Harvard Educational Review
  1. Spring 2006 Issue »

    Editor's Review - College Choices edited by Caroline M. Hoxby

    Baoyan Cheng
    College Choices: The Economics of Where to Go, When to Go and How to Pay for It
    Edited by Caroline M. Hoxby
    Chicago: University of Chicago Press, 2004. 435 pp. $52.00.

    For the past few decades, college tuition has increased steadily and dramatically in the United States. At public four-year colleges and universities, average tuition and fees increased by more than 80 percent in real terms between 1976 and 1996. At private colleges, the increase was 90 percent. Meanwhile, median family income increased by 10 percent during the same period (Redd, 1999). In 2001-02, average tuition and fees at public and private four-year colleges were $3,725 and $17,272, respectively, an increase of 38 percent for public schools and 37 percent for private schools over the 1991-92 academic year (College Board, 2002). This increase in tuition continues today: average tuition and fees among public four-year colleges and universities were $5,132 in 2004-05, an increase of 10.5 percent over the previous academic year; for private schools tuition was $20,082, an increase of 6.0 percent over 2003-04 (College Board, 2004).

    As tuition increases more rapidly than family income, the issue of providing equal educational opportunity to all qualified students, especially those from low-income families, has caught increasing attention from policymakers and researchers. As the primary means of removing financial barriers for students, financial aid policies have been at the center of this discussion. Researchers have been working hard to find evidence — both theoretical and empirical — of how individuals and institutions respond to federal and state financial aid policies. College Choices: The Economics of Where to Go, When to Go and How to Pay for It is one researcher’s efforts to present current and rigorous evidence on how financial aid policies affect students’ college-going behavior.

    This book originated from a forum held in the fall of 2000 in Cambridge, Massachusetts. At the forum, researchers and practitioners collaborated to produce relevant and rigorous contributions to the study of financial aid policy. Edited by Caroline M. Hoxby, an experienced scholar of the economics of education, this book is a collection of individual papers by distinguished economists on topics ranging from merit aid, federal tax credits, education savings program, Pell Grants, and aid packages, to the use of nonresident enrollments, differences and similarities in decisionmaking about college between inner-city and urban youth, and peer effects. For each topic, the authors provide a thorough literature review as well as detailed historical and policy contexts, which enable the reader to have a good grasp of the issue and lay a foundation for judging these studies. Further, this book is not simply a collection of individual papers; each chapter is followed by a short comment by another economist that either points out the strengths and weaknesses of the paper, raises different views, or posits overlooked aspects, thus providing a more comprehensive picture of the issues being examined.

    Contribution of the Book

    This book provides a new direction for future research in financial aid policies in American higher education. Although there are nine chapters, they have a common theme: college choice. Before this book, numerous studies examined the effect of financial aid policies; however, the majority of them were about college attendance. As editor Caroline Hoxby points out in the introduction, the effects of financial aid policies lie more in students’ college choices than in attendance decisions, because attendance decisions are much more difficult to influence. In other words, a much smaller group’s attendance decisions would be swayed by financial aid policies than those whose college choice decisions (such as two-year vs. four-year institutions, in-state vs. out-of-state) would be swayed by these policies. Therefore, this book turns over a new leaf for quantitative researchers studying the effect of financial aid policies. Furthermore, it provides a comprehensive perspective on the issue of college choice, examining it from both an individual and institutional point of view, covering various topics related to college choice.

    This book adds to the current discussion by making the following unique contributions. First, it advances current discussions on important policy issues by using new datasets. For example, the chapters by Susan Dynarski and Eric Bettinger add to the discussion of the effect of grant aid on students. Previous studies mainly examine the effect of merit aid on college enrollment, while Dynarski extends the discussion by examining its effect on college choices, as well as its differential effects along racial lines. Her findings, that merit aid shifts students from two-year to four-year colleges and appears to narrow racial and ethnic gaps, have important policy implications. Bettinger’s major contribution lies in the unparalleled dataset used in his study, which has great advantages over datasets used in previous studies, such as enabling researchers to distinguish students who withdraw from those who transfer (an important distinction to make), and having detailed financial data used by institutions to determine applicants’ eligibility for Pell Grants. Further, more than half of the studies in this book use datasets that have not been used before. These improvements in methodology and datasets allow researchers to make more accurate estimates, thus providing more scientific evidence of the effect of financial aid policies.

