The Volokh Conspiracy - Why the Texas ten percent plan is worse than traditional affirmative action:

In 1996, US Court of Appeals for the Fifth Circuit held, in Hopwood v. Texas, that the use of racial preferences to achieve "diversity" in college admissions is unconstitutional. In response, the Texas state legislature adopted the "Ten Percent Plan," a "race-neutral" way to achieve the desired proportion of minority students Texas' state universities without resorting to explicit consideration of race in admissions. The ten percent plan gives any high school student who is in the top 10% in his high school class automatic admission to any Texas state university, regardless of standardized test scores, the content of the classes he took, the strength of his high school, extracurricular activities, and other considerations.

The Ten percent plan was endorsed by then-Texas governor George W. Bush and has since been touted by the Bush Administration and others as a superior alternative to traditional affirmative action plans that rely open racial preferences; similar plans were later enacted in California and Florida. As this recent New York Times article points out, the ten percent plan succeeded in returning the percentages of African-American and Hispanic students in Texas state universities to roughly their pre-Hopwood levels - largely because many minority students attend schools where blacks or Hispanics are in the overwhelming majority. Unfortunately, however, the ten percent plan has negative side-effects and perverse incentives that are considerably worse than those of traditional affirmative action, including racial quotas.

First, it often leads universities to admit students that are probably inferior to those they would have chosen otherwise:

But the formula has also had unintended consequences . . .; it has become the tail that wagged the dog, university officials suggest. Seventy-one percent of the 6,864 Texans in the [UT Austin flagship campus] freshman class are top 10 percenters, compared with 41 percent in the first year the formula was used. That steady growth has frustrated college officials who have seen their flexibility to admit high school class presidents, high SAT scorers, science fair winners, immigrant strivers, artists and the like narrow.

"At some point you have to ask yourself, do you really want to admit your whole class on a single criteria," said Bruce Walker, the admissions director at Austin. "It doesn't give you the opportunity to recognize other kinds of merit."

To be sure, this result could happen with traditional racial preferences as well. However, the ten percent plan affects a great many more admissions decisions than even the most rigid old-style affirmative action systems do. Rarely, if ever, do traditional affirmative action plans determine the admission of more than 15-20% of a school's student body. By contrast, at the University of Texas at Austin, over 70% of the student body was admitted under the ten percent plan. While some of these students would surely have gotten in anyway, it is highly likely that the ten percent plan leads to much larger sacrifices of academic merit than do racial preferences similar to those used at most other academic institutions.

Second, and probably much worse, the article notes that the formula creates perverse incentives for students to try to game the system by transferring to weaker schools or taking easier classes. While neither the article nor other evidence I have seen provides precise data on the numbers of students who do this, the effect may well be large. When I lived in Texas in 2001-2002, I met quite a few people with high school-age children who had switched to weaker schools in order to take advantage of the plan, or were considering doing so. Obviously, there is no similar perverse incentive created by traditional affirmative action. With a system of racial quotas or "plus factors," both white and minority high school students still have incentives to go to strong schools, in order to maximize their college admissions chances.

Third, the tradeoffs inherent in the ten percent plan are less transparent to both students and the general public than those involved in racial quotas. As a result, it is more likely that harmful effects will remain unmonitored and undetected. If public universities are going to strive for racial diversity, the costs and benefits of doing so should be as transparent as possible.

Finally, the ten percent plan also has the effect of disadvantaging high-achieving minority students who go to strong schools and - in part for that reason - fall short of the top ten percent in their class. Not only are these students disfavored relative to minority students attending weaker schools, they are also disfavored compared to whites in weaker schools as well:

[T]he formula has meant that the university may neglect desirable black and Hispanic students, as well as white students, who attend lustrous high schools but may not finish in the top 10. Marcus Price, a black finance major, for example, graduated from the High School for Engineering Professions in Houston, a competitive magnet school, with a 3.4 grade point average that included three A.P. courses. But with so many college-bound students to compete with, he ranked only in the top 20th percentile.

"I thought it was funny that you could go to a less competitive school, score a total of 800 or 900 on your SATs and get into U.T. at Austin as long as you were in the top 10 percent," said Mr. Price, who scored 1200 on his SATs.

