Transcript
The Effect of Income The Effect of Income on Corruptionon Corruption
Brittni SmithBrittni SmithDepartment of Economics and ManagementDepartment of Economics and Management
Hood CollegeHood CollegeApril 16, 2010April 16, 2010
CorruptionCorruption
Misuse of public power for personal or Misuse of public power for personal or private gainprivate gain
ExamplesExamples Former Governor of Illinois Rod Blagojevich: Former Governor of Illinois Rod Blagojevich:
auctioning off President Obama’s Senate seat auctioning off President Obama’s Senate seat In 2009, China has convicted 106,000 officials for In 2009, China has convicted 106,000 officials for
corruptioncorruption Senior official accused of taking $500,000 dollars in Senior official accused of taking $500,000 dollars in
bribes from businesses seeking approval of projects. bribes from businesses seeking approval of projects. Former vice president of China’s highest court was jailed Former vice president of China’s highest court was jailed
for life for bribes totaling $600,000for life for bribes totaling $600,000
CorruptionCorruption
Seminal Work by Mauro (1995) Seminal Work by Mauro (1995) Consequence of corruption is lower investment which Consequence of corruption is lower investment which
decreases developmentdecreases development
President of the World Bank Jim Wolfensohn, President of the World Bank Jim Wolfensohn, ““Let’s not mince words, we need to deal with the Let’s not mince words, we need to deal with the
causes of corruption.” (1996) causes of corruption.” (1996)
Today considered to be one of the biggest Today considered to be one of the biggest obstacles to economic development.obstacles to economic development.
Countries in Corruption IndexCountries in Corruption Index
New Zealand
0.6
Somalia
8.9
LeastCorrupt
Most Corrupt
Spain
2.5
USAChina
MexicoRussia
HaitiAfghanistan
7.86.76.4 8.2 8.7
S. Korea
4.54
Corruption IndexCorruption Index
Source: Transparency InternationalSource: Transparency International World Bank World Bank
Perceived level of corruption in a country Perceived level of corruption in a country Based on poll-of-polls data from:Based on poll-of-polls data from:
Experts and business persons in the country and Experts and business persons in the country and abroad abroad
Independent, reputable institutions Independent, reputable institutions Ex) World BankEx) World Bank
Available from 1995-present for 178 countries. Available from 1995-present for 178 countries. Panel data: for each country there are 14 Panel data: for each country there are 14
observations, total of 2,492 observations.observations, total of 2,492 observations.
Frechette (2006)Braun-Di Tella (2004)
Treisman (2000)
Brown (2005), Kunicova-R. Ackerman (2005), Lederman (2005), Chang-Golden (2004),
Damania (2004), Dreher (2004),
Alt-Lassen (2003), Brunett-Weder (2003),
Graeff-Mehlkop (2003), Herzfeld-Weiss (2003), Knack-Azfar (2003) Person (2003),
Tavares (2003), Fisman-Gatti (2002),
Paldam (2001), Bonanglia(2001),
Swamy (2001), Abed-Davoodi (2000),
Rauch-Evan (2000), Wei (2000),
Goldsmith (1999), Ades-Di Tella (1997)
Income
Positive-SignificantNegative-SignificantVariable
Previous Studies on the Previous Studies on the Effect of Income on CorruptionEffect of Income on Corruption
Fixed-EffectsFixed-Effects Focus: Focus: time-varyingtime-varying factors of corruption factors of corruption
Ex) Income, education, etc.Ex) Income, education, etc.
Error term of model accounts for: Error term of model accounts for: Time-invariant factors that could effect corruptionTime-invariant factors that could effect corruption
Ex) Colonization, religion, geography and could affect Ex) Colonization, religion, geography and could affect corruptioncorruption
Country unobservablesCountry unobservables Could be correlated and potentially fostering corruption Could be correlated and potentially fostering corruption Accounting for the unaccountableAccounting for the unaccountable
EndogeneityEndogeneity
ititoit CorruptionIncome εββ +++= ...)(1
Income is endogenous with corruption Causal relationship
ititoit IncomeCorruption εββ +++= ...)(1
Direction of causation is not clear Do low income countries generate more
corruption? Does corruption makes countries poorer?
Frechette (2006)Frechette (2006)
ICRG index ICRG index Fixed-effects specification Fixed-effects specification Accounts for this endogenous relationship Accounts for this endogenous relationship Main findings: Main findings:
Income increases corruptionIncome increases corruption Education increases corruption Education increases corruption
Explanatory VariablesExplanatory Variables
(-)(-) IncomeIncome Real GDP per capitaReal GDP per capita
(-)(-) EducationEducation Number of pupils in primary schoolNumber of pupils in primary school
(-)(-) Share of Imports in GDPShare of Imports in GDP Merchandise trades as % of GDPMerchandise trades as % of GDP
(+)(+) Fuel, Ore, and Mineral Exports Fuel, Ore, and Mineral Exports % of Merchandise exports% of Merchandise exports
(-)(-) InternetInternet Number of UsersNumber of Users
Instrument VariableInstrument Variable
Instrument should be correlated with income but Instrument should be correlated with income but should not directly effect corruptionshould not directly effect corruption
Income
OECD Trading
Partner's Income
Haiti
U.S.
Corruption
Isolate endogenous variable
Instrument to be statistically significant
F statistic >10
First Stage Regression
Non-Linear Relationship Non-Linear Relationship Between Corruption and IncomeBetween Corruption and Income
02
46
81
0C
orru
ptio
n
0 20 40 60 80Income
Corruption Fitted values
The Effect of Income on Corruption
Bangladesh
Luxembourg
02
46
81
0C
orru
ptio
n
0 20 40 60 80Income
Corruption Fitted values
The Effect of Income on Corruption
ModelModel Specifies non-linear relationship between
income and corruption
Two-Stage Least Squares with: Panel Data Fixed-Effects Instrument Variable
Sub-Sample Income LevelsSub-Sample Income Levels
02
46
81
0C
orru
ptio
n
0 20 40 60 80Income
Corruption Fitted values
The Effect of Income on Corruption
Various Income Levels
***, **, * indicate statistical significance at the 1%, 5%, and 10% level respectively
Standard errors in parenthesis
ConclusionsConclusions
Non-linear relationship between income Non-linear relationship between income and corruptionand corruption
Subsample reveals as income increases, Subsample reveals as income increases, corruption decreases at a decreasing ratecorruption decreases at a decreasing rate
Internet reduces corruption for countries Internet reduces corruption for countries with income above $18,000.with income above $18,000.
SummarySummary
Clearly shows as income increases corruption Clearly shows as income increases corruption decreasesdecreases Opposite of Frechette (2006).Opposite of Frechette (2006).
Same technique as Frechette panel data and Same technique as Frechette panel data and fixed-effects methodfixed-effects method
Corrected for problems in past empirical Corrected for problems in past empirical research research
Endogeneity of income Endogeneity of income Non-linear relationship between income and corruptionNon-linear relationship between income and corruption
Proved using this model the results is income Proved using this model the results is income does decrease corruption but differently with does decrease corruption but differently with countries of different income levels. countries of different income levels.
top related