Segregation and the Quality of Government in a Cross-Section of
Countries
Alberto Alesina
(Harvard University and Igier)
and
Ekaterina Zhuravskaya
(Paris School of Economics, EHESS, New Economic School)
June 10, 2010
Online Appendix
1
Online Appendix
.5
ethnicity
.4
UGA
AFG
ZWE
GTM
PAK
ETH
IDN
TUR
ESP
ECU
.3
BFA
S tilde
COL
KEN
ZAF
MAR
GIN
.2
TZA
RUS
CHE
PHL
GHA
HND
BEN
TGO
BEL
ISR
BHR
PAN
PER
CIVMEX
0
.1
MKD NER
SEN
MLICAF
BGR
BLZ CHN
NPL
TWNFIN IND
GAB
MWI
UZB
BOL
CMR
USA
LTUAUS
ZMB
UKR
BLR
LKA
VNM
BRA
NZL
CRI
SAU
KAZ
KGZ
LVA
ROM
EST
GBR
QAT
DNK
NOR
CHL
CZE
KHM
HRV
JOR
PRT
FRA
TJK
AUT
RWA
ARG
SVK
PRY
SVN
CAN
ARM
ITA
BGD
DEU
SWE
ISL
NLD
IRL
JPN
GRC
HUN
KOR
LSO
0
.1
.2
S hat
.3
.4
.6
language
NGA
NAM
UGA
IDN
.4
CMR
S tilde
IND
ETH
PAK
TGO
AFG
ZWE
GTM
TUR
ESP
COLRUS
PERKGZ
MAR
GIN ZAF
HND
BEN
BEL
CHE
PAN
BLZ
GBR
MOZ
LVA
BOL
CIV
MEX
ITA
NER
MKDNIC
PRYNPL
SEN
CAF
MLI
ECU
CHN
BGR
UZB
USA
FIN
GAB
MWI
CAN LTU
UKR
AUS
EST
BLR
VNM
RWA
AUT
SAU
KAZ
HRV
THA
ROM
DNK
CHL
NOR
CZE
KHM
MUS
PRT
TJK
CRI
SVK
NZL
SVN
ARM
BGD
SWE
ISL
BRA
JPN
KOR
HUN
LSO
HTI
BFA
KEN
GHA
0
.2
ZMB
0
PHL TZA
.1
.2
.3
.4
.5
S hat
.3
religion
IDN
IND
.2
BGR
TZA
S tilde
NGA
RUS
KEN
ETH
GIN
PHLMOZ
.1
BEN
KGZ
KAZ
BFA
ROM
LKA
JPN
ISR
HUN
NLD
NAM
MUS
UZB
GHA
MDG
CIV
ZWE
TGO
EGY
THA
VNM
CHE
UGA
USA
HRV
GTM
NPL
CAN
CZE
ZMB
MEX
SVN
QAT
DOM
GBR
PAK
BRA
ZAF
LTU
RWASEN
AUT
CAF
KOR
SVK
HTI
MWI
STP
MLI
BLZ
EST
GAB
NER
PER
PRY
BGD
PRT
ARM
AUS
NZL
CHL
IRL
TUR
IRN
CMR
NIC
0
A
0
KHM
.1
.2
S hat
Figure A.1: Segregation indices Se and Sb
50
.3
.4
ZWE
GTM
AFG
UGA
TUR
ethnicity segregation (S hat)
.1
.2
.3
ECU
PAK
COL
BFA
MAR
ZAF
GINESP
KEN
IDN
MEX
CIV
TZA
HND
ETH
BEN
BEL
PAN
0
PHL
NER
GHA
CAF
MKD
MLI
BGR
IND
CHE
SEN
CHN
BOL
AUS
BLZ
FIN
UZB
ZMB LTU
BRA
BLR
CRI SAU
NZL
VNM
KAZ
UKR
GAB
NPL
GBR
DNK
CHL
NOR
MWI
ROM
EST
PRT
LVA
LSO
TJK
AUT
RWAUSA
HRV
BGD
CAN
ITA
ARM
ISL
CZE
SVK
HUN
SVN
PRY
JPN
KOR
SWE
KHM
.1
RUS
KGZ
CMR
.2
.3
language segregation (S hat)
.4
.5
.3
0
TGO
PER
IDN
religion segregation (S hat)
.1
.2
BGR
TZA
NGA
IND
KHM
RUS
KEN
GIN
ETH
BEN
BFA KAZ
HUN
MUS GHA
UZB
