CHAPTER EIGHT
THE WARMING EFFECT
AS UNIVERSITY OF CALIFORNIA administrators fretted over the first steps toward implementing Prop 209, they shared a common fear: that race-blind admissions would dramatically reduce minority enrollments. It stood to reason that fewer minorities would be admitted, especially at the most selective campuses, and that fewer would apply because once a talented black high school senior realized that her chances of getting into Berkeley had fallen from, say, 60 percent before Prop 209 to 30 percent after, it was reasonable to suppose that she should devote her energies to more promising alternatives. There was also much talk of a “chilling effect,” the idea that black and Hispanic students who were admitted might refuse offers from campuses where their numbers were in free fall. It was at least plausible that many of these students would see the UC system generally and the most selective campuses in particular as hostile places for minorities and, thus, look more favorably upon the scores of private colleges and other state universities that were waiting in the wings with offers of generous scholarships. And finally it occurred to administrators that a vicious cycle might ensue, in which minority enrollment drops would produce still greater chilling and that the cycle might end in a virtually complete disappearance of blacks and Hispanics from the most elite campuses.
It is hard to overstate how pervasive this fear was among affirmative action supporters, particularly within the university. As we mentioned in Chapter Seven, a careful estimate presented to the regents in 1995 projected that Berkeley would lose 85 percent of its black freshmen in the first year alone if the university did not institute large-scale socioeconomic preferences to offset the loss of racial ones. And that, many feared, would only be the beginning. UC Berkeley chancellor Chang-Lin Tien observed in 1996, “I’ve visited urban inner-city schools . . . many students feel, ‘Why should I work anymore? I’m a second-class citizen.’ I think 209 will have that kind of impact.” A senior administrator at Mills College, a selective private college in Oakland, commented just after the Prop 209 election that “I’ve heard that educators are worried that qualified minority students are not applying to the UC system because they feel they will be rejected or don’t like the environment.” In early 1998, pro-preference advocates pubished an entire book called Chilling Admissions, which spoke of “campus resegregation” and the “return [of our campuses] to almost total domination by the most privileged racial and ethnic communities.”
Race-blind admissions for undergraduates began with the 1997–98 season, and officials watched nervously as applications began to trickle in during the early fall of 1997. By November officials at several schools noticed unusually large numbers of incoming applications. At first glance, this was not a surprise: It made sense that lots of white and Asian high school seniors with good but not outstanding records would conclude that the end of racial preferences had significantly improved their chances. But actual counts of applications from whites and Asians did not show an increase for either group. The biggest jump in applications was coming instead from “none of the above,” applicants who declined to give any racial identification.
By January 1998 it was clear that this was turning into a banner year for the university. Applications rose at all eight UC campuses; total applications were up nearly 12 percent. This was partly because more students applied to multiple campuses, but the number of unique applicants was still up by nearly 7 percent. Because the university generally considered applications only from students who were “UC eligible” (something that a high school senior could figure out in consultation with a counselor), this meant that a much larger fraction of California seniors eligible to apply were actually doing so.
Even more astonishing was the race count. The change in total unique UC applications from 1997 to 1998 broke out like this:
Whites: down 13 percent
Asians: down 3 percent
American Indians: down 1 percent
Blacks: up 1 percent
Hispanics: up 7 percent
“Other” and “Unknown”: up 214 percent
Total: up 7 percent
It seemed likely that most of the new “unknown” race persons were whites and Asians, which would mean that applications went up for every racial group. But regardless of the true race of the “unknowns,” the message of these figures was startling. California high school seniors had applied to UC campuses in unprecedented numbers, and applications from blacks and Hispanics, rather than dropping precipitously, had gone up.
The more closely we look at these patterns, the more remarkable they become. We were able to obtain data from the College Board (which administers SAT and AP exams) on all California high school students in the years immediately before and after Prop 209 went into effect. They show that in 1997, the last year of race-conscious admissions, about 58 percent of the academically strongest blacks in California applied to Berkeley, the most elite school in the UC system and one of the top colleges in the nation. In 1998 this percentage went up to 70 percent. In 1997, with racial preferences still in place, black high school seniors in California who had a 50-50 chance of getting admitted to Berkeley had a one-in-three chance of applying there. In 1998, with no racial preferences, those who had a 50-50 chance of admission were an academically stronger group, with many more options at non-UC schools. That might lead one to suppose that fewer than one in three of this 50-50 chance group would apply to Berkeley. But in fact, one in two applied —a remarkable surge. Much the same thing happened, though not quite as powerfully, for Hispanics.
