DP-71 “A Note on How Well Available Income Information Identifies Low-Income Students,” March 2007.
This note looks at the quality of the information on family income that selective colleges rely on to increase equality of opportunity by recruiting high-ability, low-income students. Individual family income estimates embedded in the College Board’s search parameters are compared, for 635 recent Williams matriculants, with their incomes as reported on IRS Forms 1040 and, for further comparison, with self-reported incomes. The data suggest that there is considerable room for improvement and, indeed, until there is better information, that any effort to increase equality of opportunity by energetic recruitment of high-ability, low-income students will be haphazard at best.
Full Text | Return to publications page
DP-70 “Social Comparison of Abilities at an Elite College: Feeling Outclassed with 1350 SATs,” February 2006.
Two studies explored the experience and performance of students at Williams College in three-person groups that were homogeneous or heterogeneous in rated academic ability. In accord with hypotheses from Festinger’s (1954) social comparison theory, students in academically homogeneous groups had more positive experiences and performed better on measures of written and video-taped performance. These results differ somewhat from recent studies of peer effects among roommates and from a line of recent social comparison research regarding the effect of exposure to superior others on one’s own performance. In addition, students in single-sex groups had higher scores on several self-report and performance measures. Qualifying this finding were additional results showing that women did better in single-sex, while men did better in mixed-sex groups. The overall results were framed in terms of social comparison dynamics.
Full Text | Return to publications page
DP-69 “Access to the Most Selective Private Colleges by High-Ability, Low-Income Students: Are They Out There?,” October 2005.
With only a small number of their students coming from familieswith the lowest incomes (10% from the bottom two family incomequintiles), the nation’s most selective private colleges anduniversities need to know why. Two ready ideological answers are (1)that low-income high-ability students are being excluded in order tofavor the children of society’s most advantaged or (2) that very fewlow-income high-ability students exist – that by college age,low-income students have been so damaged by education, nutrition,neighborhoods, and families that few can qualify in a perfectly fairadmissions process. This paper uses the national population of highschool test-takers in 2003 to examine the national distribution overfamily incomes of high-ability students (variously defined). With thesedata, two questions can be addressed. What would be the target share oflow-income students at these schools if their student bodies were tomirror the national high-ability population? And, are they out there -do there exist enough such low-income, high-ability students to meetthose targets? It is shown that they are out there – that a somewhatlarger share of the test-taking population is made up of high-ability,low-income students than are found in these schools and that theirnumbers make it feasible for the schools to increase their enrollmentsto target that national share. Because much depends on the definitionof “high-ability” used, we consider alternative definitions but reachthe same conclusion at any reasonable level (like a minimum combinedSAT of 1300 or even 1420).
Full Text | Return to publications page
DP-68 “Institutional Ethos, Peers and Individual Outcomes,” June 2004.
In this paper, we present estimates of roommate and institutionbased peer effects. Using data from the College & Beyond survey,the Freshman survey, and phonebook data that allows us to identifycollege roommates – we estimate models of students’ politicalpersuasion and intellectual engagement. The evidence suggests that astudent’s roommate’s political sentiments have some impact on their ownpolitical views later in life. We also implement a cluster basedanalysis that attempts to answer the question: how would a student’soutcomes have changed if they’d attended a very different school? Ourfindings suggest that student outcomes are, indeed, sensitive to theschool they attend. Similar students attending schools that have adecidedly different “ethos” differ in important ways post-college.Institutional peer effects seem to have a powerful effect on studentoutcomes.
Full Text | Return to publications page
DP-67 “Who Cares? How Students View Faculty and Other Adults in US Higher Education,” April 2004.
Using Mellon Foundation’s College and Beyond survey of alumnifrom 34 colleges and universities spanning 40 years, Clotfelter foundthat those who reported that someone “… besides students [took] aspecial interest in you or your work” also reported greater generalsatisfaction with their college and, concretely, made larger alumnigifts. This paper uses those same data to see who it was who isreported to have cared – faculty, coaches, deans,… – how that differedby institutional type – public research universities, coed or women’sliberal arts colleges, Ivy universities… – and how it changed over time- for entering cohorts of 1951, 1976, 1989. Some of the results may bepredictable – for instance, that faculty are the main ‘care givers’ inall times and places – while others are unexpected – that there’s noindication of a decline in the faculty role over time, for instance, orthat athletes, while they find coaches more caring than donon-athletes, still report that faculty are more caring than coaches.
Full Text | Return to publications page
DP-66r “Affordability: Family Incomes and Net Prices at Highly Selective Private Colleges and Universities,” January 2004.
College tuition is frequently compared, in press and politics,to the US median family income. That is, however, a highly misleadingbenchmark since schools with need-based financial aid rarely chargestudents from median income families the reported sticker price.Working from the financial aid records of individual students attwenty-eight highly selective private colleges and universities (COFHEschools), we addressed two questions: what do the highly able lowincome students at these schools actually pay, net of financial aidgrants, for a year’s education and how do these schools differentiatetheir prices in recognition of the different family incomes of theirstudents – the concrete evidence of their dedication to equality ofopportunity? The answer to the first question is that while there isconsiderable variety in net prices, many of these expensive schoolscharge their low income students very little (one, less than $800 ayear for the average student in the bottom income quintile), making itquite reasonable for a highly able student to aspire to go to a veryselective private college or university regardless of family income.The second answer also reveals considerable variety among schools.Virtually all of them charge students in the bottom income quintile alower net price, on average, than they do their wealthier students, butat some, net price as a share of family income rises as incomesincrease while at others it falls. Most, however, follow pricingpolicies that embody rough proportionality between net price and familyincome over the whole range of the student incomes, including thosepaying the full sticker price. The net prices that remain to be paid byaided students are covered, of course, by direct payment and”self-help” – loans and student jobs. In these data, the error in thepopular representation of tuition and income is clear: the averagesticker price is 66% of median US family income but the average studentat that level pays just 23% of family income.