    It is worth mentioning that the findings regarding the effect of financial aid on enrollment are less than clear cut. This is a reflection of the complicated nature of financial aid policies. It is difficult to disentangle the effect of financial aid on college access and enrollment from other variables, such as family income and parental education. With the exception of the first chapter, which sets the context for the rest of the book, all of the chapters use sophisticated econometric methodology. These methodologies range from differences-in-differences analysis (Long), propensity score approach (Ma), and discontinuity regression analysis (Bettinger) to conditional logit (Avery and Hoxby) and 2SLS (Rizzo and Ehrenberg). These methods are relatively new and allow researchers to make causal inferences in a relatively convincing way. These approaches are an improvement over the traditional approach that regresses a person’s educational attainment against covariates and the aid he or she is eligible for. This is because aid eligibility is correlated with many observed and unobserved factors that affect college attendance and completion. By using an exogenous source of variation in aid, researchers were able to control for these factors, especially the unobservable ones. As a result, this book not only provides scientific evidence concerning relevant policy issues, but also sets a good example for other researchers in terms of doing the most rigorous research possible.

    In addition to elaborating on the context of the financial aid policy being examined, each study describes the dataset(s) they use in great detail, especially those that have not been used before, which is valuable information for researchers pursuing similar topics. Each study also provides a thorough explanation of the methodology it employs. Although these methods sections make College Choices a dense read, they also make the book an excellent text for students in education who are interested in pursuing econometric studies.


    The financial aid system in the United States is complicated. Its major components include programs sponsored by the federal government, such as grants, loans, and work-study, as well as those sponsored by states, private institutions, foundations, and charities. Federal financial aid is an important part of this system. Since World War II, enrollment in higher education in the United States has grown substantially; while only 2.6 million students were enrolled in the late 1940s, by 2003 this number had grown to 18.9 million (National Center for Education Statistics, 2003). Several major financial aid programs contributed to this rapid growth. First, the G.I. Bill, created after World War II, offered assistance to veterans. The Higher Education Act of 1965 further extended assistance to college students by establishing a range of programs, including loans, grants, and work-study (Herzlinger & Jones, 1993; Mumper & Ark, 1991). When signing the act on November 8, 1965, President Lyndon Johnson claimed that it would “swing open a new door for the young people of America — the most important door that will ever open — the door to education” (Johnson, 1965). Finally, in 1972, the U.S. government created the Pell Grant, originally known as the Basic Educational Opportunity Grant (Hansen, 1991). As the federal government’s first major, direct, need-based grants program, it set an example for other federal, state, and institutional aid programs (Hearn, 1998).

    Until the late 1970s, financial aid policy in the United States focused mostly on grants. However, as Ronald Reagan took office as president in 1981 with a mandate to cut domestic spending, financial aid policy began to gravitate toward loan programs (Hearn, 1998; Mumper, 1996).

    Putting the Studies in the Context of Previous Studies

    In the first chapter of College Choices, Sarah Turner presents an overview of college-going behavior in the United States over the past few decades, using data from the National Longitudinal Survey of Youth and Current Population. She finds that while the college enrollment rate has greatly increased, the completion rate has decreased among individuals in their early twenties and remains largely stagnant among somewhat older individuals. At the same time, the premium of a college degree over a high school degree has risen substantially. As a result, it is degree attainment rather than college attendance that ultimately affects the distribution of earnings. Further, Turner draws attention to the phenomenon that there is a considerable increase in the length of time taken to complete a degree. Therefore, college completion deserves more attention than researchers and policymakers currently give it. Turner devotes most of the remainder of her chapter to finding explanations for the decrease in college completion. The possible explanations she identifies are from the perspectives of economics (credit constraint and tuition increase) and psychology (uncertainty). Turner also presents possible explanations from an em-pirical perspective, including changes among students (e.g., decrease in preparedness), shifts among different types of institutions (e.g., away from four-year institutions toward two-year institutions), and state and federal policies (e.g., decreased state support). Unfortunately, Turner does not give a clear explanation and readers are left still puzzled. However, she sets the tone for the whole book and provides the context for the other studies in the book.

    The studies by Dynarski and Bettinger examine the effect of merit aid. Dynarski studies how merit aid programs in seven states affect college-going behavior, with special reference to Georgia’s Helping Outstanding Pupils Educationally (HOPE) scholarship. Largely based on data from the Current Population Survey and employing an ordinary least squares estimation with fixed effects and sensitivity analysis, Dynarski finds that these programs are effective in improving college attendance and shifting students from community colleges to four-year institutions. Furthermore, these programs appear to narrow the gap between White and minority students (Black and Hispanic). Like Dynarski, Bettinger finds that the need-based federal financial program — the Pell Grant program — has a positive impact on college persistence. Using a rich new dataset gathered by the Ohio Board of Regents, Bettinger presents convincing evidence on how Pell Grants affect students’ possibility of staying in college through two strategies — instrumental variable strategy with the panel identification, and regression discontinuity strategy with the cross-sectional identification. Both specifications find evidence that Pell Grants reduce dropout rates.