Looking across the entire pool of black and Hispanic potential applicants, race-blind admissions cut their chances of being admitted to Berkeley and UCLA by about half. Other things being equal, that should have produced a substantial drop in the likelihood of any given black or Hispanic high school senior applying to these top schools—arguably a decline on the order of 20 to 30 percent. Instead, total black and Hispanic applications to these schools went up. This means that, if we control for the “likelihood of acceptance,” black and Hispanic seniors in California became much more likely to apply to Berkeley and UCLA after race-neutral admissions began.
Similar patterns occurred at the less elite campuses. Black applications rose at seven of the eight UC campuses, and Hispanic applications rose at all eight. But it is harder to draw inferences from applications to, say, the UC Irvine campus, since minority students may have applied in higher numbers simply because they were less confident of being admitted to Berkeley or UCLA.
The implication of these application patterns is that Prop 209 did not “chill” the interest of minorities in attending the University of California; rather, if anything, Prop 209 “warmed” their interest. We cannot be sure why. But the most plausible inference from their actions is that the prospect of attending schools that admitted them without regard to their race attracted and even excited them.
An even better test of this hypothesis is the so-called yield rate—the rate at which blacks and Hispanics accepted offers of admission from the various UC schools. As college admissions officers well know, black yield rates tend to be much lower than white yield rates when we control for things like academic index because racial preferences give blacks many more options. This pattern had traditionally held across the UC campuses. In 1998, however, the black yield rate at Berkeley was nearly 52 percent, by far the highest in many years and probably an all-time record. The white rate that year was under 40 percent, slightly higher than the white average. The black yield is particularly astonishing because the black students admitted that year had, on average, far stronger academic records than their predecessors (and thus more numerous college options); race-blind admissions had greatly reduced the number of admissions of blacks with only moderately strong qualifications. More modest but still substantial improvements in yield occurred for blacks at other campuses and for Hispanics as well.
The jump in minority yield rates reinforces our impression from the application data: Blacks in particular and, to a slightly less extent, Hispanics, showed an extraordinary increase in interest in the University of California, especially in its most elite campuses, the year race-neutral admissions went into effect. Race-neutrality did not chill their interest in the UCs; instead, it drew them like a magnet.
The upshot of these shifts is that the various campuses experienced a much smaller drop in enrollment than administrators had feared. Black freshman enrollment at Berkeley in 1998 fell by 52 percent; Hispanic freshman enrollment fell by 43 percent. At UCLA the drops were 31 percent for blacks and 23 percent for Hispanics (many strong minority students cascaded from Berkeley to UCLA). For the UC system as a whole, the drops were 19 percent for blacks and 6 percent for Hispanics. And the declines in minority enrollment did not set off their own “chilling” chain reaction; in 1999 both black and Hispanic enrollment numbers rebounded modestly.
Something extraordinary seemed to be happening, but one would never know it from either the official reaction of UC officials or the media coverage. In May 1998 the New York Times ran a story entitled “Fewer Minorities Entering University of California.” The story noted that it appeared minority enrollments would “drop only slightly throughout the state system” and that “university officials . . . were heartened that the drop at [Berkeley and UCLA] had not been steeper, but they expressed concerns that the numbers foreshadowed a racially divided system.” The Times quoted UCLA’s vice chancellor, Theodore Mitchell, observing that “if this trend continues over the next five or six years,” the “University of California . . . will become a segregated system.” Another May 1998 Times story was devoted almost entirely to the anger of minority students—especially African Americans—at the post-209 policy. The reporter explained that black faculty and students, especially at the more elite campuses, “feel betrayed and insulted by what happened, and have told prospective students to beware of what could be an isolating experience.” The Los Angeles Times observed in April 1998 that “though coveted at the top two UC campuses, many high-achieving blacks and Latinos are likely to take their gifts to Stanford, the Ivy League, Michigan, and other campuses where the freshman classes intentionally will be more diverse.” Other California papers echoed the same themes.
Almost no one in either the UC administration or the media seemed to notice that, from a learning perspective, the unfolding developments of 1998 might signify a triple win for minority students and the UC system. Race-neutral admissions meant that students previously admitted to Berkeley and UCLA would “cascade” down to schools where they were likely to do better academically. (Students who previously got into less-selective UC campuses only through race-based “special admissions” would cascade into the excellent Cal State system.) At the same time, increased minority interest in the UC schools was strengthening the quality of students at every UC campus, and the rise in yield rates suggested that minority students now found the UC campuses more appealing.
