How the Changing Market Structure of US Higher Education Explains College Tuition

I. Introduction

American higher education has been changed from a chain of local autarchies to a nationally and regionally incorporated market wherein every colleges experience many possible competitors for inputs and consumers. The magnitude of the transformation is equivalent or even greater than that of other domestic industries, such as banking and retail sales. The study explores the significance of the changing market structures for tuition and students subsidies (pricing), college quality, and pupils’ college decisions.

The paper provides both theory and empirical evidence that explains how opening trade in the college markets yields a considerable increase in average college tuition. If trade is open among various autarkies, each having colleges that provide different quality of education, then theory expects a number of responses. Colleges’ decline of market power on their local consumers leads to a reduction in their rent and a subsequent rise in the mean value they supply pupils. In addition, a loss in market power yields a rise in the earnings of college inputs. Since inputs are comprised of students (who act as consumers as well), high skilled pupils are expected to be given increased subsidies after geographic market integration. Given an integrated market, the mean college quality is expected to rise, and higher quality comes with an increase in tuition with no decrease in value.

In the geographically integrated market, pupils acting as both inputs and consumers are assigned more systematically among colleges on the basis for their demand for education and ability to contribute to education production. This translates to colleges having a student body that is growingly homogenous and within-college heterogeneity drops. However, between-college heterogeneity increases as colleges yield gradually more differentiated products. The increasing distinction between colleges is reflected not only in pupils’ college decisions but also in tuitions and subsidies, which are expected to rise more variably. 

The expectations come from three connected facts: (1) pupils are both consumers and inputs; (2) they are to consume at the same college where they serve as inputs; and (3) high demand pupils are usually high quality inputs. High demand pupils are the same people whose gains benefit most from colleges’ loss of market power – their demand is concurrently induced by the loss of monopoly and market power.

II. Context

The study approaches a theory of the college education’s industrial organization. The hypothesis is placed in the context of rising college tuition, which is the most significant aspect of the college setting.

The increasing college fees in the United States have been a constant problem for education scholars, legislators, and well-known analysts. Several explanations were suggested as to why tuition seems to increase more rapidly than inflation. For instance, according to Baumol (1967), since higher education is an example of a non-traded service that gets rather few productivity benefits from enhanced technology, then its price is expected to rise quicker than that of the average goods. Rather, the price ought to increase the earning of its primary inputs, such faculty and other professional staff. Meanwhile, Ehrenberg and Murphy (1993) claim that the rise in tuition prices overstates true increase in the college price, since the need-based financial support automatically makes tuition increase quicker than tuition revenue. The discussion of Clotfelter (1990, argued in Cook and Frank, 1993) that tuition increase evince higher demand for college education due to a rise in the measured rate of return to eduction, focuses on the notion that the college education supply is inelastic. On the other side, Bennett Hypothesis, a notable explanation for growing college tuitions, suggest that federal grants and guaranteed student loans have driven the tuition hikes. The main idea of the hypothesis is that federal government is a third-party contributor, hence neither high education providers nor consumers have much motives to keep costs down. The problem with this theory is that federal monies comprise a much smaller portion of payments in college tuitions than in medical bills. As a result, empirical evidence indicates that the impact of increases in the federal Pell Grants on tuition is likely to be small.

The Clotfelter and Bennett hypotheses implicitly rely on the presence of major market imperfections in the supply of college education, for instance barriers to entry, inefficient consumer information and so on. Market imperfection is even given more weight in the rationalizing the continuing increase of tuition, which can be roughly identified as collusive behavior among colleges.

The changing market structure of college education doesn’t seem to compete with theories regarding the changing demand/supply condition for education. For example, considering the college education’s industrial organization is complementary to the Baumol and Ehrenberg/Murphy theories. The study shows the remarkable increase in private college tuition in relation to the consumer price index between 1940 and 1993, particularly from 1955 to 1965 and from 1985 to present. Focusing on the tuition revenue per pupil rather than just the tuition, it can be observed that the tuition revenue did increase more gradually than tuition.  Although this confirms the Ehrenberg/Murphy hypothesis, a considerable increase in tuition revenue per pupil is yet to be explained. Meanwhile, if faculty wages are used instead of CPI, the results show a faintly lower rate of rise in general. Although this supports the Baumol hypothesis, the theory itself does not provide explanation for the recent tuition hikes.

