Wednesday, January 30, 2008
« New York Daily News and EMI Music Offer ... |
Main |
Tests Arent Best Way To Evaluate Graduat... »
Comparative Education Review, vol. 52, no. 1
The Cost of Corruption in Higher Education
Stephen P. Heyneman, Kathryn H. Anderson, and Nazym Nuraliyeva
Electronically published November 16, 2007
Corruption was symptomatic of business and government interactions
in Russia and other countries of the former Soviet Union before and
during the economic transition of the 1990s. Corruption is difficult
to quantify, but the perception of corruption is quantifiable.
Nations can even be arranged along a hierarchy by the degree to
which they are perceived as being corrupt, for instance, in their
business practices or in the administration of public
responsibilities. Based on the Transparency International Corruption
Perceptions Index for 2005, a world map (see online appendix fig.
A1) shows how pervasive corruption remains in the public sector.1
According to this index, countries in the former USSR region,
including Central Asia and the Caucasus, were among the most corrupt
countries in the world in 2005.2
With the breakup of the USSR and decentralization, ministries and
local governments operated more independently than under the planned
economy. The central government's enforcement mechanisms weakened,
and rent-seeking (using administrative position for personal gain)
activity was not as effectively monitored as under central planning.
The result, at least in the earliest years of independence, was an
increase in overall corruption and inefficiency at many levels of
government and administration, and the education sector was not
immune from these forces (Shleifer and Vishny 1993; Shleifer and
Treisman 2005). Ministry of Education officials began to demand
bribes for accreditation and procurement.3 Administrators demanded
bribes for admission, housing, book rental, grades, exams, and
transcripts. Teachers demanded bribes for admission, grades and
exams, and book purchases.
Anecdotal evidence suggests that the level of education corruption
in the USSR was lower than in other sectors. The "fairness" of the
system, particularly to children of proletarian origins and
minorities, was manifest as a philosophy. During the economic
transition, the central authority in education broke down, and the
various agents (ministry officials, rectors, faculty, and staff) no
longer acted in concert. Decentralization and privatization did not
reduce bribe taking but may have significantly increased it. The
increase was particularly rapid if international competition from
private education providers was restricted.4 The quality of
education was likely to deteriorate during this phase because
individual rent-seeking behavior by agents increased.
Local private providers of educational goods and services entered
the market, and this sometimes included foreign providers. The new
suppliers provided education in the local language or a foreign
language. The owners were local, foreign selling a local product, or
foreign selling a "branded product."5
Today, the causes and the mechanisms of education corruption are
quite varied. Bribes may be obtained in eight different ways (table
1). In the case of procurement and accreditation, the buyer is the
educational institution, and the seller of the bribe is the
government, usually the Ministry of Education. In the case of
obtaining illegal entrance to specialized programs, having grades
raised on the grounds of an illegal payment, or paying illegally for
normal educational services (housing, library use, and
administrative procedures),6 the buyer is the student, and the
seller is either a faculty member or the administration of the
university. The agents vary but can be broadly classified as
teachers, rectors and other administrators, and the Ministry of
Education.
Table 1 Types of Corruption in Education by Buyer and Seller
From the perspective of social development, corruption in education
can be worse than corruption in the police, customs service, or
other areas because it contains both immoral and illegal elements
and involves either minors or young people. Much of education
corruption is classified under the term "professional misconduct."
Professional misconduct is behavior that breaks the code of conduct
normally pertaining to the university professorate (Braxton and
Bayer 1999). But "corruption" contains elements beyond the
professional behavior. It may include corruption on the part of
institutions. Educational institutions that pay a bribe to be
accredited (Heyneman 2003, 2004a) may turn out graduates whose
skills and professional levels could be a danger to the public.
Educational institutions that commit fraud cheat the public because
they collect an illegal rent and they commit a crime within the very
same institution that was established to select future leaders on a
fair and impartial basis (Noah and Eckstein 2001; Heyneman 2005).
Corruption can be efficiency-improving in instances in which prices
(tuition, fees, or wages) are distorted by regulation or lags in
application.7 However, the social benefits from corruption are less
likely to be observed in education because corruption affects all
the other social goals for making the education investment (Bardhan
1997). Because education serves as a way to model good behavior for
children or young adults, allowing an education system to be corrupt
may be more costly than allowing corruption in the customs service
or the police. By design, one function of education is to
purposefully teach the young how to behave in the future. If the
education system is corrupt, one can expect future citizens to be
corrupt as well. This clearly must have a cost.
Efficiency also fails within an educational institution if corrupt
officials are affected by nonpecuniary factors such as favoritism
toward one's ethnic, regional, or religious group. In this case,
university officials admit unqualified students and faculty, and
education becomes a high-price, low-quality good. Instead of
increasing the competition within the university, bribery limits
competition and reduces quality (Bardhan 1997).
If a college or university acquires a reputation for having faculty
or administrators who accept bribes for entry, grades, or
graduation, the power of the university in the labor market may be
adversely affected. In domestic labor markets, particularly those
with many state-run enterprises, the risk is less because the job
choices of graduates are fewer. But in the private sector and
particularly with companies that draw from international labor
markets, the effect of a reputation for corruption on a university
may be more serious. But how common is the perception of corruption?
And how different is the perception of corruption from the actual
experience of corruption?
*1The Web address for the map and other information about the
Corruption Perceptions Index is
http://www.transparency.org/policy_research/surveys_indices/cpi/
2005.
*2The Corruption Perceptions Index "is a composite index, a poll
of polls, drawing on corruption-related data from expert and
business surveys carried out by a variety of independent and
reputable institutions. The CPI reflects views from around the
world, including those of experts who are living in the
countries evaluated" (Transparency International 2006, 9).
*3In this article we refer to corruption as implying monetary
bribery. However, corruption also includes nonmonetary
corruption: the illegal changing of student grades or
examination scores for reasons of doing "a favor" in support of
family, friends, or important personalities.
*4Examples include Tajikistan and Uzbekistan. See Iveta Silova et
al. (2007).
*5The hierarchy in price and prestige normally includes four
categories of institutions: (i) local institutions (public or
private) that, for instance, supply a business degree in the
local language; (ii) local institutions that supply a business
degree in an international language; (iii) international
institutions with unknown names that supply a business degree in
an international language; and (iv) an international institution
with a well-known "brand name" that supplies a business degree
in an international language.