    Charles F. Manski and David A. Wise (1983) find that the Pell Grant program has had a great impact on access. They estimate that Pell Grants increased college enrollment by 21 percent; the increase is 59 percent for lower-income students, 12 percent for middle-income students, and only 3 percent for higher-income students. J. Brad Schwartz (1985) found that the Pell Grant program accounted for approximately 21 percent of enrollment at four-year colleges from lower-income families, and that students from higher-income family background were not affected as greatly. The findings confirm the general consensus in previous studies that grants tend to improve equality in college access and enrollment. Larry L. Leslie and Paul T. Brinkman (1988), by reviewing six studies, found that grant aid had substantial effects on student enrollment and that the effect was especially strong among students from low-income families.1 Based on a comprehensive literature review on studies that examine the impact of financial aid on college enrollment, Michael S. McPherson and Morton Owen Schapiro (1991) concluded that “increased student aid does raise the enrollment of lower-income students” (p. 55).
    More recent studies use other changes in aid programs to estimate their effects. Using the October Current Population Survey and the Integrated Postsecondary Education Data System, Dynarski (2000) found that each $1,000 given out through Georgia’s HOPE scholarship increased the college attendance rate in Georgia by 3.7 to 4.2 percentage points. A recent study by Dynarski (2003) analyzed the impact of the 1982 elimination of the Social Security Student benefit program in college attendance and completion using data from the National Longitudinal Survey of Youth. The finding of this study is that every $1,000 of student benefits offered induces an increase of 3.6 percentage points in college attendance.

    Two chapters in this book are among the first attempts to examine the effect of tax incentive plans. One of them is an informative study by Bridget Long on the two federal tax credit programs introduced in the Tax Relief Act of 1997, namely, the Hope Learning Credit and Lifetime Learning Tax Credit. Long finds that middle-income families are the major beneficiaries of tax credits. Interestingly, she finds no statistically significant effect either on the general enrollment or on the proportion of students that attend four-year institutions. Further, she finds that public two-year colleges respond to tax credits by raising tuition, which confirms the so-called Bennett hypothesis.2 The other chapter is written by Jennifer Ma, who, using rich data from the TIAA-CREF Survey of Participants’ Finances, analyzes how two tax-favored education savings programs — the 529 Plan and the Coverdell Education Savings Account — affect household savings.3 To control for saver heterogeneity, Ma employs two strategies, namely, using the ownership status of an IRA account to show individuals’ “taste for saving” and then estimating the model separately for those with and without an IRA account. Furthermore, she uses a recently developed technique — the propensity score approach — to reduce the selection bias. Findings show that education savings do stimulate new household savings.

    Two chapters in this book examine the effect financial aid could have from a student’s perspective. Christopher Avery and Caroline Hoxby’s study of college choice is based on new data from the College Admissions Project, in which the researchers collected information on 3,240 students from 396 high schools around the United States. Using the conditional logit technique, they examine how students’ college choice behaviors are affected by their financial packages. This study focuses on high-aptitude rather than general students because it is those high-aptitude students who face these complex choices. The results show that these students do prefer lower tuition, lower room and board charges, more selective colleges, and generous financial aid packages. This trend is shared by all high-aptitude students, except for those whose parents have high incomes or who graduated from selective colleges.
    This study enriches the body of literature that examines the effect of price on college enrollment. Economic theory tells us that students have an elastic demand for higher education, which means that they are sensitive to the price changes among colleges and universities. Many studies have examined the impact of price on college decisions and found support for this theory.4 Moreover, a number of researchers have found that low-income students may be more sensitive to price than students from families with higher incomes.5 McPherson and Schapiro (1991), using the Current Population Survey for the 1974-84 period, found that increases in the net cost of attendance have a negative and statistically significant effect on enrollment for White students from lower-income families. More specifically, they found that for students belonging in the lowest income group (annual family income less than $20,000 in 1990 dollars), every $100 increase in net cost results in a decrease of 0.68 percentage points in enrollment. Edward P. St. John (1990) found similar degrees of price sensitivity for all but the wealthiest students. St. John and Noell (1989) found that Black students are the most responsive to financial aid offers, the next group is Hispanic students, and White students are third. In an update of Leslie and Brinkman’s literature review, Heller (1997) concluded that “every $100 increase in tuition results in a drop in enrollments of 0.5 to 1.0 percentage points across all types of institution,” that “lower-income students are more sensitive to changes in tuition and aid than are students from middle- and upper-income families,” that “Black students are more sensitive to changes in tuition and aid than White students,” and that “students in community colleges are more sensitive to tuition and aid changes than are students in 4-year public colleges and universities” (p. 650).