On the contrary, there appears to have a partial conflict between the hypothesis proposed in this study and the Clotfelter and Bennett hypothesis. Even though their reasons for the increased demand are entirely compatible with a progressively competitive market for college education, their mechanisms through which higher demand leads tuition to rise are in conflict with the increasingly competitive market.

Overall, tuition increases with demand, but market power explanations imply that colleges use the higher demand to raise their rents. If, alternatively, colleges’ market power is declining then tuition is increasing owing to the fact that open market has fueled quality competition. Higher demand merely strengthens the effect of market forces.

Meanwhile, there is direct conflict between the hypothesis proposed in this study and conniving theories of college behavior. The American higher education’s changing market structure proposes that collusion should have significantly become harder and more prevalent during the post war period; however, antitrust case might have been due to increasingly competitive market structure.

III. The Changing Market Structure of American Higher Education

A. Evidence on the Geographic Integration of the College Market

From 1940, college pupils have given colleges’ geographic proximity to their homes reducing weight in their college options. Thus each college has experienced a broad market of potential pupils. The results demonstrate that the share of pupils who enrolled to colleges “in-state” dropped from 1949 to 1994. The decline was more noticeable among private colleges. 

The results also showed that colleges increasingly drew from the entire nation or a huge region. However a better measure of the capaciousness of a college’s market is a Herfindahl index of the concentration of its pupils’ residences. The index is equivalent to one if a college’s entire enrollment is from in-state, and moves closer to zero as a college’s enrollment is spread among the potential states-of-residence. From 1994, the mean private and public college’s Herfindahl indexes dropped considerably.

The statistics derived by the study showed a rising share of pupils going to four-year colleges applied to at least one college that was not merely outside their home state but also beyond the states that connect it.

The four determinants of college choice considered by the study: (1) the college’s estimated distance from the student’s home, (2) an indicator for the college being in the pupil’s state of residence, (3) the college’s tuition, and (4) the absolute disparity between the colleges’ mean joint SAT score and the pupil’s joint SAT score, are found to be statistically significant, however the geographic variables importance declines over time.  A college’s imposing high tuition, all else constant, eventually becomes a more crucial hindrance to students.

B. Evidence on the Causes of Increased Geographic Market Integration

The most significant likely causes of geographic market integration for higher education are: the advent of modern standardized admission testing in 1943–1948; the information exchange system among pupils, colleges, and scholarship benefactors that was primarily produced by the National Merit Scholarship Program in 1956–1958; the introduction of standardized financial need analysis; tuition reciprocity arrangements among states’ public college systems; deregulation in the airline and telecommunication industries led to considerably lower fees for long-distance travel and communication; and low mobility expenses for pupils who went to college on the GI Bill.

The other causes of integration rely on variation in pupils’ and colleges’ degree of information on one another.

Despite the fact that college fixed effects conspicuously absorb much of the variation in the market; the measured reasons were found to be statistically important determinants of the measures of market structure.  Colleges deal with more competitive markets when their long-distance costs drop, when their state has a tuition reciprocity arrangement (only have an effect on public school), when admission testing is implemented in their states, and when there is more corporate aid for the National Merit Program in their area.

IV. Theory

A theory of the industrialization organization of the market for college education and the impacts of having an open trade in such market is drawn.

A. Geographic Integration of Imperfectly Competitive Markets

The study explains the implications of the geographic integration of the market for college education through a simple situation: the geographic integration of autarkies, with each having a college education monopoly producer.

Given such condition, geographic integration has pro-competitive impacts since the earlier monopolists fight with each other for consumers and for inputs. This lowers price-cost margins and does well for the consumers and increases salaries of college staff, respectively. Such occurrence will be referred to as the “loss of monopoly power’ and “loss of market power”. These incidents come across the case of college education of varying quality, or vertical differentiated college education. Vertical differentiation is important since it is realistic and it yields the market power that is assumed for the autarkic colleges. As a result, it validates the assumption that the colleges in the autarkies had market power.