Full Text | Return to publications page
DP-65 “Toward a Theory of Tuition: Prices, Peer Wages, and Competition in Higher Education,” January 2003.
College tuition, as the price of higher education services,defies familiar economic analysis in important ways. It is recognizedthat tuition is a price that covers only a fraction of the cost ofproducing those educational services (about a third, nationally),creating an in-kind subsidy for students (the balance being made up bylarge flows of donative resources from gifts, appropriations, andreturns on wealth). This paper examines yet another important economicpeculiarity of tuition; it takes seriously input and output marketsimplied by Rothschild-White (1995 JPE) in which a single event – of astudent’s matriculation – is simultaneously a transaction in both aninput market (where a wage is paid for a student’s peer quality) and anoutput market (where a price is paid for the college’s educationalservices). Those two prices are obscured by the fact that the peer wageis paid in the form of a discount on the price of educational servicesas well as by the fact that the schools’ sales (tuition) revenues aresignificantly augmented by those donated resources. This framing sees aschool’s access to donated resources (wealth) critical in determiningwhich market – peer quality inputs or educational services sales – willmost influence its behavior. Apparent anomalies in the product market -like queues of unsatisfied customers that persist while schools refuseto expand capacity – disappear when they are seen to be the result ofan input market where a queue of job applicants is used to allow thefirm to select on worker – peer – quality (the result of anAkerlof-Yellen efficiency wage).
Full Text | Return to publications page
DP-64 “Peer Effects in Higher Education,” January 2003.
This paper was prepared as a chapter for College Decisions: HowStudents Actually Make Them and How They Could, edited by CarolineHoxby for publication by the University of Chicago Press for the NBER.In this chapter, we describe the potential significance of student peereffects for the economic structure and behavior of higher education.Their existence would motivate much of the restricted supply, studentqueuing, and selectivity – and institutional competition via merit aidand honors colleges – that we see in American higher education; their(appropriate) non-linearity could justify the resulting stratificationof higher education as an efficient way to produce human capital. Inaddition, we use data from the College and Beyond entering class of1989, combined with phonebook data identifying roommates, to implementa quasi-experimental empirical strategy aimed at measuring peer effectsin academic outcomes. In particular, we use data on individualstudents’ grades, SAT scores, and the SAT scores of their roommates atthree schools to estimate the effect of roommates’ academiccharacteristics on an individual’s grades. The results suggest that,for two of the three schools used, students in the middle of the SATdistribution do somewhat worse in terms of grades if they share a roomwith a student who is in the bottom 15 percent of the SAT distribution.Students in the top of the SAT distribution appear often not to beaffected by the SAT scores of their roommates. These results aresimilar to those reported in earlier research using data from Williams(Zimmerman) and Dartmouth (Sacerdote).
Full Text | Return to publications page
DP-63 “Impatience and grades: Delay-discount rates correlate negatively with college GPA,” May 2002.
Because the rewards of academic performance in college areoften delayed, the delay-discounting model of impulsiveness (Ainslie,1975) predicts that academic performance should tend to decrease aspeople place less weight on future outcomes. To test this hypothesis,we estimated (hyperbolic) discount rates for real delayed monetaryrewards ($10 to $20) using second-price auction procedures with 247undergraduates at two liberal arts colleges. College GPA was reliablycorrelated with discount rates, r = -.19 (p = .003), and remainedreliable after partialling out SAT scores. The results add to theexternal validity of the discounting model of impulsiveness, and pointto a possible contributor to academic performance of interest in thestudy of higher education.
Full Text | Return to publications page
DP-62 “Access: Net Prices, Affordability, and Equity At A Highly Selective College,” December 2001.
All of the financial aid decisions at Williams College for thepast fourteen years – nearly 14,000 of them – were used to see how muchstudents actually paid for tuition, room, board, and fees to go to thathighly selective and expensive school – their net prices. Williamspractices need blind admission with full need-based financial aid andgives neither merit nor athletic scholarships – a family’s economiccircumstances are the only reason for a price adjustment. So these datacan answer the motivating question of the study, “Can highly able lowincome students reasonably aspire to go to the best and most expensivecolleges in the country?” Does need-blind admission and full needfinancial aid, in other words, really work to serve merit and equity atthe same time? With income and net price data on all aided students andincome data for families at the 95th and 99th percentiles of the USincome distribution who pay the full sticker price, we can describe thenet price pattern across the whole student population as pricingpolicies have evolved at Williams (and similarly at other highlyselective schools). The end point – in the current academic year – seesa remarkable similarity in the shares of income paid for a year atWilliams. Aided students across the five income quintiles pay, onaverage, 11% to 19% of their pretax family incomes – the lowest incomequintile paying the smallest share – while those at the 95th and 99thpercentiles, paying full price, spend 21% and 9% of their familyincomes, respectively, for a year at Williams. One usefully concretenumber: the average student in the bottom twenty percent of the incomedistribution pays $1,683 while the full tuition is $32,470.
Full Text | Return to publications page
DP-61 “Peer effects, gender andintellectual performance among students at a highly selective college:a social comparison of abilities analysis,” October 2001.
A study was conducted to examine peer effects amongundergraduates at Williams College, a highly selective four-yearliberal arts school. Specifically, the study explored whether studentswould perform better writing about newspaper articles they read anddiscussed in academically homogeneous or heterogeneous groups of three.In homogeneous groups all three students were from either the top halfor bottom half of their class on academic ratings assigned at the timeof admission. Heterogeneous groups included students from both the topand bottom half of their class. The results showed that students in thetop and bottom half performed similarly overall, but that studentsperformed better in homogeneous groups, whether those homogeneousgroups were made up of students in the top half or the bottom half oftheir classes. This pattern of results was stronger for men subjectsthan women subjects. The results were interpreted in terms of theprinciples of social comparison theory (Festinger, 1954).
