1. education

Chapter One
Higher Education as an Associative Good
Henry Hansmann
Yale University
H
ansmann coins the phrase associative good in describing the nature of higher education. That
is, students care very much about who their fellow students are, due to the strong influence
of classmates on their educational and social experiences. Hansmann describes how this
associative characteristic affects the higher education market, given that institutions also are very concerned about who their students, or customers, are. Competition for top quality prospects has led to pronounced stratification of students as they are sorted in a largely hierarchical fashion across the nation's
colleges and universities.
Education, and particularly higher education,
nections of the colleges' other students. The reason
has an important characteristic that distinguishes it
is obvious: these and other attributes of a student's
from most other goods and services: it is an asso-
classmates have a strong influence on the quality of
ciative good.
the student's educational and social experience, the
The essential characteristic of an associative
relationships (including marriage) that the student
good is that, when choosing which producer to
will have later in life, and the student's personal and
patronize, a consumer is interested not just in the
professional reputation. In short, the thing that a col-
quality and price of the firm's products, but also in
lege or university is selling to its students is, in large
the personal characteristics of the firm's other cus-
part, its other students. Harvard College would be
tomers. And so it is with education. When choosing
nowhere near so attractive to a prospective appli-
among undergraduate colleges, for example, a stu-
cant if Harvard's faculty, curriculum, and facilities
dent is interested not just—or even primarily—in the
were to remain as they are, but its other students—
colleges' faculty, curriculum, and facilities, but also in
past, present, and future—were no different from
the intellectual aptitude, previous accomplishments,
those of an undistinguished state institution with
sociability, athletic prowess, wealth, and family con-
open enrollment.
11
Stratification
tomers is generally no different than the cost of pro-
Markets for associative goods do not function
viding it to low-quality customers. Consequently,
like markets for other goods and services. This is
nonprofit firms will charge the same price regard-
especially true when the producing firms are all
less of the quality of their customers. But, given
nonprofit or governmental, as is the case in the
that the price charged by different firms is the
upper reaches of higher education. Most impor-
same, customers would prefer to patronize the firm
tantly, when nonprofit firms produce associative
with the highest-quality customers. Since the cus-
goods, there is a strong tendency for customers to
tomers can't offer to pay a higher price, the only
become stratified across firms according to their
currency that customers can offer the firm is their
personal characteristics. Those who are most
own quality. The result is simple clustering: every-
desirable as fellow customers will tend to cluster at
one wants to patronize the firm with the highest-
one firm, the next most desirable at another, and
quality customers, but only the highest-quality
so on down.
customers will be accepted as patrons, since they
have the most to offer the firm. In effect, the high-
The reason for this stratification is that a cus-
quality customers are paying to associate with
tomer's own personal characteristics—what we
each other, using their own quality as currency. And
might term the customer's “quality”—constitute an
once the highest-quality customers cluster at a
important part of the price the customer pays for
given firm in this way, the highest-quality customers
an associative good. In choosing between two
among those that remain will cluster at a second
customers who would like to patronize a given
firm, and so forth, until customers are sorted
firm, and who are going to pay the same price for
among firms in hierarchical fashion.1
the firm's services, the firm will always prefer to
serve the customer who is of higher quality, since
This kind of stratification is very evident in high-
that will make the firm more attractive to its other
er education. The highest-quality students tend to
customers. Consequently, producers of associa-
cluster at a few elite institutions, the next-highest
tive goods have an incentive to pick and choose
stratum at another set of institutions, and so on
among their customers, serving only those that are
down. Indeed, among the elite institutions, there
of highest quality.
tends to be fairly pronounced stratification even from
institution to institution. If the readers of this essay
This incentive is particularly strong for non-
were asked to rank the eight colleges and universi-
profit firms, which are effectively constrained to
ties in the Ivy League in terms of their desirability as
charge their customers, on average, no more than
places to seek an undergraduate education, there
the cost of producing the service that the firm provides. The cost of providing a given quality of a
would probably be a fairly high degree of correlation
among the responses.