Some, perhaps including President Bush, would argue that the ten percent plan is still preferable to traditional affirmative action because racial preferences are intrinsically wrong, regardless of consequences. Perhaps they are. But if it is morally wrong to aim for a given racial balance in a state university student body by using explicit racial preferences, why is it not equally wrong to intentionally try to achieve the same effect through indirect, facially "neutral" means? In the days of Jim Crow, southern states often used facially neutral policies such as literacy tests, poll taxes, and peonage laws to disadvantage blacks. Few today would argue that these policies were somehow morally superior to those Jim Crow laws that discriminated against blacks through explicit racial classifications. If, as critics of affirmative action claim, explicit affirmative action preferences are morally wrong for the same reason that Jim Crow laws were wrong, then "facially neutral" affirmative action systems such as the Texas ten percent are wrong for the same reasons that the facially neutral means of propping up Jim Crow were.

If we want to ensure that some set percentage of university admissions slots go to particular minority groups, far better to do so through traditional affirmative action, than by means of the Texas ten percent plan.

UPDATE: I would like to briefly respond to three points raised by commenters. First, many claim that it's not possible for high school students in Texas to game the system by switching to weaker schools because of the distances involved. Texas is indeed a large state, but much of the population lives in several large cities (Houston, Dallas, San Antonio, etc.) where there are many high schools close together, and gaming the system more than possible.

Second, some claim (in contrast to the first group) that this kind of gaming would be a good thing because it might move good students to weak schools and thereby improve education for students trapped in the latter. This argument would be more compelling if it were not for the extensive empirical evidence against it. For example, numerous studies such as David Armor's Forced Justice (1992) show that efforts to bus in stronger students to poorly performing inner city schools had little or no impact on education quality. Merely introducing a relatively small percentage of superior students into a poorly performing students is unlikely to have a significant impact on those students already there. There are many better ways to improve education for those students trapped in the worst public schools, most notably school choice.

Third, several people claim that it is contradictory that I argue that the negative impact of the ten percent plan might go "undetected" yet simultaneously note several negative effects that clearly have been detected. I was guilty of a loose use of terminology. Because the ten percent plan is more opaque than traditional affirmative action, the general public (which lacks the time and incentive to follow policy issues closely) is less likely to notice its perverse effects than those of old-fashioned AA. That doesn't prevent their being noticed by experts.

NOTE: in 2001-2002, I clerked for Judge Jerry E. Smith, the Fifth Circuit judge who wrote the court's opinion in Hopwood five years earlier. I don't think this has any real connection to the merits of the ten percent plan, but I mention it to forestall the likely claim that my criticism of the ten percent plan is somehow inappropriate because I am "hiding" this fact.