Full Text | Return to publications page
DP-60 “”Grow” the College? Why Bigger May Be Far From Better,” October 2001.
This brief paper asks if the proposition that “growth is good”applies with equal force to private business and to private collegesand universities. An increasing appreciation of the fundamentaldifferences in economic structure between business firms and academicinstitutions suggests that it’s easy to make costly mistakes if thosedifferences are ignored and “expanded sales” may often be one of them.The most fundamental problem rests, simply, on the fact that since theprice paid by a college’s customers covers only a fraction of the costof providing their education, rather than yielding additional netrevenues, enrollment expansion (other things equal) will generateadditional uncompensated costs. Special circumstances can sometimesstill justify increased enrollments, but they are circumstances verydifferent from those facing a business firm.
Full Text | Return to publications page
DP-59 “Saving, Wealth, Performance,and Revenues in US Colleges and Universities,” May 2001.
Data on institutional saving in US higher education have not beenavailable until now, yet they are useful in several ways. They describehow various types of schools are doing financially, and whether theirpresent behavior is sustainable. They complete the picture of sources anduses of revenue for institutions of higher learning, which allows us topin down the degree to which the charitable mission of these schools isresponsible for their income. They describe a limit to aggressive pricereductions. And they allow for some projections of what the economicstructure of higher education will look like in the future. Financialdata for U.S. higher education institutions from the U.S. Department ofEducation’s Integrated Postsecondary Education Data System (IPEDS) areused to compute savings rates for 2109 institutions in 1995-6, and for1581 institutions for a panel of the years 1986-7, 1990-1, and 1995-6.These data are available as Excel or Stata files.
Full Text | Return to publications page
DP-58 “Economic Stratification andHierarchy Among U.S. Colleges and Universities,” November 2000.
Colleges and universities in the US differ markedly in their access toeconomic resources, hence in what they can do for their students.National (IPEDS) data are used here to describe the resulting hierarchythat’s reflected in schools’ spending on their students, the prices thosestudents pay, and the subsidies they get in consequence. Both historicaldata and projections based on recent institutional saving suggest thateconomic disparities among institutions and their students are increasing.In a final section, the paper asks what to make of this: what we can sayabout “the right degree” of institutional disparity, so whether we havetoo much, too little, or about the right amount of differentiation.
Full Text | Return to publications page
DP-57 “UndergraduateFinancial Aid and Subsequent Alumni Giving Behavior,” November 2000.
Data on 2,822 Vanderbilt University graduates are used to investigatealumni giving behavior during the eight years after graduation. A twostage model accounting for incidental truncation is used to first estimatethe likelihood of making a contribution and second estimate the averagegift size conditional on contributing. The type of financial aid receivedas an undergraduate appears to have a greater influence on subsequentalumni generosity than the amount received. Adding some scholarship to aloan-only package or eliminating all loans from a mixed loan-grant packageincreases the likelihood of a subsequent contribution. Increasing thetotal size of the package or altering the proportions of an already mixedpackage appears to be inconsequential for future donations. Students whoreceive small merit scholarships contribute more as alumni than studentswho receive either no merit scholarship or a large merit scholarship.
Full Text | Return to publications page
DP-56 “Where Is Aggressive PriceCompetition Taking Higher Education?” June 2000.
It is increasingly clear that price competition is escalating in themarket for higher education. We attempt to understand how pricecompetition would work in higher education and explore the likely long runequilibrium structure of prices in that context. We draw inferences usingboth microeconomic theory and historical parallels found in the market forgraduate education. Our analysis suggests that negative prices are likelyto prevail at the wealthiest colleges and universities. Using data fromIPEDS we estimate the resulting distribution of prices and school quality.While price competition may increase attendance by low income students atthe wealthiest colleges and universities, it is unclear how they will fareat schools with middling wealth and resources. Further, schools with lessaccumulated wealth will be particularly vulnerable to any ensuing pricecompetition. While our conclusions must be interpreted with caution,they do suggest some cause for alarm.
Full Text | Return to publications page
DP-55 “Social Comparison and PeerEffects at an Elite College,” May 2000.
A study was conducted to see whether peer effects could be observedamong undergraduates at Williams College, an elite four-year liberal artsschool. Specifically, the study explored whether students in the bottomthird of their class, with average SAT’s of about 1300, would performbetter in writing about newspaper articles they read and discussed ingroups of three if the two others in the group were academically superior– from the top third of their class, with SAT’s averaging about 1500 –rather than similar — also from the bottom third of the class. Theresults showed that women subjects performed better if their discussionpartners were from the top third of the class, but men did better if theirdiscussion partners were from the bottom third. Alternative analysescomparing subjects who had better or worse discussion partners asdetermined by the quality of their peers videotaped discussion statements,showed that across gender subjects did better written work when theirdiscussion partners were better. The results were interpreted in terms ofthe principles of social comparison theory (Festinger, 1954).
Full Text | Return to publications page
DP-54 “The Positional Arms Race inHigher Education,” April 2000.
The market for undergraduate education has many similarities to anarms race. A school’s position – relative to other schools -determines its success in attracting students and student quality.Its position, in turn, is largely determined by the size of itsstudent subsidies, the difference between its educational spending andthe net tuition it charges its students (or, much the same thing, howmuch their students have to pay for a dollar’s worth of educationalspending). High-subsidy schools spend the most per dollar of tuitionso that ‘bargain’ attracts the highest quality students. To changeits position, a school must spend more or charge less – and find theresources to support it. The positional arms race suggests whycompetition from a school further down in the hierarchy forces aresponse more effectively than competition from above and why it’sbeen typical of higher education that costs rise to reposition, butprices don’t fall.