We think the story we have just told is compelling, but there is a difference between highlighting impressive facts and analyzing a problem thoroughly with social science tools. For example, although black yield rates surged in 1998, they fell back some in 1999 (though still well above pre-209 levels). As minority applicants learned more about how race-blind admissions would work, those with particularly low chances of admission became progressively less likely to apply. It is worth standing back and asking whether a rigorous analysis of all the available data supports our hypothesis that minorities were “warmed,” not “chilled,” by Prop 209.
David Card and Alan Krueger, both renowned labor economists, published an analysis in 2005 using the College Board data mentioned above (Card generously helped us access the data as well). The College Board dataset includes detailed information on the academic background of each student and lists the schools to which she sent her SAT scores—a fairly good proxy for a student’s decision to eventually apply. Card and Krueger were interested in whether Prop 209 would cause some sort of chilling effect. They also realized that it was important to factor in the lower rate at which blacks and Hispanics would be admitted in the race-neutral regime. Their solution to this problem was to look at very strong applicants, whom they believed would be virtually certain to get into top UC schools both before and after Prop 209. They analyzed the relative probability of blacks and Hispanics, compared to whites and Asians, applying either to UC schools in general or specifically to the most selective UC schools, for 1994–1996 and 1999–2001. (They left out 1997 and 1998 on the grounds that these were transitional years, when students might be uncertain of whether racial preferences were being used, and by 1999 any “chilling effect” should have set in.) Their finding was that after the ban on racial preferences took effect, applications to UC schools from these very highly qualified blacks and Hispanics rose slightly, relative to whites and Asians. When good social scientists have a finding like this, they subject it to a variety of tests to see how “robust” it is. Card and Krueger’s finding was indeed robust.
Card and Krueger’s results actually understated the resilience of minority applications in an important sense. Even the sort of very strong minority students Card and Krueger examined faced much more competitive admissions under the race-neutral regime. To see this, suppose we divide all applicants to Berkeley and UCLA into ten equal groups (deciles), based on their academic index scores. The top three deciles are all very strong students, similar to those examined by Card and Krueger. In the three years before Prop 209’s implementation, whites and Asians in those deciles had a 75 percent chance of admission; blacks and Hispanics in those deciles were admitted 90 percent of the time. With applications going up after Prop 209, everyone faced more competitive admissions. Whites and Asians in these three deciles were admitted to Berkeley and UCLA 61 percent of the time after Prop 209, a 14 percent decline. Blacks and Hispanics were admitted 66 percent of the time, a 24 percent decline (and a more than threefold increase in the likelihood of being rejected). This means that the increase in highly qualified minority applications documented by Card and Krueger rose faster than white and Asian applications increased, even though these minority applicants were facing significantly diminished chances of admission.
So the evidence that blacks and Hispanics became more likely to apply after Prop 209, holding other factors constant, is quite strong. But what about the yield rate—their likelihood of actually attending a UC campus?
After very protracted negotiations, the relevance of which will become clearer in Chapters Nine and Ten, the University of California Office of the President (UCOP) reluctantly released to us in 2008 a fairly comprehensive dataset on students who applied to or entered the University of California as freshmen from the early 1990s through 2006. The dataset contains information on the academic characteristics of every applicant, their family background, which UC schools they applied to, which ones accepted them, and which one they chose to enroll in (if any). Though the UCOP data allows one to examine individual behavior and outcomes (critical, for example, in looking at yield issues), the data is blurred in some ways: Blacks, Hispanics, and American Indians are lumped together in a single “minority” group, and enrollment years are grouped in three-year aggregates.
Kate Antonovics, a labor economist at the University of California, San Diego, collaborated with one of us (Sander) in using UCOP’s data to study how yield rates were affected by Prop 209. Antonovics realized that gross yield rates alone could be misleading. Because stronger students would have lower yield rates (as they have more options), it was important to control for the significant strengthening of admits of all races during these years. Even trickier was the problem of adjusting for the other options students would have. For example, the black yield rate at UC San Diego might go up after Prop 209 simply because those black applicants would be less likely to get into Berkeley or UCLA. To deal with this problem, Antonovics and Sander controlled for the “choice set” that each student faced—that is, they compared similar students who had been admitted to the same set of schools before and after Prop 209.
Their analysis yielded statistically powerful results. Students of all races, (including those who did not identify their race) were somewhat more likely to accept an offer from most UC campuses in 1998–2000 (after the implementation of race-neutrality) than in 1995–1997. Yield rates generally went up by around 5 percent in the later period. But the increase was much larger for minority students (defined, as noted above, as blacks, Hispanics, and American Indians). Minority yield rates went up two or three times more than those of everyone else.
