Higher Education

NBER Reporter: Summer 2000

Higher Education

The NBER's Working Group on Higher Education, directed by Charles T. Clotfelter of Duke University, met in Cambridge on May 5. The program was:

Jeffrey Groen and Michelle J. White, University of Michigan, "In-State versus Out-of-State Students: The Divergence of Interest between Public Universities and State Governments"

Discussant: Caroline Minter Hoxby, NBER and Harvard University

Ronald G. Ehrenberg, NBER and Cornell University, and John J. Cheslock and Julia Epifantseva, Cornell University, "Paying Our Presidents: What Do Trustees Value?"

Discussant: Michael Rothschild, NBER and Princeton University

Bruce I. Sacerdote, NBER and Dartmouth College, "Peer Effects with Random Assignment: Results for Dartmouth Roommates" (NBER Working Paper No. 7469)

Discussant: David Zimmerman, Williams College

Larry D. Singell, Jr. and Joe A. Stone, University of Oregon, "The Good, the Poor, or the Wealthy: Who Benefits Most from College Financial Aid?"

Discussant: John Siegfried, Vanderbilt University

A. Abigail Payne, University of Illinois, "Does the Barrel Contain Fat or Pork? The Effect of Earmarking on Research Funding and Productivity at Universities"

Discussant: Irwin Feller, Pennsylvania State University

Jill M. Constantine, Williams College, and Laura Perna, University of Maryland, "Why Do So Many More Black Women than Black Men Enroll in College?"

Discussant: Sarah Turner, University of Virginia

Groen and White examine the divergence of interest between universities and state governments concerning standards for admitting in-state versus out-of-state students. Considering both public and private universities, the authors find that both favor in-state students in admissions. They also find that states gain more in expected future tax revenues when marginal in-state students, rather than marginal out-of-state students, are admitted to public universities. That is because in-state students' higher probability of locating in the state after attending a university there offsets their lower future state tax payments. Finally, the authors investigate whether states gain when very high ability students attend public universities. They find instead that states are better off when public universities are not highly selective, and when public universities restrict admission of out-of-state students, even those of high ability.

Ehrenberg, Cheslock, and Epifantseva use information on the changes in compensation and turnover of presidents of private universities and colleges to try to infer the factors that trustees of academic institutions value. Data on presidents' salaries and benefits have been published in the Chronicle of Higher Education for approximately 400 private institutions for academic years 1992-3 through 1996-7. The authors merge these data with information about the characteristics of the presidents, such as age and experience, obtained from surveys conducted by the American Council on Education and from the presidents' entries published in Who's Who in America. Data on various performance measures of the institution -- such as endowment growth, research funding growth, student test score changes, and changes in published rankings obtained from a variety of sources -- are also taken into account. The authors' models are very successful at predicting presidents' compensation levels, but the preliminary findings do not strongly support the notion that changes in presidents' compensation are related to observed measures of institutional performance.

Sacerdote measures peer effects among college-age roommates. He finds that among Dartmouth College freshmen roommates and dormmates, who are randomly assigned, peer effects help to determine levels of academic effort and such social decisions as whether to join a fraternity. In other major life decisions, such as the choice of college major, residential peer effects are markedly absent. Peer effects on GPA occur at the individual room level, whereas peer effects on fraternity membership occur at both the room and dorm level, Sacerdote finds. Also, freshmen with high social ability are likely to remain with their roommates in their sophomore year, but freshmen with high academic ability are less likely to keep their roommates.

Both college- and government-sponsored financial-aid programs for students now focus more on merit than on need. However, there is little empirical evidence on the distributional effects of merit-based aid -- that is, who benefits most. Singell and Jones use data for a large public university over several years and show that merit-based aid increases the likelihood of enrollment for all students. Well-to-do students benefit disproportionately, though, even with merit held constant. These findings also are supported by a natural experiment that substantially increased merit-based aid. Thus, the increased emphasis on merit in college financial aid may exacerbate the trend toward greater income inequality in the United States, even among students of equal merit.

Do universities view earmarking of federal funds as merely a result of pork-barrel politics or as another means of obtaining federal funding which may be less expensive than seeking it through the peer-review method of allocation? To analyze this issue, Payne creates a dataset that combines measures on earmarked allocations to universities, total federal research funding distributed to universities, the number of articles published and citations to articles published, and the relationship between membership on the congressional appropriations committees and universities. She finds first that earmarked funding goes primarily to the research and doctoral universities that already receive a large portion of total federal research funding. However, earmarked funding is done through agencies that are not as heavily engaged in distributing peer-reviewed funding as other agencies. Second, the link between a member on the appropriations committees in Congress and affiliation with the university receiving the earmarked funding is not as strong as may be believed. After Payne controls for differences across universities, the effect of representation in Congress on predicting whether a university receives earmarked funding diminishes. Third, the effect of earmarked funding on articles published and citations per article published is unclear. However, the effect of total federal research funding on citations per article published is lower for those universities that received several years of earmarked funding than for those with no or few years of earmarked funding.

Why do African Americans, who share similar socioeconomic backgrounds, have such different college enrollment rates by gender? For over 20 years, black women have had much higher college enrollment rates and degree attainment rates than black men. The gender gap is partially explained by different high school graduation rates for black men and women. However, the gap in college enrollments is much larger than the gap in high school graduation rates. To answer the question, Constantine and Perna use the National Educational Longitudinal Study of 1988 eighth graders, which includes data on schooling and work histories of respondents through 1994, and controls for characteristics that are likely to affect high school persistence and college enrollment, such as test scores in eighth and tenth grade. In addition to test scores and the like, though, the authors take into account measures of "social capital," that is, high school and community characteristics that influence persistence through high school and college enrollment. Their preliminary results indicate that black women are more likely than black men to enroll in a two- or four-year institution upon graduating from high school. The authors break this difference in probability of enrollment into the portion explained by characteristics (other than simply being female) and the portion explained by the effect of those characteristics on an individual's probability of enrollment. All of the difference in four-year college enrollment can be explained by differences in characteristics that lead to higher probabilities of enrollment, the authors find. However, the difference in two-year college enrollment cannot be explained by differences in characteristics.

NBER Videos

National Bureau of Economic Research, 1050 Massachusetts Ave., Cambridge, MA 02138; 617-868-3900; email:

Contact Us