CIV
TGO
NAM
0
JPN
NPL
ROM
HRV
CAN
ZMB
MEXMOZ
USAMLI
BRA
THA
GBR
HTI
MWI
VNM
CZE
LTU SENCAF
NIC
SVK
PHL
NER
KOR
EST
GAB
BGD
AUTAUS
CHE
NZL
SVN
ARM
BLZ
PRT
CHL RWA
PRY
.1
ZWE
UGA
PAK
CMR ZAF
PER
GTM
TUR
.2
.3
language segregation (S hat)
.4
.5
.3
0
KGZ
IDN
religion segregation (S hat)
.1
.2
BGR
TZA
IND
RUS
KHM
KEN
ETH
BEN
KAZ
KGZ
LKA
HUN
NLD
GHA
UZB
0
JPN
NPL
ROM
HRV
CAN
ZMB
USA QAT
MLI
BRA
GBR
SEN
MWI
VNM
CAF
CZE
LTU
SVK
NER
PHL
KOR
EST
GAB
BGD
CMR
AUT
SVN
IRL
AUS CHE
ARM
RWA
PRY
PRT
CHL NZL BLZ
0
.1
CIV
GIN
BFA
TGO
ISR
MEX
PER
ZWE
UGA
PAK
ZAF
.2
ethnicity segregation (S hat)
GTM
TUR
.3
.4
Figure A.2: Segregation indices along the three dimensions of diversity
51
ethnicity
.4
GTM
AFG
ECU
Segregation (S hat)
.2
.3
AFG
UGA
TUR
PAK
COL
MAR
BFA
ISR
PAN
PER
MEX
BHR
.1
RUS
MKD
CHE
BGR
CHN
FIN
0
AUS
UZB
LTUBLR
TWN
CRI
SAUNZL
VNM
QAT
GBR
UKR
DNK
CHL
FRANOR
LKA
JOR
ROM
EST
PRT
ARG
LSO
RWA
AUT
USA
KHM
HRV
CAN TJK
BGD
ARM
ISL ITA
CZEIRL
JPN
PRY
HUN
GRC
SVN
KOR
DEU
SWE
NLDSVK
0
BEL
.2
ETH BEN
TGO
IDN
TZA CIV
PHL
NER
GHA
CAF
MLI
IND
SEN
BOL
BLZ
BRA
KGZ
KAZ MWI
NPL
LVA
ETH
BEN
NER
MLI
IND
BOL
SEN
7
8
Log per capita GDP
9
10
R-squared = 0.1222
language
.5
.4
AFG
NAM
TGO
PAK
KGZ
MAR
HND
ZAF
BEL
PAN
CMR
ETH
GIN
BEN
GBR
LVA
CIV
IND
ITA MEX
TZA
IDN
NER
NIC
PHL CAF
ECU
BGR
MKD
MLI
BOL
CHE
CHN
SEN
BLZ
USA
KEN
AUS
FIN
RWA UZB AUT LTU BLR SAUNGA GHA
NPL
VNM
HRV
ZMBKAZ
EST BFA
GAB
DNK
CHL
NOR
MUS
THA
MWI
CAN
PRT
ROM
CRI
UKR PRY
LSO
TJK
SVN
BRA
KHM
BGD
ISL SWE
CZE
NZL SVK
HTI
JPN
KOR
HUN
ARM
.2
.4
Fractionalization
.6
MOZ
ZWE
AFG
UGA
KGZ
ETH
MAR PER
ZAF
ESP
GIN HND
CMR
BEN
BEL
PAN
MOZ
CIV
TZA
NER
MLI
MWI
IND
CAF
BOL
SEN
KEN
UZB GHA
NGARWA
NPL
VNM
ZMB BFA
TJK
KHMLSO
BGD
HTI
ARM
6
.8
TUR
NAM
RUS
COL
TGO PAK
.1
PER
ESP
UGA
0
RUS
COL
GTM
Segregation (S hat)
.2
.3
.5
.4
ZWE
TUR
Segregation (S hat)
.2
.3
CHN
UZB
language
.1
PHL
GHA
CAF
BEL
RUSMEX
MKD
BGR
CHE
AUS
BLZ
FIN
LTU
BLR
BRA
CRI
TWN
CMR
NZL
SAU
VNM
KAZ GAB
NPLKGZ
UKR
GBR
QAT
DNK
NOR
LKAJOR
FRA
ROMLVACHL
EST ARG PRT
TJK RWA BGD
AUTUSA
KHMLSO
HRVSVK
CAN
ARM
ISL
PRY
HUNCZE
SVN