Moreover, as
Figure 8.1 suggests, there were some interesting patterns across the campuses. It is hard to draw a firm statistical conclusion with only eight campuses to observe (rather than, say, one hundred). But the four most elite campuses all had big increases in black and Hispanic yield rates, whereas three of the four less-elite campuses had much smaller increases. The four top campuses, especially Berkeley and UCLA, had generally been using the largest racial preferences before Prop 209, so the change to race-neutrality was particularly dramatic for them. These patterns reinforce the general implication of our findings, that UC schools became much more appealing to minority students after preferences ended. It is plausible that the bigger the change in policy at a particular campus, the more powerful the attractive effect might be.
Source: Antonovics & Sander, “Affirmative Action Bans and the Chilling Effect” (2012), and calculations by the authors.
Figure 8.1 shows, too, the change in black and Hispanic freshman enrollment that occurred at each UC campus after Prop 209. What is striking about this pattern is its seeming irrelevance to the warming effect discussed here. Although it is possible that the aggregate effect of Prop 209 on minority enrollment on each campus mattered to black and Hispanic students assessing admissions offers, such concerns certainly do not seem to have affected cross-campus yield rates.
Another interesting barometer of sentiment comes from out-of-state applicants. After Prop 209, the number of academically gifted, out-of-state minority high school students sending their SAT scores to UC campuses jumped. From 1995–1997 to 1998–2000, score sending by academically gifted, out-of-state Hispanics to UC schools went up 12 percent. The number of gifted blacks sending scores went up 48 percent. Over 90 percent of the increase in black interest was focused on the three most elite UC schools, Berkeley, UCLA, and UC San Diego.
The warming effect is the sort of powerful finding one can obtain from a natural experiment like Prop 209. The passage of Prop 209 stoked all sorts of powerful emotions, and its implementation was undoubtedly upsetting in particular for black and Hispanic students on UC campuses who had been admitted under the old regime. But the implementation of the policy and the availability of data that lets us observe actual decisions to apply or accept an offer of admission lets us rely on people’s actual behavior to infer their fundamental attitudes. Undoubtedly, the end of preferences dismayed some blacks and Hispanics considering the UCs. Some may have decided to go to other schools with traditional preference programs. Some undoubtedly participated in protests against the very idea of Prop 209. But as high school seniors making individual decisions about what colleges were most appealing, it seems that the aura of race-neutrality attracted many, many more black and Hispanic students than it repelled.
This conclusion is strengthened when we consider what else was going on at the same time. Financial aid is important to students; could more minorities have been lured to UC campuses after Prop 209 by increased financial aid? Actually, no: Prop 209 forbade race-based scholarships as well as race-based admissions. A significant number of race-based scholarship programs halted in 1998. Administrators worked to move some funds “offshore,” contending that if the university merely provided information about minority admits to private donors, those donors could confer race-based scholarships. But this process took time and never came close to offsetting the loss of earlier race-based programs. Thus, the warming effect occurred despite a probably sizeable loss of financial aid for black and Hispanic admits, relative to whites and Asians.
What about recruitment? The chancellor at Berkeley received some media attention for his efforts to stave off the feared chilling effect by telephoning black and Hispanic admittees personally to urge them to attend Berkeley. But even if this strategy was effective, a few dozen phone calls cannot account for the breadth and depth of the warming effect over several years and many campuses. And, as we saw earlier in the chapter, many UC minority students and faculty actually engaged in “antirecruiting” after race-neutral policies arrived, warning students away from a university that they saw moving toward segregation. For example, a black Berkeley senior who served as director of the campus Black Recruitment and Retention Center reported that her message to prospective students was “It’s a very hostile environment and . . . we’re not welcome here, and they don’t want us here because they’re not letting us in. We weren’t pushing them to come to [Berkeley].”
The University did develop new and ambitious outreach plans in 1997 and 1998, and it eventually spent tens of millions of dollars trying to strengthen the pipeline of disadvantaged and minority students into college. But these were generally long-term strategies that did not meaningfully get underway before 1999 and could not plausibly have had an effect upon the first few freshman classes admitted after Prop 209. Moreover, increasing applications through “outreach” would tend to depress yield rates because those applicants are likely to have the weakest preexisting interest in attending a UC school. The simultaneous increase in applications and yield rate, which, if anything, was strongest in the first year of Prop 209’s implementation, show that long-term outreach programs cannot explain the warming effect.