service, such as education, to high-quality cus-
12
Competition
growth in the industry and the entry of thousands of
new institutions. No other industry exhibits this kind
An important consequence of stratification is to
of stability.
dampen considerably the degree of competition
among educational institutions. Although there are
Although we have been focusing so far on
more than 3,000 colleges and universities in the
undergraduate education, graduate and profession-
United States, higher education is a far less compet-
al education are associative goods too—indeed,
itive industry than such large numbers would nor-
even more clearly so than undergraduate education.
mally suggest. Very few of those 3,000 institutions
Teaching on a college or university faculty is also an
are even potential competitors for Harvard, Yale, or
associative good. Physicists and historians general-
Stanford, in the sense that they could attract stu-
ly want to be at institutions where their fellow physi-
dents away from these elite universities simply by
cists and historians are as strong as possible, and
lowering their tuition.
this associational aspect of teaching and research
A critical factor in preventing competition is the
often weighs much more heavily in choosing where
difficulty of quickly changing the character of an insti-
to work than does salary or other material forms of
tution's student body. Since undergraduates are
compensation. The result is that colleges and uni-
commonly admitted for four years, it is possible to
versities show a strong tendency to stratify, not just
change the quality of at most one-fourth of the stu-
in terms of the quality of their undergraduate and
dent body in any one year. Moreover, much of the
graduate students, but also in terms of the quality of
attraction of attending a given college or university
their faculty. Since professors like to teach good stu-
depends on the qualities of its former students, who
dents, and students like to have good teachers,
contribute strongly to the institution's (and hence all
there is also an associational link between students
future students') reputation. But an institution can do
and faculty, which tends to lead strong faculty and
virtually nothing to change the character of the stu-
strong students to cluster at the same institutions.
dents that it has already graduated in past
And since academic tenure makes it very difficult to
decades—or centuries. The consequence is a very
change the quality of a college or university's faculty
high degree of inertia in the general character of any
quickly, the associational character of faculties fur-
given institution's student body, and in turn a high
ther diminishes the effectiveness of competition
degree of inertia in the relative attractiveness of col-
among institutions of higher education.
leges and universities to prospective undergradu-
The Nonprofit Form
ates. This inertia is largely responsible for the striking
fact that the relative ranking of undergraduate insti-
I noted earlier that the tendency toward stratifi-
tutions in the United States has remained relatively
cation among educational institutions is accentuated
constant for three centuries, despite enormous
by the fact that they are nonprofit. A proprietary
13
Institutional Size
college would have a stronger incentive to use price,
rather than students' own personal qualities, as the
The associative character of higher education
basis for rationing admissions, since it might
also helps explain why private colleges and univer-
increase its profits by admitting some low-quality stu-
sities remain relatively small. When demand for
dents who were willing to pay very high tuition for the
admission to an elite institution increases—as it
privilege of associating with other students who are
has at most of the elite institutions over recent
of higher quality.
decades—these institutions rarely respond, as they
The relationship between stratification and the
could, by keeping the quality of their student body
institutions' nonprofit status also works the other
constant while increasing its size. Rather, they tend
way, however: the fact that education is an associa-
to keep the size of the student body relatively con-
tive good is a strong reason for organizing educa-
stant, and increase its quality. This reluctance to
tional institutions as nonprofit organizations. To see
expand is evidently not because the institutions
this, imagine that Harvard were suddenly to be con-
face important diseconomies of scale, in terms of
verted to a for-profit firm. Then, even if it were not to
facilities or curriculum, if they were to expand their
change the character of its student body, Harvard
student bodies. In fact, it seems likely that most pri-
would have both the incentive and the ability to raise
vate institutions operate well below the efficient
its tuition considerably, since it could do so with little
scale for an appropriately varied set of curricular
effect on the demand for admission. Many students
and extracurricular offerings. As some evidence of
would willingly pay a good deal more to go to
this, one need simply note the enormous size of
Harvard rather than to attend another institution. In
many state university campuses.
