The Real Impact of Eliminating Affirmative Action in American Law Schools: An Empirical Critique of Richard Sander's Study

Beginning of article

INTRODUCTION

I. THE EFFECTS OF ENDING AFFIRMATIVE ACTION ON THE PRODUCTION OF
AFRICAN AMERICAN ATTORNEYS
  A. The Effects on Law School Applications, Admissions, and
     Matriculation
     1. Sander's projections are based on 2001 data, which does not
     reflect current trends
     2. Sander overestimates the numbers of African Americans who would
     apply to law school, get into the law school to which they would
     apply, or choose to enroll
  B. The Effects on Law School Performance, Graduation, and Passage of
  the Bar
     1. Concerns about statistical methods
     2. Law school performance and graduation
     3. Performance on the bar examination
  C. The Bottom Line: The Net Effects on the Numbers of African
  American Lawyers
II. THE IMPACT OF ENDING AFFIRMATIVE ACTION ON THE DISTRIBUTION OF
AFRICAN AMERICAN STUDENTS AMONG LAW SCHOOLS
CONCLUSION

INTRODUCTION

In 1970, there were about 4000 African American lawyers in the United States. Today there are more than 40,000. The great majority of the 40,000 have attended schools that were once nearly all-white, and most were the beneficiaries of affirmative action in their admission to law school. American law schools and the American bar can justly take pride in the achievements of affirmative action: the training of tens of thousands of African American (as well as Latino, Asian American, and Native American) practitioners, community leaders, judges, and law professors; the integration of the American bar; the services that minority attorneys have provided to minority individuals and organizations once poorly serviced by white lawyers; and the educational benefits that law students of all backgrounds derive from studying in a racially diverse environment. (1)

But not every student admitted through affirmative action realizes his or her ambition to practice law. Of the African American students who entered law school in the fall of 1991, the one year for which we have good data, about 43% either did not graduate or graduated but had not passed a bar exam within two years of graduation. Only 17% of the white students in the 1991 cohort suffered either of these fates. (2)

In A Systemic Analysis of Affirmative Action in American Law Schools (Systemic Analysis), Professor Richard Sander argues that if affirmative action were eliminated in law school admissions, the rate at which African American students fail to graduate and pass the bar would be reduced substantially without any concomitant loss in the numbers of African Americans joining the bar. (3) He acknowledges that fewer African American students would be admitted to law school, but predicts that those who were admitted would graduate and pass the bar at much higher rates because they would no longer be attending schools where the competition was too stiff for them. Sander builds to an astonishing forecast: "that the number of black lawyers produced by American law schools each year and subsequently passing the bar would probably increase if those schools collectively stopped using racial preferences." (4) In particular, he predicts that the cohort entering law school in 2001 would have produced 7.9% more new black lawyers entering the bar. (5)

We agree with Sander that the high rate at which African American students fail to graduate and fail to pass the bar is alarming. (6) Indeed, we take the problem so seriously that despite the high value we place on racial diversity within law schools, the four of us would not support affirmative action as currently practiced in law school admissions if we believed that employing race-neutral admissions criteria would in fact lead to a net increase in the number of African Americans passing the bar. (7) We find, however, that while Sander has appropriately forced us and others to take a hard look at the actual workings of affirmative action, he has significantly overestimated the costs of affirmative action and failed to demonstrate benefits from ending it. The conclusions in Systemic Analysis rest on a series of statistical errors, oversights, and implausible assumptions. It is these empirical shortcomings that we address in this Response.

The next Part of the Response deals step-by-step with the process of becoming a lawyer, from application, admission, and enrollment in law school through graduation and sitting for the bar exam. At each stage we explain why the findings and claims in Systemic Analysis are not supported by the data. We conclude that if affirmative action was ended, there would be a substantial net decline in the number of African Americans entering the bar rather than the 7.9% increase that Sander forecasts. We cannot say precisely how severe this decline would be, but our best estimate is that it would be in the range of 30% to 40%.