The imagined chilling effect that so many supporters of racial preferences believed in (and continue to believe in) was based on the idea that black and Hispanic students would find the University of California a less welcoming place without preferences. But, of course, another possibility was at least equally plausible: that students of color would welcome the chance to attend a school without the stigma of being a suspected “affirmative-action admit.” They may have anticipated that under a race-neutral regime campus life would be easier and that white and Asian students would be less likely to stereotype them as academically weak and more likely to be friends. As we documented in Chapter Six, the perception of stigma is widespread, and the increased likelihood of cross-racial friendships in a race-neutral environment is empirically demonstrable.
Then there is the likely long-term benefit of winning a college degree that will be more valuable because it is untainted by any affirmative action stigma. A Berkeley degree is considered valuable in part because it shows to potential employers and future professional associates that one was smart and able enough to get into Berkeley. But when employers logically assume that one was admitted through a large preference, then that very credentialing effect is muffled. If it is much easier for black or Hispanic students to get into Berkeley, then won’t employers be less impressed by their Berkeley degrees?
There is empirical evidence for this idea too. Three scholars of organizational behavior conducted a series of experiments a few years ago to assess the effect of credentialing. They created and gave to a hundred business school students a series of “pitch books” describing a variety of hypothetical venture-capital investments. Each book was crammed with financial information about the investment and about the key player. The students were asked to estimate the value of the ventures. Controlling for other factors, the interaction of the educational background and the race of each key player affected in a nuanced way the valuation students gave to the venture. Companies led by white key players from elite schools received higher valuations than those led by black key players from elite schools—except that when the pitch book specified that the school in question did not use racial preferences in admission, the black-white valuation difference disappeared. The scholars concluded that their test subjects downgraded the presumed skill of black businessmen not based on their race but instead based on the assumption that their elite school credentials had been won in part by admissions preferences.
It’s true that the subjects in this study were business students rather than actual investors or employers. But that makes the study all the more relevant to the question before us. If business school students have these perceptions, it is very plausible that other students do too and that prospective black students would be aware of them.
It makes every aspect of our warming effect finding fit. Getting into Berkeley is good; getting into Berkeley without any taint of a preference is really good. The size of the warming effect should be, as it is, closely related to the reduction in racial preferences after Prop 209. Preferences fell dramatically at Berkeley and UCLA, and this had particularly impressive warming effects; they fell much less at the less-elite UC campuses.
One other manifestation of the warming effect is worth at least a brief mention: its possible influence on the ambitions of minority high school students. Many thoughtful critics of affirmative action suggest that young blacks with academic promise realize that they are a scarce commodity and that even a reasonably successful high school career will gain them entrée into a fairly elite college. An interesting question, as John McWhorter has put it in describing his own youth, is whether
the maintenance of affirmative action . . . hinders the completion of the very task it was designed to accomplish, because it deprives black students of a basic incentive to reach for that highest bar. . . . I can attest, for example, that in secondary school I quite deliberately refrained from working to my highest potential because I knew that I would be accepted to even top universities without doing so. Almost any black child knows from an early age that there is something called affirmative action which means that black students are admitted to schools under lower standards than white; I was aware of this at least [from] the age of 10.
Might, then, a ban on racial preferences produce some change in the behavior of minority high school students?
A few scholars have studied this question, and some interesting work is still in progress that specifically examines the effect of Prop 209 on minority high school achievement. The evidence thus far suggests to us that, overall, black and Hispanic high school performance (relative to other groups) may have improved in some modest ways but by no means dramatically. There does seem to be one interesting exception at the top of the achievement distribution. The College Board data described earlier allows us to track the proportion of blacks taking the SAT in California whose overall academic index in high school put them at a level comparable to the top eighth of all students—a pool similar to those ordinarily eligible for race-neutral UC admission. From 1994 (when our data begins) through 1997 the proportion of California and US blacks in this elite range tracked exactly in the neighborhood of 2.5 percent of blacks. But in 1998 the proportion of California blacks with high academic indices jumped by 20 percent over the national rate, and an upward trend begins. In 2001, the last year for which we have complete data, California blacks were achieving high academic indices at a rate 35 percent higher than the national average (meaning that 3.5 percent of blacks reached the top eighth of academic indices). This is just one piece of evidence, not “robust” in the sense of the very powerful evidence for the warming effect. But it is at least plausible that Prop 209 has stimulated black high school students aiming for the University of California to raise their sights on their own performance.
The existence of the warming effect calls into question many of the core assumptions of racial preference policies. It suggests that many talented black and Hispanic high school seniors are aware of the potential side effects of large preferences—enough so that a great many talented high school seniors would prefer to avoid them, other things being equal. That doesn’t mean that a black student will choose the local community college over Princeton simply to avoid a preference, but it certainly implies that black and Hispanic students would like to have choices among elite colleges that use smaller preferences or none at all.