effect, Harvard is a monopolist: it has a (near)
Rather, the reason for remaining small is that
monopoly on the best undergraduates in America,
increasing the size of the student body reduces its
and can thus offer prospective students a group of
quality, and this reduces the welfare of the stu-
fellow students of higher quality than competing
dents who attend the institution, of the institution's
institutions can offer. A for-profit Harvard could there-
faculty, and probably of its administration as well
fore charge a monopoly price. This would of course
(since administrators would generally like to be
make Harvard's students worse off. And this monop-
known for managing a highly selective institution).
olistic exploitation would be all the more galling to
To maintain high quality, an institution therefore
Harvard's students because the thing for which they
has an incentive to operate with a student body
would be paying a monopoly price would be their
that is well below the size that minimizes the aver-
own personal excellence! By attending a nonprofit
age cost of producing an appropriately designed
institution, Harvard's students can, at least in part,
college experience. This incentive is particularly
avoid this kind of exploitation. 2
pronounced if, as is often effectively the case, a
14
Homogeneity of Preferences
college or university cannot charge different prices
to different students according to their personal
The pattern and degree of stratification of stu-
qualities. It is even further pronounced when the
dents across institutions of higher education also
institutions are nonprofit, since a nonprofit is likely
depends on the extent to which students differ in the
to be particularly sensitive to the interests of cur-
weights they place on the various characteristics of
rently enrolled students, alumni, and current facul-
their fellow students. If those weights were the same
ty and administrators—and not to the welfare of
for everyone—that is, if all students were to place the
those students who would like to attend the institu-
same relative weight on their fellow students' apti-
tion and are willing to pay its tuition, but are reject-
tude for mathematics, knowledge of French poetry,
ed.
skill at soccer, fondness for beer, etc.—then it would
be possible to sort all students according to a simple
The Size of the Market
linear ranking from most desirable to least desirable,
Students' geographic mobility clearly affects
and stratification of students across universities
the degree of educational stratification. If students
according to that ranking would be extremely pro-
prefer not to attend an institution that is far from
nounced.
home, then stratification will tend to be limited to
As it is, of course, students do not all have the
institutions that are geographically close, and
same preferences about the fellow students they
economies of scale will force institutions to serve a
find most desirable. Some students put most weight
fairly diverse group of students.
on intellectual aptitude and motivation, others weigh
In recent years, the market for higher educa-
sociability more heavily, while yet others put special
tion has become increasingly national, as costs of
emphasis on religious commitment or theatrical and
transportation and communication have fallen and
musical performance. The result is that institutions
families have become wealthier. The result is that
tend to specialize somewhat in the types of students
the world of higher education has become much
they attract, and stratification is most pronounced
more intensely stratified. Students from Illinois who
only among institutions that specialized in similar
once would have wished to attend the best possi-
types of students.
ble college or university in that state, or in the
Here too, however, recent years seem to have
Midwest, now want to attend the best possible col-
brought increasing convergence, and hence
lege or university in the country as a whole. Elite
increasing stratification. The increasingly meritocrat-
institutions, in turn, can choose the best students
ic character of our society is one important reason
from a national pool, not just from a regional or
for this. Today, economic success, and hence social
local pool.
status, goes increasingly to individuals who have
strong intellectual aptitude and the motivation to use
15
it. Other personal attributes that once might have
tion to which they are admitted. The result is further
contributed importantly to success—such as good
stratification of students among the institutions, and
family connections, inherited wealth, sociability, and
hence a tendency for the rankings to further solidify
athletic prowess—are becoming less important.
the hierarchical ordering that they initially portray.
And, rightly or wrongly, a limited set of general meas-
The consequences of this process are conspic-
ures—such as SAT scores and grades achieved in
uous from the experience of the Yale Law School,
previous schooling—are now widely accepted as the
where I teach. For many years, the Yale Law
best measures of intellectual aptitude and motiva-
School's take-up rate—that is, the percentage of stu-
tion. The result is that it is becoming increasingly
dents who choose to attend Yale among those to
easy for all of the nation's college applicants (or
whom Yale offers admission—remained fairly con-
applicants to a given type of graduate or profession-
stant at around 50 percent. Then, in the early 1990s,
al school) to be ranked according to a common met-
the take-up rate rose rapidly to around 80 percent,
ric that is widely shared. All universities therefore
where it has remained. Given that many students
tend to rank their applicants in the same way, and all
among the 20 percent who today decline Yale's offer
applicants tend to rank the student bodies at those
of admission do so to attend graduate school or pur-
universities—and hence the universities them-
sue some other career, this means that very few stu-
selves—in the same way. A tendency toward ever
dents who are admitted to Yale choose to attend
more intense nationwide stratification of students
another law school instead.
across universities is the result.
