Freedom of Information Acts and Public Sector Corruption Monica Escaleras Shu Lin Charles Register* August 2008 Abstract: Over the past several decades a number of countries have implemented Freedom of Information acts with the hope that greater transparency would assist in controlling public sector corruption. To consider the effectiveness of these acts, we analyze annual data on 128 countries during the period 1984 through 2003 using a variety of propensity score matching techniques. Overall, we find no significant relationship between the implementation of a FOI act and corruption. While this result holds for developed countries, in the developing world, enactment of FOI acts is significantly and consistently associated with rising rather than falling levels of corruption. As such, we conclude that implementation of these acts should not be seen as a panacea for the problem of corruption in the public sector. Keywords: Freedom of information, governmental transparency, and public sector corruption. JEL Classifications: D73, D78, H11, O10 ____________ *Monica Escaleras: (Corresponding Author) Florida Atlantic University, Department of Economics, 777 Glades Road, Boca Raton, FL 33431; E-mail: [email protected]; Phone: (561) 297-1312; FAX: (561) 297-2542. Shu Lin: University of Colorado-Denver, Department of Economics, Campus Box 181, Denver, CO 80217. Charles Register: Florida Atlantic University, Department of Economics, 777 Glades Road, Boca Raton, FL 33431; E-mail: [email protected]; Phone: (561) 297-3222; FAX: (561) 297-2542.
39
Embed
Freedom of Information Acts and - FAUhome.fau.edu/mescaler/web/working papers/FOI8 18 08.pdf · Freedom of Information Acts and . ... To put teeth into this initiative, in September,
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Freedom of Information Acts and
Public Sector Corruption
Monica Escaleras Shu Lin Charles Register*
August 2008
Abstract: Over the past several decades a number of countries have implemented Freedom of Information acts with the hope that greater transparency would assist in controlling public sector corruption. To consider the effectiveness of these acts, we analyze annual data on 128 countries during the period 1984 through 2003 using a variety of propensity score matching techniques. Overall, we find no significant relationship between the implementation of a FOI act and corruption. While this result holds for developed countries, in the developing world, enactment of FOI acts is significantly and consistently associated with rising rather than falling levels of corruption. As such, we conclude that implementation of these acts should not be seen as a panacea for the problem of corruption in the public sector. Keywords: Freedom of information, governmental transparency, and public sector corruption. JEL Classifications: D73, D78, H11, O10 ____________ *Monica Escaleras: (Corresponding Author) Florida Atlantic University, Department of Economics, 777 Glades Road, Boca
Yesterday’s guilty plea by Rep. Randy “Duke” Cunningham—make that former representative, since he resigned after entering the plea—reveals the most brazen bribery conspiracy in modern congressional history. A San Diego Republican and Vietnam War veteran who served on the House Appropriations defense subcommittee and the intelligence committee, Mr. Cunningham admitted taking $2.4 million in bribes from two defense contractors angling for government contracts and from two other co-conspirators. The court papers filed in the case, jaw-dropping as they are, don’t address a critical question: How could this happen? To some extent, it’s hard to guard against out-and-out corruption and criminality by someone bent on breaking the law; as the court papers describe it, the congressman and his co-conspirators worked to “conceal and disguise” their activities “by directing payments through multi-layered transactions involving corporate entities and bank accounts.” Mr. Cunningham also lied on his financial disclosure forms and filed false tax returns.
But there are also indications that the system failed. Mr. Cunningham’s ability to pull off the caper was helped by the fact that lawmakers don’t need to list their homes or mortgage debt on financial disclosure forms; such a listing might have provided an earlier clue to the wrongdoing. More fundamentally, the papers say that Mr. Cunningham used his influence in the congressional appropriations process to benefit the contractors and “took other official action to pressure and influence” Defense Department personnel to give contracts to his co-conspirators.1
1. Introduction
While, as the author states, the case of public sector corruption described above is objectively
“brazen” it is also unfortunately not isolated. At a local level, relatively minor cases, in a monetary
sense, of providing favors to local officials for the letting of contracts for items as seemingly trivial as
garbage pick-up and disposal are quite common. At the opposite extreme, very large sums of money
often line the pockets of corrupt public officials at a national level as contracts are let for the
construction of major water treatment or electric power generation projects. Similar exist at every
intervening level of government making bribery likely the most common form of public sector
