The Early Effects of Prop 209: Background Statistics and Trends

            California’s ban on the use of racial preferences in government problems is an ideological flashpoint for the national affirmative action debate; its symbolism has often led to wild distortions about its actual effects.  This background paper provides some hard data on actual changes over time. 

Table 1:  Trends in Freshman Enrollment at the University of California, 1989-2013

  Total UC freshman enrollment from California high schools in:
  1989 1997 1998 2006 2013
African Am 1,177 917 739 1,072 1,333
Hispanic 2,991 3,131 2,948 5,481 9,322
Asian Am 4,275 6,909 6,979 11,334 9,973
“unknown” 784 774 3,441 1,496 1,047
White 10,314 9,451 8,257 10,687 7,933
Total 20,534 23,682 24,877 33,540 33,135

Source: UC Application, Admission and Enrollment of California Residential Freshmen for Fall 1989 through 2013 (UCOP, 1/14).  Columns do not sum to total because some smaller racial categories are omitted.

It is often claimed, for example, that the number of underrepresented minorities (a term of art that generally includes African-Americans, American Indians, and Hispanics, and is abbreviated “URM”)  at UC was irretrievably damaged by Prop 209, which went into effect in 1998.  Table 1 shows otherwise.  Note, first, that during the eight years prior to Prop 209 (1989-1997), black UC enrollment was actually declining, and Hispanic enrollment was flat in absolute terms and declining in percentage terms.  There were two reasons for this.  On the one hand, UC administrators recognized that in the 1980s, they had used excessively large racial preferences; “special admissions” had been used to admit black and Hispanic students who had terrible prospects of academic success.  UCLA, which had the highest black freshman enrollment among UC campuses during much of this period, had extremely low URM graduation rates:  black four-year graduation rates, for example, averaged about 13% in the early 1990s.[1]  Administrators were thus steadily tightening their use of racial preferences even before Prop 209 came along.  On the other hand, however, UC during this period was not doing much to improve the pipeline for minority students; there was little engagement between UC schools and California’s high schools.

            The first year of implementation of Prop 209 (in 1998) did produce a further drop in URM enrollment – about a 20% decline in black enrollment and a 6% decline in Hispanic enrollment.  But it also catalyzed a series of important changes in the way UC did business.

            First, in 1997-98 the UC system undertook a major examination of the ways the university could improve outreach and improve the outcomes of disadvantaged students in California’s K-12 system, especially its high schools.  Over the next fifteen years, UC spent on the order of half a billion dollars on these efforts[2] – a huge investment from a university system that receives only about $2.4 billion per year in state funding.[3]  Many of these efforts have not been rigorously evaluated, but some of them clearly had important effects.  To give one example, high school completion rates for California students in general, and URMs in particular, started to rise sharply soon after UC’s efforts began:

  Table 2.  High School Completion Rates for California 20-year-olds in 2000 and 2009, by Race

Racial group Proportion of 20-year olds in California who had completed high school in:
2000 2007-11 (average)
African-American 74% 84%
Hispanic 58% 78%
Asian 91% 96%
All groups 76% 87%

Source: 2000 PUMS data and 2007-11 American Community Survey.

            In the fifteen years after Prop 209, black and Hispanic students not only completed high school at higher rates, but they were more likely to successfully complete the “A-G” courses required for admission to the UCs, and to actually apply for admission.  This helps to explain the phenomenal increase in URM applications to UC schools in the years after Prop 209:

Table 3.  UC Application Volume Before and After Prop 209

  Unique applications from California high school students to UC in:
  1989 1997 1998 2006 2013
African Am 2,191 2,141 2,151 3,307 5,978
Hispanic 5,273 6,933 7,285 13,656 31,908
Asian Am 8,165 12,367 12,205 18,742 22,180
“unknown” 1,567 1,738 7,556 3,263 2,979
White 22,307 20,870 18,149 24,926 26,881
Total 41,460 49,030 52,301 71,011 99,447

Source:  UC Application, Admission and Enrollment of California Residential Freshmen for Fall 1989 through 2013 (UCOP, 1/14)

            As Table 3 makes clear, black applicant volume to UC schools was flat in the eight years before Prop 209, and Hispanic applicant volume rose very modestly; after Prop 209, applications from both groups increased at an accelerating rate – far greater than the overall rate of growth in applications generally, and greater than the rate of growth of the underlying URM high school population.