	[Ilya Somin, December 16, 2006 at 2:20am] Trackbacks Why the Texas ten percent plan is worse than traditional affirmative action: In 1996, US Court of Appeals for the Fifth Circuit held, in Hopwood v. Texas, that the use of racial preferences to achieve "diversity" in college admissions is unconstitutional. In response, the Texas state legislature adopted the "Ten Percent Plan," a "race-neutral" way to achieve the desired proportion of minority students Texas' state universities without resorting to explicit consideration of race in admissions. The ten percent plan gives any high school student who is in the top 10% in his high school class automatic admission to any Texas state university, regardless of standardized test scores, the content of the classes he took, the strength of his high school, extracurricular activities, and other considerations. The Ten percent plan was endorsed by then-Texas governor George W. Bush and has since been touted by the Bush Administration and others as a superior alternative to traditional affirmative action plans that rely open racial preferences; similar plans were later enacted in California and Florida. As this recent New York Times article points out, the ten percent plan succeeded in returning the percentages of African-American and Hispanic students in Texas state universities to roughly their pre-Hopwood levels - largely because many minority students attend schools where blacks or Hispanics are in the overwhelming majority. Unfortunately, however, the ten percent plan has negative side-effects and perverse incentives that are considerably worse than those of traditional affirmative action, including racial quotas. First, it often leads universities to admit students that are probably inferior to those they would have chosen otherwise: But the formula has also had unintended consequences . . .; it has become the tail that wagged the dog, university officials suggest. Seventy-one percent of the 6,864 Texans in the [UT Austin flagship campus] freshman class are top 10 percenters, compared with 41 percent in the first year the formula was used. That steady growth has frustrated college officials who have seen their flexibility to admit high school class presidents, high SAT scorers, science fair winners, immigrant strivers, artists and the like narrow. "At some point you have to ask yourself, do you really want to admit your whole class on a single criteria," said Bruce Walker, the admissions director at Austin. "It doesn't give you the opportunity to recognize other kinds of merit." To be sure, this result could happen with traditional racial preferences as well. However, the ten percent plan affects a great many more admissions decisions than even the most rigid old-style affirmative action systems do. Rarely, if ever, do traditional affirmative action plans determine the admission of more than 15-20% of a school's student body. By contrast, at the University of Texas at Austin, over 70% of the student body was admitted under the ten percent plan. While some of these students would surely have gotten in anyway, it is highly likely that the ten percent plan leads to much larger sacrifices of academic merit than do racial preferences similar to those used at most other academic institutions. Second, and probably much worse, the article notes that the formula creates perverse incentives for students to try to game the system by transferring to weaker schools or taking easier classes. While neither the article nor other evidence I have seen provides precise data on the numbers of students who do this, the effect may well be large. When I lived in Texas in 2001-2002, I met quite a few people with high school-age children who had switched to weaker schools in order to take advantage of the plan, or were considering doing so. Obviously, there is no similar perverse incentive created by traditional affirmative action. With a system of racial quotas or "plus factors," both white and minority high school students still have incentives to go to strong schools, in order to maximize their college admissions chances. Third, the tradeoffs inherent in the ten percent plan are less transparent to both students and the general public than those involved in racial quotas. As a result, it is more likely that harmful effects will remain unmonitored and undetected. If public universities are going to strive for racial diversity, the costs and benefits of doing so should be as transparent as possible. Finally, the ten percent plan also has the effect of disadvantaging high-achieving minority students who go to strong schools and - in part for that reason - fall short of the top ten percent in their class. Not only are these students disfavored relative to minority students attending weaker schools, they are also disfavored compared to whites in weaker schools as well: [T]he formula has meant that the university may neglect desirable black and Hispanic students, as well as white students, who attend lustrous high schools but may not finish in the top 10. Marcus Price, a black finance major, for example, graduated from the High School for Engineering Professions in Houston, a competitive magnet school, with a 3.4 grade point average that included three A.P. courses. But with so many college-bound students to compete with, he ranked only in the top 20th percentile. "I thought it was funny that you could go to a less competitive school, score a total of 800 or 900 on your SATs and get into U.T. at Austin as long as you were in the top 10 percent," said Mr. Price, who scored 1200 on his SATs. Some, perhaps including President Bush, would argue that the ten percent plan is still preferable to traditional affirmative action because racial preferences are intrinsically wrong, regardless of consequences. Perhaps they are. But if it is morally wrong to aim for a given racial balance in a state university student body by using explicit racial preferences, why is it not equally wrong to intentionally try to achieve the same effect through indirect, facially "neutral" means? In the days of Jim Crow, southern states often used facially neutral policies such as literacy tests, poll taxes, and peonage laws to disadvantage blacks. Few today would argue that these policies were somehow morally superior to those Jim Crow laws that discriminated against blacks through explicit racial classifications. If, as critics of affirmative action claim, explicit affirmative action preferences are morally wrong for the same reason that Jim Crow laws were wrong, then "facially neutral" affirmative action systems such as the Texas ten percent are wrong for the same reasons that the facially neutral means of propping up Jim Crow were. If we want to ensure that some set percentage of university admissions slots go to particular minority groups, far better to do so through traditional affirmative action, than by means of the Texas ten percent plan. UPDATE: I would like to briefly respond to three points raised by commenters. First, many claim that it's not possible for high school students in Texas to game the system by switching to weaker schools because of the distances involved. Texas is indeed a large state, but much of the population lives in several large cities (Houston, Dallas, San Antonio, etc.) where there are many high schools close together, and gaming the system more than possible. Second, some claim (in contrast to the first group) that this kind of gaming would be a good thing because it might move good students to weak schools and thereby improve education for students trapped in the latter. This argument would be more compelling if it were not for the extensive empirical evidence against it. For example, numerous studies such as David Armor's Forced Justice (1992) show that efforts to bus in stronger students to poorly performing inner city schools had little or no impact on education quality. Merely introducing a relatively small percentage of superior students into a poorly performing students is unlikely to have a significant impact on those students already there. There are many better ways to improve education for those students trapped in the worst public schools, most notably school choice. Third, several people claim that it is contradictory that I argue that the negative impact of the ten percent plan might go "undetected" yet simultaneously note several negative effects that clearly have been detected. I was guilty of a loose use of terminology. Because the ten percent plan is more opaque than traditional affirmative action, the general public (which lacks the time and incentive to follow policy issues closely) is less likely to notice its perverse effects than those of old-fashioned AA. That doesn't prevent their being noticed by experts. NOTE: in 2001-2002, I clerked for Judge Jerry E. Smith, the Fifth Circuit judge who wrote the court's opinion in Hopwood five years earlier. I don't think this has any real connection to the merits of the ten percent plan, but I mention it to forestall the likely claim that my criticism of the ten percent plan is somehow inappropriate because I am "hiding" this fact.