1 Brazen Conspiracy, Staff Editorial, The Washington Post, November 29, 2005, p. A20.
1
corruption in that private parties have monetary interests in nearly all decisions made at all levels of
government. Thus, cases of public officials trading the power afforded them by their office for
personal gain can be found involving all levels of government, many if not most economic sectors,
and in all or nearly all countries, regardless of their level of development.2
Of course, bribery is but one common form of public sector corruption. If the specific case
referred to in the introductory piece is considered “brazen”, it pales in comparison to the numerous
instances of state leaders plundering their country’s resources and coffers. Consider, for example, the
case of the former Haitian leader Jean-Claude Duvalier who, it is estimated, stole what amounts to
about 3 percent of that country’s meager GDP each year while in office. Similar cases of looting
include: $15-35 billion taken by former Indonesian dictator Mohamed Suharto; $5-10 billion by the
Philippines’ Ferdinand Marcos; $5 billion by Mobutu Sese Seko of the Congo; $2.5 billion by Sani
Abacha of Nigeria; $1 billion by Serbia’s Slobodan Milosevic; and, about $600 million by Peru’s
Alberto Fujimori.3
The problem of corruption in government has been seen to be so severe and pervasive that
international actors such as the World Bank and the United Nations, among others, have taken
affirmative steps to attempt to reign in these practices. For example, the World Bank has a
Department of Institutional Integrity as part of its focus on improved governance. The department is
charged with investigating allegations of corruption in Bank-related operations. Since 2001, the
department has handled nearly 3,000 cases of alleged corruption resulting in sanctions being placed
2 Rose-Ackerman (1999) offers a good treatment of the literature on public sector corruption. Further, numerous case
studies on corruption can be found in Transparency International’s annual Global Corruption Report, published by London’s
Pluto Press.
3 Data taken from Transparency International’s Global Corruption Report 2004, Special Focus: Political Corruption.
2
on some 340 companies and individuals.4 For its part, the United Nations adopted a Convention
against Corruption in September, 2003, which took effect in December, 2005.5 This convention
includes a chapter specifically addressing the issue of the recovery of assets stolen by government
officials such as those noted above. To put teeth into this initiative, in September, 2007 the World
Bank and the United Nations established a joint program, the Stolen Assets Recovery Initiative,
charged with recovering the billions of dollars illegally looted annually from the public assets of
developing countries.6,7
While the actions of international players like the United Nations and the World Bank in
fighting public sector corruption are growing in strength, equally important are steps being taken at
the national level in many countries. Generally speaking these focus on enhancing the transparency of
governmental actions. The underlying idea here is simple: In the case of a public official being
bribed for the right, as examples, to erect a housing complex or water treatment facility in a
sub-standard way or perhaps in which the bribe targets a favorable zoning ruling, enhanced
transparency affords all interested parties the opportunity to see the corruption and, if empowered to
4 Further information on this and other recent World Bank activities in the area can be found at
http://go.worldbank.org/CI3TOJK4I0.
5 Available at: http://unodc.org/unodc/en/treaties/CAC/index.html.
6 Additional information about this program can be found at http://abcnews.go.com/International/wireStory?id=3616330.
7 A great deal of research has been directed toward identifying ill-effects of public sector corruption and shows, among
other things, that it significantly: 1) lowers investment and limits economic growth (Mauro 1995, Mo 2001, Pellegrini and
Gerlagh 2004, and Meon and Sekkat 2005); 2) causes countries to receive less foreign aid (Alesina and Weder , 2002); 3)
causes countries to be less attractive targets for foreign direct investment (Habib and Zurawicki, 2002); 4) increases fatalities
from traffic accidents (Anbarci, Escaleras, and Register, 2006); 5) increases fatalities from earthquakes (Escaleras, Anbarci,
and Register, 2007); and, 6) restricts access to clean drinking water and improved sanitation (Anbarci, Escaleras, and
Yemen … … 2.81 Zambia … … 2.54 Zimbabwe 2002 2002 2.83 0.00 Notes: … indicates a country did not adopted/implement FOI laws during 1984-2003 *** indicates a country adopt/implement FOI laws before 1984. a World Bank classification of Develop and Developing countries.