Why did Yale suddenly emerge as everyone's
This process is receiving a strong boost today
top choice among law schools? I would like to think
from the advent of prominently published rankings of
that the answer lies at least partly in a real change in
academic institutions, such as those put out by U.S.
the quality of the institution as compared to its com-
News and World Report. Partly as a reflection of the
petitors. Some of the popularity may also be due to
stratification of educational institutions that has
publicity about the Yale Law School's graduates that,
already occurred, and partly perhaps because
whatever it says about the school's real quality, has
everyone enjoys a contest and because of the need
focused considerable attention on the institution. But
for simplification in popular journalism, these rank-
I suspect that a particularly important factor was the
ings typically involve little more than simple listings
advent of U.S. News and World Report's nationwide
of institutions in linear order from best to worst.
rankings of law schools. When those rankings were
Students who have no better information will of
first published, in 1987, the Yale Law School was
course have an incentive to apply to the highest-
ranked number 2, behind Harvard. When the rank-
ranked institutions that they believe might accept
ings were next published, in 1990, Yale was ranked
them, and then to attend the highest-ranked institu-
number 1, while the Harvard Law School had the
16
misfortune to be ignominiously dropped to number 5.
tended to attract the more commercially minded stu-
In the rankings published in 1991 and annually since
dents. The rankings may be breaking down this seg-
then, Yale has held steady at number 1, while
mentation, causing students to sort themselves
Harvard rebounded to the number 2 spot and has
among schools simply according to their Law School
likewise remained there. Not surprisingly, the big
Aptitude Test Scores and college grade point aver-
jumps in Yale's take-up rate came with the classes
ages. Of course, the Yale Law School can try to pre-
entering in 1992 and 1993.
serve its special character by giving preference in
Before the U.S. News rankings, the choice
admissions to students who appear more academi-
between going to the Yale Law School and the
cally minded or more oriented toward public interest
Harvard Law School was a complex decision on
careers. But admissions applications offer only
which opinions might reasonably differ. Once the
crude evidence of an applicant's real interests and
rankings were published, however, students were
ambitions. Truly effective sorting on these dimen-
choosing between attending the best and the sec-
sions must be done by the applicants themselves.
ond-best law schools—or at least between the
The rankings may be undercutting that kind of self-
schools that other people thought were the best and
sorting, thus depriving educational institutions of an
second-best law schools. Who wouldn't choose
important degree of diversity.
(what other people thought was) the best?
To be sure, published evaluations of academic
Moreover, the rankings made choosing Yale quite
institutions could in principle be subtle and complex,
rational. Whether or not Yale was the best law
providing a wealth of detail about individual institu-
school before the rankings, after the rankings were
tions that actually promotes rather than defeats the
published Yale would surely attract an even larger
tendency of institutions to differentiate themselves
share of the best applicants, and hence it would
along a number of dimensions. And perhaps as the
become the best by virtue of having the best collec-
evaluations mature they will place less emphasis on
tion of fellow law students. The rankings are a self-
simple hierarchical rankings of institutions. But I sus-
fulfilling prophecy.
pect that the economics of information is such that
These rankings have been a mixed bless-
simple rankings will have unusual salience for some
ing even for the Yale Law School. Before the rank-
time to come, that students will tend to sort them-
ings, students would sort themselves among law
selves among institutions according to that measure,
schools according to their particular interests. Very
and that the quality of the students at an institution—
roughly speaking, for example, Yale Law School
as measured by some simple linear index—will
tended to attract students with more academic inter-
therefore itself become an ever greater element of
ests in the law or whose orientation was more
the published rankings of institutions. Consequently,
toward careers in the public interest, while Harvard
stratification of students among institutions will con-
17
tinue to be a strongly self-reinforcing process. The
world-wide market that is highly stratified. And, given
result may be a degree of stratification that is far
the dynamics of hierarchical competition that I
higher than one would choose if one could assign
described earlier, those positions at the top of the
students to institutions simply with an eye to maxi-
hierarchy may prove to be secure for a long time.
mizing the quality of the education they receive.
