%20%28without%20 appendices%29.pdf (visited Sept 23, 2007) (warning that unless the performance plan was implemented the office would be held back by its culture).
76. Conflicts of interest typically arise when two defendants in the same case have interests that diverge. In these circumstances, the CCPD will represent one defendant and court-appointed counsel (typically a private attorney) will represent the other.
77. Felony attorneys will occasionally take on a misdemeanor case, but these instances are rare.
78. The CCPD records the defendant's race, but it is missing in 20 percent of the data.
79. We chose the date range based on the case opening date rather than closing because the random assignment process occurs according to when cases begin. One potential difficulty with this method of case selection is truncation bias; that is, especially long cases will be excluded. In order to try to mitigate this potential source of bias we have chosen December 2005 as the end date for case initiation. Since the mean case duration is five months with a standard deviation of five months, over 90 percent of cases initiated in December 2005 are resolved in our data set. The fraction of cases begun earlier will be much higher, so overall, truncation bias should not impact the findings.
80. There were several other steps in the data cleaning process that reduced the size of the data set. A full description of the process is available from the authors.
81. There is one caveat: the office has an internal rule that all cases that proceed to trial have two attorneys. The second attorney serves as the second chair, solely to provide assistance to the lead attorney should the case proceed to trial. This feature should not affect our findings substantially, given that a very small percentage of felony cases result in trial. Further, unless there is a systematic bias to assignment of second chairs, this should bias our estimates towards zero.
82. When a defendant is placed in custody, he has a right to a preliminary hearing within fifteen days (the "speedy trial" rule). See Nev Rev Stat § 171.196 (1998). If the defendant is placed out of custody, the court sets the preliminary hearing within forty-five days.
83. There are a few caveats to this process. As mentioned above, newly hired attorneys are not part of the felony division, and the CCPD does not randomly assign capital and child sex crime cases. Second, the team chief may occasionally reassign cases when a disproportionate number of cases occur on a given preliminary hearing date; the court tries to avoid setting too many cases for the same day, but it is sometimes unavoidable given the number of cases that arise from the "speedy trial" rule.
84. As part of this study, we contacted over fifty public defender offices across the United States. The overwhelming majority of these offices assign cases nonrandomly.
85. Although we only test for random assignment using a limited number of case characteristics, we take this as evidence of random assignment on all observed and unobserved variables, since case characteristics are highly correlated.
86. We control for judge even though cases are randomly assigned to public defenders because public defenders often appear before the same judge repeatedly.
87. For example, assume drug cases increase over time, and that they result in a higher conviction rate. Then one would find a spurious negative correlation between tenure and incarceration rate, as newer PDs simply get a case mix with a higher mean incarceration rate.
88. See Katharine G. Abraham and Henry S. Farber, Job Duration, Seniority, and Earnings, 77 Am Econ Rev 278, 278-79 (1987) (presenting evidence that wages do rise with seniority, but at a lower rate than previously estimated when accounting for unobserved quality).
89. We do not include age because there is not sufficient variation to avoid a collinearity with tenure.
90. We discarded the data from the single American-Indian attorney in the CCPD.
91. If there are multiple charges, the relevant plea is the top count.
92. The CCPD does not keep formal statistics, but the public defender informed us that far less than 5 percent of felony cases go to trial.
93. See, for example, Christian Dustmann and Costas Meghir, Wages, Experience and Seniority, 72 Rev Econ Stud 77, 77-79 (2005) (reporting that skilled workers experience positive returns to wages from experience and firm tenure).
94. See also Daron Acemoglu and Jörn-Steffen Pischke, The Structure of Wages and Investment in General Training, 107 J Polit Econ 539, 560 (1999) (describing how firms create de facto "specific" rather than "general" skills, which may encourage workers to remain at the firm); Abraham and Farber, 77 Amer Econ Rev at 279 (cited in note 65) (finding that individuals who remain with the same employer have higher earnings than others with similar job sector experience but lower firm tenure).
95. But see Joseph G. Altongi and Robert A. Shakotko, Do Wages Rise with Job Seniority?, 54 Rev Econ Stud 437, 454 (1987) (finding that the effect of tenure on wages is small and that most wage growth can be explained by the general labor market and switching of jobs).
96. See Thomas E. MaCurdy, An Empirical Model of Labor Supply in a Life-Cycle Setting, 89 J Polit Econ 1059, 1060-66 (1981) (constructing an economic and empirical model of employment that incorporates long-term as well as short-term incentives).
97. See also George Baker, Michael Gibbs, and Bengt Holmstrom, The Wage Policy of a Firm, 109 Q J Econ 921, 923 (1994) (describing how in addition to wage differences within individuals over time, there are also wage differences across cohorts).
98. See NALP Foundation and American Bar Foundation, After the J.D. at 68 table 9.3 (cited in note 10) (listing the average salaries of attorneys by racial group). The overall medians are as follows: Asians, $80,000; whites, $73,000; Hispanics, $71,000; blacks, $65,000.