Full Text | Return to publications page
DP-53 “So You Want to Earn a Ph.D.in Economics: How Long Do You Think It Will Take?” March 2000.
The elapsed time taken to earn a Ph.D. in economics is analyzed withdata from 620 (of about 950) 1996-97 Ph.D.s. The median is 5.3 years. Aduration model indicates that those students at several of the most highlyregarded programs, those supported by no-work fellowships, and thoseholding a prior masters degree finish faster than others. Americans, thosewho start jobs before completing their degree, and those who have childrentake longer. Kids slow the progress of women, but not of men. The onlydifference among fields is a longer time required for industrialorganization and international economics. There is no difference intime-to-degree between men and women, married and single students, olderand younger students, and those enrolled in larger or smaller Ph.D.programs. Fellowship support is more important for speeding the progressof women than that of men.
Full Text | Return to publications page
DP-52 “Peer Effects in AcademicOutcomes: Evidence from a Natural Experiment,” November 1999.
This paper uses data from Williams College to implement aquasi-experimental empirical strategy aimed at measuring peer effectsin academic outcomes. In particular, data on individual student’sgrades, SAT scores, and the SAT scores of their roommates are used.It is argued that first year roommates are assigned randomly withrespect to academic ability. This allows measurement of differencesin grades of high, medium, or low SAT students living with high,medium or low SAT roommates. With random assignment these estimateswould provide compelling estimates of the effect of roommates’academic characteristics on an individual’s grades. The paper alsoconsiders the effect of peers at more aggregated levels; inparticular, the effects associated with different “academicenvironments” in clusters of rooms that define distinct social units.The results suggest that peer effects are almost always linked morestrongly with verbal SAT scores than math SAT scores. Students in themiddle of the SAT distribution may do somewhat worse in terms ofgrades if they share a room with a student who is in the bottom 15percent of the verbal SAT distribution. Students in the top of theSAT distribution are least affected by the SAT scores of their (roomor entry) peers. The effects are not large, but are statisticallysignificant in many models.
Full Text | Return to publications page
DP-51 “Peer Influences Among CollegeStudents: The Perils and the Potentials,” September 1999.
Students’ intellectual, social and personal development is highlyinfluenced by peers during the college years. These changes can beunderstood in terms of social comparison theory, which outlines theconsequences for group dynamics of people’s need to evaluate theiropinions and abilities. Discussion aimed toward opinion consensus andcompetition aimed toward improving ability levels promote thedevelopment of intellectual capacities and a range of otherabilities. Discussion and competition also promote the definition andpolarization of values. An expanded account of social comparisonprocesses considers the further group consequences of the need forself-esteem. The distinction between informational and normativesocial influence underlines the importance of people’s standing ingroups for their self-concepts and self-esteem. Social identitytheory expands these accounts to consider the implications ofself-esteem needs for intergroup competition, discrimination andhostility. Leadership within groups is critical in countering thedestructive consequences of tendencies toward fragmentation of largergroups into smaller homogeneous groups which think and act in extremeways and which enact ingroup favoritism and outgroup discrimination.
Full Text | Return to publications page
DP-50 “Students Educating Students:The Emerging Role of Peer Effects in Higher Education,” April 1999.
The quality of the education a student gets at a college oruniversity depends both on the school’s resources – faculty,facilities, libraries – and importantly on the quality of his or herfellow students. He or she simply learns more – better, faster, moredeeply – in the company of able students than with weak ones. Putthat way, the proposition seems reasonable, persuasive, and appealing- we can usually get by simply by asserting it. But as we’ve lookedmore closely at those “peer effects,” we have encountered anincreasingly complicated, subtle, and often slippery set of issues: atbase, not much is known about peer effects in higher education,despite their potential importance. The purpose of this paper is, ina sense, to describe the structure of our ignorance – what it lookslike, why it matters, and what might be done to overcome it – aresearch agenda.
Full Text | Return to publications page
DP-49 “For-Profit Higher Education:Godzilla or Chicken Little?” November 1998.
Student subsidies are large, ubiquitous, and very unevenlydistributed in US higher education – covering, on average, two-thirdsof a student’s educational costs and ranging from $2,600 in the bottomdecile of schools ranked by subsidy size to $24,000 in the top. Sodata on the distribution of those subsidies among colleges anduniversities identifies the schools that are most vulnerable to theemergence of an accredited, degree-granting for-profit sector: profits(price minus cost) are simply negative subsidies (cost minus price).Roughly a third of private two-year colleges, and doctoral andcomprehensive universities are badly protected by their meager studentsubsidies that put them in the bottom ten percent. In the topdecile, with large entry barriers erected by large student subsidies,schools can worry less about survival but no less about for-profit”cherry picking” within their curricula and programs as firms move totake over those courses that give students the smallest subsidies.
Full Text | Return to publications page
DP-48 “The Effect of HistoricallyBlack Colleges on Wages of Black Students: An Analysis by Gender,”October 1998.
This study considers the effect of attending Historically BlackColleges and Universities (HBCUs) on wages of black students. A modelis developed to estimate reduced form wages equations conditioned onthe decision to attend a four year HBCU, non-HBCU or no four yearinstitution. Models are then estimated separately for men and women.Men and women both benefit in terms of wages, conditional on thedecision to attend an HBCU. However, HBCU attendance may be beneficialto a broader population of men than women.
