Elsevier

Labour Economics

Volume 15, Issue 5, October 2008, Pages 1023-1039
Labour Economics

Subsidizing Enjoyable Education

https://doi.org/10.1016/j.labeco.2007.09.002Get rights and content

Abstract

College education is not only an investment; for many people it also generates consumption benefits. If these benefits are normal goods, then the rich attend college at higher rates than the poor. Furthermore, the marginal poor student is smarter than the marginal rich student. Colleges aiming to attract smart students may therefore charge lower tuition to poorer students, even when the colleges lack market power. Moreover, when the social return to education exceeds the private return, allocative efficiency requires government grants to students to be means-tested.

Introduction

Much literature on higher education concerns the empirical pattern that the poor invest less in education than do the rich. At a National Press Club event in the United States, a former College Board official claimed that “The fact is, the dumbest rich kids have as good a chance of going to college as the smartest poor kids".1 For statistical data, consider students with test scores in the top third of the class of 1992. Only 68 percent of the low-income, high-test-score youth went on to a four-year college within twenty months of high school graduation, compared with 84 percent for youth with the same test scores from high-income families (Ellwood and Kane, 2000). The same pattern holds for the middle and bottom test-score groups.2

A common explanation for low college attendance by the poor is capital market imperfections–the poor cannot borrow to finance education.3 Some empirical work (see, for example, Checchi, 2003) finds that credit constraints do limit education achievement among the poor; but Checchi (2003) also finds that participation in education, particularly among women, also increases with family wealth independently of financial constraints. Other evidence casts doubt on the importance of credit constraints. For example, Keane and Wolpin (2001) estimate that borrowing constraints little affect school attendance decisions, but that transfers from parents to children do. Cameron and Heckman (1998), Shea (2000), and Cameron and Taber (2004) also find little evidence for credit constraints affecting schooling.

Here we extend Keane (2002) to examine an additional explanation for low college attendance by the poor which can complement borrowing constraints. Many people find the time they spent at school, and particularly at college, as some of the happiest years of their lives. Some of the pleasure undoubtedly has to do with youth. But some comes from the environment–attractive members of the opposite sex, with many opportunities for meeting them; the opportunity to consume services that appeal to youth (such as concerts, movies, plays, football games, and athletic facilities that appeal to twenty-year olds); the beauty of the physical surroundings on many campuses; and so on. In short, attending college is not only an investment, but also a consumption good. If the consumption goods available on campus are normal goods, then the rich are more willing to pay for them than are the poor, and so the rich will attend college at higher rates than the poor. Moreover, this consumption benefit may cause some high-ability persons from poor families to skip college, while low-ability people from rich families do attend college.

The idea that education is not merely an investment but also provides consumption benefits is widely acknowledged. Some empirical evidence shows a consumption value of higher education. Lazear (1977), using data on young males in the United States, finds that individuals with much education (M.A.'s and Ph.D's), pursue education beyond the level that maximizes the present value of future income, suggesting that education has consumption value. The reverse holds for lower levels of education. Heckman et al. (1999), using data on male earnings in the United States, find that individuals in the second-highest ability quartile enjoy large nonpecuniary benefits from attending college; individuals in the other quartiles suffer non-pecuniary costs.4 Using a larger dataset, Carneiro et al. (2003) estimate that, when ignoring psychic gains, fourty percent of college attendees would regret it. Once they account for psychic benefits and costs of attending college, only 8 percent of college graduates regret attending college. The authors conclude, therefore, that much of the gain from college is nonpecuniary.5 Using Dutch data, Oosterbeek and van Ophem (2000) find evidence that schooling is a good that raises future income and generates utility. Alstadsæter (2004) provides similar evidence for Norway.

We shall apply these ideas to address a puzzling phenomenon in higher education–subsidies to the poor. College tuition and government grants to students are commonly means-tested.6 The literature offers three main arguments for such means-testing: capital market imperfections, redistribution, and price discrimination by monopolistic colleges. Although all these arguments are appealing, none is fully satisfactory. First, as argued above, little empirical evidence shows capital market imperfections. Moreover, means-tested tuition fees or grants are inefficient ways to correct capital constraints. Lending to students (with repayment conditional on future income) is the efficient and more equitable way to deal with missing capital or insurance markets (see Jacobs and van Wijnbergen (2007)). Second, though optimal redistribution may require means-tested grants (Dur et al., 2004), the redistribution argument cannot explain why private colleges in a competitive education market charge different tuition to students with different incomes. Third, exploitation of monopolistic power by colleges is unlikely the full explanation for why tuition varies with income. As Epple et al. (2006, p. 889) note, “The stylized fact that colleges can extract so much revenue from higher income households is clearly an empirical puzzle given many colleges competing for students. More future research is needed to find other compelling explanations for this puzzle.”

We provide a new rationale for means testing of college tuition and of government grants to students. Recall that our model implies that a rich person with low ability may be willing to pay for college, while a poor person with high ability would not. So colleges aiming to attract smarter students may charge poor students a lower price than rich students. Moreover, when the social return to education exceeds the private return, allocative efficiency requires government grants to students to be means-tested. As we will see, our argument for means-tested tuition requires that the average ability of enrolled students declines with income. For means-tested government grants, it suffices that the marginal poor student is smarter than the marginal rich student, which arises naturally in our model when education provides a consumption benefit.