            The second fundamental change prompted by Prop 209 was a significant improvement in URM performance and outcomes – in particular, student grades, persistence in STEM fields, and four-year graduation rates.  This happened partly because of a reduction in the “mismatch” phenomenon.[4]  Students who attend schools where their credentials are far below those of their classmates are likely to experience a variety of academic problems.  At the UCs in the 1980s and early 1990s – particularly at Berkeley and UCLA — huge credential gaps (equivalent to hundreds of SAT points) were produced by very aggressive racial preferences.  As UC moderated the size of these preferences (even before Prop 209), mismatch declined and minority graduation rates improved.  This trend was reinforced by Prop 209.  Perhaps an equally important development, however, was that UC administrators began to pay much more attention to the graduation problem after Prop 209.  Recognizing that they could not arbitrarily admit as many URM students as they desired, administrators instituted a variety of changes to improve the success odds of students they did admit.   As a result, the number of “underrepresented minorities” who entered as freshmen and graduated within four years increased by about 60% between the 1995-97 cohort and the 2001-03 cohort.  The number of URMs in STEM fields who graduated in four years rose by nearly 90%.  (See Table –)  Almost certainly, these graduation gains are even larger for the 2004-06 cohort – but over the past decade, the university has made it so much harder to obtain data that we do not have comparable data for that cohort.

Table 4.  Underrepresented minorities (“URMs”) entering and graduating from the UC system, 1992-2003

Cohort Matriculants 4-Year B.A.s Total B.A.s 4Y Grad Rate Tot Grad Rate
1992-94 12,228 2,630 8,286 21.5% 67.8%
1995-97 13,291 4,155 9,572 31.3% 72.0%
1998-2000 12,577 4,597 9,413 36.6% 74.8%
2001-03 15,793 6,841 11,830* 43.3% 74.9%*

Table 5.  Underrepresented minorities in STEM entering and graduating from the UC system, 1992-2003

Cohort Matriculants 4-Year B.A.s Total B.A.s 4Y Grad Rate Tot Grad Rate
1892-94 2,299 379 1,566 16.5% 68.1%
1995-97 2,409 520 1,689 21.6% 70.1%
1998-2000 2,445 696 1,744 28.5% 71.3%
2001-03 3,142 953 2,100* 30.3% 66.8%*

*Note:  2001-03 counts of total bachelor’s degrees appear to be understated because the data was produced in 2010.  Source for Tables 4 and 5:  analysis of UC student database disclosed in 2008-10.

            A third fundamental change prompted by Prop 209 was a shift in admissions methods.  As labor economists Kate Antonovics (UCSD) and Ben Backes[5] have shown, the various UC colleges made changes in their admissions procedures immediately after Prop 209 that softened the effect of race-neutrality on URMs.  For example, several campuses decreased the weight given to scores on the SAT I in math (where black-white gaps, for example, were particularly large) and increased the weight given to high school grades.  They also generally increased the weight given to socioeconomic factors, such as parental educational attainment and income.  Antonovics and Bakke estimate that the total effect of these changes significantly offset the impact of race neutrality on most UC campuses.  For example, they find that, for the first three admissions cycles of “race-neutrality” (1997-98 through 1999-2000), changes in the weight given to non-racial admissions factors raised the rate at which URMs were admitted to UCLA by 6 points relative to whites (e.g., from 14% to 20%);[6] for the next three admissions cycles (2000-01 through 2002-03) these changes raised the rate at which URMs were admitted by 11 points relative to whites.[7]

Table 6.  Illustrative Effects of Campus Admission Changes on
the Relative Rate of URM Admissions

Campus Percentage point increase in URM admission rate relative to whites, due to race-neutral changes in admissions policies, in:
  1998-2000 2001-03 2004-06
Berkeley 5 9 7
UCLA 6 11 11
UCSD 5 9 10

Source:  Antonovics & Backes (2014)

            These various changes not only cushioned the short-term loss of URMs in the wake of Prop 209; they also had the effect of expanding socioeconomic diversity within the university.  Taking personal disadvantage into account, rather than race, nicely complemented the university’s expanded outreach efforts at struggling high schools.  It also meant that over most of the last decade, UC schools have easily led the nation’s most selective schools in the proportion of students receiving federal Pell grants (a widely-used proxy for assessing a school’s socioeconomic diversity).  For example, more than 35% of UCLA’s students typically receive Pell grants, compared to under 15% at many elite public universities and Ivy League schools.[8]

            The fourth change prompted by Prop 209 (and the final one we shall discuss) was a redistribution of URMs from the two most elite UC campuses – Berkeley and UCLA – to the other six (later seven) UC undergraduate campuses.  This shift is illustrated in Table 5, below. 

Table 7.  California URM enrollment at the “most elite” and other UC colleges, 1989-2013

  Total UC freshman enrollment from California high schools at Berkeley and UCLA in:
  1989 1997 1998 2006 2013
African Am 606 453 260 243 318
Hispanic 1,353 1,034 700 1,122 1,587
  Total UC freshman enrollment from California high schools at all other UC campuses:
  1989 1997 1998 2006 2013
African Am 571 464 479 829 1,015
Hispanic 1,638 2,097 2,248 4,359 7,735

Source:  See Table 1.