GRC ITA
JPN
IRL
KOR
DEU
SWE
NLD
6
GTM
0
ESP
BHR ISR
PAN
PER
CIV
TZA
MWI
R-squared = 0.1789
0
ZAF
HND
TGO
ZMB
1
MAR
GIN
IDN
ZMB
CMR
.8
ECU
COL
GAB
.4
.6
Fractionalization
TUR
KEN
BFA
0
HND
ZAF
KEN
GIN
GTM
UGA
PAK
.1
ESP
ZWE
Segregation (S hat)
.2
.3
.4
ethnicity
ZWE
7
GBR
ITA
LVA
MEX
IDN
NIC
PHL
ECU
CHN
BGR
MKD
CHE
BLZ
AUS USA
FIN
LTU
BLR
AUT
SAU
HRV
KAZ GAB
EST
DNK
CHL
PRY
MUS
THA
CAN NOR
ROM
PRT
CRI
UKR
SVN
BRA
SVK
ISL
NZL SWE
HUNCZE
KOR
JPN
8
Log per capita GDP
9
10
R-squared = 0.0518
R-squared = 0.1320
religion
.3
.3
religion
IDN
IDN
TZA
NGA
IND
KHM
RUS
KEN
GIN
ETH
KGZ
LKA
UZB
EGY
NAM
ISR
MUS
NGA
KEN
0
.2
.6
.8
RUS
KHM
GIN
BEN
BFA
NLD
MDG CIV
GHA
TGO
KGZ
LKA
KAZ
MUS HUN
NLD
UZB
CIV GHA ZWE EGY
TGO
NAM
ISR
UGA
JPN
NPL
ROM HRV
CAN
ZMB
MOZ
MEX
MLI
USA
QAT
BRA
SENHTI
PAK
THA
GBR
GTM
MWI NER
CAFSTP
VNM
GAB LTU
ESTSVK CZE
KOR
ZAF
BGDCMR ARM NIC PHLPER
AUT
CHE
DOM
SVN
NZL AUS
IRL
RWA
PRYBLZ
IRN
TUR CHL
PRT
ZWE
UGA
JPN
CAN
ZMB
MOZ
USA
BRA
GTMGABKOR
VNM
HTI
MWI
CZE GBR CAF
NIC
SVK
STP
EST
ZAF
AUT
CHE
SVN
BLZ CMR AUSNZL
CHL DOM
RWA
.4
Fractionalization
IND
ETH
MDG
0
0
ROMNPL
MLIHRV MEX QAT
PAK
SEN
THA
LTU
NER
PHL
PERBGD
ARM
IRN
PRTIRL
TUR
PRY
BEN
KAZ
BFA
HUN
BGR
TZA
Segregation (S hat)
.1
.2
Segregation (S hat)
.1
.2
BGR
6
R-squared = 0.0001
7
8
Log per capita GDP
9
10
R-squared = 0.0973
Figure A.3: Segregation and fractionalization (left); Segregation and per capita GDP (right)
52
1
CHL
ETH
TZA
-.5
Rule of law
0
.5
TWN
BEN
VNM
MLI
JOR
CHN
BEL
SEN FRA
MWI
MARESP
HUN
NZL
AUS
PRT
UGA
AUT NLD
CZE
IRL
RWA
GBR
SVN
CRI
ZMB SWE
IND
SVK
CAN
GHACHE
KGZ NER
HRV
NPL
FIN EST BFA
KHM
PHL
BHR
DNK LTUBGR
PAN
LKA QAT ARM
USA
LVA
TJK
TGO
JPNCIV
SAU
BOL
NOR
ROM
KEN
DEU
LSO
UZBISL
BRA ITA
GAB
CMR
GIN
GRC IDN
MEX
ECU
PER
HND
CAF
ISR
BGD KAZ KOR MKD
PAK
BLZ
TUR GTM
UKR
AFG
ARG
PRY
ZAF
RUS
COL
ZWE
-1
BLR
-.2
-.1
0
.1
ethnicity segregation (S hat)
.2
.3
coef = -1.4209862, (robust) se = .4719442, t = -3.01
1
CHL
TZA
ETH
-.5
Rule of law
0
.5
VNM
BEN
HUN
MLI
MWI
CHN
SEN
AUS MAR
PRT
NZL
BEL ESP
RWA CZE
GBR
UGA
CRI MOZ
GHA AUT
ZMB
CAN
SVK
