Content uploaded by Philip A. Trostel
Author content
All content in this area was uploaded by Philip A. Trostel on Mar 31, 2020
Content may be subject to copyright.
The Fiscal Impacts of College Attainment
Philip A. Trostel
Received: 17 December 2008
ÓSpringer Science+Business Media, LLC 2009
Abstract This study quantifies one part of the return to U.S. public investment in college
education, namely, the fiscal benefits associated with greater college attainment. College
graduates pay much more taxes than those not going to college. Government expenditures
are also much less for college graduates than for those without a college education. Indeed,
over an average lifetime, total government spending per college degree is negative. That is,
direct savings in post-college government expenditures (conservatively, about $85,000 per
four-year-equivalent degree over an average lifetime) are greater than government
expenditures on higher education (generously, about $74,500 per degree). Plus, the direct
extra tax revenues from college graduates alone (roughly $471,000 per degree over a
lifetime) are more than six times the gross government cost per college degree. The
average real fiscal internal rate of return on government investment in college students is
conservatively estimated to be 10.3%.
Keywords College education Fiscal Public finance Tax revenues Expenditures
Government investment Rate of return
Introduction
U.S. government support for higher education has, arguably, decreased sharply in recent
years.
1
In fiscal year 1984, nationwide net state funding for higher education was 4.1% of total
This paper was written while serving as a visiting scholar at the New England Public Policy Center at the
Federal Reserve Bank of Boston.
P. A. Trostel (&)
Margaret Chase Smith Policy Center, School of Economics, University of Maine, Orono, ME, USA
e-mail: philip.trostel@maine.edu; trostel@wisc.edu
P. A. Trostel
Wisconsin Center for the Advancement of Postsecondary Education, University of Wisconsin,
Madison, WI, USA
1
According to data from State Higher Education Executive Officers’ State Higher Education Finance
project, national real state and local appropriations for postsecondary education in FY 2005 were 3.6% lower
123
Res High Educ
DOI 10.1007/s11162-009-9156-5
state government spending (derived from data from the U.S. Census Bureau’s State and Local
Government Finances). In 1994, this proportion was 2.4%, and in 2004, it was 1.8%. Public
investment in higher education is evidently a falling priority.
2
This has fueled an increasing
number of studies highlighting the benefits of investment in college education. The private
benefits of college attainment (higher earnings, lower unemployment, better health, etc.) have
been overwhelmingly demonstrated and widely publicized.
3
The wider social benefits from
higher education attainment (higher volunteering and civic participation, lower crime, etc.)
are not quite as overwhelmingly demonstrated and known, but have become increasingly
publicized in recent years.
4
Prior to this literature, the argument for public support of higher
education was based largely on the presumed existence of beneficial externalities that were
generally vague, nebulous, and unquantified.
This study complements this growing literature on the wider private and social values of
higher education attainment by quantifying the fiscal impacts of college attainment. That is,
this study focuses on one specific public benefit of higher education, the greater tax revenues
from and reduced government spending on college graduates. The broad unanswered issue
that this study addresses in part is the fiscal rate of return to public investments in higher
education. As noted above, there is a huge literature on the private monetary rate of return to
investments in education. These studies have generally found a high real rate of return (10%
or so in the U.S.). A growing literature also shows significant non-monetary returns to
education, both in private and social terms. But there is relatively little work quantifying the
returns to government investments in higher education. The fiscal impacts of college
attainment are starting to be quantified in some relatively recent studies, but mostly only in a
rather piecemeal and superficial way. Given that 2005 U.S. National Income and Product
Accounts indicate net federal, state, and local government investment in higher education was
$109 billion, i.e., 1.0% of national income, this is an important unanswered policy question.
This study addresses a crucial first part of the broader issue of the fiscal rate of return to
public investments in higher education. It quantifies the fiscal impacts of higher education
attainment. This project is narrower than the broad question for two reasons. The first
reason is that educating students in a state does not necessarily create a corresponding
increase in the state’s education attainment. Many college graduates migrate to other states,
thus creating an interstate fiscal externality (see Strathman 1994). This issue, however, has
been quantified in Trostel (2007), thus it is straightforward to take this into account in this
study. The second reason is that changes in public funding for higher education do not
necessarily induce proportionate changes in college attainment. The causal effect of state
funding for higher education is unclear. Moreover, the marginal and average effects
probably differ. However, although this issue has not been tested directly, recent work by
Bound and Turner (2007) suggests a strong causal effect of public funding on college
attainment, and that the effect is close to proportionate.
Footnote 1 continued
than in FY 2000 (these appropriations have increased in more recent years, though), while there were
increases in state income and in the number of college students during this period. For detailed discussion of
this issue see Longanecker (2006) and Trostel and Ronca (2009).
2
This is at least partly due to increasing fiscal pressures from slow growth in state tax bases and increasing
state spending on Medicaid and corrections (see Hovey 1999; Kane et al. 2005).
3
A few of the many examples are Card (1999), Mortenson (2001), Harmon et al. (2003), Baum and Payea
(2004), Institute for Higher Education Policy (2005), and Barrow and Rouse (2005).
4
Some examples are Haveman and Wolfe (1984), Wolfe and Haveman (2003), Baum and Payea (2004),
Lochner and Moretti (2004), McMahon (2004), Rizzo (2004), Topel (2004), and Institute for Higher
Education Policy (2005).
Res High Educ
123
While this is not the first paper to show fiscal benefits from public investment in college
students, this study quantifies these effects more completely and precisely than in the small
previous literature. Unlike most of this literature, practically all of the important direct
fiscal benefits of college attainment are systematically quantified here. Also unlike most of
the small literature, this study accounts for the timing of the fiscal effects. This study also
takes advantage of a significantly better dataset than has been used in previous work, thus
providing better estimates of the effects on various tax revenues and various government
expenditures. In addition and unlike the previous work, this study estimates the effects
from different types of college degrees (i.e., associate’s, masters, etc.), separates state and
local fiscal effects from federal fiscal effects, and carefully examines the fiscal cost of
public support for higher education. Finally, this study is the first to calculate a fiscal rate of
return to government investment in college education.
Consistent with the small previous literature, public support for higher education is evi-
dently a sound investment of tax dollars. Indeed, the fiscal payoff to public investment in
college students appears to be greater than suggested in the earlier work (although the limited
previous work is not directly comparable). Greater college attainment has numerous positive
fiscal repercussions: increased tax revenues from increased income and reduced low-income
tax credits, and decreased expenditures on Medicaid, Medicare, public healthcare, correc-
tions, supplemental security, unemployment compensation, worker’s compensation, public
assistance, food stamps, the WIC program, housing subsidies, energy assistance, transpor-
tation assistance, child-care assistance, and school lunches. Added together, there is a sizable
payoff to the government. The average real internal rate of return on government investment
in college education appears to be, conservatively, at least 10%. The fiscal rate of return to the
federal government is particularly high; conservatively, about 25%. The average real fiscal
rate of return to state and local governments, which provide most of the public funding for
higher education, is, conservatively, about 3%.
Even if the lack of careful targeting of public support for college education causes its
marginal effect on degree attainment to be, say, only half of proportionate, and thus the
marginal rate of return to be half of the average rate, the fiscal return to additional public
investment in higher education appears to be relatively high. Moreover, this project only
quantifies the direct government fiscal benefits from college attainment in public colleges.
It does not measure indirect effects on tax revenues and government expenditures through
higher education’s effect on economic growth. The estimated fiscal return also does not
include any economic benefits from graduates from private colleges, publicly sponsored
university research, from university public service and extension activities, or from the
effect of public colleges and college education on entrepreneurial activity and job creation.
Various social benefits such as higher civic involvement, lower crime, and greater toler-
ance are not quantified either.
Literature Review
As noted above, there is some, but not much, previous research that has highlighted some
of the fiscal benefits of college attainment.
5
Moreover, this literature only quantifies some
5
There is also some similar literature that has quantified the long-term fiscal impacts of government
investment in pre-school programs (e.g., Currie 2001; Heckman and Masterov 2004; Lynch 2004; Belfield
et al. 2006), and some recent literature that has quantified the fiscal effects from high school completion
(e.g., Krop 1998; Vernez et al. 1999; Goldhaber and Player 2003; Brady et al. 2005; Levin et al. 2007).
Res High Educ
123
of the various fiscal effects, and most of this literature does not account for the potentially
important issue of the timing of the fiscal effects. Most of the studies discussed below only
noted some of the fiscal benefits of degree attainment as a part of their work on related
issues. Unlike the present study, quantifying the average fiscal effects of college attainment
was not the primary purpose of previous studies.
Mortenson (1994) estimated amounts of federal income taxes paid by education-
attainment categories by combining published reports on average income by education and
average taxes paid by income level. This study did not estimate the additional revenues
created by college attainment, though.
Trostel (1996) used a stylized model to show that government spending for education
can lead to additional tax revenues that more than offset the spending in the long run, i.e.,
government spending for education can be more than self financing. But this study also did
not estimate the additional revenues created by college attainment.
Trostel (2003) estimated average additional state- and local-government tax revenues
from attaining a bachelor’s degree by deriving an average tax rate and applying it to an
estimated average income difference between those with a bachelor’s degree and those
with a high school diploma as their highest academic qualification. This study also esti-
mated government cost savings in unemployment insurance benefits using the same
approach. But this study did not account for the timing of the additional tax revenues and
reduced unemployment insurance.
Baum and Payea (2004) estimated average additional federal-, state-, and local-gov-
ernment tax revenues from attaining various education qualifications by combining pub-
lished reports on average income by education qualification and average federal, state, and
local taxes paid by income level. This study also did not account for the timing of the
additional tax revenues.
Institute for Higher Education Policy (2005) showed significant differences in the
receipt of public assistance between those with a bachelor’s degree and those with a high
school diploma as their highest academic qualification. Although significant government
cost savings were suggested, they were not quantified in the report.
