Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Inequality, Corruption and Development
Jong-sung YouAustralian National University
Eunro LeeCharles Darwin University
MotivationPrevious studies:• Inequality Growth (-)• Corruption (institutional quality) Growth (-)
My own study:• InequalityCorruption (+)
Taken together,• InequalityCorruption (+) Growth (-)
(institutional quality)No previous study tested this hypothesis.
Literature Review
Inequality Growth (-): Growing consensus (including researchers at World Bank, IMF and OECD)-Robustness, still being debated:-Corruption, or instituonal quality as a causal mechanism (Easterly 2007): Not adequately tested.
Inequality Corruption (+): You (2015)-causal mechanism: elite capture, clientelism-cross-national evidence-comparative historical evidence: KOR, TWN & PHL
Causality from Inequality to Corruption
Source: You (2015) Democracy, Inequality and Corruption: Korea, Taiwan and the Philippines Compared. Cambridge University Press.
Corruption (institutional quality) Growth
A number of studies (Mauro 1995; Knack & Keefer 1995; Keefer & Knack 1997; Kaufmann and Kraay 2002; ----)- Robustness? (Svensson 2005; Glaeser and Saks 2006; Kurtz &
Schrank 2006)
Doubts on instruments:• Ethnolingusitic fractionalization, Legal origin: Weak identification at
the first stage (Shaw et al. 2011)• Predicted trade share (Shaw et al.): Valid?
Inadequate longitudinal data:• Mauro (1995): BI (1980-83) Growth (1960-85)• Knack & Keefer (1995): BERI (1972), ICRG (1982) Growth (1974-89)
Research strategies
• Cross-National Longitudinal Study:Inequality (1960) → Corruption (1980)→ Growth (1980-2010)
Instruments for inequality: Mature cohort size (You & Khagram 2005)
• Comparative Historical Study (You 2015)– East Asia: South Korea, Taiwan, and the
Philippines
OLS & IV regressions of corruption
Income Gini: Solt (2014), Standardized World Income Inequality DatabaseLand Gini: Deininger and Olinto (2000), Frankema (2009)Family Farms: Vanhanen (2003)ln Bribe (2004-10):TI’s Global Barometer SurveysControls: ln GDPpc (1960), Polity (1960s), Ethnolinguistic fractionalization
OLS & IV regressions of institutional quality
Dependent variables: ICRG QoG (1984-85), Evans & Rauch’s Meritocratic Bureaucracy (1970-90), QoG Survey’s Professional Bureaucracy (2010)Controls: ln GDPpc (1960), Polity (1960s), Ethnolinguistic fractionalization
OLS regressions of growth on corruption and bureaucratic quality
CPI 1980-85 ER Meritocratic
Corruption/Bureaucracy 0.34432 0.35810 3.84687 4.50250
(t-stat) 2.87 2.78 2.64 3.7
ln GDPpc 1980 -1.66871 -1.67179 -0.99021 -0.83551
(t-stat) -3.77 -3.65 -2.42 -2.18
Schooling 1980s 0.07178 0.14928 0.19486 0.18232
(t-stat) 0.61 1.31 1.49 1.47
Fractionalization -2.34307 -2.92683 -3.11232 -3.65347
(t-stat) -2.36 -2.73 -2.26 -2.7
Polity 1980s -0.02432 -0.02823 -0.01387 -0.00733
(t-stat) -0.5 -0.57 -0.32 -0.18
Income Gini 1980s -4.72281 -0.378682
(t-stat) -2.66 -0.13
N 49 49 30 32
R-squared 0.4968 0.4535 0.5175 0.5719
ER Meritocratic
(Dependent variable=Average annual GDP per capita growth between 1980 and 2010)
Mediation Structural Equation Modeling of corruption explaining the
impact of inequality on growth
Mediation Estimates of Corruption (1980s) Explaining Inequality(1960s)’s