SVN EST
NER
SWE
MUS IND
CHEHRV
KHMFIN
KGZ
USA
PAN
BFA
SAU
DNK
LTU BGR
TJK
UZB
ARM
TGO
LSO
PHL
LVA
NPL
ROM
KEN THA
CIVJPN
BRA
BOL NOR
GAB
ISL
ITA
MEX
CMR
BGD
CAF
GIN
NIC
IDN
PER
ECU
HND
KOR
KAZ
NAM
PAK TUR
NGA MKD BLZ
AFG
UKR
PRY
COL
HTI
BLR
ZAF
GTM
ZWE
-1
RUS
-.2
0
.2
language segregation (S hat)
.4
coef = -1.3428299, (robust) se = .43677639, t = -3.07
1
CHL
EGY
.5
VNM
TZA
ETH
-.5
Rule of law
0
SEN
MLI
PRT
HUN
AUS
MWI
IRL
UGA
NZL
BEN
AUT NLD
QAT
RWA
ZMB
GBR
CZE GHA
NER
HRVCAN
KHM
MDG
SVK
CHE
IND
EST
THA
USASVN
LKA
MOZ PHL
KGZ
BFA
STP
KEN
NPL
MEX
MUS
BRA
ARM
UZB GAB
ROM
JPN
PER
LTU
BGD
GIN
TGO
NAM
CIV
CMRPAKCAF NIC
KOR IRN
ISR
DOM
TUR
PRYKAZ
GTM
BLZ
ZAF
BGR
IDN
NGA
ZWE
HTI
-1
RUS
-.1
0
.1
religion segregation (S hat)
.2
coef = -.2467093, (robust) se = .99137285, t = -.25
Figure A.4: Residual scatter plots for rule of law and segregation (OLS)
53
ethnicity
ethnicity
ZWE
.3
.3
ZWE
ECU
ISR
ESP
RUS
BFA
ZAF PAK
HND
PAN
MAR
EST
KEN
BENTGO
GIN
FIN
PER
LTU
TWN
MEX
HUNBLR
ETH
AUS ISL
MKD CHN
DEUCHE
KHM DNK
LVAROM
CAF
UKR
GBR
ARM
GHA UZB
NLD
NZLCZEKOR
HRV
NOR
LSO
IDN
SVN
TZA
CANVNM
SVK
QAT
CRI
IRL
BLZ
JPN
SWE
PRT
KAZ
NER
PRY
GRC FRA AUT
ARG
CHL ZMB
JORLKA
USA
TJK
IND
MLI
KGZ
ITA
PHL BGD
SENMWI
SAU
RWA
BOL
NPL
GAB
BRA
CMR
BEL
BHR
BGR
-.1
CIV
Segregation (S tilde)
0
.1
.2
UGA
TUR
GTM
ESP
TUR
AFGCOL
RUS
BFA
ETH
PAK
ECU
HND
PAN
ESTFIN
BEL
TWN
MAR
CHE
LTU ZAF
TGO
TZA KEN BEN
BLR
HUN
PER
ISL
MKD GIN
GHA MEX
CZE
KHM DNK
LVA
AUS
UKR
NLD
ROM
UZB
ARM
DEU HRV
GBR
CHN LSO
QAT
CAF
NZL
SVN
BLZ
CAN
NOR
PRT
IRL
PRY
CRI
SVK
KOR
SWE
VNM
GRC
USA
FRA
NER
LKA
JOR
ARG
JPN
PHL
AUT KAZ
MWI
CHL
SEN
BGD
NPL
ITA
GAB
TJK
MLI
RWA
KGZSAU ZMB
BOL
IND
CMR
BRA
ISR
IDN
BHR
BGR
CIV
-.2
-.2
-.1
Segregation (S hat)
0
.1
.2
GTM
AFG
UGA
COL
-.1
0
.1
.2
Predicted Segregation (S hat)
.3
.4
-.1
0
.1
.2
Predicted Segregation (S tilde)
language
.3
.4
.4
.4
language
GTM
NAM
ZWE
NGA
Segregation (S hat)
0
.2
UGA
COL
RUS
TGO
KGZ
TUR
AFG
LVA
PER
BEL
HRV
AUT
USA
RWA
BGD
-.2
-.2
PAK
GBR HND PAN
ESP
ZAF
CMR
BEN FIN
MAR
BGR
GIN NICLTU KHM
DNK BLZ
BFA
KOR AUS
ECU
BLR
ETH