Brady et al. (2005) contained estimates of additional state- and local-government tax
revenues and reduced spending on welfare and incarceration from increased education
attainment at various levels. The estimates were for the state of California. Like some of
the other studies, this report estimated an average tax rate and applied it to estimated
average incomes across education levels. This report also used average spending amounts
per welfare recipient and per prisoner and applied them to estimated incidence rates across
education levels. But also like the other studies, it did not account for the timing of the
fiscal effects. Moreover, although this study quantified the fiscal impacts of different
aggregate education-attainment scenarios, it did not report estimates of the impacts per
college degree.
The most in-depth previous work on the fiscal effects of college attainment is the RAND
study (Krop 1998; Vernez et al. 1999). This study estimated almost all of the important
direct fiscal benefits of attaining bachelor’s degrees (as well as for high school completion
and some college less than a four-year degree) using data from the Survey of Income and
Program Participation. Also, unlike the other previous literature, it accounted for the
timing of the fiscal benefits. The emphasis of the RAND study, however, was on the fiscal
implications of narrowing racial gaps in education attainment. The overall fiscal effects of
college attainment were not reported.
The present study goes beyond the RAND study and the rest of the literature in
numerous ways. Like the RAND study, this study estimates many of the direct fiscal
Res High Educ
123
benefits of college attainment. Specifically, it quantifies the fiscal effects in the RAND
study plus the effect on worker’s compensation and the effect on public healthcare
spending due to the lack of health insurance. Also like the RAND study, it accounts for the
timing of the fiscal effects. In addition, this study uses a better dataset than used in previous
work. Recent data from the U.S. Census Bureau’s Current Population Survey (CPS)
provide better estimates of effects on tax revenues, as well as effects on Medicare and
Medicaid spending. The CPS is also a much larger sample than that in the RAND study,
thus enabling reasonably precise estimation of fiscal effects in all individual states. This
study also estimates separate effects from different levels of college attainment (i.e.,
associate’s and bachelor’s degrees as well as advanced degrees), calculates the fiscal cost
of publicly supporting postsecondary education, and separates federal fiscal effects from
state and local fiscal effects. Also, this study is the first to estimate a fiscal rate of return to
government investment in college students.
Methodology
Data
The main data source for this study is the 2006 March Annual Social and Economic
Supplement of the CPS, a yearly sample of about 210,000 non-institutionalized individ-
uals. These data are for calendar year 2005. Because this study focuses on outcomes after
high school, only data from those age 19 and older are analyzed. Age is truncated at 80,
thus only observations from those 79 and younger are analyzed in this study. This yields a
sample of 136,514 observations. The CPS measures college attainment in terms of degrees
(except for those who have some college but no degree). Thus, this project focuses on the
fiscal returns from earning degrees. It would be preferable to quantify the effects in terms
of credit-years of college education, but the data do not allow it. Holders of professional
and doctorate degrees are small proportions of the sample (1.4 and 1.2%, respectively).
Thus, these degrees are lumped together to reduce the problem of small cell sizes.
Timing
The fiscal benefits of public investment in college students occur well after the fiscal cost.
College students also generally earn less and pay less in taxes while in school. Because
these fiscal opportunity costs occur up-front, they are relatively important in present dis-
counted value. These obvious and important points are frequently neglected in the nascent
literature. Most of the small previous literature also neglects the fact that the college
earnings premium, and thus the college fiscal premium, is not constant over the lifecycle. It
is well known that the earnings premium is the smallest immediately after graduation, and
then increases gradually at a decreasing rate. Figure 1shows average labor earnings in
2001–2005 (in 2005 dollars) for each age from 25 through 64 for each education quali-
fication. The average lifecycle of labor earnings is clearly steeper early in the careers of
those with the most education.
6
Although the average bachelor’s degree premium for those
within the ages of 25–64 in 2001–2005 is $25,122 annually, the premium is clearly much
less than this figure early in the work lifecycle.
6
Following standard practice, the absence of cohort effects is implicitly assumed. That is, average earnings
across ages in a given year are implicitly assumed to reflect average earnings over individual’s lifecycles.
Res High Educ
123
Instead of following the frequent implicit assumption of constant degree premia over the
lifecycle, this study quantifies the degree premia over the lifecycle as the differences in
lifecycle paths estimated from a fourth-order age polynomial (following Murphy and Welch
1990). Although the CPS sample is big, it is not big enough to prevent some relatively large
sampling variation in the averages at each year of age. Such sampling variation is seen clearly
in Fig. 1, particularly for professional and doctorate degrees because they are a relatively
small fraction of the sample. Moreover, the problem is more severe when using fewer years
of data. Thus, an age polynomial is used to capture life-cycle variation while smoothing the
sampling variation. In other words, rather than follow the usual practice of calculating the
degree premia as simply the differences in the averages across education levels, this study
first uses a simple regression analysis to predict the smoothed averages over the life cycle for
each education level, and then calculates the degree premia as the differences in these
averages across the life cycle and education levels.
Specifically, each fiscal amount X
Da
for each education qualification, D, at each age, a,
is first estimated as
X^
Da ¼b^
0þb^
1aþb^
2a2þb^
3a3þb^
4a4;
where the bs are estimated from an ordinary least-squares regression for each variable. An
example is illustrated in Fig. 2which charts estimated state income taxes paid at each age
by high school graduates (and no college) and bachelor’s degree holders (without an
advanced degree). Each degree premium, P
D
, is then calculated as
PD¼X79
a¼19 X^
Da X79
a¼19 X^
Ca;
where Cis prerequisite qualification (i.e., the prerequisite qualification for associate’s and
bachelor’s degrees is a high school diploma or equivalent, and the prerequisite qualification
for master’s, professional, and doctorate degrees is a bachelor’s degree). This is the
Average Life Cycle of Labor Earnings in 2001-05
$0
$20,000
$40,000
$60,000
$80,000
$100,000
$120,000
$140,000
25 30 35 40 45 50 55 60 65
Age
Average Labor Earnings
High School Associate's Degree Bachelor's Degree Master's Degree Professional & Doctorate Degree
Fig. 1 Average life cycle of labor earnings in 2001–2005
Res High Educ
123
cumulative difference between the two sets of plots (negative at first, then positive for the
rest of the life cycle) shown in Fig. 2. And the present discounted value of each degree
premium at the start of college at age 19, PV
D
,is
PVD¼X79
a¼19 1þrðÞ
a19ðÞ
X^
Da X79
a¼19 1þrðÞ
a19ðÞ
X^
Ca;
where ris the discount rate.
This study makes the simple assumption that the average career paths of graduates are
those of traditional students; that is, students progressing straight from high school to college
and from matriculation to graduation in the standard number of years. To be specific, the work
career is assumed to begin at age: 19 for high school graduates, 21 for associate’s graduates,
23 for bachelor’s graduates, 25 for master’s graduates, and 27 for professional and doctorate
graduates. Associate’s and master’s degrees are assumed to take two additional years of
education, while bachelor’s, professional, and doctorate degrees are assumed to average four
additional years of education. This study also makes the simple assumption that students
create no fiscal impact while in college other than the direct public cost of higher education.
That is, college students are assumed to pay no taxes, and to receive the average level of
social-insurance payments during college as before and after college.
7
$0
$500
$1,000
$1,500
$2,000
$2,500
19 23 27 31 35 39 43 47 51 55 59 63 67 71 75 79
Annual State Income Taxes Paid
Age
Estimated State Income Taxes over the Adult Life Cycle
High School Bachelor's Degree
Fig. 2 Estimated state income taxes over the adult life cycle
7
To be specific about assumed social-insurance receipts during college, degree holders are assumed to
receive the level of benefits received by average graduates with: high school diplomas at age 19, associate’s
degrees at age 21, bachelor’s degrees at age 23, master’s degrees at age 25, and the interpolated values at
ages 20, 22, 24, and 26. For example, the imputed social-insurance benefits during college for a master’s
degree is B
19
H
?(B
19
H
?B
21
A
)/2 ?B
21
A
?(B
21
A
?B
23
B
)/2 ?B
23
B
?(B
23
B
?B
25
M
)/2, where B is the average
annual level of benefits, the superscripts H, A, B, and M, respectively, denote high school, associate’s
degree, bachelor’s degree, and master’s degree, and the subscripts denote age.
Res High Educ
123
Obviously many college (and high school) students take more than the standard num-
bers of years to graduate (and some take less than the usual number). Thus, for many
graduates the benefits of college attainment occur later than here, and the assumption that
students are traditional on average overstates the present value of the fiscal benefits. On the
other hand, many students work while in college and/or work while taking time out from
college, and hence pay taxes before graduation. Indeed, presumably these are the main
reasons why many students take longer than usual to graduate. Thus, the assumption that
students are traditional on average also understates the total present value of the fiscal
benefits. The implicit assumption here is that the latter effect largely offsets the former. To
try to quantify these opposing effects would add more in complexity and uncertainty than
in precision. But it should be clearly acknowledged that this study really quantifies the
fiscal impact of traditional college students.
Tax Rates
8
The typical approach in the nascent literature assumes that taxes are a constant percentage
of income. Average tax rates, however, are not constant across education categories.
Income taxes are generally progressive. Payroll, property, and sales taxes are generally
regressive on an annual basis. But most of these are easy to account for using recent CPS
data. Beginning in 2005, the CPS has included estimates of individuals’ federal income and
payroll taxes and state and local income and property taxes.
9
State and local average sales (and excise) tax rates across education categories are
computed using data generated by the Institute on Taxation and Economic Policy’s Mi-
crosimulation Tax Model (McIntyre et al. 2003). To be specific, McIntyre et al.’s estimates
of average sales and excise tax burdens for each income quintile in each state are matched
with individuals’ incomes (excluding transfers) in the CPS. Although the McIntyre et al.
estimates are based on 2002 state tax laws and 2000 income data, this method is far
superior to simply applying a uniform average sales tax burden across education catego-
ries. McIntyre et al. show substantial within-state variation in effective average sales tax
rates.