B. Geographic Integration of Markets with a Vertically Differentiated Product

Take into account a local autarky wherein there are a few colleges producing education of different quality. Production of college education has several significant attributes. Colleges have high fixed expenses of furnishing education with respect to their variable costs. Variable costs are roughly flat for a substantial region, up to the position where the college must do a major extension in their physical plant or faculty. Generally, the ratio of the high quality colleges’ to low quality colleges’ fixed costs is greater than the ratio of their variable costs. Likewise, the ratio of pupils’ eagerness to pay for a high quality college to a pupils’ eagerness to pay for a low quality college exceeds the ratio of high and low quality colleges’ variable expenses.

If pupils have heterogeneous demands for college quality, the output is a market in which colleges generate educational services at a variety of quality levels. The finite number of quality levels provided is contingent on the distribution of pupil demands, but does not rely on the market scale.  Several colleges do not have direct competitors at the level of quality they belong to. Price competition is slackened by the vertical differentiation of colleges – this is a natural oligopoly.

Going back to market power of colleges in the autarkies, this time not from assumption but a natural aspects, provided that autarkies have initial distribution of pupil demands  that overlap, the output of geographic integration will be more straightforward competition among colleges. Such integration will place more colleges in the propinquity of each quality level provided. This will reduce the efficiency of vertical differentiation for relaxing competition, and colleges will suffer the loss of monopoly and market power.

Furthermore, geographic integration has an added impact in the case of natural oligopolies with vertical integration. As autarkies join together, the mean college quality rises since they experience greater marginal returns to expenses on quality enhancements. Thus for a given increase in quality, a college can draw in more pupils. There is also a mean price hike that corresponds to the mean quality developments, but the price hike is coextensive and not welfare-reducing.

C. The Industrial Organization of the Market when Students are Inputs as well as Consumers

The college’s quality is partially identified by the peers of a potential student. Hence, pupils are both inputs and consumers in the production of college education. In addition, they should be inputs at the same college where they act as consumers. Lastly, pupils having high demand for quality education are potential high quality inputs.

According to Rothschild and Stiglitz (1995), since pupil double-hat as consumers and inputs, net tuition (tuition minus any grants) merges the price that a pupil pays and the salary he is given.  In a simpler view, colleges might impose tuition on pupils whose input quality was negligible and provide scholarship to other pupils. The student’s wage may be regarded as the entire subsidy he is given: his education costs subtracted by tuition and institutional grants. Every pupil at a college is a possible recipient of subsidy typically through tuition that does not cover educational cost.

An immediate impact after integration is the increase of quality demanded by high quality (and high demand) pupils. This is owing to the fact that they are the primary receivers of the loss of market power – the increase in “wage” meant being able to afford more college quality.

If colleges provide vertically differentiated services and offer similar “wage” for a certain input quality, then high demand/high input quality pupils will go to high quality colleges. Consequently, their presence will improve the colleges’ quality. Such is the McPherson-Whinston multiplier effect.

Regarding how the multiplier effect turns price competition rather unsuccessful, assuming that a college competes by reducing the tuition it imposes for its service quality level, then the college draws in new pupils, but a disproportionate number of those students have low demand and low quality. The college’s quality declines due to peer effects, and equilibrium is restored with the college at a lower fee and lower quality.

Now considering a college that competes in quality through offering high “wages” or merit scholarships to draw in pupils it could not attract, otherwise. Although a college may improve its quality, it is implausible to achieve moving up in levels of quality via this mean, since a college need to outlay more for every high input quality pupil than a college that provides high quality educational services. In equilibrium, the college that started with high fixed costs and high service quality has the possibility of maintaining its position and its high quality pupils.