UZB
ISL
HUN
TZA
JPN
MEX
CHE
ITA CHN
VNM
EST
MKD
SWE
CZE
ROM
GHACIV
NOR
IND
MOZ
ARM
CAF
SVNNGA ZMBNERTHA
BGD LSO
SVKUKR
CRI
KAZ
KEN BOL
SAU
NZL
MUS
BRA
IDNPRT
SEN
HTI MLI
TJK
MWI PRY
NPL
CHLPHL
GAB
CAN
Segregation (S tilde)
0
.2
ZWE
NAM
-.2
0
.2
Predicted Segregation (S hat)
.4
GTM
TGO
ESP
CMR RUS
COLUGA
KGZ
IDN
AFG
BFA
LVA
GBR PAN
HND
PER
MAR FIN
BGR
PAK ETHNIC
ISL BLZ GHA BEL
HRV
ITA MKD
BLR
LTU ZMB
UZB
IND
GIN
CZE
ARM
CHE
KHM
ESTROM
HUN
BEN
DNK
KOR
MEX
SWE
SVN
ECU
KEN
AUT
NOR
PRT
PRY
CHN
MUS
JPN
UKR
AUS
SVK
TZA
CRI
PHL
ZAF
BOL
LSO CHL
VNM
NZL
NPL CIVTHARWA
KAZ
HTI
SAU
NER
GAB
SEN
CAN BRA
MOZ
TJK
CAF MLI
USA
MWI
-.5
.6
TUR
0
.5
Predicted Segregation (S tilde)
1
religion
.2
.2
religion
IDN
IDN
BGR
BGR
IND
Segregation (S tilde)
0
.1
NGA
INDGIN KEN
NLD
KGZ
NAM
IRL
AUT
CHE
KHM
MUS
QAT
BEN
CAN STP
PRT HUN
BFA
MEX
HTI CHL
HRV
LKA
PRY
GTM
NIC
MDG
ISR
DOM
EGY
NZL
BLZ
SVK
PER ZWE
TGO GAB
EST
SVN
NPL
GHA
LTU
GBR
CAF
CIV
IRN
PHL
UZB
ZMB
USA
TUR
RWA
PAK
SEN
BRA
JPN
UGA
MLI
MWI
CZE
KOR
MOZ
ZAF
ROM
NER
AUS
ARM
CMR
THA
BGD
RUS
ETH
KAZ
-.1
-.1
Segregation (S hat)
0
.1
TZA
VNM
-.2
0
.2
Predicted Segregation (S hat)
.4
NGA
GIN KEN
NAM
PHL
KGZ
QATMOZ
CHE
NLD
NIC
IRL
ISR
ETH
STP AUT
HRV MUS
ROM
BLZ BFA PRT HUN
CANBEN
PRY
CMR
EST
LKA
GAB
SVN
GTM
ZWE
SVK
DOM
EGY
UZB
JPN
MEX
HTI
CHL
NZL
LTU
GHA
PERCAF
MDG
TGOZMBNPL
IRN
RWA
CZE
SEN
GBRPAK
THA TUR
USACIV
UGA
NER
ARM
MWI
ZAF MLI
KORAUS
BRA
BGD
RUS
KHM
-.2
.6
TZA
VNM
0
.2
Predicted Segregation (S tilde)
Figure A.5: Predictive power of the instrument conditional on all controls
54
KAZ
.4
.6
ethnicity
1
CHL
-.5
Rule of law
0
.5
ETH
TZA
TWN
BEN
VNM
MLI
JOR
CHN
FRASEN
MWI MAR
ESP
HUN
AUS
PRTNZL
UGA
AUT
NLD
IRLCZE
RWA
GBR
SVN
CRI
SWE
ZMB
IND GHA CHE
NER
KGZCAN SVK
HRV
BFA
NPL
KHMFIN
EST
PHLARM
QAT
BHR
LTUDNK
LKA
LVAPANUSA
TJK
TGO
JPN
SAU
BOL
ROM
KEN NOR
DEU
BRA
ITA LSO
ISL UZB
GAB
CMR
GIN
IDN
GRC
MEX
PER
HND
CAF
MKD KAZ KOR
PAK
BGD
BLZ
UKR
AFG
PRY
COL
ZAFBLR
BEL
BGR
CIV
ECU
ISR
TUR
GTM
ARG
RUS
-1
ZWE
-.05
0
.05
.1
Fitted value of S hat from stage 1
.15
.2
language
1
CHL
-.5
Rule of law