*6Students are sometimes required to show their teacher that they
bought the textbook authored by the teacher (rather than using a
borrowed book) from a store before they are allowed to take the
final examination. Postgraduate students may be asked for a
bribe before a member of the dissertation committee will agree
to sign off on their doctoral thesis.
*7Christopher Ahlin and Pataki Bose (2007) demonstrate that
partially honest bureaucrats for whom bribery is
efficiency-improving in a static framework contribute to
economic inefficiency when corruption is modeled as a dynamic
process, and consumers have to reapply continuously for permits
and licenses. Among other things, this suggests that when
normative (often centralized) structures are loosened, wide
variations in cultural definitions of corruption emerge.
Corruption and Higher Education in Europe and Central Asia
Incidence of Corruption: Empirical Evidence
Surveys of university students from Serbia, Croatia, Bulgaria, and
Moldova were sponsored by local student organizations. Random
samples of university students were asked about their knowledge of
corruption in admissions, grading, and housing and their attitudes
toward corruption. A large number of these questions were identical
in all four surveys, but in Serbia in particular, additional
questions were included to better identify the corrupt agents and
buyers within universities.
We also examine more limited information from Kazakhstan and the
Kyrgyz Republic in Central Asia. The Kazakhstan data are from one
foreign-managed university and are the teaching evaluation records
for all faculty in 2001/2 and 2004/5. Of the five questions asked,
one focused on bribery. The second source of information is surveys
of higher-education institutions in the Kyrgyz Republic conducted by
the Eberts Fund. We do not have the individual evaluations from
these institutions, however, and rely on tabulations from public
reports on corruption within these institutions.8
Students at universities in six countries through the
Anti-Corruption Student Network in Southeastern Europe surveyed
randomly selected samples of students at institutions throughout
their countries on their experiences with and attitudes toward
corruption in education.9 In Bulgaria, Serbia, Croatia, and Moldova,
significant corruption is evident in admissions and for grades (see
table 2). Of the students in Bulgaria, Moldova, and Serbia, between
79 and 84 percent were aware of the practice of illegal bribes to
gain admission (i.e., 16-20 percent had "never heard" of such
bribes). Between 35 and 45 percent thought that the official
selection process could be bypassed, and between 28 and 36 percent
thought that admission test scores could be changed. On average,
between 18 and 20 percent of the students in Bulgaria, Croatia, and
Serbia and 40 percent of the students in Moldova reported that they
had used some illegal method to gain admission to their university.
Table 2 University Student Opinions about Corruption (%)
Many of the students in all four countries had a rather fatalistic
sense about corruption. Few students said that they would report
cheating if they observed it, most would cheat if they could get
away with it, and few would feel badly about it (see table 3).
Nevertheless, even in circumstances in which cheating is the "norm,"
between 11 and 14 percent of the students and faculty "resist" the
temptation and consider corrupt behavior to be "unacceptable."
Table 3 How Do You Feel about Cheating? (Percent Answering Yes)
What should be done with those who are caught? Thirty-four percent
of the students in Bulgaria thought that they should be expelled; 36
percent of students in Serbia approved of a severe penalty such as
expulsion, legal prosecution, jail, and public humiliation. However,
90 percent of the students did not know the penalty for faculty who
are caught accepting a bribe, and 87-90 percent of them did not know
the penalty for students who cheat (Posliyski and Vatev n.d.).
In a study of students at local universities in the Kyrgyz Republic,
the majority of students in most of the universities described their
institutions as being "bribable" (see table 4). A private university
managed by foreign owners, in accordance with international
standards of professional conduct, might be expected to have
significantly less corruption than in government universities.10
This is supported by the data in table 4, where only 5.1 percent of
the students at the American University of Central Asia report that
their university is "bribable" as opposed to between 50 and 70
percent in local government universities (Ebert Fund, March 2,
2006).11
Table 4 University Students in the Kyrgyz Republic Who Describe
Their University as Bribable
This latter finding is corroborated by evidence from Kazakhstan.
Students at a Turkish-owned university were asked if teachers were
willing to take gifts and other payments and if the course was
professional.12 An average of 7 percent of students in 2001 reported
that the faculty accepted gifts. However, when the survey was
repeated 4 years later the proportion declined to 3.3 percent. There
was a significant decline in the percentage of students reporting
gift taking or bribery among faculty after instructors knew that
this behavior would be reported. Like the American University of
Central Asia, the Kazak-Turkish University has a lower incidence of
bribery than local state institutions because of the application of
external standards of professional conduct to the behavior of
faculty and administrators.
*8None of the surveys collected information on bribery where the
buyer was the institution. We therefore have no information on
corruption in accreditation or procurement.
*9The following link provides information on this organization
and its programs:
http://www.soros.org/initiatives/hesp/focus/sesi/network_anti.
Students were responding to different surveys asking about
cheating. The surveys were not coordinated from a single source,
and hence the students might have differing concepts about the
meaning of cheating. However, all seemed to know that, though
common, it was an infraction of the rules.
*10A foreign university must adhere to two sources of rules and
regulations: from the country in which they are situated and in
their home country.
*11In this context "local" implies locally as opposed to
internationally accredited universities. It does not imply a
subnational geographical unit as in county or region.
*12The professional quality of the course and the instructor was
defined in terms of the quality of lectures and seminars,
respect from students, and objectivity in evaluation of a
student's progress. The average score of the four professional
quality questions was the instructor's rating on professional
skill. The faculty did not know in advance that the question on
gift taking would be included on the student evaluation form. In
the 2004/5 survey of students the faculty knew that the question
on gift taking would be included; the gift taking question was
changed so that bribery could be separately identified from
giving gifts without requiring a favor from the instructor. The
revised question was "Did the teacher demand gifts or bribes
from you?"
Which Disciplines Are More Likely to Accept Bribes?
Within higher-education institutions, there was wide variation in
corruption across departments and fields. This is evident in the
data from the Kazak-Turkish University (see table 5). In 2001 gift
taking was highest in chemistry, world history, Kazakh literature,
language, political science, and economics and lowest in physics,
Russian language, and fine arts. Bribery was reported by 16.8
percent of the students in the organic and physical chemistry
departments and 8.1 percent in economics. Four years later, the
incidence declined in all departments except English language,
economics, management, graphics, and law. Reported bribery in
economics increased from 8 to 10 percent and in law from 5.4 to 12.8
percent.