Full Text | Return to publications page
DP-47 “What’s Been Happening toHigher Education? Facts, Trends, and Data. 1986-7 to 1994-95,” March1998.
There’s a new “global” way of organizing the economic informationabout an individual college or university that leads to a new way oftracking and understanding the changes that have been overtakinghigher education, nationally. The paper gives a simple introductionto this way of looking at colleges and universities, defining themajor relationships between prices, costs, and subsidies that governthese schools, then applying this structure to the analysis ofnational economic data from 1986-87 to 1994-95 for most of thecolleges and universities in the U.S.. Many of the results revealedby this approach are already quite commonly known (like the fact thatsticker prices are rising), but take on new meaning in this context(rising sticker prices are used by different schools for verydifferent purposes).
Full Text | Return to publications page
DP-46 “A Guide to Measuring CollegeCosts,” January 1998.
This paper reviews the major conceptual and practical problemsthat emerge in estimating the cost of producing a year ofundergraduate education. The three major areas discussed are thecomplicated issues in estimating the yearly cost of physical capital(the cost of using land, buildings, and equipment), the treatment ofstudent financial aid (as a legitimate cost of producing education orsimply a price discount), and the cost allocation problems inherent inseparating out those costs generated by undergraduate education in acomplicated “multi-product” university.
Full Text | Return to publications page
DP-45 “College Costs: Subsidies,Intuition, and Policy,” November 1997.
College students pay only a fraction of the cost of theireducation. Large student subsidies – paying two-thirds of the cost atthe average US college or university – are the central fact in theeconomics of higher education. Because prices (net tuitions) don’tcover costs, the economics of colleges and universities is differentin quite fundamental ways from that of the for-profit business firmsthat shape our national intuition and define our common sense. So itis often the case that what’s true about colleges and universities andhigher education is counter-intuitive and what’s sensible is simplywrong. This paper describes the economic structure of a college oruniversity, presents national data on the size and distribution ofstudent subsidies, and demonstrates the often-perverse effects theycreate for public policies and expectations that ignore them.
Full Text | Return to publications page
DP-44 “Thinking Seriously About Payingfor College: The Large Effects of a Little Thought,” September 1997.
The high price of attending college has generated a great deal ofdiscussion and some heated controversy in recent years. Popularopinions generally depict college prices as unreasonable, unjustified,and unpayable. In the context of thinking seriously about the wayscolleges and universities ought to approach the difficult issue ofpricing in higher education, we survey students at variousinstitutions in order to seek insight into how students think aboutthe amount they pay, and what they get for their money. Morespecifically, we ask students to seriously consider the subsidyassociated with their educational program, and the effect that achange in the size of this subsidy might have on this program. As aresult of this exercise, we find that students tend to regard theprice of their education as significantly more reasonable than theyhad prior to thinking seriously about its cost.
Full Text | Return to publications page
DP-43 “Financing Undergraduate Education: Designing National Policies,” September 1997.
In this paper we summarize our recent work analyzing pricing, aid, accessand choice in American higher education, and we draw out implications from those findings for national higher education policy. We find that real increases in net tuition have impaired access and choice principally for students from low-income families. The Clinton administration education proposals, rather than addressing the needs of this group, focus on providing tax benefits to middle and upper-middle income families. We argue that the nation needs a higher education program that provides more assistance to the students for whom the issue of college affordability is the most pressing.
Full Text | Return to publications page
DP-42 “Why Can’t a College Be MoreLike a Firm?” May 1997.
A sophisticated and widespread intuition is supported by ourexperience with business firms. And it is confirmed, influenced, andexpanded by the formal microeconomic analysis of those firms and theirmarkets. This paper asks if that theory and intuition are helpful forunderstanding colleges and universities and the higher educationindustry. The answer, even at this early stage of investigation, hasto be “Yes, but…” Six quite fundamental economic characteristics ofcolleges and universities make the familiar analogy with firmssometimes inappropriate and sometimes simply wrong. Most fundamentalare: that the price a customer pays for the product covers, onaverage, only one-third of the cost of its production — the averagestudent in the US pays 30 cents for a dollars worth of highereducation –; that those student subsidies are very different atdifferent institutions, creating the highly skewed but stablehierarchy that is the dominant feature of the industry; and thatfellow-student quality is a key input to educational quality –students educate students. A start is made here on developing theimplications of those facts but much remains to be done.
Full Text | Return to publications page
DP-41 “Subsidies, Costs, Tuition andAid in US Higher Education: 1986-87 to 1993-94,” April 1997.
Data from a panel of 2,269 colleges and universities track the majorchanges in educational costs, prices, subsidies, and financial aidover the seven eventful years from 1986-87 to 1993-94. The ability togive student subsidies is recognized as a central determinant of aninstitutions economic circumstances and strategy. Subsidy resourcesallow a school to sell its educational services at a net price belowthe costs of their production. So prices are always less than costs– how much less depends on a schools resources.
Using a global accounting frame, the paper emphasizes theinter-relatedness of institutional decisions on enrollments, subsidyresources, sticker prices, financial aid, and general subsidies.Public and private sectors faced very different circumstances andbehaved very differently. Within each sector, Carnegie school typeslived in different worlds. Public Research Universities faced sharplyreduced public support but countered it with restricted enrollmentsand higher tuitions that allowed them to maintain and even expandeducational quality. Public Two-year Colleges, in contrast, were wellprotected by public policy so that even as they absorbed a twenty-fivepercent increase in enrollments, they maintained educationalexpenditures with only very modest increases in net tuition ($62 overseven years). The prices that students paid for a dollars worth ofeducation changed between public and private institutions, betweenResearch Universities and Two-year Colleges within the public sector,and between high- and low-subsidy schools in both sectors — and thesechanges appear to have influenced students enrollment choices.