0
.5
ETH
TZA
VNM
BEN
HUN
MLI
SEN CHNMWIAUSMAR
NZL PRT
ESP
CZE
GBR
CRI ZMBGHAUGA
MOZ
CAN
EST
SVK
SVN
NER
SWE
MUS
CHE
KHM
IND
FIN
PAN
BFA
SAU
DNK
LTU
BGR
TJK
UZB
ARM
LSO
PHL
LVA
NPL
ROM
KEN
NOR
CIV
THA
BRA
JPN
BOL
GAB
ISL
ITA
MEX
CMR
BGDIDN CAF
GIN
NIC
PER
ECU
HND
KOR
PAKKAZ NAM
NGA
GTM
BLZ
MKDUKR
AFG
PRY
COL
HTI
BLR
ZAF
BEL
RWA
AUT
HRV
KGZ
USA
TGO
TUR
ZWE
-1
RUS
-.05
0
.05
.1
Fitted value of S hat from stage 1
.15
religion
1
CHL
.5
VNM
TZA
EGY
ETH
-.5
Rule of law
0
SEN MLI
PRT
HUN
MWI AUS
IRL
UGA
NZL
BEN
AUTNLD
QAT
RWA
ZMB
GBR
CZE
GHA
NER SVK
KHM
HRV
MDG
CHE
CAN
IND
EST
THA
USA
SVN
LKA
MOZ
PHL
KGZ KEN
BFA
STP
NPL
MEX
MUS
BRA
ARM ROM JPN
UZB
IDN
GAB
PER
LTU
BGD
GIN
TGO
NAM
CIV
CAF
CMR PAK NIC
KOR
ISR
IRN
DOM
TUR
PRY
GTM
NGA
BLZ
ZWE
HTI ZAF
BGR
KAZ
-1
RUS
-.05
0
.05
Fitted value of S hat from stage 1
.1
Figure A.6: Residual scatter plots for the rule of law and segregation (second stage of 2SLS)
55
Table A.1: Sources of data on group composition
Country
Afghanistan
Argentina
Armenia
Australia
Austria
Bahrain
Bangladesh
Belarus
Belgium
Belize
Benin
Bolivia
Brazil
Bulgaria
Burkina Faso
Cambodia
Cameroon
Canada
Central African Rep.
Chile
China
Colombia
Costa Rica
Cote D’Ivoire
Croatia
Czech Rep.
Denmark
Dominican Rep.
Ecuador
Egypt
Estonia
Ethiopia
Finland
France
Gabon
Germany
Ghana
Greece
Guatemala
Guinea
Haiti
Honduras
Hungary
Iceland
India
Indonesia
Iran
Ireland
Israel
Italy
Japan
Jordan
Kazakhstan
Kenya
Korea
Ethnicity
Lang
INDEC (www.indec.mecon.ar)
Census (www.armstat.am)
Lang
NSO (www.statistik.at)
Census (www.bahrain.gov.bh)
NSO (www.bbsgov.org)
Census (www.ipums.umn.edu)
Lang
Census (www.statisticsbelize.org.bz)
DHS (www.measuredhs.com)
Census (www.ine.gov.bo)
Census (www.ipums.umn.edu)
Census (www.nsi.bg)
DHS (www.measuredhs.com)
Lang
DHS (www.measuredhs.com)
Census (www.statcan.ca)
DHS (www.measuredhs.com)
Census (www.ine.cl)
Census (www.ipums.umn.edu)
Census (www.ipums.umn.edu)
Census (www.ipums.umn.edu)
DHS (www.measuredhs.com)
Census (www.dzs.hr)
Census (www.czso.cz)
Council of Europe report
.