Table 5 Bribery by Departments in the Kazak-Turkish University, 2001 and 2005
In general the disciplines most likely to be characterized by
bribery seem to be those in highest demand--law, economics, finance,
and criminology--where the competition to enter is greater, the fees
and tuitions are larger, and the stakes for graduating are higher.13
Senior officials interviewed in Russia reported that the faculty
salary was 50 percent of the wage paid in the private sector for
PhD-trained employees in the humanities and 30 percent or lower of
the private-sector wage for PhD-trained employees in economics and
law.14 The pattern of corruption in the Kazak data is consistent
with the evidence from other sources. The probability that an
instructor accepts a bribe increases as the difference in the market
value of the instructor's wage and the university wage widens. If
bribery cannot be detected easily or if sanctions are weak, the
probability of bribe taking is even higher (Becker and Stigler 1974;
Rose-Ackerman 1975; Shleifer and Vishny 1993; van Rijckeghem and
Weber 2001; Sosa 2004). The sanctions imposed for revealed
corruption are severe at foreign-operated universities such as the
American University of Central Asia or the Kazak-Turkish Institute
and include expulsion of students and firing of faculty.15 However,
in many other universities, corruption is openly practiced, and
faculty and students are rarely sanctioned for corrupt behavior.16
A similar pattern across the disciplines is observed in Bulgaria and
Serbia. Figure 1 shows the percentage of students who bought a
textbook as a condition for passing an exam in Bulgaria. It is
highest in medicine (58 percent) and dentistry (54 percent),
followed by chemistry (50 percent), law, history,
sociology/political science, economics, and engineering/electronics
(40-50 percent). In Serbia the largest sources of corruption also
appear to be in medicine, economics, and law (see table 6). Students
in these Serbian university departments were more likely to pay
other students or faculty to take an exam, to pay a bribe, or be
forced to buy a personal textbook by a faculty member, enroll
illegally, or experience corruption among administrators than within
the university as a whole.
Fig. 1.-- Was student required to buy professor's textbook as a
condition of passing an exam? Responses by Bulgarian students in
2003. Source: Posliyski and Vatev (2003).
Table 6 Corruption by Faculty, Serbia, 2003
*13Tuitions are often higher in vocational certification programs
(such as for the customs service), where the opportunities for
bribes are greater. Bribery may also be involved when
universities seek to be accredited by the Ministry of Education
to offer those particular programs.
*14In the Kyrgyz Republic, e.g., the average faculty salary in
universities in the capital city, Bishkek, is about $50 a month.
This information was obtained from conversations with faculty at
several universities. In the poorest region of the
country--Naryn oblast--the average take-home faculty salary is
$30 a month, which, according to one professor of economics, was
enough to cover only 2 weeks of living expenses in Naryn.
*15Here we refer to foreign, publicly operated universities. We
might speculate that a similar pattern might occur with foreign,
privately operated universities, but we have no firsthand
evidence on this as yet.
*16In several universities, students reported to us that some
professors selected a class representative to collect bribes
from other students in the class before exams were graded. This
differs from a "grading fee" known in other regions because the
fees pocketed by the faculty member in this case are done so
illegally. All students in the class were aware that a bribe was
expected. In other cases, parents sent bribes to professors
before their children took their oral exams. The difficulty of
these exams depended on the amount of the bribe paid by parents
before the exam. The incidence of bribery was close to 100
percent if the course was a correspondence course.
Corruption and Higher-Education Quality
Corruption, Higher-Education Quality, and Private Economic Returns
Education potentially serves two functions in the labor market. It
is an investment in human capital and develops labor productivity.
We expect to find that each additional year of higher education
increases earnings, because the student acquires new knowledge,
skills, and attitudes that are transferable to employment. Both the
student and society gain from private investment in higher
education. Second, completion of education can signal to employers
that the student is of high ability and integrity and has great
potential to be a productive employee (Heyneman 2002/3) or a status
group selectivity marker (Collins 1979). Only the most able students
are able to complete higher education because the costs of
completion (time, money, and effort) are lower than for other
students. Higher education may not increase the productivity of
labor directly, but it may help employers sort out the most
productive and reliable from other more costly workers. If education
signals ability in the labor market, individual investors gain from
investments in education, but the benefits to society are small.
Monetary returns associated with different levels of educational
attainment are the most common method to gauge economic impact
(Griliches and Mason 1972). But educational quality varies
significantly from one part of the world to another (Heyneman
2004b), and reliance on measures of educational attainment to
quantify impact may be economically misleading (Behrman and Birdsall
1983). Yet valid and reliable measures of higher-education quality
are difficult to find for most countries (James et al. 1989).
Nevertheless, progress has been made in terms of hammering out
common definitions of quality (Welch 1966; Solmon 1975) and
exploring the relationship between earnings, educational quality,
and individual ability.17 Because it is difficult to directly
measure college quality, it has been common to utilize proxy
measures. These have included college prestige and selectivity18
type or purpose (Solmon and Wachtel 1975) and one's individual
course of study (Link 1973; Rumberger and Thomas 1993). While most
studies have found that higher quality is associated with higher
salaries, some have wondered whether the effect on earnings is
through the use of college as a screening device (Psacharopoulos
1974; McGuinness 2003a) and the result of natural ability rather
than college quality (Hause 1971, 1972; Griliches and Mason 1972).
Whichever proxy has been chosen, college quality has generally been
found to be positively associated with lifetime earnings.19 In one
study, the rate of return to investment in college varied from a low
of 2.5 percent to a high of 15.6 percent depending on the measure of
college quality (Ono 2004, 612).
Corruption affects the private and social return to education
investment through both of these paths. If students purchase grades,
for example, students have less incentive to learn. Advancement is
less correlated with knowledge and the acquisition of skill than
with wealth and the ability to pay for achievement. The private and
social returns to higher education are degraded. The signaling
function of education is also reduced if there is significant
corruption among faculty and administrators. Completion of education
cannot be closely linked to the ability of students if entry into
programs and high grades are for sale. The employer does not know
whether the student completed the program and did well because she
was a high-ability, low-cost student or because she acquired grades
illegally or unfairly. The variance in ability of students
completing a corrupt program is higher than for students who do not
complete a corrupt program. Even if an individual student from a
relatively corrupt institution is honest and of high ability, the
signaling value of the degree is reduced.20 An employer with a
choice of candidates reduces the risk of hiring an unproductive
employee by avoiding graduates of corrupt institutions and programs
and hiring only students from institutions, departments, or programs
with a reputation for honesty. For this employer to hire a student
from a corrupt program, the student would have to accept a
significantly lower salary and prove his or her economic value to
the employer through on-the-job experience.