In the final Part, we shift to a related question: without affirmative action, how would African Americans be distributed across the range of American law schools? Sander acknowledges that the numbers of African Americans at the dozen or so most elite schools would be reduced by at least three-fourths, but expects that most other schools would have as many African American students as they do today. We disagree. We believe that the numbers of African Americans would decline substantially at the great majority of the nation's fifty to eighty most selective law schools and expect that this decline would be followed in turn by a decline in the number of African Americans attaining the sorts of leadership positions that graduates of these schools attain today.

As we begin, we want to emphasize the limited scope of our Response. First, Sander confines his analysis to African Americans, and we have done the same. His findings and ours might be quite different for Latinos, Native Americans, and other groups that have benefited from affirmative action. Second, Sander addresses more aspects of the affirmative action system than we examine here. We focus solely on the likely consequences of ending affirmative action because we agree with Sander that it is a "central question." (8) Indeed, it is almost certainly the central question of interest to policymakers and the public that his article raises. We want to make clear, however, that our silence on other claims Sander makes, such as his claims regarding the evidence before the Court in Grutter v. Bollinger on the University of Michigan Law School's admissions procedures (9) or his analysis of the job market for African American graduates, (10) does not mean that we agree with Sander. Had we been allowed more space, we would have disputed aspects of these claims as well.

Indeed, space prevents us from being as detailed as we would like in dealing with some aspects of Systematic Analysis we do address. For those readers who desire a finer-grained analysis, we have created a longer version on the Web. (11) It is on the Web also that we will respond point by point to the counterclaims that Sander makes in this issue.

I. THE EFFECTS OF ENDING AFFIRMATIVE ACTION ON THE PRODUCTION OF AFRICAN AMERICAN ATTORNEYS

A. The Effects on Law School Applications, Admissions, and Matriculation

Part VIII of Systemic Analysis estimates the impact on African American enrollments in law school if affirmative action were ended tomorrow. (12) Sander's estimate is built of the following steps: (1) an assumption that there would be no decline in African American applications to law school; (2) an estimate that there would be only a 14% decline in the numbers of African American applicants who would be admitted to at least one school and an assumption that those eliminated would be the 14% of current African American law students with the lowest entry credentials; (3) an assumption that among those admitted, African Americans would maintain current matriculation rates (i.e., that "cascading" to lower schools would not reduce the rate at which admitted African Americans chose to enroll in law school); and hence, (4) a forecast that there would be only a 14% decline in the total number of African Americans matriculating in American law schools. We believe each of these assumptions and predictions is unsound, and that all of them err in the direction of overestimating the probable levels of matriculation by African Americans.

Sander rests his conclusion that ending affirmative action would produce only a 14% decline in African American matriculation to law school on the research of Linda Wightman, who directed the Bar Passage Study for the Law School Admission Council (LSAC). (13) Using what she referred to as the "grid" method, which applies white admission rates to African Americans with similar LSAT scores and similar undergraduate grade point averages (UGPA), Wightman concluded that, in 2001, if African American law students had been admitted in the same proportions as whites with similar credentials, 14% of the African American students who received at least one offer of admission would not have received any offers at all, even if they had applied to a wide range of schools to which they never actually applied. (14) Sander accepts Wightman's 14% figure as a realistic estimate of the probable decline in African American admissions. For two different sets of reasons, the actual decline in matriculation by African American students would be much greater.

1. Sander's projections are based on 2001 data, which does not reflect current trends

Sander bases his predictions on data from the year 2001, which was the most recent year available to Wightman when she wrote her article. While Sander treats 2001 as representative of what would happen if affirmative action ended at law schools today, (15) no single year can serve that function. Further, 2001 turns out to have been one in a group of adjacent years when white and overall application levels to law school were comparatively low.

In Table 1, we provide for each year from 1991 through 2004 grid model estimates based on exactly the same procedure that Wightman used for 2001. (16) The table reveals that the projected size of the decline in African American admissions in any given year is strongly tied to the size of the overall applicant pool. It is, in particular, tied to the volume of applicants with high LSATs and UGPAs. (17) In the "dot com" boom years of 1997 through 2001, young white college graduates in much larger than usual numbers took jobs or pursued other schooling opportunities rather than apply to law schools. While African American applications to law school grew slightly during this period, total applications to law schools declined from a high of 99,000 in 1991 to a low of 72,000 in 1998. By 2001, they had risen slightly to 77,000, and, by 2004, they had returned to the levels of 1991.