Section snippets

Literature

Since Schultz (1960) and Becker (1964) developed the theory of human capital, economists have largely neglected the consumption benefits from education. Exceptions are Alstadsæter (2003) and Malchow-Møller and Skaksen (2004), who study optimal taxation and financing of education when education yields both a pecuniary and a non-pecuniary return. Both papers employ a representative agent framework and so abstract from heterogeneity in ability and in wealth among agents, which are crucial in our

Assumptions

We suppose college students differ in two ways. First, students differ in ability, denoted by a. Second, they differ in initial wealth, w. Each person knows his own ability, but colleges or the government do not; the colleges and government can only observe a student's wealth.7

College attendance

A person with ability a and wealth w attends college ifv[a+p(a)-t+w]+bv(a+w).

Let α(w) denote the ability of a person with wealth w, who, in equilibrium, is indifferent about attending college. The following equation describes, for each level of wealth, the people who attend college:v{α(w)+p[α(w)]-t+w}+b=v[α(w)+w].Since smarter students enjoy a higher return to education, p′(a) > 0, a person with wealth w and with ability a  α(w) attends college.10

Means-tested tuition

We turn to the behavior of colleges. The topic becomes interesting if a college prefers to enroll smart students. Such a preference can arise for many reasons. 1) Peer group effects within colleges can make increased attendance by smart students benefit other students (see Rothschild and White, 1995, Epple and Romano, 1998). 2) Faculty may find it more pleasant or interesting to teach smart students, and so a college may attract better faculty, or attract a given quality of faculty at lower

Government means-tested grants

So far, we ignored externalities from education. Suppose now that, in addition to the private return p(a), education generates a public return λp(a). Of course, only the private return, not the social return, affects an individual's decision to attend college, or affects a college's tuition policy. Subsidies can correct the resulting underinvestment in human capital. We shall see in this section that the consumption benefit from education implies that optimal subsidies are means-tested rather

Generalization: Ability correlated with wealth

We assumed that the distribution of ability is independent of wealth (f (a, w) = f (a) for all w). Clearly, relaxing this assumption may affect our results.

Consider first our result on a college's tuition policy. We saw that when college is enjoyable, poor students are on average smarter than rich students, and so profit-maximizing colleges make tuition increase with student's wealth. When ability is positively correlated with wealth, such discriminatory pricing need not be profitable. Though the

Conclusion

We showed that the consumption benefit from attending college makes rich students, who are most willing to pay for the consumption benefit, especially eager to attend college. Among the poor, only the brightest attend college. Hence, when colleges prefer to enroll smart students, in the market equilibrium tuition will be means-tested. Rich students are nevertheless over-represented in college, and the marginal rich student has lower ability than the marginal poor student. To maximize the social

References (53)

  • BoadwayRobin et al.

    Investment in education and the time inconsistency of redistributive tax policy

    Economica

    (1996)
  • BovenbergA. Lans et al.

    Redistribution and educations subsidies are Siamese twins

    Journal of Public Economics

    (2005)
  • CameronStephen V. et al.

    Life cycle schooling and dynamic selection bias: Models and evidence for five cohorts of American males

    Journal of Political Economy

    (1998)
  • CameronStephen et al.

    Should college attendance be further subsidized to reduce rising wage inequality?

  • CameronStephen V. et al.

    Estimation of educational borrowing constraints using returns to schooling

    Journal of Political Economy

    (2004)
  • CarneiroPedro et al.

    Estimating distributions of treatment effects with an application to the returns to schooling and measurement of the effects of uncertainty on college choice

    International Economic Review

    (2003)
  • ChecchiDaniele

    Inequality in incomes and access to education: A cross-country analysis (1960–95)

    Labour

    (2003)
  • De FrajaGianni

    Reverse discrimination and efficiency in education

    International Economic Review

    (2005)
  • Department of Education and Science
  • DurRobert et al.

    Are education subsidies an efficient redistributive device?

  • DurRobert et al.

    Should higher education subsidies depend on parental income?

    Oxford Review of Economic Policy

    (2004)
  • DynarskiSusan

    Hope for whom? Financial aid for the middle class and its impact on college attendance

    National Tax Journal

    (2000)
  • EllwoodDavid et al.

    Who is getting a college education: Family background and the growing gaps in enrollment

  • EppleDennis et al.

    Competition between private and public schools, vouchers, and peer-group effects

    American Economic Review

    (1998)
  • EppleDennis et al.

    Admission, tuition, and financial aid policies in the market for higher education

    Econometrica

    (2006)
  • FernandezRaquel et al.

    To each according to...? Markets, tournaments, and the matching problem with borrowing constraints

    Review of Economic Studies

    (1999)
  • Cited by (5)

    We are grateful for useful comments and suggestions by the editor, Fabien Postel-Vinay, two anonymous referees, Alessandro Balestrino, Gianni de Fraja, Bas Jacobs, Thijs van Rens, participants at the 2005 CESifo Area Conference on Public Sector Economics in Munich, the 2005 Econometric Society World Congress at University College London, and seminar participants in Rotterdam. Dur gratefully acknowledges financial support from NWO, KNAW, and VSNU through a Vernieuwingsimpuls grant.

    View full text