Corruption in undergraduate institutions affects the probability
that a student can obtain a graduate degree. Graduate schools,
particularly in Western universities, discount applicants from
institutions in which corruption is perceived as common. Applicants
from corrupt programs are less likely to be selected because grades
and test scores do not represent their ability to do graduate-level
work.
When employers know the level of corruption within higher-education
institutions and across programs within these institutions, then we
expect employment and earnings to be positively affected by
attending a university with a low level of corruption among its
faculty and staff.21 The productivity and signaling functions are
illustrated through a standard earnings functions given in equation
(1) below: The logarithm of earnings
of employee i at time t is a function of education (S), market
experience (X), and other observable characteristics (Z) of i that
affect employment and wage decisions. The term b[1] is the annual
rate of return to each year of investment in education if S is
measured as years of schooling; b[1] is affected by corruption (C)
within the education institutions attended by i. If corruption
lowers the quality of education and, therefore, the productivity of
labor, then corruption lowers the return to education. If, however,
corruption in education helps to "grease the wheels" in the labor
market and provides workers with entry into more lucrative, and
possibly corrupt, jobs, the return to education could be higher if
corruption in education is pervasive. If education has less value in
the labor market if it is obtained through corruption, then
employers may place a higher weight on other noneducational
productivity-related characteristics of workers such as previous
work experience, health, and marital status in their employment
decisions. The returns to experience and other productivity
characteristics (b[2]) may increase in the presence of corruption.
In general, , , , and if S and X increase the earnings of employee
i. This model is similar to the education quality models of Charles
Link (1973) and others.
With information on individual earnings, completion of education,
corruption in education, and other characteristics of individuals,
equation (1) can be estimated with linear regression with inclusion
of interactions between the quantity and quality (corruption) of
education to measure the impact of corruption on education returns.
The empirical model is given in equation (2): where a,
, and ( , 2, 3) are estimated regression coefficients. The return to
years of education in the absence of corruption is (> 0), and the
presence of corruption lowers this return to if and C is positive.
If employers put more weight on other productivity-related
characteristics (X) in the presence of corruption ( ), then the
impact of X on the log wage increases to .22 For example, employers
may place greater value on previous experience if corruption in
education is widespread because these workers have a productivity
track record. Or employers may place greater weight on the youngest
workers because they have had less experience with corruption than
older workers. We try to empirically determine whether these changes
occur.
*17See, e.g., Griliches and Mason (1972), Solmon and Wachtel
(1975), Kingston and Smart (1990), Brewer et al. (1999), and
Dale and Krueger (2002).
*18See, e.g., Mueller (1988), Kingston and Lewis (1990), Kingston
and Smart (1990), Brewer et al. (1999), and Strayer (2002).
*19See, e.g., Wales (1973), Hilmer (2002), Strayer (2002), and
McGuinness (2003b).
*20While corruption may lower the signaling value of a degree in
the private sector, a degree creates a different kind of signal
in the marriage market in some of the rural, more traditional
areas of the region. In interviews with local officials in Osh
and Djalalabad oblasts in the Kyrgyz Republic, we were told that
parents bought college degrees, including master's degrees, for
a daughter in order to signal that she was from a wealthier
family. Education degrees increased her value in the marriage
market, and her family would benefit from making a better match
for their daughter.
*21Not all employers view education corruption in an identical
fashion. International companies with transnational access to
employees might be most concerned and, hence, the most
selective. Private non-state-owned enterprises similarly would
be concerned. State-owned enterprises and the public sector
might be the least concerned. Our interviews with employers were
drawn from the second category.
*22Equations (1) and (2) assume that the level of corruption is
exogenous to the wage and is not affected by the unobserved
characteristics of workers such as their inherent honesty or
ability. This is not likely to be the case for several reasons.
Those students who are willing to pay a bribe are more likely to
be selected from the lower tail of the ability distribution and
have other sources of income, possibly obtained through illegal
transactions or corruption in the market. Professors who accept
bribes are also not randomly selected from the population of
professors but are more likely to be drawn from fields that are
in high demand, pay a low salary relative to their value in the
outside market, and are from the lower tail of the teaching
ability distribution. If a professor graduates highly productive
students, this signals her productivity to the private sector
and can lead to additional income through outside contracts,
grants, and part-time employment. The professors who are of
lower ability and have low value in the outside market are more
likely to demand bribes to compensate them for their lack of
skill. In this case, regression estimation of eq. (2) produces
biased estimates of returns to education with and without
corruption. If the fixed-effect component of the residual is
correlated with corruption perceptions, then the fixed-effects
regression model we estimate minimizes this bias. The
Transparency International data that we use do not contain any
variables that we can use to otherwise identify corruption
perceptions through instrumental variables estimation. In future
surveys of the perception of education corruption, uniform
definitions would help to reduce the risk of having variations
in interpretation.
Data and Estimation
To estimate model (2), we use data on corruption in education from
the Transparency International (TI) Global Corruption Barometer
2005. The data were collected from public opinion surveys between
May and October 2005 conducted by the Gallup Organization for
Transparency International. Over 55,000 people in 69 countries were
interviewed face-to-face or over the telephone.23 To the extent
possible, samples were randomly selected and, in most countries,
were conducted nationally. In some countries only persons in urban
areas (e.g., Thailand) or the main city (e.g., Georgia) were
included in the survey. Francis Hutchinson et al. (2005) provide
more information on the surveys and the survey results.
The dependent variable in equation (2) is the log of income.
However, in the TI data, income is a categorical variable and can
take one of three values: 1 = low to medium-low income, 2 = medium
to medium-high income, and 3 = high income. We estimate two versions
of model (2). In the first version, we try to predict whether a
person will have high income. The dependent variable is a dummy
variable equal to one if the individual reports high income and zero
for low or medium income. In the second version of the model, we try
to predict whether a person will have a low income. The dependent
variable is a dummy variable equal to one if the individual reports
low income and zero for high or medium income. By estimating both
models, we can get a better picture of how corruption in education
affects economic well-being.