Consider, for example, the internationally dominant
position that Oxford and Cambridge maintained for
Internationalization
decades after the rest of the English economy had
become relatively uncompetitive in world markets.
The increasing hierarchical stratification of higher education has to date been largely a national phe-
Is Stratification a Good Thing?
nomenon. But higher education is rapidly becoming
an international industry, and the hierarchical stratifi-
Once we recognize that the associative char-
cation of institutions is becoming international as
acter of higher education has a strong tendency to
well. To continue with the example I know best, one
drive the industry toward hierarchical stratification,
sees this process even in a field as traditionally
it is natural to ask whether that stratification is
parochial as law. Both the number and quality of
desirable, and thus whether public policy—and
applicants to the Yale Law School's graduate pro-
educational institutions themselves—should assist
gram, from countries all over the world, have gone
or resist it.
up noticeably in recent years, as has the take-up
One potentially unfortunate consequence of
rate among those who are admitted. When asked
hierarchy is that—as I have already noted—it tends
why they wish to come to the Yale Law School, what
to reduce substantially the degree of effective com-
do foreign applicants say? A surprising number men-
petition among institutions of higher education, and
tion the rankings in U.S. News and World Report.
this in turn provides room for a good deal of slack in
This rapid internationalization of the market for
the management of those institutions. An elite col-
higher education is, of course, a very good thing for
lege or university can survive many years of
the United States. The best institutions of higher
mediocre management without losing its ability to
education in the world, in nearly every discipline, are
attract top-flight students and faculty, since the stu-
today found in the United States. This is probably a
dents and faculty are largely coming for each other
consequence, in large part, of the highly competitive
and not for the other amenities and services that the
character of U.S. higher education, which differenti-
university's administration provides them, including
ates it strongly from the centralized systems found in
instruction. This may be one reason why profession-
most other nations. But, whatever the reason for
alization of the administration of higher education
their current strength, the result is that U.S. institu-
has come relatively slowly, even though—when
tions are likely to end up at the top of an emerging
measured by assets, income, or employment—the
18
leading universities rank among the largest corpora-
if strong students gain the most from being with
tions in the United States.
other students who are strong—then the kind of sorting that results from unregulated competition is a
But even in terms of the sorting of students
among universities, the tendency toward hierarchi-
good thing. As it is, we simply do not know what is
cal stratification may have some unfortunate con-
the optimal degree of mixing, and thus whether we
sequences. One of those consequences, which I
have too much hierarchical stratification in higher
have already noted, is that the limited information
education.
available to student applicants may cause stratifi-
Another problem is distributional. Even if a high
cation of institutions to become excessively unidi-
degree of hierarchical stratification of students
mensional.
across universities maximizes the average efficiency
Another potential problem is that, even if stu-
of higher education, it may strongly reinforce social
dents are perfectly informed about the qualities of
inequality. Students who are intellectually talented,
different universities, they may have strong incen-
highly motivated, well organized, and disciplined are
tives to stratify themselves too severely across insti-
likely to end up in the higher reaches of American
tutions. Suppose, for example, that while all students
society no matter where they go to school. If all the
find it advantageous to attend an undergraduate
students who are strongest in these respects cluster
institution where their classmates are strong stu-
at the same colleges and universities, they will not
dents, the degree of this advantage is greater for rel-
only reinforce their prospects for success but also
atively weak students than it is for strong students.
form an elite that is strongly socially connected. And,
That is, students whose educational background,
for better or for worse, the elite institutions that edu-
motivation, or even aptitude is relatively weak may
cate those students will come to play an ever more
gain more by going to school with classmates who
important role in society.
