When Lawrence H. Summers, president of Harvard University, recently announced a plan to deal with the barriers to economic diversity at Harvard, he put the spotlight on higher education's dirty little secret: While colleges have admirably focused on creating racially integrated student bodies, they haven't given comparable help to economically disadvantaged students.

Summers says he plans to expand financial aid for low-income students and to actively recruit them in what could become the new frontier of affirmative action. The announcement comes on the heels of Texas A&M University's decision to eliminate legacy preferences -- a practice that Sen. John Edwards rightly says is "a birthright out of 18th-century British aristocracy, not 21st-century American democracy."

To date, selective colleges have failed to make economic diversity a priority, a point documented by a study included in a new book that I edited, America's Untapped Resource: Low-Income Students in Higher Education. Conducted by Anthony P. Carnevale, vice president for assessments, equity, and careers at the Educational Testing Service, and Stephen J. Rose, a senior research economist at ORC Macro, a research, consulting, and survey company, the study found that the underrepresentation of low-income students at elite colleges is many times greater than it is for underrepresented minority students who are the subject of the wars over affirmative action.

Using data from the top 146 colleges, representing the most selective 10 percent of four-year colleges as defined by Barron's Profiles of American Colleges, the researchers found that students in the highest economic quartile take up 74 percent of the available slots, while students in the bottom quartile fill just 3 percent. Economically disadvantaged students are 25 times less likely to be found on elite college campuses than economically advantaged students -- and yet this phenomenon receives none of the attention or moral outrage associated with efforts to curtail racial preferences.

Most colleges say they already provide a leg up to low-income applicants and would admit more as students if they could handle the work. In their amicus brief in the University of Michigan affirmative-action cases, Harvard, Princeton, and six other elite institutions said they "already give significant favorable consideration" to socioeconomic status. William G. Bowen and Derek Bok, former presidents of Princeton and Harvard, respectively, argue in their seminal book, The Shape of the River: Long-Term Consequences of Considering Race in College and University Admissions (Princeton University Press, 1998), that it is not "realistic" to admit more disadvantaged students. "The problem is not that poor but qualified candidates go undiscovered, but that there are simply too few of these candidates in the first place," they say.

But Carnevale and Rose's research has revealed otherwise. If low-income students routinely received a break in admissions, as many colleges suggest, we would expect to see them overrepresented compared with their academic records. Racial preferences, for example, increase the proportion of black and Latino students at the nation's top colleges from 4 percent, under a strict system of admissions based on grades and test scores, to 12 percent, according to Carnevale and Rose. But they found that the representation of low-income and working-class students today is actually slightly lower, not higher, than if grades and test scores were the sole basis for admissions.

The researchers simulated a pool consisting of students who had good grades and test scores above 1300 on the SAT (or the ACT equivalent), plus economically disadvantaged students with grade-point averages of at least 3.0 in core courses, strong teacher recommendations or involvement in extracurricular activities, and test scores between 1000 and 1300. The cutoff was 1000 because students have a good chance of succeeding when they score above that point.

Students were considered disadvantaged if they were in the bottom 40 percent by socioeconomic status -- defined as parents' income, education, and occupation -- or if they attended high schools with a small proportion of graduates going on to four-year colleges or a large proportion of students who are eligible for free or reduced-price lunches. Carnevale and Rose assumed that all students within the pool had an equal chance of admission. The researchers estimated that the preference implied under their model is roughly half the size currently used for race, based on an aggregate of the 146 institutions.

The study demonstrated that if colleges admitted applicants from such a pool, they would have far more economic diversity than they now have -- roughly 38 percent of the students admitted would come from the bottom economic half of the pool compared with 10 percent today. Moreover, if colleges also eliminated other preferences (legacy, athletics, and the like), overall academic quality would not suffer. Indeed, the researchers found that graduation rates would actually climb, from 86 percent today to almost 90 percent.

One reason to be confident in this simulation is that economic affirmative action is not meant to be a challenge to merit, but rather a better approximation of it. A 3.6 grade-point average and SAT score of 1200 surely means something more for a low-income first-generation college applicant who attended terrible schools than for a student whose parents have graduate degrees and pay for the finest private schooling. There is also some evidence that a student who has done well despite having to overcome serious obstacles is likely to have greater long-run potential. One study of Harvard students in the 1950s, '60s, and '70s found that blue-collar students with more modest SAT scores were the most successful adults, as measured by income, personal satisfaction, and community involvement.