Census (www.ipums.umn.edu)
.
Census (http://pub.stat.ee)
DHS (www.measuredhs.com)
Lang
INED, Population, 2004 (www.ined.fr)
DHS (www.measuredhs.com)
NSO (www-ec.destatis.de)
DHS (www.measuredhs.com)
NSO (www.statistics.gr)
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
.
Census (www.ine-hn.org)
Census (www.nepszamlalas.hu)
NSO (www.statice.is)
DHS (www.measuredhs.com)
Census (www.bps.go.id)
.
Census (www.cso.ie)
NSO (www1.cbs.gov.il)
NSO (www.dossierimmigrazione.it)
Census (www.stat.go.jp)
Census (www.dos.gov.jo)
NSO (http://en.government.kz)
DHS (www.measuredhs.com)
NSO (www.kosis.kr)
Language
NSO (www.mrrd.gov.af)
.
Census (www.armstat.am)
Census (www.abs.gov.au)
NSO (www.statistik.at)
.
Ethn
Census (www.ipums.umn.edu)
www.eurolang.net
Census (www.statisticsbelize.org.bz)
Ethn
Census (www.ine.gov.bo)
Census (www.ipums.umn.edu)
Census (www.nsi.bg)
DHS (www.measuredhs.com)
Census (www.ipums.umn.edu)
DHS (www.measuredhs.com)
Census (www.statcan.ca)
Ethn
Ethn
Ethn
Ethn
Census (www.ipums.umn.edu)
Ethn
Census (www.dzs.hr)
Ethn
Ethn
.
Census (www.ipums.umn.edu)
.
Census (http://pub.stat.ee)
Ethn
NSO (www.stat.fi)
.
Ethn
.
DHS (www.measuredhs.com)
.
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
Ethn
Census (www.nepszamlalas.hu)
Ethn
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
.
.
.
NSO (www.dossierimmigrazione.it)
Ethn
.
Ethn
DHS (www.measuredhs.com)
Ethn
Continued to the next page
56
Religion
.
.
DHS (www.measuredhs.com)
Census (www.abs.gov.au)
NSO (www.statistik.at)
.
NSO (www.bbsgov.org)
.
.
Census (www.statisticsbelize.org.bz)
DHS (www.measuredhs.com)
.
Census (www.ipums.umn.edu)
Census (www.nsi.bg)
DHS (www.measuredhs.com)
Census (www.ipums.umn.edu)
DHS (www.measuredhs.com)
Census (www.statcan.ca)
DHS (www.measuredhs.com)
Census (www.ine.cl)
.
.
.
DHS (www.measuredhs.com)
Census (www.dzs.hr)
Census (www.czso.cz)
.
DHS (www.measuredhs.com)
.
DHS (www.measuredhs.com)
Census (http://pub.stat.ee)
DHS (www.measuredhs.com)
.
.
DHS (www.measuredhs.com)
.
DHS (www.measuredhs.com)
.
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
.
Census (www.nepszamlalas.hu)
.
DHS (www.measuredhs.com)
Census (www.bps.go.id)
NSO (www.sci.org.ir)
Census (www.cso.ie)
NSO (www1.cbs.gov.il)
.
Census (www.stat.go.jp)
.
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
NSO (www.kosis.kr)
Continued from the previous page
Country
Kyrgyzstan
Latvia
Lesotho
Lithuania
Macedonia
Madagascar
Malawi
Mali
Mauritius
Mexico
Morocco
Mozambique
Namibia
Nepal
Netherlands
New Zealand
Nicaragua
Niger
Nigeria
Norway
Pakistan
Panama
Paraguay
Peru
Philippines
Portugal
Qatar
Romania
Russia
Rwanda
Sao Tome
Saudi Arabia
Senegal
Slovakia
Slovenia
South Africa
Spain
Ethnicity
DHS (www.measuredhs.com)
NSO (www.csb.lv)
Lang
NSO (www.stat.gov.lt)
NSO (www.stat.gov.mk)
.