Causation
A potentially important problem facing any effort to quantify the effects of educational
attainment is the issue of causality. Although earnings are highly correlated with education
attainment, this does not prove that more education causes earnings to be greater. Higher-
ability and/or higher-motivation individuals generally obtain more education and have
greater earnings potential independent of their levels of education. Education does not
necessarily cause earnings or the other outcomes correlated with it. The observed corre-
lations may be the result of omitted-variables bias (also often referred to as ability bias in
this context).
A large literature has developed to try to identify the causal effect of education on
earnings.
10
A growing literature has also developed trying to identify the causal effect of
8
With the exception of payroll taxes, the burden of taxes is assumed to fall on the payers. The real burden
of employers’ share of federal payroll taxes is assumed to be borne by workers.
9
See O’Hara (2004) for information on the CPS procedures used to estimate individuals’ tax burdens.
10
See Card (1999) and Harmon et al. (2003) for recent reviews of this literature.
Res High Educ
123
education on health, mental health, and mortality.
11
Lochner and Moretti (2004) attempted
to identify the causal effect of education on criminal behavior, arrests, and incarceration.
Despite the plausibility of the ability-bias hypothesis, these literatures have generally
demonstrated that the observed correlations are evidently causal effects of education. In
fact, many, if not most, of the studies in these literatures find slightly larger causal effects
of education than the simple correlations. Thus, it appears that the simple correlations are
not misleading; if anything, they appear to be on the conservative side.
This does not mean that the issue of causation can be completely dismissed here. This
study examines the correlations between higher education and various outcomes, such as
unemployment, welfare participation, and incarceration where causation has either not
been tested or tested to only a limited extent (but, as will be shown below, these fiscal
effects are relatively small anyway). Thus, a potentially important limitation of this project
is that it does not demonstrate causal effects of higher education. Unfortunately, the data
used in this project are insufficient to allow for causality testing. Previous research on
several different outcomes, however, is suggestive that this may not be an important
limitation.
College Earnings Premia
The discussion thus far implicitly assumes that college premia are constant. That is, the
average earnings differentials (and differentials in other outcomes) between education
qualifications are independent of the relative numbers of people with different education
levels. A simple supply-and-demand framework, however, suggests that this assumption
may be problematic. An increase in the relative supply of college-educated labor pre-
sumably creates downward pressure on their relative earnings. Moreover, a relative
increase in the supply of college-educated labor presumably implies digging deeper into
the talent pool of potential graduates. Thus, the college premia may not constant as the
proportion of college graduates varies. This is an old concern dating back to the 1976 book,
The Overeducated American, by Richard Freeman.
Contrary to the predictions in the 1970s, the return to higher education did not fall as
more and more Americans obtained college degrees. In fact, the economic return to higher
education rose in the 1980s and early 1990s, and has remained roughly stable since then.
12
Card and Lemieux (2001) and Fortin (2006), however, found that a relative increase in
college graduates in a state causes a statistically significant negative effect on the state’s
college wage premium. On the other hand, using a less-restrictive framework, Trostel
(2007) found this effect to be miniscule and not statistically different from zero. Juhn et al.
(2005) also found the effect on the college earnings premium to be very small. Although
the issue is not yet settled and is the subject of ongoing research, the effect of the relative
supply of college-educated labor on the college earnings premium appears to be no larger
than a small second-order effect.
Interstate Migration
As mentioned earlier, interstate migration of college graduates can cause some state and
local investments in college students to end up creating fiscal benefits in other states. A
11
See Groot and Maassen van den Brink (2004, unpublished manuscript), Cutler and Lleras-Muney (2006),
and Chevalier and Feinstein (2006) for recent reviews of this literature.
12
For surveys of this literature, see Katz and Autor (1999) and Autor et al. (2005).
Res High Educ
123
state’s production of college graduates does not necessarily have a corresponding impact
on the state’s college attainment. Thus, the fiscal return on a state’s investment in higher
education is reduced by the extent of the net interstate migration of its college graduates.
The national fiscal return is obviously unaffected by such interstate migration, but this
interjurisdictional spillover reduces the fiscal return to individual states.
Recent research by Trostel (2007) estimates the extent of this interstate spillover of
college graduates and thus quantifies the extent that fiscal return to individual states needs
to be adjusted downward. That study indicates that the average net loss of a state’s new
bachelor’s degrees to other states is about 7.1%. For new associate’s degrees, the net
interstate leakage is estimated to be about 3.4%. For new master’s degrees, the net leakage
appears to be about 7.9%. The net leakage of professional and doctorate degrees to other
states is roughly 9.6%.
13
These adjustments to states’ fiscal benefits of college attainment
are applied only to the end calculations of the overall post-college fiscal effects and the
fiscal rate of return. Thus, it should be kept in mind that these adjustments are not included
in the initial tables showing the various fiscal benefits of college attainment.
The Effect of Public Support
Perhaps the most problematic issue confronting an effort to quantify the fiscal return to
public investment in college students is the causal effect of public support on college
attainment. Many graduates who take advantage of public financial support would have
still earned their college degrees without the public support. Indeed, it has been persua-
sively argued that public subsidies for higher education often benefit those who would have
gone to college anyway.
14
Thus, the causal effect of public support on college attainment
may be significantly less than suggested by the correlation between public support for
higher education and the number of graduates from public colleges. In other words,
because public higher education subsidies are generally not well targeted at those on the
margin of college attendance, the marginal fiscal effect per public dollar invested in higher
education may be substantially less than the average fiscal effect.
Although this issue is not addressed directly in a recent study by Bound and Turner
(2007), their results suggest a roughly proportionate marginal causal effect of state support
for higher education on bachelor’s degree production in the state. Their findings indicate that
the average relationship between public support and bachelor’s degree attainment is not
misleading about the causal effect. To be specific, they find that exogenous increases in the
number of potential college graduates in a state (high school graduates 4 years earlier)
increases state funding for higher education by about only about 60% of the increase in the
number of potential college students (i.e., funding per student falls by 40%), and it also
decreases the number of bachelor’s degrees awarded relative to the number of potential
college graduates in the state by roughly 40%. Thus, the natural experiment created through
changes in cohort size indicates that bachelor’s degree attainment in a state changes about
13
These adjustments probably err on the high side (i.e., conservative in estimating the fiscal return to
states). Trostel (2007) also found that the net interstate leakage of new college graduates is evidently less for
the case of public colleges than for private colleges. With the exception of professional and doctorate
degrees, there appears to be no net interstate loss of new graduates from public institutions. The net loss of
new professional and doctorate graduates to other states is evidently about the same for public and private
institutions. Because most state support for college students is clearly directed toward those in public
institutions, the emphasis in this study is on graduates from public colleges. Nonetheless, the more con-
servative approach is taken.
14
For example, see Hansen and Weisbrod (1969) and Fernandez and Rogerson (1995).
Res High Educ
123
proportionately with state funding per potential college graduate. Hence, it tentatively
appears that the average correlation between public support and college attainment is not
misleading about the causal impact. But unless additional evidence can confirm Bound and
Turner’s result, uncertainty about the magnitude of the causal effect of public funding on
degree attainment remains an important caveat to keep in mind for the following results.
Tax Revenues
Table 1presents estimated lifetime (actually, through age 79) taxes across education
categories. This table also shows these lifetime taxes paid in present discounted value (at
the start of college at age 19) using a 3% real interest rate. In addition, they show the total
and present-value lifetime degree premia in taxes; that is, the differences in lifetime tax
paid for each degree level relative to that paid by high school graduates (bachelor’s
Table 1 Estimated lifetime taxes across education categories
High school Associate’s
degree
Bachelor’s
degree
Master’s
degree
Professional and
doctorate degree
State and local
Income taxes
Sum $34,044 $50,241 $89,667 $119,138 $168,449
Present value $15,898 $23,378 $39,760 $49,146 $70,521
Degree premium—sum $16,197 $55,624 $29,470 $78,782
Degree premium—PV $7,480 $23,862 $9,386 $30,761
Property taxes
Sum $88,536 $112,789 $123,078 $147,270 $146,109
Present value $37,549 $46,791 $50,256 $58,025 $55,632
Degree premium—sum $24,253 $34,542 $24,192 $23,031
Degree premium—PV $9,242 $12,707 $7,768 $5,375
Sales taxes
Sum $57,266 $71,554 $85,307 $93,802 $120,907
Present value $29,398 $35,427 $40,431 $42,921 $52,261
Degree premium—Sum $14,288 $28,042 $8,494 $35,600
Degree premium—PV $6,029 $11,033 $2,490 $11,830
Federal
Income taxes
Sum $111,132 $194,896 $367,132 $466,111 $733,428
Present value $47,255 $82,866 $155,699 $190,303 $300,808
Degree premium—sum $83,764 $255,999 $98,980 $366,297
Degree premium—PV $35,611 $108,444 $34,604 $145,109
Social security payroll taxes
Sum $169,398 $223,516 $289,111 $322,247 $482,088
Present value $86,642 $110,896 $138,423 $148,746 $209,136
Degree premium—sum $54,118 $119,713 $33,136 $192,977
Degree premium—PV $24,254 $51,781 $10,323 $70,714
The estimates are derived using CPS data in 2005 to estimate the average amount of each variable at each
age between 19 and 79 and summing. Present values are calculated using a 3% real interest rate
Res High Educ
123
graduates in the advanced degree cases). As discussed earlier, these estimates of lifetime
taxes are calculated assuming that no taxes are paid from ages 19 through 20 for associate’s
degrees, from 19 through 24 for master’s degrees, etc.
As expected, college education creates substantial state and local tax revenues. Each
bachelor’s degree leads to roughly $55,600 in additional state income taxes over a lifetime.