Increases in sticker prices either reallocate subsidies orraise net tuitions.
Full Text | Return to publications page
DP-40“The Economic Structure of HigherEducation: Subsidies, Customer-inputs, and Hierarchy”
Misunderstanding its economic structure will make it difficult topredict the effects of changes that are sweeping higher education:increasing price competition, the weakening of tenure, taxpayerrevolts, new technologies, the reduction in research support, etc.And while there is considerable utility in the parallels that seecolleges as firms selling educational services in a market tostudent-customers, colleges and universities remain different in somevery fundamental economic ways from the for-profit firms that informour intuitions and economic theories. This paper follows Hansmann,James, Rothschild-White, Baku, and Clotfelter, inter alia, to describethe economic structure of higher education and identify its uniquecharacteristics and circumstances. Chief among these: only itscustomers can provide key inputs to production so the firm cares aboutwho it sells to; firms always sell their product for much less thanthe average or the marginal cost of its production, subsidizing theircustomers from donated resources; and the ability to support suchsubsidies differs dramatically among firms, creating a stronglyhierarchical market. Rather than push a rigorous formal economicmodel as far toward this reality as possible, in this paper the stickis picked up from the other end to push the realities of highereducation as far toward the economists framing as possible. Theresult appears to be both preliminary and highly promising.
Full Text | Return to publications page
DP-39 “Subsidy Shock: ReshapingJudgments of College Sticker Prices”
Three studies tested the hypothesis that 1) people know very littleabout the extent to which colleges and universities subsidize theirstudents and 2) providing people with subsidy information leads themto judge the prices that such schools charge as more reasonable. Theresults offered qualified support for both hypotheses. First, peoplewere generally aware of the subsidies provided by public institutionsbut were in many cases unaware of the subsides provided by privatecolleges and universities. Second, after receiving subsidyinformation, people increased their ratings of the reasonableness ofthe prices of private schools but not public schools. This was trueeven when they did not initially underestimate the private schoolssubsidy. More research is needed to explain precisely why peoplejudge a private schools price as more reasonable after they receivespecific information about its subsidy.
Full Text | Return to publications page
DP-38 “Impact of College QualityChoices on Wages: Are There Differences Among Demographic Groups?”
There has been a rising interest in understanding better the impact ofcollege choices on wages that has been motivated by concerns aboutincreasing wage inequalities, about increasing costs of elite collegesand about the perceived increasing roles of highly educatedindividuals in maintaining international competitiveness. Previousstudies on the impact of college quality on wages generally have beensubject to two limitations: (1) They have not considered thedifferential impact for different demographic groups identified byrace and gender. (2) They have not considered the choice dimensionswith regard to time in college and college quality (with two recentexceptions). This study addresses both of these limitations.Estimates are obtained of the impact of time in college and of collegequality choices on wages for four demographic groups based on anexplicit dynamic framework. These estimates suggest that treatingtime in college and college quality as choices within a dynamicframework affects fairly substantially the estimated effects. Basedon these results, most previous studies may have overestimated theimpact of both college quality and time in college because of thefailure to deal with such choices and, in a lesser but stillsubstantial number of cases, because of the failure to control for thequality dimensions of pre-college education. The estimates alsosuggest that there are important demographic differences. Theestimated wage benefits from higher college quality and more time incollege tend to be highest for nonwhite males and next for nonwhitefemales, then white females and least for white males. For somemembers of all groups there appear to be some incentives forincreasing college quality (by paying somewhat higher tuitions orwhatever), more so for those who are longer in college because ofinteractions between college quality and time in college. But theseincentives differ strongly across groups, being much higher fornonwhite males and much lower for white males, with the females inbetween. For nonwhites (again stronger for males than females) therealso appear to be incentives to increase time in college. For whites,in contrast, the estimated net gains do not appear to create suchincentives at least at a 4% real discount rate — a result thatcontrasts with common interpretations of semilog wage relations. Theresults for nonwhites raise important questions regarding why there isunrealized potential for reaping net benefits from more time incollege or attending higher quality colleges. A number of possibleanswers to this question — poor information about the impact ofcollege, imperfect capital markets for financing college, poorinformation about college quality differentials, discriminatoryadmissions — may have important policy implications on both equityand efficiency grounds.
Full Text | Return to publications page
DP-37 “The Economic Analogy”
This paper examines the strengths and weaknesses of the analogy ofhigher education to a for-profit industry where schools are seen asfirms and students as customers. It looks at two areas of collegefinance. First, the authors examine the surprising fact that nearlyall colleges, equally public and private, use non-tuition resources toheavily subsidize the cost of a students education. So in the averagecollege or university, on the basis of 1991 US data, a student paid$3,101 for a $10,653 education, getting a $7,551 subsidy that coveredmore than 70% of costs. This is important in understanding a collegesstrategic decisions about size, tuition, sticker price, and aid and indetermining how successful the school will be in attractinghigh-quality students. A school with more non-tuition resources, allelse the same, can offer students a better deal for the sameeducation, which drives demand and allows the school to select forstudent quality.
The authors go on to discuss the strategic nature of financial aiddecisions. While in the recent past financial aid was a ratherseparate item from enrollment planning, it is increasingly thought ofas a critical tool in helping a college meet its admissions andrevenue goals. Admissions and aid statistics for a mythicaluniversity are presented and the consequences of variousadmissions-financial aid scenarios are explored.