The measure of education corruption (C) that we use from this survey
is based on perceptions of corruption in education. The question was
asked: To what extent do you perceive the education sector in this
country is affected by corruption? The scale is from 1 to 5, where 1
= not at all corrupt and 5 = extremely corrupt. There is no
information in the survey on the respondent's actual experience with
corruption.
The data on education and other personal characteristics of workers
are limited and categorical. Education (S) is measured with two
dummy variables: completion of secondary education and completion of
higher education relative to no or basic education. Other
characteristics include age (less than 30, 51-65, and greater than
65 in comparison to ages 31-50), gender (1 = male, 0 = female), and
region (Western Europe, Central and Eastern Europe, Latin America,
Asia and Pacific, Middle East, Africa, and North America). We also
include two variables that control for the overall corruption
environment in each country. The first variable measures whether
corruption affects political life (1 = not at all, ... , 5 =
extremely corrupt), and the second variable measures whether
corruption decreased over the last 3 years (1 = increased a lot, ...
, 4 = decreased a lot).
We include in equation (2) interactions between secondary and higher
education and education corruption (C). The coefficients on the
interactions between corruption and education indicate whether
corruption in education affects the impact of investments in
secondary and higher education on income (c[1]). Equation (2) also
includes interactions between corruption and age and gender, and the
coefficients on the interactions indicate whether corruption in
education affects the impact of experience and gender on income
(c[2]). Our sample includes 36,404 persons from 68 countries who are
under the age of 65 and have complete data on income, corruption,
education, and age; no one from the Chinese sample is included
because data on income were not collected. About 45 percent of the
people in our sample report low income or medium income, and almost
15 percent have high income. About half of the sample has completed
secondary education, with an additional 18 percent with only basic
education and 31 percent with higher education. The average
education corruption perceptions score is 3.3.
Figures 2 and 3 illustrate the variation in perceptions of
ccorruption in education across regions and, specifically, within
Eastern and Central Europe and Central Asia (ECA). In the
highest-income countries, less than 40 percent of respondents report
high levels of corruption (4 or 5). Corruption in education is
significantly higher in the other regions: in Latin America and the
Middle East over 60 percent of respondents report a high level of
corruption. Within the EC region, corruption is highest among the
countries of the Commonwealth of Independent States (CIS) (Russia,
Moldova, Ukraine) and Georgia and lowest among the countries of the
European Union (Lithuania, Czech Republic, Poland) and in EU
accession countries (Bulgaria, Romania).24
Figure thumbnail Fig. 2.-- (30 KB)
Fig. 2.-- Perceptions of corruption in education across regions
Figure thumbnail Fig. 3.-- (22 KB)
Fig. 3.-- Perceptions of corruption in education in Eastern Europe,
Central Europe, and Central Asia
Because the income variable is a categorical variable, we estimate
equation (1) with a linear probability model. The term
is the income indicator variable and is equal to one if individual i
in country j reports either high income or not poor (medium or high
income) and zero otherwise. We regress Y on two education variables
(S[sed], S[hed]), age, gender, political corruption, and the change
in corruption (Z), and the interactions between education corruption
and education, age, and gender. The random error is assumed to have
two components: a person-invariant, country fixed effect ( ) and an
individual-specific random error ( ). The country fixed effect
measures tastes and cultural, social, and economic norms in country
j that are common to all persons in j and likely affect perceptions
of corruption, educational achievement, and the reported level of
income. To eliminate the influence of these country effects on the
estimate of the coefficient vector B, we estimate our model as a
fixed-effects model with robust standard errors. The fixed-effects
model is given in equation (3) below: We report the marginal effects
of educational attainment on the probability of high income (table
7) and on the probability of low income (table 8) when there is no
corruption and when corruption is maximum (5).25 The marginal
effects of secondary education and higher education are relative to
primary education. The cells are in boldface if corruption in
education directly changes the impact of secondary or higher
education relative to primary education on income.26 We compare the
effects of secondary and higher education with no corruption in
education to the effects under maximum corruption (5). We expect to
find that those individuals with higher education are more likely to
be poor and less likely to have high income when corruption in
education increases.
Table 7 Educational Corruption, Educational Attainment, and the
Probability of High Income
Table 8 Educational Corruption, Educational Attainment, and the
Probability of Low Income
In high- and lower-income Asian countries, Eastern and Central
Europe, and Western Europe, secondary education over basic education
in the absence of corruption in education increases the probability
of high income, and the effect ranges from .05 higher in the
lower-income Asian countries to .18 higher in the high-income Asian
countries. The secondary education effect is .05 over all countries;
on average, secondary education increases the probability of high
income by .05 or 5 percent over basic primary education. In all
regions except the Middle East and in the absence of corruption,
completion of higher education has a large impact on the probability
of high income. On average, higher education in comparison to
secondary education increases the probability of high income by
about .2. The marginal effect of higher education27 is highest in
North America and the high-income countries of Asia. Within the ECA
region, the marginal effect of higher education is also .20 and is
slightly lower among the CIS countries at .18.28 The highest benefit
from higher education in the ECA region is found in the EU accession
countries.
The education results for the low-income models mirror the
high-income results, but the effects are larger. In all regions
except Africa, completion of secondary education over primary
education decreases the probability of being poor by .18 on average,
ranging from a low of .09 in lower-income Asian countries to a high
of .28 in high-income Asian countries. In all regions, higher
education decreases the probability of being poor relative to the
completion of secondary education on average by .17 or about 4.25
percent per year of higher education. Within the ECA region, higher
education is most effective in keeping people out of poverty in the
CIS region, where higher education decreases the probability of
being poor by about 10 percent per year of investment.
Higher education becomes a less effective means to high income if
there is significant perceived corruption in education in Africa,
Western Europe, North America, and the high-income countries of
Asia. In Africa, the marginal effect of higher education on the
probability of high income falls by about 70 percent from .21 to .06
under maximum corruption (
). In Western Europe the marginal effect of higher education on the
probability of high income falls by about 25 percent from .165 to
.125 under maximum corruption.