In present value using a 3% discount rate, bachelor’s degree holders pay, on average,
2.5 times as much state income tax as high school graduates without college. Holders of
professional and doctorate degrees pay 4.4 times as much state income tax in present value
as high school graduates not going to college. But college education does not create nearly
as much property tax revenues. In present value, the average bachelor’s degree holder pays
only 34% more property tax as the average high school graduate without college. Pro-
fessional and doctorate degree holders pay only 48% more property tax over their lifetimes
in present value than high school graduates with no college. Sales and excise taxes increase
with college attainment in similar proportions as property taxes. In present value, bache-
lor’s degree holders pay 38% more and professional and doctorate degree holders pay 78%
more sales and excise taxes as high school graduates. Lifetime total state and local taxes
increase by an average of nearly $55,000 per associate’s degree, more than $118,000 per
bachelor’s degree, more than $62,000 per master’s degree, and more than $137,000 for
each professional and doctorate degree. The present values (using a 3% discount rate) of
these degree premia in state and local tax revenues are, respectively, about $23,000,
$48,000, $20,000, and $48,000.
The degree premia in estimated lifetime federal income tax revenues across college
education categories are more dramatic than for state income tax revenues. More specif-
ically, both the absolute and relative sizes of the degree premia for federal income tax are
larger than for state income tax. Each bachelor’s degree leads to an additional $256,000 in
federal income tax revenues over a lifetime. In present value using a 3% discount rate,
bachelor’s degree holders pay 3.3 times as much federal income tax as high school
graduates. Holders of professional and doctorate degrees pay 6.4 times as much federal
income tax in present value as high school graduates. Because of the ceiling on income
subject to payroll taxation, the degree premia in lifetime Social Security payroll tax are not
as pronounced as for the income tax. In present value, holders of bachelor’s degrees pay
only 60% more payroll tax as holders of high school diplomas. Professional and doctorate
degree holders pay 141% more payroll tax over their lifetimes than high school graduates
with no college. In additional lifetime total federal taxes, an associate’s degree contributes
almost $138,000, a bachelor’s degree contributes almost $376,000, a master’s degree leads
to more than $132,000, and each professional and doctorate degree leads to more than
$559,000. The present values (using a 3% discount rate) of these degree premia in federal
taxes are about $60,000, $160,000, $45,000, and $226,000, respectively.
Government Expenditures
The effects of college attainment on various public-assistance programs are quantified in
Table 2. For each education category it shows estimated lifetime (through age 79) income
from six public-assistance programs: food stamps, school lunches,
15
various types of public
15
School lunches are for the family rather than for the individual. It seems appropriate to include the value
of school lunches for children.
Res High Educ
123
cash assistance, energy assistance, housing subsidies, and Medicaid.
16
The estimates are
for the cash values of these programs in 2005. The estimates reported in Table 2are
somewhat conservative in that they do not include any public costs in administering these
programs. That is, they show the value to the recipients rather than the total fiscal cost. In
addition, the estimates do not take into account any variation in the use of health care paid
with Medicaid. Given that health problems vary inversely with education attainment, the
public cost of Medicaid across college attainment should vary more than its insurance
value.
17
Not surprisingly, college attainment generally leads to lower levels of public assistance.
Each bachelor’s degree reduces lifetime receipts of food stamps by almost $7,100, chil-
dren’s school lunches by more than $2,400, public cash assistance by almost $1,500,
energy assistance by almost $400, housing subsidies by almost $300, and Medicaid by
more than $23,000. While the most important types of public assistance are those shown at
the top of Table 2, they are not the only ones. The CPS contains information on receiving
four other forms of welfare: Special Supplemental Nutrition Program for Women, Infants,
and Children (WIC), childcare assistance, transportation assistance, and participation in
work programs. Unfortunately, it does not contain information on their cash value or fiscal
cost. The lifetime incidence of these small in-kind public-assistance programs across
college qualifications is at the bottom of Table 2. The pattern of decreasing participation in
these programs as higher education qualification increases is similar to the other welfare
programs. Compared to the average person with a high school diploma as their highest
education in 2005, the average person with a bachelor’s degree is 21% as likely to receive
WIC, 45% as likely to receive childcare assistance, 24% as likely to receive transportation
assistance, and 30% as likely to participate in a work program in order to receive cash
assistance.
Most, but not all, public assistance is funded by the federal government, while most
federal public-assistance programs are administered through state governments. As a result
of the substantial intergovernmental transfers in welfare spending, a precise decomposition
of the fiscal premia into federal and state components is not possible. State and local
expenditure data are categorized differently than federal spending data (and they use
different fiscal years). State-level data are classified by function, while federal data are
classified by agency. Moreover, the state-level data are not disaggregated by program.
Approximate federal/state shares of welfare spending are constructed by combining the
U.S. Census Bureau’s Consolidated Federal Funds Report,Federal Aid to States, and State
and Local Government Finances.
18
This indicates that about 78.5% of total non-Medicaid
welfare spending is funded at the federal level in fiscal year 2004 (the last year that the
16
Medicaid benefits are for the family rather than for the individual, because unlike other forms of public
assistance, eligibility rules differ for parents and children, and it seems appropriate (and consistent) to
include the Medicaid benefits of children. Also, the CPS measures Medicaid benefits as its insurance market
value as opposed to the value of health care bought with Medicaid.
17
The CPS question about health condition suggests that this could be important. Among recipients of
Medicaid age 27 and older, 38% of those with only a high school education report their health to be very
good or excellent, compared to 54% for those with a bachelor’s degree as their highest qualification.
18
To be more specific, federal spending on welfare is derived as the sum of direct payments for food stamps
and housing assistance in Consolidated Federal Funds Report plus grants to state and local governments
from the Department of Agriculture’s Food and Nutrition Service (mostly child nutrition, food stamps, WIC)
and from the Department of Health and Human Services’ Administration for Children and Families
(Temporary Assistance to Needy Families, Head Start, and various other programs) in Federal Aid to States.
State and local welfare spending is derived as public welfare less vendor payments (i.e., Medicaid) in State
and Local Government Finances less the federal grants above.
Res High Educ
123
Table 2 Estimated lifetime public-assistance effects across education categories
High
school
Associate’s
degree
Bachelor’s
degree
Master’s
degree
Professional and
doctorate degree
Cash values
Food stamps
Sum $8,601 $4,188 $1,513 $1,458 $1,472
Present value $5,401 $2,489 $934 $935 $873
Degree premium—sum -$4,413 -$7,088 -$54 -$40
Degree premium—PV -$2,912 -$4,468 $2 -$61
School lunches
Sum $3,645 $2,376 $1,239 $928 $844
Present value $2,192 $1,388 $625 $430 $351
Degree premium—sum -$1,269 -$2,406 -$311 -$395
Degree premium—PV -$804 -$1,567 -$195 -$274
Cash assistance
Sum $1,714 $1,354 $248 $166 $561
Present value $1,130 $758 $173 $141 $373
Degree premium—sum -$360 -$1,466 -$82 $313
Degree premium—PV -$372 -$957 -$33 $200
Energy assistance
Sum $506 $273 $117 $126 $86
Present value $259 $150 $63 $61 $45
Degree premium—sum -$232 -$388 $9 -$31
Degree premium—PV -$109 -$196 -$2 -$18
Housing subsidy
Sum $386 $242 $123 $96 $95
Present value $211 $120 $63 $58 $57
Degree premium—sum -$144 -$263 -$27 -$27
Degree premium—PV -$91 -$148 -$5 -$7
Medicaid
Sum $42,227 $30,786 $19,191 $14,462 $14,154
Present value $22,912 $15,913 $9,751 $7,565 $6,727
Degree premium—sum -$11,440 -$23,035 -$4,730 -$5,038
Degree premium—PV -$6,999 -$13,161 -$2,187 -$3,025
Rates of receipt
WIC (%) 1.95 1.09 0.41 0.19 0.05
Degree differential (%) -0.86 -1.54 -0.22 -0.36
Childcare assistance (%) 0.40 0.47 0.18 0.06 0.06
Degree differential (%) 0.07 -0.22 -0.12 -0.12
Transportation assistance (%) 0.31 0.26 0.08 0.05 0.07
Degree differential (%) -0.05 -0.23 -0.03 0.00
Work program (%) 0.09 0.08 0.03 0.00 0.00
Degree differential (%) -0.01 -0.07 -0.03 -0.03
The estimates are derived using CPS data in 2005 to estimate the average amount of each variable at each
age between 19 and 79 and summing (the rates of receipt are averaged across ages, not summed). Present
values are calculated using a 3% real interest rate
Res High Educ
123
federal data are available). This percentage evidently varies considerably across states,
though. In fiscal year 2004, 59.9% of the national cost of Medicaid was picked up by the
federal government (Kaiser Commission on Medicaid and the Uninsured 2004). Thus,
about 40% of the estimated fiscal premia accrue to state governments, although this per-
centage also varies among states.
Table 3reports estimated family market values
19
of lifetime Medicare benefits across
education levels. It also reports estimated lifetime Social Security retirement benefits.
These estimates should be interpreted with additional caution. Compared to high school
graduates, college graduates generally live longer (see Lleras-Muney 2005). Thus, all else
the same, increasing education attainment translates into longer periods of receiving
Medicaid and Social Security benefits, and consequently, higher fiscal costs (as well as also
longer periods of paying income, property, and sales taxes, and hence, greater total fiscal
benefits). This study does not account for differences in mortality rates, thus this may cause
the fiscal costs of these retirement programs to be biased downward as college attainment
increases. This may not make much difference, though, in present value at age 19.