    Because of the varying impact of price changes on different income groups, it is of great significance to study how financial aid programs impact those groups in order to know whether these programs are able to remove financial barriers by providing equal opportunities across income groups. The study by Christopher Avery and Thomas Kane, which examines the effect of financial aid from a students’ perspective, bears important implications for how to improve access to college for low-income students. Avery and Kane compare data on the participants of the College Opportunity and Career Help (COACH) program — a program that enables students from Harvard University to help senior students in three low-income, inner-city Boston public schools with their future plans and college applications — with the data on one suburban, non-COACH school. They find that these two groups of students share similar aspirations and perceptions of college costs and payoffs, although their participation rate in college differs significantly. They argue that the hurdles for inner-city youth lie in the application process, such as GPA, registering for and taking the SAT, and writing essays. These findings have important policy implications because it suggests that setting up programs that help students with their college application might be more effective in improving college access for students from low-income families than increasing financial aid.

    The remaining two chapters explore peer effect as a partial explanation and justification for financial aid policies. From an economic perspective, students of higher quality can pay less because they can potentially bring better influence to the institution and students of lower quality have to compensate for their low quality by paying more. Michael Rizzo and Ronald Ehrenberg explore how institutions respond to changes in federal and state support through tuition and enrollment strategies, such as out-of-state tuition and nonresident enrollment, using data from various sources.6 They find that public institutions do not use nonresident enrollment to generate revenues but to raise peer quality. Gordon Winston and David Zimmerman find that peer effect does exist when measured with grade performance and academic ability. However, the direction in which peer influence takes place among groups of students with different abilities still remains unclear.

    Limitations of the Book

    While these studies are rigorous and informative, their sophisticated methodologies may prove challenging to lay readers. The findings, furthermore, are rarely clear-cut. Therefore, this book might frustrate readers who are looking for unequivocal policy recommendations. Also, some of the longer chapters could benefit by being shorter because it would make them more focused and highlight the main message researchers want to convey.

    In addition, this book shows that econometric research in education still has room to grow. As rigorous as the studies are, the limitations of data and methods leave some questions unresolved, despite the best efforts of the researchers. For example, the findings in Ma’s paper need to be viewed with reservation because the participants in the data used in this study are not representative — they are older, more educated, and tend to have higher income than the general population. And Winston and Zimmerman use grade performance and academic ability to measure peer effects, even though they only represent one dimension of student behavior. A final example is Avery and Kane’s chapter; the comparison group consists of only one high school, which might not be as representative of urban youth as the researchers intend it to be.

    This book is an important addition to the body of literature that examines the effect of financial aid on students’ opportunity to higher education. It develops the discussion by extending the current dialogue from mostly focusing on financial aid’s effect on college attendance and enrollment to college choice. Furthermore, it improves current studies by adopting sophisticated econometric methodology and by using the latest available datasets. The studies in this book provide great resources and examples for researchers who have been studying similar topics, as well as researchers who have just started exploring this complicated research area. Also, it makes an excellent textbook for professors who teach advanced quantitative methodology courses at the graduate level.

    Reviewer: Baoyan Cheng
  2. Share

    Spring 2006 Issue


    Community Forces, Social Capital, and Educational Achievement
    The Case of Supplementary Education in the Chinese and Korean Immigrant Communities
    By Min Zhou and Susan S. Kim
    What the Resistance of New Teachers Reveals about Professional Principles and Prescriptive Educational Policies
    By Betty Achinstein and Rodney T. Ogawa
    An Interview with Khalil Mahshi
    No Longer Overlooked and Undervalued?
    The Evolving Dynamics of Endogenous Educational Research in Sub-Saharan Africa
    Richard Maclure
    Editor's Review - College Choices edited by Caroline M. Hoxby
    Baoyan Cheng

    Book Notes

    Teaching Social Studies That Matters
    By Steven J. Thornton

    Becoming Adult Learners
    By Eleanor Drago-Severson

    NCLB Meets School Realities
    By Gail Sunderman, James S. Kim, and Gary Orfield

    Compelled to Excel
    By Vivian S. Louie

    Inequality in America
    By Benjamin M. Friedman