Corruption in education adversely affects the relative ability of
higher education to keep people out of poverty in Western Europe,
low-income Asia, Africa, Latin America, and ECA. In the ECA region,
we find this effect most pronounced in the CIS countries. At maximum
corruption within the CIS, corruption increases the probability that
those with higher education are poor. With no corruption, higher
education reduces the probability of low income by 40 percent in
comparison to completion of secondary education. With maximum
corruption, the marginal effect of higher education on the
probability of low income falls to -.11; this is a 70 percent
decrease in the ability of higher education over secondary education
to keep people out of poverty, and the number of highly educated
persons who live in poor households increases if there is
significant corruption in education. Higher education is also a less
effective means to reduce poverty when there is significant
corruption in education in Africa and the lower-income countries of
Asia, but this occurs because corruption makes secondary education
more attractive. In Latin America, corruption has the opposite
effect on higher education and makes it less likely that highly
educated persons will live in low-income households.
The results suggest that corruption in education does change the
ability of education to increase income. In general, corruption in
education has its largest effects on high income in the
higher-income countries of Asia, North America, and Western Europe,
while in lower-income countries corruption has its largest effects
on low income. With a few exceptions, higher education increases
income, but this effect is significantly lower when corruption in
education is pervasive.29
We next turn our attention to the effect that education corruption
might have on the economic returns to investments in education. The
literature on the internal rates of return to education in ECA is
summarized in Kathryn Anderson and Michael Cain (2006). From that
review, we assume a benchmark internal rate of return to higher
education of .2 when there is no significant corruption in higher
education. In the Kyrgyz Republic, the typical secondary school
graduate earns about $600 a year, 4 years of college education
increases earnings 20 percent, and a college graduate earns on
average $720 per year with 4 years of college education. If earned
income grows at 3 percent per year, the discount rate is 3 percent,
and he works for 40 years, he accumulates about $28,800 in lifetime
earned income with college education, or $4,800 more than with
secondary education.30 This calculation assumes no change in the
value of experience with corruption. If pervasive corruption reduces
the marginal return to higher education by 70 percent from .2 to
.06, the college graduate earns on average 6 percent more than the
typical secondary school graduate, or $636 a year. Over 40 years he
accumulates $25,440, or $1,440 more than those with secondary
education. Corruption reduces lifetime income accruing to the
investment in higher education by $2,360, or 50 percent of the
uncorrupted higher-education advantage.
This calculation is not based on the direct estimation of economic
cost. The result of this experiment, however, is consistent with
evidence from the Kyrgyz Republic obtained through recent interviews
with firm managers, government officials, and university faculty in
five diverse oblasts.31 One high-technology firm required employees
to have a college degree and paid about $200 a month for an
entry-level position. Employees were recruited from four
universities: American University of Central Asia, Manas University
(Turkish-Kyrgyz University), Slavonic University (Russian-Kyrgyz
University), and Kyrgyz National University. Three of these
universities were connected to foreign institutions and had the
lowest levels of reported corruption. Students from regional and
smaller state universities were not considered. In addition, this
firm and all other firms that recruited college-trained employees
required a 3-month apprenticeship or internship period during which
managers indicated that students who had little ability would be
dismissed. In some of these firms and at least one government
ministry, written tests were administered before anyone was
employed. The test separated students from state universities who
had bought their degrees from students who were truly educated.
Finally, most of these firms reported that they use performance
incentives when possible. They reward their employees with bonuses
or piece-rate pay in addition to monthly salary income.
The use of these sorting devices imposes additional costs on firms
and is related to the uncertainty attached to the quality of
education attained by students.32 Most important, among the firms in
which we conducted interviews, students from highly corrupt
universities were not considered for technical and professional
private-sector jobs and even some government jobs and were screened
out of jobs in foreign enterprises. If the best alternative for
students from corrupt programs is employment in the government with
starting salaries of $50-70 a month, the wage losses are enormous.
If students from corrupt educational programs sort into government
jobs with the potential for bribes, the private income costs of
corruption are reduced, but the social costs remain.
*23The countries from the Eastern European and Eurasian region
are Bosnia-Herzegovina, Bulgaria, Croatia, the Czech Republic,
Georgia, Kosovo, Lithuania, Macedonia, Moldova, Poland, Romania,
Russia, Serbia, and Ukraine.
*24Pairwise difference in variance tests (Levene and Brown and
Forsythe robust tests) indicate significant differences in the
spread within the regional distributions. The ECA region
displays more variation in corruption perceptions than Western
Europe, Latin America, lower-income Asia, and the Middle East
and less variation than the other regions. Within the ECA
region, variation in perceptions is lowest in the CIS countries
and highest in the countries of the former Yugoslavia.
*25In the Middle East and the new European Union, EU accession,
and former Yugoslavian countries, there was no effect of
education corruption on high or low income.
*26We assume that an insignificant coefficient has a value of
zero in the calculations in tables 7 and 8.
*27The marginal effect of higher education is defined relative to
the next-lowest level of completed education--secondary
education. The higher-education regression coefficient is
interpreted relative to basic primary education.
*28The CIS consists of the countries of the ECA region except the
three Baltic states of Lithuania, Latvia, and Estonia.
* 29It is possible that other institutional quality factors may
influence this result and could be the focus of future analysis.
* 30A 3 percent growth in earnings per year is a common estimate
of the effect of experience in the transitioning states.
* 31All the firms interviewed received management consulting
through the Employment Development Program sponsored by the U.S.
Agency for International Development through Pragma Corporation
and constitute a convenience sample of enterprises.
* 32The next stage in the effort to research the cost of
higher-education corruption might be a carefully considered
calculation of the monetary cost of these selection mechanisms
to the private and public sectors.
Summary and Policy Implications
Although educational corruption existed under the Soviet Union, we
hypothesize that it was modest by comparison to the level today.
Corruption in education increased because of decentralized decision
making, which makes agents more difficult to control, the
proliferation of university owners and purposes, and the
inexperience with enforcing professional norms without traditional
sanctions. Moreover, private education and tuition-based public
education are new, and there is little experience in managing the
expectations of fee-paying students.33
This article illustrates the extent of higher-education corruption
by citing student surveys in six countries--the Kyrgyz Republic,
Kazakhstan, Croatia, Moldova, Serbia, and Bulgaria. These surveys
suggest that corruption varies in accordance with the market demand
for the subject of study, with higher frequencies of corruption
found in the subjects in highest demand. Also, corruption is more
likely to be found in local universities with local professional
codes of conduct and less likely to be found in universities
accredited in Europe or North America.