As one can see, there was a large drop in both African-American and Hispanic entry into Berkeley and UCLA when Prop 209 was implemented and black enrollment has never fully recovered to its pre-209 levels.  For critics of Prop 209 who rely on real, as opposed to imaginary, statistics, this has been the one they most focus upon.  Again, however, some context is enormously helpful.  Note, first, that both black and Hispanic enrollment at UCLA and Berkeley were declining steadily in the eight years before Prop 209.  This is because these two schools were using by far the most aggressive racial preferences in the 1980s, and therefore were those who tried to improve the outcomes of their minority enrollees by reducing preferences during the early and mid-1990s.  Second, note that, even in 1997 (when URM enrollment at Berkeley and UCLA had been falling in relative or absolute terms for several years), half of all African-Americans attending UC, and a third of Hispanics, were enrolled at Berkeley and UCLA, compared to just over a quarter of all other students.  In other words, URMs were considerably overrepresented at the most elite campuses prior to Prop 209, relative to other groups’ distribution across the UC campuses.[9]

            By ending or at least dramatically reducing racial preferences, Prop 209 thus inevitably had its largest effects at Berkeley and UCLA, which had been using the most aggressive racial preferences.  The impact on the other campuses was fully offset by the redistribution of some URM students who would, without Prop 209, have attended UCLA or Berkeley, and because of the various admissions countermeasures discussed above.   URM enrollment at these other campuses rose immediately after Prop 209, and at an accelerating rate as outreach and K-12 improvement programs took hold.  Blacks and Hispanics (as measured by the index of dissimilarity[10]) were not “segregated” at less elite UC campuses after Prop 209.  Indeed, black and Hispanic freshmen were decidedly more evenly integrated across campuses after Prop 209 than before.

            In sum, much of the conventional wisdom about Prop 209 is inaccurate.  UC administrators, especially at Berkeley and UCLA, had recognized well before Prop 209 that they were using excessively large racial preferences with often-disastrous consequences for student academic outcomes.  Prop 209 led to major institutional changes that improved the pipeline of disadvantaged students and racial minorities to college in California, increased the socioeconomic diversity of UC students, substantially improved outcomes for URM students at UC campuses (especially in the sciences and engineering), and improved racial integration across UC campuses. 

[1] The university’s office of the President (“UCOP”) for many years posted these and a good deal of other data online, in a program called UC Statfinder.  Around the time scholars started publishing research showing the positive effects of Prop 209, however, the university (perhaps coincidentally) took Statfinder off the web, and there is no comparable public source available for detailed information on student outcomes.  I have posted some graduation data at [  ]; see also Sander & Taylor, Mismatch, Chapter 8 (2012).

[2] In an April 2014 op-ed largely devoted to criticizing Prop 209, UC President Janet Napolitano referenced “more than half a billion dollars” spent by the university on programs and policies to achieve race-blind diversity.   Napolitano may have been including some financial aid expenditures in her analysis, but UC budget documents show that some $240 million was spent on the new outreach programs just in the 1998-2002 period, when the efforts were most intense.  See; Karl Pister, UC Outreach: Systemwide Perspective and Strategic Plan (Sept. 1998 report to Board of Regents).

[3] According to a university website,, UC received $2.374 billion in state funding in 2011-12.

[4] By far the most definitive analyses of Prop 209’s short-term effects on graduation rates, and the operation of mismatch upon graduation and STEM-major outcomes, are two papers from a team of Duke scholars: Peter Arcidiacono, Esteban Aucejo, Patrick Coate, and V. Joseph Hotz, “Affirmative Action and University Fit: Evidence from Proposition 209,” (May 2014 draft) and Arcidiacono, Aucejo, and Hotz, “University Differences in the Graduation of Minorities from STEM Fields: Evidence from California,” (May 2014), selected for a revise-and-resubmit by the American Economic Review.  Both papers are available at

[5] Kate Antonovics and Ben Backes, “The Effect of Banning Affirmative Action on College Admissions Policies and Student Quality,” 49 Journal of Human Resources (Spring 2014).  Antonovics and Backes used extensive data on applicants and admissions decisions for eight UC campuses from 1995 through 2006, disclosed by UCOP in response to a public records request submitted by Sander and several other economists in 2007.  The authors used the data on applicant characteristics and admissions outcomes to infer the weight given to various factors in admissions.

[6] That is, Antonovics and Backes are referring to absolute increases in admissions rates.

[7]These changes helped Asian students, too, though to a much smaller degree than they helped URMs.  It appears Asians benefited

[8] The most recent US News ranking, using 2011-12 data, reported that 38% of UCLA undergraduates received Pell Grants; the median rate for the “top 25” research universities was 15%.  See

[9] See data source for Table 1.

[10] The index of dissimilarity, the principal measure of segregation used in studies of school and housing segregation, is calculated by determining the proportion of Population A (e.g., UC African-Americans) at each location (here, the 8 UC undergraduate campuses in 1997), calculating the same proportions for Population B (e.g., all non-black students), taking the absolute difference between each campus level proportion of A and B, summing the results and dividing by 2.  A “0” on the resulting scale means that A and B have matching proportions at all campuses; a “1” means that the campuses are completely segregated by race.  The cross-campus index of dissimilarity has generally been around .2 for both blacks and Hispanics; both numbers fell in 1998 and afterwards.