Now that the Supreme Court has put race-sensitive admissions on solid legal footing -- at least for the next 25 years -- one can hope that colleges will take on the larger issue of economic diversity, as it is clear that racial preferences alone are inadequate.

In the University of Michigan undergraduate case, Gratz v. Bollinger, Justice Ruth Bader Ginsburg, joined by Justices David H. Souter and Stephen Breyer, supported affirmative action with data finding that African-American and Hispanic students have higher poverty rates than white students (22.1 percent and 21.2 percent compared with 7.5 percent), and that black and Latino students "are all too often educated in poverty-stricken and underperforming institutions."

But in practice, affirmative action in higher education does little to reach those low-income and working-class minority students. Bowen and Bok, strong supporters of affirmative action, found that 86 percent of black students who enrolled in the 28 selective universities they studied were middle class or upper-middle class. And race-based programs do nothing to help students from white and Asian families living in poverty or attending "poverty-stricken and underperforming institutions."

Some higher-education leaders have argued that wealthy minority students aren't admitted to elite colleges as a matter of fairness or reparations but because they add to the student body. The whole concept of "deserving" or "earning" a spot is considered naïve by many members of academe. Students are admitted because they fit the needs of the college and the society at a particular point in time, not because there is anything intrinsically worthy about them. But for many Americans, students are admitted to elite colleges not just to offer a diversity of viewpoints that make for an interesting class discussion for other students and professors for four years. Admission to college also provides an opportunity to enjoy greater professional success for the following 40 years -- so fairness is important.

Moreover, if colleges are truly concerned about diversity, attention to race alone is insufficient. If diversity is defined broadly, to value differences in both economic and racial backgrounds -- so that it values kids from trailer homes, ghettoes, and barrios, not just those minority students from affluent suburbs -- economic affirmative action would add a great deal that race-based affirmative action alone does not provide.

The Supreme Court's majority opinion in Grutter v. Bollinger does leave the door open for an eventual shift to economic affirmative-action programs. The court declares that colleges must engage in "periodic reviews to determine whether racial preferences are still necessary to achieve student body diversity." The opinion then points to colleges in California, Washington, and Florida that are "engaged in experimenting with a wide variety of alternative approaches" -- of which economic affirmative action is an important component. The University of California uses "comprehensive review," thereby examining academic accomplishments in light of such obstacles as "low family income, first generation to attend college," and "disadvantaged social or educational environment." The University of Washington looks at academic achievement in the context of such factors as family income, the number of children in a family, the parents' educational level, and the percentage of high-school students who receive free lunch. And as part of the State University System of Florida's "Profile Assessment" program, the "family educational background, socioeconomic status, special talents, or the high school or geographic location of the applicant" will be considered.

Of course, the Supreme Court decision in Grutter will not end the debate over race-based affirmative action, but merely move it from the legal to the political sphere. Already, Ward Connerly, the author of successful anti-affirmative-action initiatives in California and Washington, has vowed to take his fight to Michigan and elsewhere. That should give concern to the supporters of affirmative-action policies who have focused only on racial preferences. A wide body of public-opinion research over the past several years has found that public support for racial affirmative action in higher education is thin, whereas support for consideration of economic obstacles is wide and solid.

Three polls from January and February 2003 are illustrative. A Los Angeles Times survey found that 55 percent of Americans agreed with President Bush's opposition to the University of Michigan's racial-preference plan compared with just 27 who disagreed. (Even Democrats narrowly supported Bush.) A Newsweek poll revealed that 68 percent of Americans opposed preferences for black students, while only 26 percent supported them. And a survey by a Michigan-based polling company, EPIC/MRA, showed that 63 percent of Michigan residents opposed the University of Michigan's affirmative-action plan, compared with only 27 percent who supported it.

By contrast, Americans in the same polls supported preferences for low-income or economically disadvantaged students of all races by 65 percent to 28 percent (Newsweek), 60 percent to 31 percent (Los Angeles Times), and 57 percent to 36 percent (EPIC/MRA survey of Michigan residents alone). Internationally, many other countries pursue such economic affirmative-action policies over racial affirmative action.