28
Appendix 2. Definition and Sources of Variables Variable
Definition
Source
FOI Implemented (FOII)
A country has adopted and implemented a Freedom of Information law
Basinar (2006)
FOI Adopted (FOIA)
A country has only adopted a Freedom of Information law
Basinar (2006)
CORRUTION-ICRG Index (COR1)
Annual surveys with 6 indicating least corruption and 0 most corruption
International Country Risk Guide
CORRUPTION-TI Index (COR2)
Annual surveys with 10 indicating least corruption and 1 most corruption
Transparency International
DEMOCRACY (DEM)
Scale of 0-10 with higher values indicating more thoroughgoing democratic institutions
Polity IV database
CIVIL LIBERTIES (CIVIL)
Scale of 1-7 with higher values indicating less freedom
Freedom House
FEDERALISM (FED)
Dichotomous variable with a value of one when government is relatively centralized and zero otherwise
Treisman (2000)
POTESTANT (PROT)
Share of the total population of Protestant faith in 1980
Treisman (2000)
BRITISH LEGAL ORIGIN (BRIT)
Dummy variable indicating that the legal origin of the country is British
Treisman (2000)
GDP per capita (GDPPC)
Real GDP per capita, expressed in constant (2000) U.S. dollars
World Development Indicators
OPENESS TO TRADE (TRADE)
Sum of exports and imports of goods and services measured as a share of GDP
World Development Indicators
URBAN POPULATION (URBP)
The share of the total population living in areas defined as urban
World Development Indicators
POPULATION DENSITY (PDEN)
Population of a country per square kilometer
World Development Indicators
29
Table 1. Descriptive Statistics
Variable Mean Std. Dev Minimum Maximum
1COR 3.136 1.392 0 6
2COR 4.633 2.392 0 10
FOII 0.188 0.391 0 1
FOIA 0.198 0.399 0 1
DEM 5.255 4.110 0 10
CIVIL 3.586 1.801 1 7
FED 0.181 0.385 0 1
PROT 13.850 22.911 0 97.8
BRIT 0.286 0.452 0 1
GDPPC 6,630.4 9,231.5 82.158 54,778.65
TRADE 76.597 47.357 10.831 473.510
30
Table 2. Sample Characteristics by Treatment Status
Variable Mean Std. Dev Minimum Maximum
Treated
DEM 9.301 1.331 0 10
CIVIL 1.898 1.215 1 7
FED 0.236 0.425 0 1
PROT 26.677 32.069 0 97.8
BRIT 0.258 0.438 0 1
GDPPC 14,683.030 11,240.520 274.371 40,947.42
TRADE 80.212 36.772 17.175 184.121
Control
DEM 4.483 4.008 0 10
CIVIL 3.934 1.702 1 7
FED 0.163 0.370 0 1
PROT 10.203 17.946 0 96.6
BRIT 0.293 0.455 0 1
GDPPC 4,800.599 7,604.916 82.158 54,778.65
TRADE 75.775 49.419 10.831 473.510
31
Table 3. Probit Estimates of Propensity Scores Baseline model
(FOII as dependent variable) FOIA as dependent variable Adding more controls to Baseline model
Pooled Developing Developed Developing Developed Developing Developed
Notes: Constant terms are included but not reported. Robust standard errors are reported in parenthesis. *, **, and *** indicate the significance level of 10%, 5% and 1%, respectively.
32
Table 4. Matching Estimates of Average Treatment Effect (Baseline Model) Matching Method
Radius Matching
Nearest Neighbor Matching
3-Nearest Neighbor Matching r=0.025 r=0.05 r=0.1
Local Linear Regression Matching
Kernel Matching
-0.009 -0.015 -0.002 0.002 0.037 0.050 0.003 Pooled
Notes: A 0.06 fixed bandwidth and an Epanechnikov kernel are used for kernel and local linear regression matching. Bootstrapped standard errors are reported in parenthesis. They are based on 500 replications of the data. *, **, and *** indicate the significance level of 10%, 5%, and 1%, respectively.
33
Table 5. Matching Estimates of Average Treatment Effect (using more restrictive sample) Matching Method
Notes: A 0.06 fixed bandwidth and an Epanechnikov kernel are used for kernel and local linear regression matching. Bootstrapped standard errors are reported in parenthesis. They are based on 500 replications of the data. *, **, and *** indicate the significance level of 10%, 5%, and 1%, respectively.
34
Table 6. Matching Estimates of Average Treatment Effect (using alternative measure of corruption) Matching Method
Notes: A 0.06 fixed bandwidth and an Epanechnikov kernel are used for kernel and local linear regression matching. Bootstrapped standard errors are reported in parenthesis. They are based on 500 replications of the data. *, **, and *** indicate the significance level of 10%, 5%, and 1%, respectively.
35
Table 7. Matching Estimates of Average Treatment Effect (using date of adoption, FOIA, as dependent variable) Matching Method
Notes: A 0.06 fixed bandwidth and an Epanechnikov kernel are used for kernel and local linear regression matching. Bootstrapped standard errors are reported in parenthesis. They are based on 500 replications of the data. *, **, and *** indicate the significance level of 10%, 5%, and 1%, respectively.
36
Table 8. Matching Estimates of Average Treatment Effect (adding more explanatory variables to the Baseline Model) Matching Method
Notes: A 0.06 fixed bandwidth and an Epanechnikov kernel are used for kernel and local linear regression matching. Bootstrapped standard errors are reported in parenthesis. They are based on 500 replications of the data. *, **, and *** indicate the significance level of 10%, 5%, and 1%, respectively.
37
38
Table 9. Results from Control Function Regressions No controls Control for estimated propensity scores Further control for country and year fixed effects
Developing Developed Developing Developed Developing Developed Developing Developed Developing Developed
Notes: Constant terms are included but not reported. Robust standard errors are reported in Parenthesis. *, **, and *** indicate the significance level of 10%, 5% and 1%, respectively.