We hypothesize that the cost to a student who attends a university
characterized by a high level of corruption would be the equivalent
of sacrificing the economic impact of higher-education quality.
Using data from Transparency International on perceptions of
corruption in education in 68 countries, we find that a nation's
perceived corruption significantly reduces the payoff to higher
education; when corruption is pervasive, highly educated persons are
much less likely to report high income in high-income countries and
Africa and are more likely to be poor in ECA, Africa, and
lower-income countries of Asia. The change in these high-income
returns to higher education ranges from 25 to 70 percent.
Many of the countries of the ECA region are participating in the
Bologna Process with members of the European Union. One objective of
that process is to make university degrees equivalent in hopes of
facilitating transfer students and greater mobility in the labor
market. Universities or university systems with reputations for
corruption, whether experienced or perceived, will likely end the
Bologna Process. Were this process to actually take effect, it would
constitute the educational equivalent in the European Union of
unilateral disarmament. It is difficult to imagine why a country or
a university with a high reputation would allow its degrees to be
made equivalent to those of a university or a university system with
a reputation for corruption.
A final implication of education corruption is for development
assistance agencies, many of which make investments in higher
education. These agencies may have to rethink their strategies when
it is understood that the impact of their investments could be
significantly reduced if made in higher-education systems with high
levels of perceived corruption.
There are many mechanisms that a country or a university needs to
adopt to lessen the possibility of corruption and to lower the
perception that it is corrupt. These include codes of conduct for
faculty, administrators, and students; statements of honesty on
public Web sites; university "courts" to hear cases of misconduct;
and annual reports to the public on changes in the number and types
of incidents. These mechanisms may well be requirements for
universities in those parts of the world hoping to have their
degrees declared equivalent to those of universities in the European
Union or having the support of international development assistance
agencies.
However, the first step to effective policy intervention is to
acquire information about the experience and cost of corruption. We
recommend regular surveys of students such as those reported here.34
In one country, with surveys at two points in time, the decline in
corruption was significant, suggesting that when the possibility of
exposure and professional embarrassment is real, the propensity to
engage in corruption declines.
* 33Where professional integrity is not based on tradition,
"private" university may imply that all educational products are
for sale. Codes of conduct and legal agreements spelling out the
rights and obligations of students and faculty are only now
beginning to appear.
* 34A complete list of interventions to lower the incidence of
education corruption can be found in Stephen Heyneman (2004a).
Appendix
Figure thumbnail Fig. A1.-- (68 KB)
Fig. A1.-- Corruption perceptions world map. Source: Transparency
International (2005).
References
*Ahlin, Christopher, and Pataki Bose. 2007. "Bribery,
Inefficiency and Bureaucratic Delay." Journal of Development
Economics 84 (1): 465-86.
First citation in article, CrossRef
*Anderson, Kathryn, and Michael Cain. 2006. "The Importance of
Social Investments for Economic Growth and Governing in
Transitioning States: A Survey of Recent Evidence." Report
prepared for U.S. Agency for International Development through
Aguirre International, Washington, DC.
First citation in article
*Bardhan, Pranad. 1997. "Corruption and Development: A Review of
Issues." Journal of Economic Literature 35 (3): 1320-46.
First citation in article
*Becker, Gary S., and George J. Stigler. 1974. "Law Enforcement,
Malfeasance, and Compensation of Enforcers." Journal of Legal
Studies 3 (1): 1-19.
First citation in article, Abstract
*Behrman, Jere R., and Nancy Birdsall. 1983. "The Quality of
Schooling: Quantity Alone Is Misleading." American Economic
Review 73 (5): 928-46.
First citation in article
*Braxton, John M., and A. E. Bayer. 1999. Faculty Misconduct in
Collegiate Teaching. Baltimore: Johns Hopkins University Press.
First citation in article
*Brewer, Dominic J., Eric R. Eide, and Ronald G. Ehrenberg. 1999.
"Does It Pay to Attend an Elite Private College? Cross-Cohort
Evidence on the Effects of College Type on Earnings." Journal of
Human Resources 34 (1): 104-23.
First citation in article, CrossRef
*Collins, Randall. 1979. The Credential Society: An Historical
Sociology of Education and Stratification. New York: Wiley.
First citation in article
*Dale, Stacy B., and Alan B. Krueger. 2002. "Estimating the
Payoff to Attending a More Selective College: An Application of
Selection on Observables and Unobservables." Quarterly Journal
of Economics 117 (4): 1491-1527.
First citation in article, CrossRef
*Griliches, Zvi, and William M. Mason. 1972. "Education, Income
and Ability." Journal of Political Economy 80 (3, pt. 2):
S74-S103.
First citation in article
*Hause, John C. 1971. "Ability and Schooling as Determinants of
Lifetime Earnings or If You're So Smart, Why Aren't You Rich?"
American Economic Review 61 (2): 289-98.
First citation in article
*Hause, John C. 1972. "Earnings Profile: Ability and Schooling."
Journal of Political Economy 80 (3, pt. 2): S108-S138.
First citation in article
*Heyneman, Stephen P. 2002/3. "Defining the Influence of
Education on Social Cohesion." International Journal of
Educational Policy, Research and Practice 3 (Winter): 73-97.
First citation in article
*Heyneman, Stephen P. 2003. "Education and Misconduct." In
Encyclopedia of Education,vol. 5, ed. James Guthrie. New York:
Macmillan.
First citation in article
*Heyneman, Stephen P. 2004a. "Education and Corruption."
International Journal of Educational Development 24 (6): 638-48.
First citation in article, CrossRef
*Heyneman, Stephen P. 2004b. "International Education Quality."
Economics of Education Review 23 (4): 441-52.
First citation in article, CrossRef
*Heyneman, Stephen P. 2005. "Organizations and Social Cohesion."
Peabody Journal of Education 80 (3-4): 1-8.
First citation in article
*Hilmer, Michael J. 2002. "Human Capital Attainment, University
Quality, and Entry-Level Wages for College Transfer Students."
Southern Economic Journal 69 (2): 457-69.
First citation in article, CrossRef
*Hutchinson, Francis, Tom Lavens, and Marie Wolkens. 2005.