Full Text | Return to publications page
DP-36 “Cohort Size Effects on US EnrollmentDecisions”
Although considerable effort has been expended on measuring thereturns to education in the U.S. and on modeling the individualdecision-making process in human capital investment, surprisinglylittle work has been done in terms of attempts to forecast collegeenrollment rates in the U.S. population, despite the obvious need forsuch forecasts. In the relatively small subset of forecastingliterature, cohort size variables tend to play a significant role, andone of the most successful models, in terms of forecasting accuracy,appears to be the work of Ahlburg, Crimmins and Easterlin (1981),which predicted an increase in enrollment rates throughout the 1980swhen most other models were predicting a decline. It is important tore-assess these models now, since for the first time since the 1960sboth enrollment rates and the size of college-aged population appearto be on the rise. Although total enrollments for women increasedsubstantially during the 1970s and 1980s, for males the largefluctuations in enrollment rates and in population size largelycanceled each other out, so that there was virtually no change intotal enrollments over those two decades.
The purpose of this study has been to draw together this literaturein a more unified framework where it is possible to compare seeminglydifferent results and make use of the information they provide todevelop an improved model. In doing so it has been possible not onlyto demonstrate that cohort size effects have indeed been significantover the past forty-five years, but also to explain the mechanismunderlying these effects. Projections of enrollment rates made usingan updated version of the Ahlburg-Crimmins-Easterlin model indicatethat total enrollments of U.S. residents aged 18-24 may increase by30% of the next decade, from the current 8.8 million to 11.4 million.
Full Text | Return to publications page
DP-35 “Physical Capital and CapitalService Costs in US Colleges and Universities: 1993”
This paper reports on the distribution of capital stocks andthe costs of capital services in 3,148 colleges and universities in1993. The $387 billion in physical capital estimated for theseinstitutions imply that $40 billion in yearly capital service costsare incurred in US higher education, adding roughly 31% to reportedcurrent costs. While private Research-I Universities, among majorCarnegie types, have the most capital per student ($143,954) andpublic Two-Year Colleges have the least ($14,831), a greater disparityin capital stocks appears when schools are differentiated by wealth –their average subsidies per student. The top decile of privateschools average over $150,000 of physical capital per student and thebottom decile of public and private schools have less than $10,000.The distortion of educational costs that results from omitting capitalservices range from 25% for Research-I Universities, both public andprivate, to more than 40% for private Liberal Arts-I Colleges andpublic Comprehensive-II Universities. Clearly, capital service costsare large and unevenly distributed within higher education, creatingserious distortions in any economic analysis that ignores them. Thedata are available as FoxPro files.
Full Text | Return to publications page
DP-34 “Are We Keeping CollegeAffordable? Student Aid, Access, and Choice in American HigherEducation”
In our 1991 Brookings book, Keeping College Affordable:Government and Educational Opportunity, we examined whether ournation’s colleges and universities were affordable for Americans ofall economic and social backgrounds, and outlined policies aimed atthe efficient allocation of government and private resources towardsthat aim. In this paper we review, update, and expand our earlieranalysis. Of particular interest is how the combination of governmentfunding and institutional financial and scholarship aid combine toexplain observed trends in student access and choice.
We begin with an overview of changes over time in the finance ofAmerican colleges and universities, focusing on the role ofgovernments, institutions and families in meeting college costs. Wethen turn to a consideration of the implications of these recentfinancing trends for the issue of access to college for people of alleconomic backgrounds. Our focus here is on the bearing of theserecent trends in enrollment and pricing on our understanding of theimpact of prices and student aid on the demand for college enrollment.We proceed next to examine evidence on the enrollment destinations ofstudents from different income groups and find that students’ choicesabout where to go to school seem to be increasingly constrained byfinances. We conclude by speculating about the future and making someobservations about public policy.
Full Text | Return to publications page
DP-33 “Capital andCapital Service Costs in 2700 US Colleges and Universities”
Using data generated for a study of student subsidies(in WPEHE Discussion Paper No. 32), this paper reports on thedistribution of capital stocks and the costs of capital services in2700 colleges and universities in 1991. The $330 billion in physicalcapital estimated for these institutions imply that $40 billion inyearly capital service costs are incurred in US higher education,adding roughly 33% to reported current costs. While privateResearch-I Universities have the most capital per student ($143,557)and public Two-Year Colleges have the least ($14,540), a greaterdisparity in capital stocks appears when schools are differentiated bywealth — their average subsidies per student. The top decile ofprivate schools have over $150,000 of physical capital per student andthe bottom decile of public schools have less than $10,000. Thedistortion of educational costs that results from omitting capitalservices range from roughly 25% for Research-I Universities, bothpublic and private, to more than 40% for private Liberal Arts Collegesand public Comprehensive Universities. Clearly, capital service costsare large and unevenly distributed within higher education, creatingserious distortions in any economic analysis that ignores them. Thedata are available as FoxPro files.
Full Text | Return to publications page
DP-32 “Costs, Prices, Subsidies, andAid in U.S. Higher Education”
Studies of student subsidies in US higher education — how muchmore it costs to educate a student that he or she pays — have focusedon the distribution of subsidies by student characteristics : Thispaper turns to the very different question of institutional strategieswith respect to price, costs, and aid to ask , Enrollment andfinancial data from 2687 US colleges and universities for 1991 areused to describe price, costs, and aid patterns for public and privateinstitutions by Carnegie type and, most important, by sizedistribution of the subsidies among schools. These appear to bedefining characteristics of both individual colleges and universitiesand, more fundamentally, of the economic structure of highereducation.
Full Text | Return to publications page
DP-31 “Skills, Innovations and Values:Future Needs for Postsecondary Education”
We discuss the current context in which colleges and universitiesfind themselves and then speculate about the future state of thenation and of its educational needs. The current context issummarized in an examination of trends in revenue sources and chargesfor different groups of higher education institutions, trends in thetypes of tuition discounting, and trends in the postsecondary educatondestinations of students from different income backgrounds. Thefuture speculation falls in to three broad headings: skill developmentand professional training; research and innovation; and valueseducation and social criticism.