What will happen to racial diversity if Connerly and colleagues are successful, and colleges turn to economic affirmative action as an alternative? Carnevale and Rose found in their simulations that, at the top higher-education institutions, a race-blind economic affirmative-action program -- based on parental income, education, and occupation, and the high school's socioeconomic status -- would boost African-American and Latino admissions from only around 4 percent to 10 percent, down from the current 12 percent.

The current level of racial diversity would not be sustained, in part because data show that, on average, even middle-class African-American students lag in achievement. The researchers point out that black and Latino students are more likely to attend elementary and secondary schools where poverty is concentrated than are white students of similar income.

Those racial patterns reflect differing concentrations of neighborhood poverty, which, in turn, often reflect housing discrimination. One recent study found that black families with incomes in excess of $60,000 live in neighborhoods with higher poverty rates than white families earning less than $30,000. Other research has revealed that middle-class black people are generally newer arrivals to middle-income status than middle-class white people, and that even among white and black people of similar income, black people have fewer financial assets. While the median income of black people is about 62 percent of the median income of white people, the median net worth of black people is just 12 percent of the median net worth of white people, according to Edward N. Wolff's Top Heavy: The Increasing Inequality of Wealth in America and What Can Be Done About It (New Press, 2002). Whereas income reflects a snapshot in time, wealth measures the accumulation of income, or debt, over generations and signals the legacy of slavery and segregation.

But economic affirmative action could capture both forms of discrimination indirectly. In defining obstacles, a college should also consider such factors as neighborhood economic status and net worth. Counting those additional factors should boost racial diversity beyond the 10-percent figure that the more limited definition of economically disadvantaged yields. For example, in the fall 2002 class of the University of California at Los Angeles School of Law, African-American students were 16 times as likely to be admitted under a program counting wealth and single-parent-family status alongside other traditional socioeconomic factors as they were under other programs, and Latino students were 6.8 times as likely to be admitted.

In a sense, the debate over economic affirmative action at selective institutions reflects in miniature the larger debate about equity in higher education. Will we ever squarely face up to the fundamental inequalities rooted in economic class? The Congressional reauthorization of the Higher Education Act, and the continuing debates in state capitals and within individual colleges, will help answer that enduring question.

What is required is a fundamental change in the culture of higher education: Economic diversity needs to become a measure of a college's quality, just as racial diversity is today. Larry Summers may have begun that process, potentially doing for economic inclusion what his predecessor Derek Bok did for racial inclusion.

Whether Summers ultimately fulfills this role remains to be seen. Some are skeptical of Harvard's leadership on this issue, given its poor history on economic diversity. Just 6.8 percent of Harvard students were eligible for Pell Grants in 2001-2, compared with 32.4 percent at the University of California at Berkeley and 15.8 percent at Amherst College. And although Summers announced that parents in families who earn less than $40,000 a year would no longer be asked to pay anything toward their children's education, his plan didn't commit to any particular goal -- say, tripling the number of Pell students. It just said Harvard "is intensifying" outreach efforts and is "re-emphasizing" a policy of taking special note of applicants who have overcome obstacles. But Summers appears deeply committed to this issue and identified the growing gap between rich and poor children as "the most serious domestic problem in the United States today." He is likely to have some healthy competition for leadership on economic diversity. Amherst's new president, Anthony W. Marx, for example, is beginning discussions with his board of trustees about increasing his college's representation of low-income and working-class students beyond its current numbers.

To make economic diversity a reality at other colleges that lack Harvard's or Amherst's financial resources, however, Congress will need to step in to the debate. It could begin by restoring the purchasing power of the Pell Grant to its mid-1970s level. Doing so would cost about $12-billion, according to data in America's Untapped Resource, but that is less than the annual cost of just the dividends and capital-gains component of the recent tax cut. In addition, Sen. Edward M. Kennedy's proposal to require universities to disclose their use of preferences for legacies could help reduce the prominence of today's biggest affirmative-action program for the well-to-do.

The debate on these questions is long overdue. Talented poor and working-class students represent America's untapped resource. They must be allowed to realize their full potential -- for their sake, and for ours.

Richard D. Kahlenberg is a senior fellow at the Century Foundation. He is the editor of America's Untapped Resource: Low-Income Students in Higher Education (Century Foundation Press, 2004), from which this essay is adapted, and the author of The Remedy: Class, Race, and Affirmative Action (Basic Books, 1996).