In 2004, as Table 1 shows, we estimate that ending affirmative action would have cut by about 32.5% the numbers of African Americans who would have been admitted to any accredited law school. Because of improvements in African American entry credentials over the years and a small increase in the number of law schools, the projected decline for 2004 is smaller than the projection had been in 1991, when total applications were about the same, but 32.5% is still an enormous reduction, much higher than the estimate of 14% for 2001. The overall pattern from 1991 through 2004 suggests that the impact of ending affirmative action on potential African American admissions to law school would vary across years, but that in most years the negative impact would be substantially greater than it would have been in 2001. Indeed, the numbers lost would be so great that even if Sander were correct that the remaining black students would graduate and pass the bar at the same rate as their white classmates (and we explain later why he is not), there would have been a net loss in 2004 of about 21% in the number of African American lawyers produced under Sander's model, and from early indications, nearly the same loss in 2005 as well. (18)

2. Sander overestimates the numbers of African Americans who would apply to law school get into the law school to which they would apply, or choose to enroll

The grid model is useful solely for suggesting how many African Americans might have been admitted to some law school somewhere without affirmative action, if they had chosen to apply to the school that would admit them. It offers an upper-bound estimate of the numbers of African Americans who could enter law school under race-neutral criteria. (19) Wightman, from whom Sander borrowed his grid approach, made clear that the grid model cannot tell us whether African American students would actually apply to significantly lower-ranked law schools to which they never applied in real life, and she cautioned against the very use Sander makes of the model's approach. (20) Nor can the grid model tell us whether African Americans, even if their law school aspirations were not dampened by the diminished prestige of the schools they might attend, would successfully identify and apply to schools that would admit them. In short, the grid model cannot provide even a loose estimate of how many African Americans would in fact matriculate in law school, but Sander, though recognizing that the model cannot tell us what African Americans would actually do, in the end treats it as if it does. (21) We no more than Sander can state precisely how many African Americans would enter law school in a world without affirmative action, but we can offer reasons, supported by evidence and common sense, why the number Sander gives us is a substantial overestimate.

First, Sander incorrectly believes that, if affirmative action were ended, law would remain as appealing to African Americans for a career as it is today. He acknowledges that an African American college student "attracted to the law but not desperate to have a legal career might have second thoughts if she faced the prospect of attending a fortieth-ranked school instead of one ranked fourteenth." (22) He nonetheless guesses that there would be no decline in law school applications because African Americans will learn of his findings and recognize that they will, in general, have a better chance of passing the bar by going to the fortieth-ranked school. (23) Our estimate is that many of the African Americans who now secure admission to the fourteenth-ranked school could, in the absence of affirmative action, at best expect admission only to a school in the sixtieth- to eightieth-rank range, (24) and we expect that whether it is the fortieth- or the eightieth-ranked school that would admit them, many African Americans who now opt to attend elite law schools will turn to other careers.

Even today, for many African American students applying to law school, other career paths appear to be nearly as attractive as law. (25) A large proportion of applicants to law school (of all backgrounds) are tentative in their commitment to law school, much more tentative than, say, applicants to medical school. (26) Among the respondents to the Bar Passage Study, for example, 54% of African Americans and 52% of whites said that they had considered applying to graduate and professional programs other than law in the preceding two years. A less robust commitment to applying to law school among African Americans is also evident in that black students apply later in the admissions cycle compared to whites, apply to fewer schools on average than whites (4.2 versus 4.7 in 1999-2003), and take the LSAT later in the admissions cycle. (27) For some African Americans, the ending of affirmative action would probably be the "tipping point" away from law school and toward other career paths. (28)