Moreover, for two reasons these estimates of the public costs of retirement programs might
not be biased downward as college attainment increases. As noted in the discussion of
Medicaid, the estimates do not take into account variation in the use of health care paid
with Medicare. Given the positive relationship between health and college attainment, this
should cause the public cost of Medicare to decrease with education attainment more than
Table 3 Estimated lifetime market value of Social Security across education categories
High
school
Associate’s
degree
Bachelor’s
degree
Master’s
degree
Professional and
doctorate degree
Medicare benefits
Sum $183,452 $173,941 $179,536 $167,034 $165,884
Present value $50,592 $42,617 $42,334 $35,515 $32,816
Degree premium—sum -$9,512 -$3,917 -$12,502 -$13,652
Degree premium—PV -$7,976 -$8,259 -$6,818 -$9,517
Social Security benefits
Sum $180,557 $174,288 $173,404 $161,467 $183,655
Present value $44,186 $38,558 $35,763 $30,014 $31,971
Degree premium—sum -$6,269 -$7,153 -$11,937 $10,251
Degree premium—PV -$5,628 -$8,423 -$5,749 -$3,792
Total benefits net of payroll taxes
Sum $194,611 $124,713 $63,829 $6,254 -$132,549
Present value $8,137 -$29,721 -$60,326 -$83,217 -$144,349
Degree premium—sum -$69,898 -$130,783 -$57,574 -$196,378
Degree premium—PV -$37,858 -$68,463 -$22,891 -$84,023
The estimates are derived using CPS data in 2005 to estimate the average amount of each variable at each
age between 19 and 79 and summing. Present values are calculated using a 3% real interest rate
19
As with Medicaid, Medicare benefits are for the family rather than for the individual because of the
importance of spousal benefits, and Medicare benefits are measured as their insurance market value.
Res High Educ
123
the insurance value of Medicaid decreases with education attainment.
20
Also, retirement
age and the start of receiving retirement benefits increase with education attainment.
Hence, it is not necessarily the case that those with more education receive greater
retirement benefits because of their longer life expectancies, particularly in present value.
Therefore, the net bias is unclear. But the results for Medicaid and Social Security are more
uncertain than for the other results.
The estimates reported in Table 3reveal relatively small differences in the lifetime
market value of Medicare across education levels. The differences are a little more
noticeable in present value (at age 19 using a 3% discount rate), though. Because average
retirement age increases with education attainment, each degree level creates present-value
fiscal savings in Medicare between $6,800 and $9,500 per degree. The levels and pattern of
lifetime and present-value Social Security benefits are similar to those for Medicare.
Average total lifetime (through age 79) benefits are roughly stable across education levels.
The average present value of these benefits, however, generally decreases slightly with
education levels. The present-value fiscal premium in Social Security for each degree is
between about $3,800 and $8,400. The sum of the fiscal effects of college attainment on
Social Security and Medicare benefits and taxes are shown at the bottom of Table 3. These
net-of-tax college premia in Social Security and Medicare are large. Each bachelor’s
degree creates net fiscal savings in Social Security and Medicare of almost $131,000 over a
lifetime. In present value, this net fiscal savings is about $68,500. Each professional and
doctorate degree creates a lifetime net fiscal savings of more than $196,000, and about
$84,000 in present discounted value.
Estimated average lifetime receipts of Supplemental Security Income (SSI),
21
unem-
ployment insurance (UI) compensation, and worker’s compensation across college quali-
fications are reported in Table 4. The patterns are generally similar to that for the various
public-assistance programs. The fiscal burden of the federal SSI program falls with college
attainment. As is the case for most of the public-assistance programs, most of the impact of
college education in SSI payments is for the undergraduate degrees. In present value,
average lifetime SSI for bachelor’s degrees holders receive is 26% as much as average
lifetime benefits for high school diploma holders with no college. In present value, each
associate’s degree reduces SSI payments by about $2,400, and each bachelor’s degree
reduces SSI by approximately $3,300. Advanced degrees, however, reduce the present
value of SSI by only an additional $100.
The incidence of unemployment decreases dramatically with college attainment. The
unemployment rate in 2005 among those with a high school education and no college was
4.55%. The unemployment rate of those with an associate’s degree as their highest edu-
cation qualification was 3.21%. The unemployment rate of those with a bachelor’s degree
and no advanced degree was 2.64%. The unemployment rate of master’s degrees holders
was 2.03%, and for professional and doctorate degrees it was 1.55%.
22
The relative
magnitude of the fiscal savings in UI is not as great as suggested by the differences in
20
The CPS question about health condition again suggests that this could be important. Among recipients
of Medicare, 33% of those with only a high school education report their health to be very good or excellent,
compared to 50% for those with a bachelor’s degree as their highest qualification.
21
The relationship between college education and work disabilities is similar to the relationship between
college education and health. CPS data indicate that for those aged 27 and older, 24% of those with only a
high school education report a disability or health problem that limits their ability to work, compared to 15%
of those with a bachelor’s degree as their highest qualification.
22
These unemployment rates across education categories are calculated for those within the ages of 27 and
79 using the CPS outgoing rotations groups.
Res High Educ
123
unemployment rates, though, because UI compensation is partially tied to pre-unem-
ployment earnings, and earnings are clearly related to college attainment.
23
These esti-
mates are derived under the assumption that no UI compensation is received while in
college. Given the simplifying assumption that degrees are earned by traditional full-time
students, it would be inconsistent to assume that some students receive UI benefits by
actively seeking employment after losing a job while in college. Unlike the fiscal degree
premia in SSI, the fiscal degree premia in UI is small for associate’s degrees, but relatively
large for advanced degrees. The present value of lifetime UI compensation for those with
bachelor’s degrees as their highest education qualification is 66% as much as that for those
with high school as their highest qualification. The present value of average UI compen-
sation for those with professional and doctorate degrees is 12% as much as that for those
with high school diplomas only. The present value of the lifetime fiscal premium in UI is
almost $1,200 per bachelor’s degree, and more than $1,800 per professional and doctorate
degree.
Fiscal savings from college attainment are also created in through worker’s-compen-
sation programs. In some states, though, worker’s compensation is an off-budget item.
Some states run worker’s compensation through private insurance companies dealing
directly with employers. The effects, however, are the same whether on- or off-budget.
Thus, worker’s compensation is treated as an implicit fiscal item in all states. These
estimates are derived under the same assumption as for UI compensation; that is, no
worker’s compensation is received while in college. The estimated fiscal degree premia in
worker’s compensation are relatively the largest for bachelor’s degrees and professional
and doctorate degrees. The present value of the lifetime fiscal premium in worker’s
Table 4 Estimated lifetime social insurance across education categories
High
school
Associate’s
degree
Bachelor’s
degree
Master’s
degree
Professional and
doctorate degree
Supplemental Security income
Sum $9,386 $5,141 $3,330 $3,116 $1,724
Present value $4,433 $2,077 $1,137 $1,038 $677
Degree premium—sum -$4,245 -$6,056 -$214 -$1,606
Degree premium—PV -$2,355 -$3,296 -$98 -$460
Unemployment compensation
Sum $6,689 $6,690 $4,846 $3,508 $880
Present value $3,443 $3,355 $2,262 $1,616 $417
Degree premium—sum $1 -$1,842 -$1,338 -$3,967
Degree premium—PV -$88 -$1,182 -$646 -$1,845
Worker’s compensation
Sum $4,226 $3,472 $2,591 $2,654 $781
Present value $1,671 $1,527 $1,073 $997 $300
Degree premium—sum -$754 -$1,635 $63 -$1,810
Degree premium—PV -$144 -$599 -$76 -$772
The estimates are derived using CPS data in 2005 to estimate the average amount of each variable at each
age between 19 and 79 and summing. Present values are calculated using a 3% real interest rate
23
Unemployment taxes are ignored in the calculations because employers bear the statutory liability.
However, the real incidence of the tax may fall on employees to at least some extent.
Res High Educ
123
compensation is about $600 per bachelor’s degree, and almost $800 per professional and
doctorate degree.
Table 5reports estimated fiscal impacts on two activities that could not be estimated
directly using CPS data: corrections and public healthcare. Combining data from the
Bureau of Justice Statistics’ Prisoners in 2005 (Harrison and Beck 2006), the Census
Bureau’s State and Local Government Finances: 2004–2005, and the Office of Manage-
ment and Budget’s Budget of the United States Government (Table 3.2) indicates that
corrections expenditure per inmate was $29,877 in fiscal year 2005.
24
Cost per federal
prisoner was $35,203, and cost per state and local prisoner was $29,538. A Bureau of
Justice Statistics report (Harlow 2003) using 1996 and 1997 data indicates that 0.115% of
the adult population with a bachelor’s degree or higher is incarcerated. Among those with
some college or with associate’s degrees, the incarceration proportion is 0.317%. For those
with a high school diploma only, the proportion is 1.191%, which is 10.4 times higher than
the proportion for bachelor’s graduates. Thus, the differential in corrections costs is $321
annually per bachelor’s degree [$29,877 9(0.0191 -0.0115)].
25
The estimates for
associate’s degrees use the incarceration probability for some college and associate’s
degrees, which should understate its fiscal premium somewhat (presumably holders of
associate’s degrees have a lower probability of incarceration than those with some college
but no degree). Estimates for bachelor’s and advanced degrees are lumped together. Given
the magnitude of the average incarceration cost for this group, this probably makes little
difference.
Unlike most of the other fiscal effects which accrue more at the federal level, the large
majority of the fiscal benefits from college education accrue to state and local govern-
ments.
26
Direct federal corrections expenditure is only 9.0% of the national total in 2005.
The issue of causation may be more problematic for corrections than for the other
college differentials. That is, those in prison may be disproportionately less educated
because they were incarcerated, and/or the underlying reasons for imprisonment and low
education are the same. Most of those in prison started there at college age or younger.
Lochner and Moretti (2004), however, present evidence that the effect of education on the
incidence of incarceration (and criminal behavior) is indeed causal. Moreover, Harlow
(2003) reports that many prisoners take advantage of college and vocational courses
offered in prisons, thus it is possible that the simple correlation between education
attainment and incarceration understates the causal effect.