"Report on the Transparency International Global Corruption
Barometer 2005." Policy and Research Department, Transparency
International-International Secretariat, Berlin.
First citation in article
*James, Estelle, Nabeel Alsalam, Joseph C. Conaty, and Duc-Le To.
1989. "College Quality and Future Earnings: Where Should You
Send Your Child to College?" American Economic Review 79 (2):
247-52.
First citation in article
*Kingston, Paul William, and Lionel S. Lewis. 1990.
"Introduction: Studying Elite Schools in America." In The
High-Status Track: Studies of Elite Schools and Stratification,
ed. Paul William Kingston and Lionel S. Lewis. Albany: State
University of New York Press.
First citation in article
*Kingston, Paul William, and John C. Smart. 1990. "The Economic
Payoff of Prestigious Colleges." In The High-Status Track:
Studies of Elite Schools and Stratification, ed. Paul William
Kingston and Lionel S. Lewis. Albany: State University of New
York Press.
First citation in article
*Link, Charles R. 1973. "The Quantity and Quality of Education
and Their Influence on Earnings: The Case of Chemical
Engineers." Review of Economics and Statistics 55 (2): 241-47.
First citation in article, CrossRef
*McGuinness, Seamus. 2003a. "Graduate Overeducation as a
Sheepskin Effect: Evidence from Northern Ireland." Applied
Economics 35 (5): 597-608.
First citation in article, CrossRef
*McGuinness, Seamus. 2003b. "University Quality and Labour Market
Outcomes." Applied Economics 35 (18): 1943-55.
First citation in article, CrossRef
*Mueller, Ralph O. 1988. "The Impact of College Selectivity on
Income for Men and Women." Research in Higher Education 29 (2):
175-91.
First citation in article, CrossRef
*Noah, Harold J., and Max A. Eckstein. 2001. Fraud and Education:
The Worm in the Apple. Lanham, MD: Rowman & Littlefield.
First citation in article
*Ono, Hiroshi. 2004. "College Quality and Earnings in the
Japanese Labor Market." Industrial Relations 43 (3): 595-617.
First citation in article, CrossRef
*Posliyski, S., and V. Vatev. 2003. "Survey Report: Corruption in
Higher Education." Photocopy, Anti-Corruption Student Network in
Southeastern Europe, Sofia, Bulgaria.
First citation in article
*Psacharopoulos, George. 1974. "College Quality as a Screening
Device." Journal of Human Resources 9 (4): 556-58.
First citation in article, CrossRef
*Rose-Ackerman, Susan. 1975. "The Economics of Corruption."
Journal of Public Economics 4 (2): 187-203.
First citation in article, CrossRef
*Rumberger, Russell W., and Scott L. Thomas. 1993. "The Economic
Returns to College, Major, Quality, and Performance: A
Multilevel Analysis of Recent Graduates." Economics of Education
Review 12 (1): 1-19.
First citation in article, CrossRef
*Shleifer, Andrei, and Daniel Treisman. 2005. "A Normal Country:
Russia after Communism." Journal of Economic Perspectives 19
(1): 151-74.
First citation in article, CrossRef
*Shleifer, Andrei, and Robert W. Vishny. 1993. "Corruption."
Quarterly Journal of Economics 108 (3): 599-617.
First citation in article, CrossRef
*Silova, Iveta, Mark Johnson, and Stephen Heyneman. 2007.
"Education and the Crisis of Social Cohesion in Azerbaijan and
Central Asia." Comparative Education Review 51 (2): 159-80.
First citation in article, Abstract
*Solmon, Lewis C. 1975. "The Definition of College Quality and
Its Impact on Earnings." Explorations in Economic Research 2
(4): 537-87.
First citation in article
*Solmon, Lewis C., and Paul Wachtel. 1975. "The Effect on Income
of Type of College Attended." Sociology of Education 48 (1):
75-90.
First citation in article, CrossRef
*Sosa, Luis A. 2004. "Wages and Other Determinants of
Corruption." Review of Development Economics 8 (4): 597-605.
First citation in article, CrossRef
*Strayer, Wayne. 2002. "The Returns to School Quality: College
Choice and Earnings." Journal of Labor Economics 20 (3):
475-503.
First citation in article, Abstract
*Transparency International. 2006. "Transparency International
Corruption Perceptions Index 2006." Press release,
http://www.transparency.org/policy_research/surveys_indices/cpi/
2006.
First citation in article
*van Rijckeghem, Caroline, and Beatrice Weber. 2001.
"Bureaucratic Corruption and the Rate of Temptation: Do Wages in
the Civil Service Affect Corruption and by How Much?" Journal of
Development Economics 65 (2): 307-31.
First citation in article, CrossRef
*Wales, Terence J. 1973. "The Effect of College Quality on
Earnings from the NBER-Thorndike Data." Journal of Human
Resources 8 (3): 306-17.
First citation in article, CrossRef
*Welch, Finis. 1966. "Measurement of the Quality of Schooling."
American Economic Review 56 (3): 379-92.
SIGN UP and GET POSTS DELIVERED TO YOUR EMAIL
ON THIS PAGE....
ARCHIVES
|
| Sun | Mon | Tue | Wed | Thu | Fri | Sat |
| 30 | 31 | 1 | 2 | 3 | 4 | 5 |
| 6 | 7 | 8 | 9 | 10 | 11 | 12 |
| 13 | 14 | 15 | 16 | 17 | 18 | 19 |
| 20 | 21 | 22 | 23 | 24 | 25 | 26 |
| 27 | 28 | 29 | 30 | 31 | 1 | 2 |
| 3 | 4 | 5 | 6 | 7 | 8 | 9 |
| August, 2008 (30) |
| July, 2008 (71) |
| June, 2008 (65) |
| May, 2008 (124) |
| April, 2008 (103) |
| March, 2008 (57) |
| February, 2008 (94) |
| January, 2008 (102) |
| December, 2007 (113) |
| November, 2007 (274) |
| October, 2007 (47) |
SEARCH THIS BLOG
Navigation
SUBSCRIBE TO CATEGORY FEEDS
Blogroll
ABOUT
Disclaimer
The opinions expressed herein are of the writers personal opinions and do not represent Educational CyberPlayGround™ views in anyway.
© Copyright 2008, edu-cyberpg.com
E-mail
Tag Cloud
Sign In