In a nutshell, the theory expects at least eight responses to geographic market integration:

1. A college’s loss of monopoly power, yields increased value for pupils as consumers;

2.  A college’s loss of market power, yields greater subsidies for pupils being high quality inputs;

3. A boost in the mean college quality due to investments in quality earn greater returns;

4. A rise in the mean college tuition is proportionate to the increase in average quality;

5. Augmented allocation of pupils among colleges in accordance to their demand for quality;

6. A greater increase in quality as well as tuition, for colleges that were initially of high quality, because high demand pupils have their demand heightened by the income impact of the loss of market power;

7. Odd sustainability of quality competition in comparison to price completion because of the multiplier effect favoring the former;

8. Raising variety among colleges in terms of student ability, quality, tuition, and subsidies.

D. A Note on Colleges’ Other “Consumers”

Some colleges provide a subsidy to each of their pupils and such grants are possible since colleges have “consumers” other than their existing pupils. For private colleges, the “other consumers” encompass of sponsors of research, such as foundations and donors, such as former pupils. For public schools, aside from sponsors and donors, the primary “other consumer” is usually the state government.

For the purpose of this study, it is sufficient to acknowledge that the actions of other consumers is similar to the conduct of high demand pupils (as consumers) and the data may overstate the subsidies furnished to high quality pupils. Similar colleges that supply high subsidies to high ability pupils will be providing   high pay to other inputs that are being acquired by the colleges’ “other consumers”.

E. The Differing Effects of Geographic Market Integration on Private and Public Colleges

Public colleges in the United States are considerably more restrained than private colleges in their admission and tuition policies, college scale, use of funds, and ability to price discriminate among students. Due to the constraints, public colleges are incapable to compete in the upper region of quality level. On the other hand, they are also shielded from the revenue losses related to losing high demand, high quality pupils that some colleges can suffer from.

Selective public colleges in autarky experience a difficulty to shift to geographically integrated markets since many of their pupils are high demand consumers likely to be attracted to colleges that are less restricted on increasing quality, admissions selectivity, and tuition as a reaction to market integration.

F. A Note on Horizontal Differentiation

Another possible response of colleges to geographic integration may be through horizontal differentiation, which relaxes direct price and quality competition of colleges having relatively similar service quality. For instance, a college finding a comparative advantage may offer programs to a particular market niche.

V. Data

Since the market structure and price theory mentioned so far are relevant to baccalaureate education in the United States, therefore each college included in the analysis has been constantly baccalaureate-granting from 1940 to present. 

A panel of 1,221 colleges, with 731 private and 390 public, covering the period from 1940 to 1991 was used. Every American baccalaureate-granting college in the study had constant data on tuition, tuition revenue, college expenses, and students’ residence from 1940 to present, and also data on admission test scores from 1960 onwards.

For the period from 1966 and beyond, the primary sources of enrollment and financial data on the college where Higher Education General Information Survey (HEGIS), the Integrated Postsecondary Education Data Systems (IPEDS), and CASPAR, which is a panel version of selected variables from HEGIS and IPEDS.

Sources of data on admission tests scores of colleges’ enrolled pupils are from Cass and Birbaums’ Comparative Guide to American Colleges, Peterson’s Guide to Undergraduate Study/ Peterson’s Guide to Four Year Colleges, and Barron’s Profiles of American Colleges. Lastly, some unpublished data was supplied by The College Board.

Meanwhile information on the college students’ state-of-residence is derived from a set of “Residence and Migration” surveys carried out by the Department of Education.

The main source of the test score, the pre-1972 residence, and pre-1966 financial data was the American Council on Education’s (ACE) American College and Universities, an array that encompasses every ACE-accredited college. Minor sources such as Lovejoy’s College Guide/Complete Guide to American Colleges and Universities and others were able to help complete the data.

All National Merit data were derived from the National Merit Scholarship Corporation’s annual reports. Census of Population and Housing supplied the data on the median income and share of adults having 16 years of schooling.

VI. Evidence

It is important to mention that since there is no measure that fully explains the market structure in higher education, the regressions are expected to underestimate the true impact of market structure on tuition and other dependent variables. The statistics and regressions are all weighted by enrollment in order to accurately (as possible) give a national representation of the market for baccalaureate education.