are already strong in these respects than would
Public Versus Private Education
other strong students, since strong students will generally learn quickly in any environment. The aggre-
Future decades will probably bring substantial
gate effectiveness of education, then, will be
privatization of our university system (and of those in
maximized by some mixing, within individual univer-
other countries), as public institutions come to be
sities, of students of varying strengths. But the
operated more like private ones, with higher fees
dynamics of competition for an associative good like
and greater autonomy, and as private institutions
education, particularly when the providers are non-
come to replace public institutions. 3 I explore these
profit, will tend to sort the students quite strictly
issues at greater length in “The Changing Roles of
according to their qualifications, and thus frustrate
Public, Private, and Nonprofit Enterprise in Education,
the optimal mixing. Of course, if it is the other way—
Health Care, and Other Human Services.”4 A likely
19
consequence will be a dramatic increase in the ten-
the educational hierarchy are relatively rare. One
dency toward hierarchical stratification in higher edu-
thinks of Stanford in the 1950s and 60s, Brown in the
cation. The state-owned universities, which today
1970s, and New York University in the 1980s and
account for nearly 80 percent of American higher
90s. Because the associative aspect of higher edu-
education, have served as a check on the overall
cation imposes such immense inertia on the quality
amount of stratification in the U.S. higher education
of colleges and universities, an institution that wish-
system. Whether it is because very large institutions
es to advance itself must adopt bold and imaginative
are more cost efficient (because of economies of
strategies.
scale), more educationally efficient (because mixing
One strategy for improving faculty, for example,
students of different abilities increases the average
is to give prominence to lackluster departments by
rate at which students learn), or less distributionally
seeding them with academic superstars who are
distasteful (because they avoid concentration of
offered exceptional salaries and perquisites. Another
elites), the state universities are generally much larg-
is to hire faculty in groups through coordinated offers
er, and their student bodies much less stratified, than
to professors at other institutions who would like to
is the case with the private institutions.
work together.
The problems of equity and efficiency that are
The quality of an institution's student body can
presented by stratification are therefore likely to
sometimes be improved by artful gaming of pub-
become ever more pressing. For example, we can
lished rankings. When, some years ago, guides for
expect debate in the world of higher education—fol-
prospective students included, as an index of admis-
lowing the lines of the debate already taking place in
sions standards, only the average aptitude test
primary and secondary education—about the desir-
scores of the institutions they were ranking, it was
ability of tying public subsidies to a willingness on the
said that some colleges and professional schools
part of private institutions to accept a diverse student
strategically chose, for several years, to admit stu-
body. But, even more for higher education than for
dents almost exclusively on the basis of those test
primary and secondary education, diversity goes
scores. The result was a drop in the actual quality of
well beyond the issues of ethnicity, gender, and par-
the institution's student body—since test score alone
ents' wealth and social status that are the usual foci
is a worse predictor than test score together with a
of attention under that label. In future years, the most
measure of previous accomplishment—but a dra-
important and controversial dimension of diversity is
matic increase in the apparent quality of the student
likely to be intellectual aptitude and accomplishment.
body as shown in the consumer guides. The end
Competitive Strategies
result was then a substantial increase in the actual
Examples of private colleges and universities
quality of the student body, as truly qualified students
that have successfully bootstrapped themselves up
were induced to come to the institution by its
20
improved ranking.
strong. It is interesting to ask, then, what is the moti-
A change in the nature of the institution and its
vation for these agreements. Consider, for example,
students can also be effective. The quality of
the Ivy League schools. One can well understand
Brown's student body shot up in the 1970s as a con-
why Harvard would want Dartmouth to agree not to
sequence of its more flexible “new curriculum,”
offer merit scholarships, since by doing so
which cost the university nothing but proved highly
Dartmouth renounces one of the few means it has
attractive to many of that generation's independent-
for competing effectively with Harvard for exception-
minded students. And the talented students that
ally talented students. But why would Dartmouth
were thus attracted to Brown in turn made Brown
wish to go along?
more attractive to students of all interests, and
A plausible answer is that the no-merit-scholar-
allowed the university to leapfrog up the convention-
ship agreement permits the Ivy League schools as a
al status hierarchy.
group to improve the quality of their student bodies.