In addition to Medicaid being negatively correlated with college attainment, the lack of
any health insurance is negatively correlated with college attainment. Moreover, the
uninsured impose significant fiscal costs, although it is difficult to identify these in gov-
ernment finance data. The fiscal costs created by the uninsured are implicitly small
24
This cost per prisoner is calculated using all corrections costs, including probation. Thus, the resulting
estimates of fiscal premia are for corrections costs rather than just prison costs. This interpretation imposes
the implicit assumption that the probabilities of being on probation are roughly proportional to the prob-
abilities of being incarcerated.
25
Moreover, prisoners (and the homeless) are not included in the CPS. CPS sample selection is not entirely
independent of college attainment. Thus, the average high school graduate pays disproportionately less tax
revenues than shown earlier. As a result, the preceding estimates of fiscal premia in tax revenues are
understated to the extent that the CPS sample is not completely random with respect to college education.
26
Separate estimates for federal and state corrections cost per inmate are derived using different federal and
state incarceration proportions. Adult population proportions in state and local prisons and jails are 1.122,
0.289, and 0.098% for high school diploma, college below a bachelor’s degree, and bachelor’s degree or
higher, respectively.
Res High Educ
123
fractions of various spending categories (Medicare, Medicaid, Hospitals, Public Health,
etc.). Hadley and Holahan (2003), however, provide an estimate of the governmental costs
of the uninsured. They estimate the total government cost per uninsured to be $823 (in
2005 dollars). The cost to the federal government is roughly $535 per uninsured, and the
cost to state and local governments is about $288 per uninsured. Given that healthcare
costs have generally risen faster than inflation, these estimates are probably on the con-
servative side.
The incidence of uninsurance across education categories can be computed from the
CPS. These estimates are reported in Table 5. Among those age 27 and older, 16.8% of
high school graduates with no college reported having no health insurance, more than
double the percentage of those with bachelor’s degrees without an advanced degree.
Table 5also reports estimates of the college premia in public healthcare costs from the
lack of health insurance by assuming an average cost of $823 per uninsured. However,
given the positive relationship between college attainment and health, these estimates are
particularly conservative.
27
That is, it is likely that the cost per uninsured is decreasing in
attainment, but this is not taken into account in the estimates. As is the case for most
public-assistance programs, most of the estimated fiscal impact on public healthcare occurs
for undergraduate degrees. The results in Table 5reveal that estimated average lifetime
public-healthcare cost for holders of associate’s degrees is 60% as much as for high school
graduates with no college. The average for bachelor’s degrees is 49% as much as the
average for high school diplomas. In present value, each associate’s degree reduces public-
healthcare cost by almost $2,300, and each bachelor’s degree reduces it by more than
$2,900.
Table 5 Estimated lifetime corrections and public-healthcare costs across education categories
High
school
Associate’s
degree
Bachelor’s
degree
Master’s
degree
Professional and
doctorate degree
Corrections costs
Sum $21,702 $6,166 $2,697
Present value $10,202 $3,102 $1,576
Degree premium—sum -$15,536 -$19,004
Degree premium—PV -$7,100 -$8,626
Uninsured percentage 16.8 10.5 8.3 5.0 5.7
Public-healthcare costs
Sum $9,220 $5,488 $4,548 $3,255 $3,715
Present value $5,811 $3,539 $2,892 $2,229 $2,536
Degree premium—sum -$3,733 -$4,673 -$1,293 -$833
Degree premium—PV -$2,272 -$2,918 -$664 -$357
Average corrections costs are estimated using incarceration rates from Harlow (2003) and the author’s
calculations of cost per prisoner using various sources. Corrections costs for bachelor’s degree are actually
for bachelor’s degree and higher. The estimates of uninsured percentages are derived using CPS data in 2005
from those between ages 27 and 79. The estimates of public-health care costs are calculated using the
estimated uninsurance rates and Hadley and Holahan’s (2003) estimate of governmental cost per uninsured.
Present values are calculated using a 3% real interest rate
27
Among the uninsured age 27 and older, 52% of those with only a high school education report their
health to be very good or excellent, compared to 65% for those with a bachelor’s degree as their highest
qualification.
Res High Educ
123
Public Cost per Degree
Government spending on higher education is taken from Table 3.16 of the National Income
and Product Accounts (NIPA). Arguably better data for state and local appropriations for
higher education are available from the State Higher Education Executive Officers’ State
Higher Education Finance (SHEF) project. Moreover, most government spending on higher
education is at the state level.
28
Nonetheless, this study relies on the NIPA data for two
reasons. First, the NIPA data include all federal spending on higher education, and measures it
and all state and local spending on higher education consistently. Second, the NIPA measure
is somewhat greater (specifically, 15.9% greater) than the SHEF measure, thus it produces a
more generous estimate of the public cost per degree and hence a more conservative estimate
of the fiscal rate of return. The main reason for the discrepancy between the NIPA and SHEF
measures of states and local funding for higher education is that the NIPA measure implicitly
includes expenditures financed from revenues from public college endowments. The SHEF
measure only includes state and local government appropriations for higher education. Thus,
in terms of appropriations, the NIPA measure overstates the fiscal cost (evidently by about
16%). Rather than evaluate the relative merits of the different underlying concepts of public
opportunity cost, this study simply chooses the more generous measure (i.e., more conser-
vative measure in showing the fiscal return).
29
The NIPA (and the SHEF) measure of the fiscal cost of college degrees is generous in
that it includes the costs of university research and service activities. Given that this study
only attempts to quantify the fiscal payoffs from attainment (rather than the fiscal payoffs
from the higher education sector), the costs of research and service should be not be
included. Data from the National Center for Education Statistics for academic year 2001
(the latest available) indicate that research and service accounts for 19.3% of total edu-
cational and general expenditure. On the other hand, the NIPA data are current expendi-
tures only. NIPA does not separate capital expenditures in higher education from primary
and secondary education. Data from the Census Bureau’s State and Local Government
Finances on public higher education expenditures for fiscal years 2002 through 2005
indicate that, at the state and local level, capital outlays are 13.5% as large as operating
expenditures. Thus, taking the opposing effects of the inclusion of research and service
costs and the exclusion of capital costs into account suggests that the measure of gov-
ernment spending on higher education is overstated by roughly 5.8%.
Data on college degrees for each state are calculated using the National Center for Edu-
cation Statistics’ Integrated Postsecondary Education Data System. Although most public
funding for higher education is clearly directed toward students in public institutions, sig-
nificant funding, particularly federal funding, also goes to students in private higher educa-
tion. Unfortunately, the government cost data cannot be separated into the amounts going to
students in public colleges and to students in private colleges. Thus, to err on the conservative
side this study compares public funding for higher education to degrees granted from public
institutions only. This imposes the implicit assumption that financial aid to students in private
colleges does not lead to any additional college attainment. The estimated fiscal return from
public investment in college education is also conservative in that it ignores any fiscal benefits
from those with some college but no degree. This group compared to high school graduates
28
The NIPA data for calendar years 2002–2005 indicate that the federal share is about 19%.
29
One could reasonably argue that expenditures financed from publicly owned endowments are public
contributions. One could also argue that some of these endowments are established through private
donations.
Res High Educ
123
with no college earns $5,652 more annually. The fiscal benefits from some college experience
without earning a degree are ignored because of the difficulty in assigning a fiscal cost to this
education category. Unlike new college degrees, determining the amount of new ‘‘some
college’’ per year is problematic.
Calculating cost per degree is also problematic because higher education costs are not
assigned by degree level. Moreover, degrees are not earned within a fiscal year. The latter
problem is not particularly troublesome, though. The multi-year nature of degrees should be
adequately taken into account by averaging government spending on higher education over
the relevant preceding years. This study matches degrees awarded in academic year 2005 to
average government funding for higher education in the preceding four fiscal years, 2002
through 2005 (thus putting the emphasis on four-year degrees rather than two-year degrees).
To deal with the former problem, instead of trying to assign separate costs to the
different degree levels, this study makes the simple assumption that each year of college
creates the same fiscal cost. As discussed below, this is somewhat conservative in esti-
mating the fiscal rate of return to public investment in college students. It is assumed that
associate’s and master’s degrees take two additional years of education, while bachelor’s,
professional, and doctorate degrees average four additional years. Associate’s and master’s
degrees thus count as half of a four-year degree. The resulting estimates of total, federal,
and state and local government costs per four-year-equivalent degree from public insti-
tutions are presented at the top of Table 6. Given that these degrees are assumed to take
four years, this table also reports the estimated costs per degree in terms of present value at
the beginning of college (at the beginning of graduate school for the advanced degrees).
The relative importance of the separate degree levels is account for by weighting the
corresponding fiscal effects. That is, the fiscal weights for the college premia at each
degree level are their proportions of total four-year-equivalent degrees.
30
The above framework does not account for the fiscal cost of each year of college
education increasing with the level of college education. The yearly fiscal cost of a doc-
torate degree is clearly greater than for an associate’s degree. Not accounting for this is
conservative in estimating the fiscal return to public investment in college education
because this then puts insufficient fiscal weight on the highest degrees, and the largest fiscal
impact per degree year is for professional and doctorate degrees, followed by bachelor’s
degrees. If the increasing fiscal cost with degree levels were taken into account, greater
fiscal weight would be placed on the degrees with the higher fiscal return per degree year.
Average Fiscal Rate of Return
Applying the fiscal weights just discussed to the fiscal premia for each degree estimated
earlier yields the estimated fiscal impacts per four-year-equivalent degree. These are
reported in Table 6. This table shows the weighted averages of the various fiscal effects
presented earlier. The middle columns report these weighted-average effects accruing at
the federal level, and the right two columns show these weighted-average effects accruing
to state and local governments.
31
As discussed earlier, the state-level fiscal effects shown in
30
Specifically, the fiscal weights are 19.7% for associate’s degrees, 65.3% for bachelor’s degrees. 10.2%
for master’s degrees, and 4.7% for professional and doctorate degrees.
31
To be consistent with this study’s measure of cost per degree, Table 6shows the present values of the
fiscal premia for the graduate degrees relative to the assumed beginning of graduate school (i.e., the present
values at age 23).