A. The Effect of Geographic Integration on the Distribution of Students Among Colleges

Theory predicts that geographic integration of the American market for college education should raise the between-college variance and reduce the within-college variance in pupils’ admission test scores. This implies that consumers will gather more homogenously anchored on their demand for quality. In addition, since students act as both inputs and consumers, they would tend to increase the vertical differentiation among colleges given the process above.

The findings on the distribution of combined SAT scores between colleges illustrate a decline in the average SAT scores and the variance have increased quickly. Much of the rise in between-college variance in average test scores was observed among private schools. 

The results demonstrate private colleges obtained between-college variance by extending out the tails of the distribution, in contrast to public colleges which expanded the distribution of test scores in the middle of the distribution. This is because public schools experience constraints on their admissions policies brought about by state legislatures. As a consequence, they are not seen extreme points on the admission test scores distribution. Constrained public colleges that are initially selective can face trouble in keeping their mean admissions test scores, while unconstrained private colleges may admit some of their best potential pupils.

The Herfindahl indices derived by the study show that the concentration of National Merit Scholar has not been so high, but has grown over time – from a Herfindahl index equivalent to equally distributing scholars to more than a hundred colleges in 1956 to a Herfindahl index translating to the allocation of Scholars equally among less than fifty colleges in 1991. 

The statistics obtained also presented a break in the trend towards more concentration between 1966 and 1971; the break was due to a deliberate alteration in the method of scholarship allocation. The change was intended to distribute the scholarship more equally among students’ state-of-residence – thus a reducing the concentration of scholars.

Meanwhile, the results show that the within-college variance in students’ admission test scores substantially decreased from 1966 to 1991.

Further analysis has the market structure measured by two alternative measure of concentration (not competition): the share of pupils from “in-state” and the Herfindahl index of the concentration of students’ state of residence.  Note that the rise in geographic integration of the college market and in competition is expressed through a decline in the measures of concentration.

It seems that given a market with greater geographic integration and (thus) more competitors, the colleges’ mean SAT scores have the tendency to rise, and particularly more so if it was initially a high selectivity college. An effect of greater competition is it increases the average college test scores. Another is the increasing effect it has on the difference between the test scores of pupils who attend the highest quality schools to those who attend the lowest quality schools.

In addition, the expected impacts of geographic integration on the average test score and between-college spread in test scores are larger for private than for public colleges. Public colleges struggle to keep their average SAT scores when they’re experiencing growing competitive market structures. At first, comparatively elite public colleges admit pupils who have the most potential to go to private colleges or other public colleges when the market turns competitive.

On the other side, growing geographic integration in a market causes the participating colleges’ student body to become less heterogeneous.

All the preceding results confirm that heightened geographic integration and competition in the market are likely to improve the average college’s student quality, raise the student ability between-colleges, and decline the student ability within-colleges. The results suggest that the college quality has become more variable since 1940. It is possible that the value of educational services provided has become more variable over the period since colleges compete for high quality students. 

B. The Effect of Geographic Integration on College Tuition

Since 1940, the tuition was observed to have increased rapidly, with the public colleges’ in-state tuition consistently, considerably lower than that of private colleges’. Meanwhile the public colleges’ tuition distribution remained significantly narrower than that of private colleges’ tuition, especially during 1971 to 1991. 

Even though college quality has grown on the average, the vertical differentiation has been observed more significantly in private, rather than in public higher education. However it has yet to be proven if the college market integration has been a substantial determinant of the general increase in tuition and the broadening of private colleges’ tuition distribution.

The output indicates that private colleges in an increasingly integrated and competitive market impose higher tuitions. Deeper analysis finds the changing market structure to be accountable to a considerable share of increases in private colleges’ average tuition and in the variance of private colleges’ tuition.