By virtue of the agreement, Dartmouth agrees not to
Coordination of Tuition and
Admissions Policies
compete with Harvard on the basis of price for students who are both talented and prosperous. As a
One method of competing for the best students
consequence (with apologies to Dartmouth, which is
is to offer special scholarships to individual students
a superb institution), the best of those students will
of exceptional promise. This form of competition is
generally go to Harvard, and Harvard will charge
pervasive among graduate schools in the arts and
them a high price. Then, with the proceeds of that
sciences. It was long rare, however, in undergradu-
high price, Harvard can offer large scholarships to
ate education. Indeed, from the 1950s until recently,
highly talented students who are relatively impecu-
various groups of institutions that compete with each
nious, who will also come to Harvard (since they
other for students—including, conspicuously, the Ivy
would be offered a deal that is no more attractive at
League schools—entered into formal agreements in
Dartmouth). This means that Harvard will get all the
which they committed themselves not to offer merit
most talented students, regardless of income. But
scholarships, but rather to offer scholarships based
this also means that Harvard will not need to reach
only on the students' ability to pay.
so far down in the talent hierarchy among prosper-
An important consequence of such agreements
ous students in order to fill its classes. Dartmouth will
is to lock in the status hierarchy among the schools
thus have no competition for the latter students, to
involved. Since, owing to the agreements, any given
whom it will be able to charge full tuition—since
student will pay the same tuition regardless of which
other schools in the Ivy League that might, in turn,
institution she attends, the incentives for strict strati-
compete on price for those students with Dartmouth
fication of students across institutions become very
have also agreed not to do so. And Dartmouth, too,
21
terms of talent.5
will be able to take advantage of the high tuition it
charges prosperous students to offer more scholar-
Conclusion
ships to the talented poor.
The end result is that Dartmouth consigns itself
Although the upper reaches of higher education
to a student body that is clearly below Harvard's on
are becoming increasingly stratified, the same does
the status hierarchy, but by this means ends up with
not seem to be happening at the lower levels of the
a higher-quality student body overall than it could
higher education system, and particularly where stu-
have obtained through unrestricted competition. In
dents are non-resident. Among these institutions,
effect, the Ivy League schools are agreeing to
education is increasingly becoming a commodity.
charge monopoly prices to talented students who
Moreover, the individual course—rather than the
are prosperous, and using the proceeds to improve
four-year degree—is becoming the basic unit of con-
their student bodies by attracting larger numbers of
sumption and of competition among institutions.
talented students who are impecunious.
Students take one course at one college and another course at another, as convenience dictates. And
This works, however, only because the Ivy
League schools as a group have some monopoly
then, when they have taken enough courses, they
power that derives from the stratification of higher
submit the accumulated credits for a degree at one
or another institution. As a consequence, the asso-
education. If there were other institutions outside
of the Ivy League's agreement that were close
ciative aspect of education at these institutions is rel-
competitors to in the eyes of applicants, the Ivy
atively inconspicuous. A commuter student taking a
single evening class from an institution has much
League schools could not get away with charging
monopoly prices to talented students who are
less interest in the qualities of the institution's other
prosperous. Indeed, the Ivy League schools were
students than does a student who is entering a fouryear residential college. The result is a system in
always frustrated that they could not induce
Stanford, which they saw as a close competitor, to
which large numbers of colleges and universities—
join their agreement.
public, nonprofit, and for-profit—compete for students intensely with respect to price, curriculum,
Since 1991, the courts and the Congress have
facilities, and faculty.