Res High Educ
123
Table 6are reduced by an average of 6.6% to account for interstate migration of college
graduates.
32
Table 6shows that the largest fiscal payoff is in the additional tax revenues. Indeed,
85% of the $556,000 estimated cumulative fiscal benefits over the lifetime of a four-year-
equivalent degree is from additional tax revenues. In terms of present value using a 3% real
discount rate, the additional tax revenues are 78% of the $253,000 total fiscal benefit per
degree. Although the effects of college attainment on various expenditures are small
relative to the effects on taxes, they are not small relative to government expenditure on
higher education. Indeed, the $84,700 sum of expenditure savings over a lifetime exceeds
the government cost of $74,500. The net public cost of higher education is negative. In
present value, however, the expenditure savings are less than the government investment
per four-year-equivalent degree from public institutions.
Moreover, it should be kept in mind that these effects are estimated under numerous
conservative consumptions. These estimates understate the fiscal effects for several rea-
sons. The estimates ignore fiscal benefits from college that does not lead to degrees.
Reductions in mortality rates associated with higher education are also ignored, as are
intergenerational effects (children of college-educated parents are much more likely to
become college graduates
33
and create resulting fiscal benefits). Savings in the
Table 6 Estimated lifetime fiscal effects per four-year-equivalent degree
Total Federal State and local
Sum Present
value
Sum Present
value
Sum Present
value
Cost -$74,573 -$71,378 -$14,007 -$13,407 -$60,566 -$57,971
Income taxes $290,296 $122,904 $237,819 $100,569 $48,741 $20,699
Payroll taxes $115,442 $49,528 $115,442 $49,528
Property taxes $38,170 $14,022 $35,439 $12,842
Sales taxes $27,373 $10,787 $25,357 $9,872
Welfare $10,218 $6,544 $8,018 $5,135 $2,073 $1,331
Medicaid $20,763 $12,021 $12,442 $7,203 $7,814 $4,534
Medicare $9,518 $10,619 $9,518 $10,619
Social Security $9,104 $9,248 $9,104 $9,248
Supplemental Security income $5,749 $3,129 $5,749 $3,129
Unemployment compensation $1,665 $1,054 $1,552 $988
Worker’s compensation $1,438 $506 $1,350 $476
Corrections $21,385 $9,726 $1,925 $876 $18,322 $8,351
Public healthcare $4,828 $2,974 $3,139 $1,933 $1,593 $983
Totals $481,376 $181,682 $389,149 $174,832 $81,675 $2,105
Internal rate of return (%) 10.3 24.8 3.1
These are the weighted averages of the degree premia reported in Tables 1,2,3,4, and 5. Present values are
calculated using a 3% real interest rate. State and local post-college fiscal effects are reduced to account for
interstate emigration of graduates using estimates from Trostel (2007), thus the totals exceed the sum of
federal and state and local
32
This number is the sum of the estimated net migration rates from Trostel (2007) times their respective
weights.
33
See, for example, Haveman and Wolfe (1995).
Res High Educ
123
administration of expenditure programs are not included. Some welfare programs are not
included in the calculations. And only the direct fiscal effects from college attainment are
captured (as opposed to potential indirect effects from higher education generally through
innovation and growth). Moreover, the estimates of government cost per degree are gen-
erous. Public costs of private college education are included, while those degrees are
ignored. Expenditures financed from public college endowment revenues are included, as
are expenditures on research and service activities. Also, the assumption that all years of
college impose the same public cost causes too little weight to be placed on the fiscal
benefits from degrees (professional and doctorate) that have the largest fiscal effects.
On the other hand, it should also be kept in mind that these fiscal effects might not be so
conservative for three reasons. First, the observed correlations might not be completely
causal effects. Second, the estimates for the effects on Social Security and Medicare
spending are particularly uncertain, and these are fairly large relative to the other spending
programs. And third, the estimates are for traditional college students, and the fiscal returns
are lower for older college graduates. This is shown by comparing the results in Table 6
and Fig. 3to those from an illustrative case of part-time degree attainment shown in
Table 7and Fig. 4.
34
The fiscal returns in this illustrative case of half-time college
attendance are slightly lower than in the full-time case.
The results reported in Table 6show that the largest fiscal benefits accrue in federal
programs. Moreover, the government cost is concentrated at the state level. Thus, the fiscal
return on investment in college education is much higher at the federal level than at the
state level. Comparing the left and middle columns of Table 6indicates that 72.5% of the
total fiscal premium accrues at the federal level. In present value, 74.4% of the total fiscal
premium is federal. But the federal government only picks up 18.8% of the estimated total
public cost. In total undiscounted lifetime effects, the federal fiscal benefit of $403,000 per
degree is almost 29 times greater than the federal cost of $14,000 per public-college
degree. The average fiscal impact at the state level is not nearly as dramatic, although the
net lifetime impact of state and local government investment in college students is still
positive. In total undiscounted lifetime effects, the state and local fiscal benefits of
$142,000 are 2.35 times greater than the state and local cost of $60,500 per degree from
public colleges. Again, it should be kept in mind that these fiscal effects are conservatively
estimated.
The bottom of Table 6reports estimates of the real average fiscal internal rate of return
implied by the cumulative fiscal effects. The state internal rate of return estimates take into
account average net interstate migration of college graduates. The estimated overall fiscal
rate of return is, conservatively, 10.3%. At the individual state level, the average internal
rate of return is estimated to be 3.1%.
35
Without the adjustment for interstate migration of
college graduates, the average state rate of return per public degree is 3.5%.
The time required to recoup the public investment in college students is relatively short.
Actually, given the high fiscal payoffs to college attainment shown earlier, perhaps this is
to be expected. For all levels of government combined, the public investment is recovered
in less than 9.5 years after traditional four-year-equivalent graduation. Public investment
34
The results in Table 7and Fig. 4are derived assuming half-time work and half-time college attendance
beginning at age 19. Thus, an associate’s degree is assumed to take 4 years, a bachelor’s degree takes
8 years (i.e., graduation at age 27), and so on. During college these students are assumed to pay half as much
taxes, receive half as much unemployment insurance on average, etc. as full-time workers.
35
The average state fiscal rate of return varies across states. Among the six New England states, for
example, the average fiscal rate of return ranges from 3.0 to 4.7%.
Res High Educ
123
Table 7 Estimated lifetime fiscal effects per four-year-equivalent degree earned part-time
Total Federal State and local
Sum Present
value
Sum Present
value
Sum Present
value
Cost -$74,573 -$67,398 -$14,007 -$12,660 -$60,566 -$54,738
Income taxes $279,079 $114,038 $228,751 $93,378 $46,830 $19,222
Payroll taxes $108,286 $44,069 $108,286 $44,069
Property taxes $37,540 $13,775 $35,026 $12,780
Sales taxes $25,510 $9,429 $23,773 $8,750
Welfare $9,500 $5,885 $7,454 $4,618 $1,923 $1,193
Medicaid $19,505 $10,887 $11,688 $6,524 $7,331 $4,098
Medicare $5,403 $6,880 $5,403 $6,880
Social security $4,150 $4,742 $4,150 $4,742
Supplemental security income $5,600 $2,993 $5,600 $2,993
Unemployment compensation $1,591 $981 $1,477 $915
Worker’s compensation $1,409 $475 $1,321 $445
Corrections $20,928 $9,304 $1,884 $838 $17,916 $7,976
Public healthcare $4,625 $2,788 $3,006 $1,812 $1,524 $920
Totals $448,553 $158,849 $362,215 $153,194 $76,556 $1,561
Internal rate of return (%) 9.5 20.3 3.1
These are derived assuming half-time work and half-time college attendance beginning at age 19. Present
values are calculated using a 3% real interest rate. State and local post-college fiscal effects are reduced to
account for interstate emigration of graduates using estimates from Trostel (2007), thus the totals exceed the
sum of federal and state and local
-$100,000
-$50,000
$0
$50,000
$100,000
$150,000
$200,000
$250,000
$300,000
$350,000
$400,000
$450,000
$500,000
19 23 27 31 35 39 43 47 51 55 59 63 67 71 75
Cumulative Fiscal Effect
Age
Cumulative Fiscal Effect per Four-Year-Equivalent Degree
Fig. 3 Cumulative fiscal effect per four-year-equivalent degree
Res High Educ
123
in a traditional college student is fully recovered just after age 31. At the state level,
however, it takes more than 24 years (age 46) to recover the state and local investment
(after accounting for net migration of graduates). Figure 3illustrates the time path of the
total fiscal effects per four-year-equivalent degree begun at age 19. The government
investment per degree reaches $90,000 at college graduation (government spending per
degree plus the reduced tax revenues while in college). That investment is recovered at age
31.5. The cumulative total fiscal premium per 4 year-equivalent degree reaches $100,000
at age 40, $200,000 at age 48, $300,000 at age 55, etc.
Conclusion
College education clearly creates substantial fiscal benefits for governments. This study has
quantified these fiscal effects of college attainment more thoroughly than in the small
previous literature. It quantifies more fiscal effects using a more systematic methodology
and better data than in previous work. This study also carefully accounts for the timing of
the fiscal benefits. Fiscal effects accruing to state and local governments are separated from
those accruing to the federal government. Unlike in any of the previous work, this study
estimates the separate fiscal effects from different types of college degrees, as well as their
overall weighted average. The fiscal cost per degree is also quantified, which allows a fiscal
rate of return to be calculated.
Consistent with the small previous literature, public support for higher education appears
to be a sound use of tax dollars. Although all of the fiscal effects of college education cannot
be estimated with great precision due to the numerous complicating factors, the return on
public investment is evidently substantial. The average real fiscal rate of return appears to be
at least 10%. Moreover, this estimate is conservative for numerous reasons.