Impact on Growth(1980 -2010)
Comparative historical evidence:Korea, Taiwan, and the Philippines (You 2015)
• Most similar cases: Initially, similarly poor, corrupt & unequal– Inequality: PHL>>KOR>TWN (since 1950s/60s)– Corruption: PHL>>KOR>TWN (since 1960s/80s)
Korea Taiwan PhilippinesGDP per capita 1953 (2005 constant PPP $) 1,586 1,243 1,730Primary enrollment 1950 (%) 83 79 91Secondary enrollment 1950 (%) 16 11 27Tertiary enrollment 1950 (per 10,000 population) 18 9 88Urban population 1950s (%) 18 - 15Non-agricultural population 1950s (%) 30 - 29Population 1950 (thousand) 20,846 7,981 21,131
Land reform, industrial policy and state autonomy/capture
13
The Critical Role of Land Reform
• Success & failure of land reform: Determined by external factors (Threats from communist neighbors, US policy)
• High inequality in PHL: • Low inequality in KOR, TWN: gradual development of programmatic
politics, meritocratic bureaucracy, and state autonomy
Consequences of land reformThe Trends of Land Gini
• Land reform → Low inequality → Expansion of education → Meritocratic bureaucracy; Separation of political and economic elites
• No land reform → High inequality → Landed elite diversified into commerce, industry, finance, and politics → Clientelism & capture
Trends of CCI, 1996-2011
Real GDP per capita of KOR, TWN & PHL, 1953-2007
0
5000
10000
15000
20000
25000
30000
1953
1955
1957
1959
1961
1963
1965
1967
1969
1971
1973
1975
1977
1979
1981
1983
1985
1987
1989
1991
1993
1995
1997
1999
2001
2003
2005
2007
Korea
Taiwan
Philippines
Summary• Cross-national longitudinal evidence:Inequality of income & land (1960)Corruption (1980 and later)Growth (1980-2010)
• Comparative historical evidence:Land reform InequalityCorruptionDevelopment
• Vicious cycle can be broken: KOR, TWNBut, increasing inequality (esp. in KOR), a great concern.
19
The relationship between inequality and corruption, by regime types
TGO
GEO
IRNMOZMRT
CUB
GINMMRIRQ
LBRCIV
MAR
MYS
TJK AZE
BIH
KHM KGZBLR
LSO
PAK
BFA
JOR
LAOTKM
SGP
UZB
GMBRUS
SOM
DJI LBN
TUN
GAB
TCD
SWZDZA
CMR
CHN
ZWEHTI
EGYKAZ UGA
RWA
BWA
COG
NAM
ETH VNM
SYC
TZAGUY
ZAF
YEMBDINPL
SDN
ZMB MWI
SVK
TWN
PNG
LUX
BRA
MUS
USAJPN
MLIPHL
CAN
GRC
NLD
ROM TTO
ISL
FRA
MLT
THA
HND
POL
COL
PRT
SUR
ITA
FIN
DEU
BRBSVN
MKD
CAF
SLVIND
ISR
TUR
NER
NZL
KENNIC
GTM
LTU
LKA
AUS
KOR
ESTCYP
AUT
CHE
CHLESP
PRYBOLMDG
BGD
CZE
PER
BEN
BEL
BGR
VEN
DNK
GHA
ECU
MEXDOM
NOR
NGA
URY
ARMJAMMDA
PAN
GBR
HUN
SENALBUKR
SLE
ARG
HRV
SWE
YUG
GNB
CRILVA
MNGIDN
IRL
05
10
.2 .4 .6 .8 .2 .4 .6 .8
0. Dictatorship 1. Democracy
CPI Fitted values
Gini
Graphs by Democracy
Inequality & corruption, by duration of electoral democracy
IRQ
CUB
AFGGNQ MMRSYR
SAUCPV
BFAKWT
BHR
SDNPAKLBRSWZCHNBDI
JOR OMN
RWABIHTUN
LBY
BTNQAT
COG
ARE
ERIMAR
LVAGAB
AZE
POLTZA
FJIGMBDZA PHLETHMNGMOZNPLRUS
HUN
MWINGA MRTCAF
THA SURGTM
SVK
BOL
CHL
UGAIRN ZMBHRV
YUGCMR
GHACZE
CIVLSO
TGOROMJAM
IDNALBHTIKEN
ARMNAM
BGD KGZSLEBENBGR
TWN
MKDUKR
LTUSVN
MDA
BRB
TURSLV
URY
ARGBLZ
MDG
PRT
DOM
ESP
PERCOL PAN
SGP
ZWETTOECU
KOR
BHS
PRYPNG
MLT
GRC
CYP
SAMMUS
HNDBRA EGY
MYS
SENNIC
BWA
MEXZAF
FIN
ITA
NLDIRL
LKA
FRAAUT
AUSNZL
VEN
ISR
ISL
CRI
DNK
IND
LUXNORGBRCANBEL
SWE
USA JPN
05
100
510
30 40 50 60 30 40 50 60
0 1
2 3
CPI Fitted values
Estimated household income inequality
Graphs by Duration_Electoral_Democracy