In contrast, the impact of market integration in tuition for public colleges is more modest. Growing market integration induces public colleges (particularly the highly selective) to hike their tuition on average as well as to shrink the tuition distribution (through low selectivity public colleges increasing their tuition almost as fast as high selectivity colleges). The reason for this is the fact that public colleges that are initially low selective are the same colleges that conventionally maintained very low in-state tuitions. Such methods turn out to be unmanageable (since residency requisites for in-state tuition are hard to monitor and pupils try to trade the subsidy in a highly integrated college market) and unfamiliar (since middle-class taxpayers’ children, conventionally the main beneficiaries of public colleges, are unwilling to be limited to choosing only amongst in-state public colleges) as college markets integrate and pupils grow to be more mobile.  Such students do not want geographic location to be the only determining factor of their college choice, consequently, given the expectation that these students are unlikely to use up the benefits of a highly subsidized public college system, the taxpayers opt not to support high subsidies but rather agree on increasing tuition.

The growingly competitive market structure of American higher education seems to account for the large share of the rise in college tuition, particularly among private colleges, the expansion of the private college tuition distribution, and the shrinking in the public college tuition distribution. 

C. The Effect of Geographic Integration on Subsidies to Students

First off, per-pupil subsidy which unlike per-student scholarship is insensitive to a college’s financial support policy is merely the differences between college expenditure per student and tuition revenue per student.

The theory expects that if market integration raises average college quality as well as college tuition, then the colleges offering the biggest subsidies will be the same colleges that would experience those increases since they draw pupils whose “wage” is high.

One of the observations is that the median public subsidy is consistently greater (and increased faintly faster) than the median private subsidy. Although it is erroneous to associate the entire rise in subsidy with pupils, it is evident that some pupils go to colleges wherein their tuition payments cover just a part of their education costs.

Another observation is that market integration induces subsidies to grow by a tiny amount in the typical private college, and by a substantial amount in colleges that are initially highly selective, while no rise in subsidies for the case of private colleges that initially had low selectivity.

A notable outcome is that the colleges that increase their student quality and tuition the highest owing to a more competitive market are also the same colleges that increase their subsidies the most. However this particular fact is in conflict with several collusive theories of tuition hikes. Majority of the collusive theories suggest that colleges that can impose higher tuitions must gather higher rents and do not offer greater subsidies.

Meanwhile, a decline in market concentration yields statistically insignificant effect in the per-student subsidy for the common public college and low selectivity public colleges.  These results imply that market integration has a certain strong competitive impact on initially selective public colleges, pressuring them to contend against private colleges and other public colleges for their mobile, high quality, potential pupils.

VII. Conclusion

The increasing competitive market structure brought about by geographic integration of previously remote markets, has led American colleges to increase their quality, their tuitions, and their costs. It also induced colleges to be more varied, in order for the distribution of quality, tuition, expenditure, and student quality among colleges to be considerably broadened. On the contrary, the student ability distribution within any individual college has shrunk.  The changing market structure was found to account for roughly 50 percent of the tuition increase of typically selective private colleges since 1950, as well as for the smaller tuition hikes for public colleges and less selective private colleges.

On the other hand, the changes in tuition equivalent to commensurate changes in college quality are implied by the theoretical and empirical evidence. The empirical results suggest that the average student has higher utility in the face of geographically integrated market regardless paying higher tuition.

The growingly competitive market may likely explain the colleges’ increasing horizontal differentiation – moving away from generic curricula and towards “market niches”, to which they serve the students’ particular needs. However, evidence for horizontal differentiation is simply subjective.

The changes in the market structure are partially the result of the fundamental and irreversible changes in pupils’ costs of geographic mobility and the amount of information that pupils and colleges have regarding one another. The progressively strong market forces in college education limit the growing share of colleges’ choices. It is worth mentioning that much of the former literatures on college behavior regard colleges as independent decision makers and not as organizations whose decisions are progressively determined by factors outside their control. Thus, policy makers must acknowledge that arbitrary restraints on some colleges’ prices or expenses can potentially induce reactions from pupils as inputs and consumers.

Source:

Caroline M. Hoxby, “How the Changing Market Structure of US Higher Education Explains College Tuition” NBER Working Paper 6323. (December 1997).