Furthermore, the 10% estimate measures only the direct fiscal return from college
attainment. Indirect effects on tax revenues and government expenditures through higher
-$100,000
-$50,000
$0
$50,000
$100,000
$150,000
$200,000
$250,000
$300,000
$350,000
$400,000
$450,000
$500,000
Cumulative Fiscal Effect
Age
Cumulative Fiscal Effect per Four-Year-Equivalent Degree Earned Part-Time
19 23 27 31 35 39 43 47 51 55 59 63 67 71 75
Fig. 4 Cumulative fiscal effect per four-year-equivalent degree earned part-time
Res High Educ
123
education’s effect on economic growth are not included. The estimated fiscal return also
does not include any benefits from publicly sponsored university research, from university
public service and extension activities, or from the effect of public colleges and college
education on entrepreneurial activity and job creation. Nor does the return quantified in this
study include the value of various other social benefits such as knowledge creation and
dissemination, lower crime, higher civic participation, and so on. In addition, the 10%
estimate is the average rate of return. The marginal fiscal payoff to public support for
higher education targeted at those on the margin of college attendance is probably con-
siderably higher.
The caveats noted earlier should not be forgotten, though. If marginal public funding for
college students is not well-targeted toward those on the college-going margin, then the
marginal fiscal return is probably lower than the average return shown here. The magnitude
of the causal effect of public funding on college attainment is uncertain. Moreover, this
study could not prove that the observed correlations between college attainment and taxes/
spending are causal effects. Also, the preceding estimates are for traditional college stu-
dents. The fiscal return to public investment in non-traditional students is probably
somewhat lower. Issues such as these are left for future research.
Acknowledgments I am grateful for the feedback and support from New England Public Policy Center at
the Federal Reserve Bank of Boston. For constructive comments I am also grateful to Sara Goldrick-Rab,
Lee Hansen, Nik Hawkins, Eric Olds, Darcy Saas, Alicia Sasser, Robert Tannenwald, and the anonymous
referees.
References
Autor, D. H., Katz, L. F., & Kearney, M. S. (2005). Trends in U.S. wage inequality: Re-assessing the
revisionists. National Bureau of Economic Research Working Paper No. 11627.
Barrow, L., & Rouse, C. (2005). Does college still pay? Economists’ Voice, 2(4), 3.
Baum, S., & Payea, K. (2004). Education pays: The benefits of higher education for individuals and society.
Washington, DC: College Board.
Belfield, C. R., Nores, M., Barnett, S., & Schweinhart, L. (2006). The High/Scope Perry Preschool Program:
Cost-benefit analysis using data from the age-40 followup. Journal of Human Resources, 41(1), 162–
190.
Bound, J., & Turner, S. (2007). Cohort crowding: How resources affect collegiate attainment. Journal of
Public Economics, 91, 877–899.
Brady, H., Hout, M., & Stiles, J. (2005). Return on investment: Educational choices and demographic
change in California’s future. Berkeley, CA: Survey Research Center.
Card, D. (1999). The causal effect of education on earnings. In O. Ashenfelter & D. Card (Eds.), Handbook
of labor economics (Vol. 3A). Amsterdam: Elsevier.
Card, D., & Lemieux, T. (2001). Can falling supply explain the rising return to college for younger men? A
cohort-based analysis. Quarterly Journal of Economics, 116(2), 705–746.
Chevalier, A., & Feinstein, L. (2006). Sheepskin or Prozac: The causal effect of education on mental health.
Center for the Economics of Education working paper.
Currie, J. (2001). Early childhood education programs. Journal of Economic Perspectives, 15(2), 213–238.
Cutler, D. M., & Lleras-Muney, A. (2006). Education and health: Evaluating theories and evidence. National
Bureau of Economic Research Working Paper No. 12352.
Fernandez, R., & Rogerson, R. (1995). On the political economy of education subsidies. Review of Eco-
nomics Studies, 62(2), 249–262.
Fortin, N. M. (2006). Higher-education policies and the college wage premium: Cross-state evidence from
the 1990s. American Economic Review, 96(4), 959–987.
Freeman, R. (1976). The overeducated American. New York: Academic Press.
Goldhaber, D., & Player, D. (2003). Analytical framework for assessing the potential return on a federal
investment in the Alliance for Excellent Education’s Every child a graduate. Washington, DC: Alliance
for Excellent Education.
Res High Educ
123
Hadley, J., & Holahan, J. (2003). How much medical care do the uninsured use, and who pays for it? Health
Affairs—Web Exclusive, 22(2), 66–81.
Hansen, W. L., & Weisbrod, B. A. (1969). Benefits, costs and finance of public higher education. Chicago,
IL: Markham.
Harlow, C. W. (2003). Education and corrections populations. Washington, DC: Bureau of Justice
Statistics.
Harmon, C., Oosterbeek, H., & Walker, I. (2003). The returns to education: Microeconomics. Journal of
Economic Surveys, 17(2), 115–156.
Harrison, P. M., & Beck, A. J. (2006). Prisoners in 2005. Washington, DC: Bureau of justice statistics.
Haveman, R. H., & Wolfe, B. L. (1984). Schooling and economic well-being: The role of nonmarket effects.
Journal of Human Resources, 19(3), 377–407.
Haveman, R., & Wolfe, B. (1995). The determinants of children’s attainments: A review of methods and
findings. Journal of Economic Literature, 33(4), 1829–1878.
Heckman, J. J., & Masterov, D. V. (2004). The productivity argument for investing in young children. Invest
in Kids Working Group Working Paper No. 5.
Hovey, H. A. (1999). State spending for higher education in the next decade: The battle to sustain current
support. Washington, DC: National Center for Public Policy and Higher Education.
Institute for Higher Education Policy. (2005). The investment payoff: A 50-state analysis of the private and
public benefits of higher education. Washington, DC: Institute for Higher Education Policy.
Juhn, C., Kim, D. I., & Vella, F. (2005). The expansion of college education in the United States: Is there
evidence of declining cohort quality? Economic Inquiry, 43(2), 303–315.
Kaiser Commission on Medicaid and the Uninsured. (2004). Medicaid facts. Washington, DC: Kaiser
Commission on Medicaid and the Uninsured.
Kane, T. J., Orszag, P. R., & Apostolov, E. (2005). Higher education appropriations and public universities:
Role of medicaid and the business cycle. In G. Burtless & J. Rothenberg Pack (Eds.), Brookings-
Wharton papers on urban affairs 2005. Washington, DC: Brookings Institution.
Katz, L. F., & Autor, D. H. (1999). Changes in the wage structure and earnings inequality. In O. Ashenfelter
& D. Card (Eds.), Handbook of labor economics (Vol. 3A). Amsterdam: Elsevier Science.
Krop, R. A. (1998). The social returns to increased investment in education: Measuring the effect of
education on the cost of social programs. Santa Monica, CA: Rand.
Levin, H., Belfield, C., Muennig, P., & Rouse, C. (2007). The costs and benefits of an excellent education for
all of America’s children. New York, NY: Center for Benefit-Cost Studies of Education.
Lleras-Muney, A. (2005). The relationship between education and adult mortality in the United States.
Review of Economic Studies, 72(1), 189–211.
Lochner, L., & Moretti, E. (2004) The effect of education on crime: Evidence from prison inmates, arrests,
and self-reports. American Economic Review, 94(1), 155–189.
Longanecker, D. (2006). A tale of two pities. Change, 38(1), 14.
Lynch, R. G. (2004). Exceptional returns: Economic, fiscal, and social benefits of investment in early
childhood development. Washington, DC: Economic Policy Institute.
McIntyre, R. S., Denk, R., Francis, N., Gardner, M., Gomaa, W., Hsu, F., et al. (2003). Who pays? A
distributional analysis of the tax systems in all 50 states (2nd ed.). Washington, DC: Institute on
Taxation and Economic Policy.
McMahon, W. (2004). The social and external benefits of education. In G. Johnes & J. Johnes (Eds.),
International handbook on the economics of education. Cheltenham: Edward Elgar.
Mortenson, T. (1994). Federal income taxes paid by college educated workers 1970 to 1991. Postsecondary
Education Opportunity, 28, 1–9.
Mortenson, T. (2001). Individual economic welfare in the human capital economy 1973 to 2000. Post-
secondary Education Opportunity, 114, 1–16.
Murphy, K. M., & Welch, F. (1990). Empirical age-earnings profiles. Journal of Labor Economics, 8(2),
202–229.
O’Hara, A. (2004). New methods for simulating CPS taxes. U.S. Census Bureau Technical Paper.
Rizzo, M. (2004). The public interest in higher education. In Federal reserve bank of Cleveland, education
and economic development. Cleveland, OH: Federal Reserve Bank of Cleveland.
Strathman, J. G. (1994). Migration, benefit spillovers and state support of higher education. Urban Studies,
31(6), 913–920.
Topel, R. (2004). The private and social values of education. In Federal reserve bank of Cleveland,
education and economic development. Cleveland, OH: Federal Reserve Bank of Cleveland.
Trostel, P. A. (1996). Should educaiton be subsidized? Public Finance Review, 24(1), 3–24.
Res High Educ
123
Trostel, P. A. (2003). The long-term economic effects of declining state support for higher education: Are
states shooting themselves in the foot? Wisconsin Center for the Advancement of Postsecondary
Education.
Trostel, P. A. (2007). The impact of new college graduates on intrastate labor markets. Wisconsin Center for
the Advancement of Secondary Education Working Paper No. 11.
Trostel, P. A., & Ronca, J. M. (2009). A simple unifying measure of state support for postsecondary
education. Research in Higher Education, 50(3), 215–247.
Vernez, G., Krop, R. A., & Rydell, C. P. (1999). Closing the education gap: Benefits and costs. Santa
Monica, CA: Rand.
Wolfe, B. L., & Haveman, R. H. (2003). Social and nonmarket benefits from education in an advanced
economy. In Y. Kodrzycki (Ed.), Education in the 21st century: Meeting the challenges of a changing
world, 2003.
Res High Educ
123