been struggling to decide whether the kind of pricefixing that the Ivy League has pursued should be
An interesting question is whether the increas-
permitted under the antitrust laws. From the preced-
ing technological sophistication of our society, which
ing discussion, we can see why the issue is so diffi-
is fueling the trend toward stratification among the
cult. For it depends, among other things, on the
elite educational institutions, will someday produce
virtues and vices of permitting our system of higher
technologies that make it less important for elite
education to become ever more highly stratified in
higher education to be a residential experience, and
22
hence remove much of the associative character of
they graduate? One plausible answer is that the
higher education. If that happens, then the elite insti-
donations are part of an implicit loan program,
tutions, like the non-elite institutions today, may
whereby the college effectively lends impecunious
come to be much less clearly stratified and more
undergraduates part of the cost of their education, in
highly competitive. Until that happens, however, it
return for the implicit promise that the students will
appears that we will have a dual system of higher
voluntarily repay the loan later if they are successful
education, with ever more uniform competition at the
in life. See Henry Hansmann, “The Role of Nonprofit
lower end and ever more hierarchical stratification at
Enterprise,” Yale Law Journal 89 (1980): 835, 859-
the top.
62, reprinted in Susan Rose-Ackerman, ed., The
Economics of Nonprofit Institutions (Oxford: Oxford
Endnotes
University Press, 1986). This means, of course, that
1
For more detailed analysis of markets for associa-
as the availability and flexibility of educational loan
tive goods and the organization of the firms that pro-
programs improves, the need for donative financing,
duce them, see Henry Hansmann, “A Theory of
and hence for the nonprofit form, diminishes.
Status Organizations,” Journal of Law, Economics,
and Organization 2 (1986):11930. See also Henry
3
Hansmann, The Ownership of Enterprise, chapter
large market share held by the public institutions is
10 (Cambridge, Mass.: The Belknap Press of
largely a product of the surge in the demand for
Harvard University Press, 1996). For analysis of
higher education that the country experienced in the
associative phenomena in primary and secondary
decades immediately following World War II. Until
education, see Dennis Epple and Richard Romano,
1950, the percentage of college and university stu-
“Competition Between Private and Public Schools,
dents enrolled in public institutions had long held
Vouchers, and Peer-Group Effects,” American
steady at roughly 50 percent. The post-war demand
Economic Review 88 (1998): 33-62.
surge was the result of the baby boom and the
There are several reasons for this. One is that the
simultaneous increase in the percentage of high
are, of course, other reasons why institutions
school graduates going on to college and graduate
of higher education are nonprofit. The most obvious
school. That increase in demand could not be
is to be able to attract donations. Why donative
accommodated by the existing private nonprofit
financing is so pronounced among these institutions
institutions, which lacked the capital necessary for
is an interesting question, particularly since nearly all
rapid expansion. Moreover, even if government had
of the donations that a college receives come from
offered them the capital—which to some extent it
its graduates – i.e., its (former) customers. Why don't
did—the private institutions also had very little incen-
the colleges just raise their prices, and get the
tive to expand, partly because of the associative
money from the students up front rather than after
phenomena we have been exploring here.
2 There
23
Consequently, the best way to triple the capacity of
Term Care in America (University of Chicago Press,
the country's higher education system—which is
1996).
what happened between 1950 and 1975—was for
the state governments simply to build and operate
5
the campuses themselves. But such a rapid surge in
Hansmann and Alvin Klevorick, “Competition and
demand is unlikely to be repeated in coming
Coordination in Markets for Higher Education and
decades.
Other Associative Goods” (working paper, 1998).
Another reason why public institutions will be
less important in the future is that, in education as in
health care, we are replacing supply-side subsidies
(such as the Hill-Burton capital grants to hospitals in
the 1940s and 1950s, and the federal grants for university construction that were common in the 1960s)
with demand-side subsidies (such as the Medicare
and Medicaid health insurance programs, and the
now-extensive federally funded student grants and
loans for higher education and the competitive
research grants given to faculty by the NSF and the
NIH). Since students and faculty can take those
demand-side subsidies to any institution they wish—
whether public, private nonprofit, or even private forprofit—the effective demand for private colleges and
universities is now high, and is likely to be much
higher in coming years as state governments, too,
move more toward demand-side subsidies.
4
Henry Hansmann, “The Changing Roles of Public,
Private, and Nonprofit Enterprise in Education,
Health Care, and Other Human Services,” in Victor
Fuchs, ed., Individual and Social Responsibility:
Child Care, Education, Medical Care, and Long-
24
These issues are discussed further in Henry