Full Text | Return to publications page
DP-30 “Measuring the Effect ofAttending Historically Black Colleges and Universities on Future Wagesof Black Students”
This paper estimates the effect of attending historically blackcolleges and universities (HBCUs) on future wages of black students.The analysis attempts to determine if HBCUs have a causal effect onwages by first modelling and estimating all the choices available toblack high school graduates. Wages are then estimated conditional onthe determinants of this choice. Data from the National LongitudinalSurvey of the Class of 1972 (NLS-72) are used to estimate the models.Students that attend HBCUs appear to come from lower in the potentialwage distribution than students that attend mixed or historicallywhite institutions. The value added, in terms of future wages, fromattending appears to be over 30% for some individuals. These resultssuggest HBCUs played an important role in the labor market success ofblack students in the 1970s.
Full Text | Return to publications page
DP-29 “College Choice and FamilyIncome: Changes Over Time in the Higher Education Destinations ofStudents From Different Income Backgrounds”
McPherson and Schapiro analyze data from the American FreshmanSurveys of 1980, 1989, and 1993 to determine how family income hasaffected choice of institution, and how this has varied over time. Amajor topic of analysis is middle income melt, a popular conceptionwhich holds that rising net tuition at private institutions has forcedmiddle-income students to switch to less costly institutions. Theyfind that although the percentage of students attending privatecolleges from middle-income families has fallen, this is only becausethe percentage of students in higher education in general frommiddle-income families has fallen; the proportion of middle-incomestudents attending private universities and 4-year colleges hasremained relatively stable. Instead, at least for the four-yearprivate colleges, the decline in the proportion of high-income,full-pay students may explain their apparent financial pressures.High-income students have shifted to public and private universitiesand to public four-year colleges; along with the middle-incomestudents, they have shifted away from public two-year colleges whichhas made these schools increasingly concentrated with low-incomestudents.
Full Text | Return to publications page
DP-27 “Expenditures and Revenues inAmerican Higher Education”
This paper analyzes changes in expenditure patterns and revenuesources for the majority of private and public non-profit colleges anduniversities from the years 1986-87 to 1988-89 and 1990-91. Itcompares the changes between public and private schools by Carnegietypes, uncovering similar trends in all types. Net spending perstudent is greater at private institutions than at public ones–and isincreasing relative to them, largely because of declines in state andlocal appropriations to public schools. Their reaction to cuts inappropriations has been to cut spending with future benefits (i.e.,library, plant maintenance) rather than spending on current students.Financial aid is increasing dramatically at all types of institutions,but net tuition still increases, posing concerns about what ishappening to the access to higher education of those less able toafford it.
Full Text | Return to publications page
DP-26 “Predicting Higher EducationEnrollment in the United States: An Evaluation of Different ModelingApproaches”
The purpose of this paper is to assess the state of the art inmodel-based enrollment prediction for U.S. higher education. Wereview available studies, consider methodological anddata-availability issues raised by the approaches reflected in theliterature, and report on a study comparing the forecast performanceof several alternative models. We conclude that combining the resultsfrom disaggregated forecasting models and trying alternativeapproaches is a much better option for predicting higher educationenrollments than searching for a universal model that works for allgroups at all times.
Full Text | Return to the publications page
DP-18 “The College Investment Decision:Direct and Indirect Effects of Family Background on Choice ofPostsecondary Enrollment and Quality”
This paper focuses on two components of the postsecondaryschooling decision process. The first concern is to investigate theimplications of the endogeneity of high school scholastic achievementin analyzing postsecondary school choice. The second concern is toincorporate an explicit analysis of choice of institutional qualityinto the investigation of postsecondary enrollment behavior. Ourtheme in both these components is the role of family background. Ourbasic data source is the National Longitudinal Study of the HighSchool Class of 1972 (NLSHS72). Family background characteristicsshow the expected effect on scholastic achievement: students fromfamilies with high income, better educated parents, or parents withhigher socioeconomic status generally do better on tests of scholasticachievement. Scholastic achievement is positively related toattending a four-year postsecondary school, but not to attending atwo-year school. The estimated effects of achievement are larger withachievement treated as endogenous than when achievement is treated asexogenous. We find similar influences of family background directlyon postsecondary enrollment and quality.
Full Text | Return to publications page
DP-8 “Projections of College Costs and Affordability: 1990-2010”
This paper presents alternative forecasts of future trends in theaffordability of higher education. In presenting simulations offuture college prices, student aid, and family income, we consider theeffects of variation in economic growth, governmental support, andcosts of providing higher education. Two principal findings are (1)that the higher education sector is highly vulnerable to persistentweak performance of the economy and (2) that under a variety ofcircumstances the relative affordability of private four-yearinstitutions compared to public four-year and two-year institutions islikely to decline.
Full Text | Return to publications page
DP-3“The Implications of GradingPolicies for Student Course Choice”
Utilizing panel data, we measure the responsiveness of studentcourse choice to grades and asses the impact on the distribution ofenrollment across departments of differences in grading policies. Weshow that grades strongly influence course choice; this influenceremains powerful after accounting for student responsiveness tosignals of comparative advantage contained in grades. Enrollments arebeing skewed towards high-grading departments. Finally, we presentevidence that, if the aim of grading is to convey information aboutstudents relative strengths and weaknesses, greater uniformity ingrading policies should be achieved by lowering grades in high gradingdepartments.
Full Text | Return to publications page
DP-2 “Progression to Graduate School fromthe ‘Elite’ College and Universities”
Full Text | Return to publications page
Comments? Sendqueries and suggestions to Rachel Louis, Project Coordinator.