Even those African American students who could still get into one of the nation's most selective law schools might find attending law school less attractive than they do today. By Sander's own estimates, without affirmative action African Americans would constitute only about one to two percent of the student bodies at the most elite law schools. (29) Today, the top thirty law schools in U.S. News & World Report (U.S. News) have student bodies that are, on average, 8.1% African American (excluding the three schools where affirmative action has been prohibited by law). (30) Many African American students care about attending a law school that has other minority students. On the Bar Passage Study survey, 68% of African American students at the two most elite tiers of schools said that the numbers of minority students at the school they were attending was a very important or somewhat important reason for applying. (31) We thus expect that some African American students who could still get into an elite law school will choose not to apply at all, rather than be a part of a tiny minority. (32)

Second, Sander assumes that so long as an African American considering law school could get into some law school, she will apply to that law school regardless of where it is in the United States. Although large numbers of law students, including African American students, travel substantial distances from home to attend the nation's most selective law schools, most students who attend lower-tier schools are from the same or an adjacent state.

The question that Sander's imagined future poses is whether African American students now traveling afar to attend relatively prestigious schools would be willing to travel similar distances to attend lower-tier schools. Sander believes the question is of minimal significance because there are plenty of lower-tier law schools in the states where most African Americans already live. While it is true that lower-tier law schools are located throughout the country, we are quite uncertain exactly what admissions landscape African Americans now at higher-tier law schools would face in a world without affirmative action. It is important to remember that if affirmative action ended, African Americans who applied to a nearby lower-tier school with credentials within that school's range that might secure admission will not necessarily be accepted. If race is irrelevant to admissions, the lower their credentials are within the pool of admissible applicants, the more they will have to offer other strong qualities apart from race to secure admission. (33) African Americans who are not admitted to the nearby lower-tier schools will have to turn elsewhere, and a disproportionate number of the lower-tier schools that might have space for them are located in states in the Great Plains, Rocky Mountains, Southwest, Pacific Northwest, and rural New England, where few African Americans go to law school today (34) and where African Americans from other parts of the country may be reluctant to move, especially if the schools in these other locations primarily place their graduates in locations where African Americans are unlikely to want to live and practice.

Third, Sander acknowledges that the availability of financial aid can affect decisions about attending law school, but points to the "After the JD" study to show that African American students receive about three times as much in "grants and aid" from law schools as do students of other races and concludes that financial considerations will not reduce post-affirmative action law school enrollment estimates. (35) His forecast is doubtful. If African American students currently receive grants in part through race-conscious programs not solely related to need, these programs are likely to end with the end of affirmative action. If the reason they receive more grants is because they have greater need, then that need will continue even if affirmative action is ended.

Today, even with the availability of scholarships, more African Americans than whites borrow to attend law school (95% versus 84%), and those who borrow borrow as much on average as white students. (36) Thus, in deciding whether to attend a lower-tier law school, an African American student who could attend a more elite school today is likely to be affected by his estimate of the size his educational debt will be in relation to the earnings he can expect to receive, and the earnings of graduates of lower-tier schools are in general much lower than the earnings of the graduates of elite schools. (37) Sander argues that these status-associated differences would be more than made up for by the better grades the student would receive at the lower-tier school because grades are more important than prestige in predicting earnings. We strongly doubt his conclusions in this regard, especially as they apply to African Americans attending elite law schools. (38) As David Wilkins points out in this issue, law school prestige is a much more conspicuous long-range signal in the labor market than grades. (39)

We have suggested several reasons why, if affirmative action were ended, fewer African Americans would apply to law school than do today. We also expect that many African Americans who could get in somewhere would apply only to law schools that would not admit them. Even with affirmative action in place, hundreds of African Americans with solid credentials are currently rejected by every school to which they apply. (40) An end of affirmative action, by restricting greatly the range of schools available to most African American applicants, would surely increase the number of futile applications. Thus, Sander's posited national admissions market, where, without affirmative action, the vast majority of African Americans would smoothly "cascade" down a tier or two, is quite implausible. (41) Many African American students who would be admitted to some law school in an imagined world where they would be willing to go anywhere will, in the real world where they choose five …