Lang
DHS (www.measuredhs.com)
.
Census (http://censos.ccp.ucr.ac.cr)
Lang
.
.
Census (www.cbs.gov.np)
NSO (www.cbs.nl)
Census (www.stats.govt.nz)
.
NSO (www.stat-niger.org/)
.
Census (http://statbank.ssb.no)
Lang
Census (http://censos.ccp.ucr.ac.cr)
Census (http://www.dgeec.gov.py)
DHS (www.measuredhs.com)
Census (www.ipums.umn.edu)
NSO (www.sef.pt)
Census (www.planning.gov.qa)
Census (www.ipums.umn.edu)
Census (www.perepis2002.ru)
DHS (www.measuredhs.com)
.
NSO (www.cds.gov.sa)
DHS (www.measuredhs.com)
Census (http://portal.statistics.sk)
Census (www.stat.si)
Lang
Lang
Sri Lanka
Sweden
Switzerland
NSO (www.statistics.gov.lk)
NSO (www.ssd.scb.se)
Lang
Taiwan
Tajikistan
Tanzania
Thailand
Togo
Turkey
NSO (http://eng.dgbas.gov.tw)
NSO (www.stat.tj)
DHS (www.measuredhs.com)
.
DHS (www.measuredhs.com)
Multu, Servet (Int. J. Middle East
Stud., 28, 1996)
NSO (www.ubos.org)
Census (www.ukrcensus.gov.ua)
Council of Europe report
Census (www.census.gov)
DHS (www.measuredhs.com)
Census (www.ipums.umn.edu)
DHS (www.measuredhs.com)
Lang
Uganda
Ukraine
United Kingdom
Usa
Uzbekistan
Vietnam
Zambia
Zimbabwe
Language
Religion
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
NSO (www.csb.lv)
.
DHS (www.measuredhs.com)
.
Ethn
NSO (www.stat.gov.lt)
Ethn
.
.
DHS (www.measuredhs.com)
NSO (www.nso.malawi.net)
DHS (www.measuredhs.com)
Ethn
DHS (www.measuredhs.com)
NSO (www.gov.mu)
NSO (www.gov.mu)
Census (http://censos.ccp.ucr.ac.cr)
Census (http://censos.ccp.ucr.ac.cr)
Census (www.statistic-hcp.ma)
.
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
Census (www.npc.gov.na/census/index.htm)
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
.
NSO (www.cbs.nl)
Census (www.stats.govt.nz)
Census (www.stats.govt.nz)
Census (http://censos.ccp.ucr.ac.cr)
Census (http://censos.ccp.ucr.ac.cr)
Ethn
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
Ethn
.
Census (www.statpak.gov.pk)
Census (www.statpak.gov.pk)
Ethn
.
Census (http://www.dgeec.gov.py)
Census (http://www.dgeec.gov.py)
Census (www.inei.gob.pe/)
Census (www.inei.gob.pe/)
Census (www.ipums.umn.edu)
Census (www.ipums.umn.edu)
Ethn
NSO (www.sef.pt)
.
Census (www.planning.gov.qa)
Census (www.ipums.umn.edu)
Census (www.ipums.umn.edu)
Census (www.perepis2002.ru)
Census (www.perepis2002.ru)
NSO (www.statisticsrwanda.gov.rw)
NSO (www.statisticsrwanda.gov.rw)
.
Census (www.ine.st)
Ethn
.
Ethn
DHS (www.measuredhs.com)
Ethn
Census (http://portal.statistics.sk)
Census (www.stat.si)
Census (www.stat.si)
Census (www.ipums.umn.edu)
Census (www.ipums.umn.edu)
Centro De Investigaciones Sociologi.
cas (www.cis.es)
.
NSO (www.statistics.gov.lk)
Ethn
.
Piguet, E. and Wanner P., Population
Piguet, E. and Wanner P., Population
Studies 31, 2000.
Studies 31, 2000.
.
.
Ethn
.
Ethn
DHS (www.measuredhs.com)
Census (web.nso.go.th)
Census (web.nso.go.th)
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
Ethn
DHS (www.measuredhs.com)
Ethn
Census (www.ukrcensus.gov.ua)
Council of Europe report
Census (www.census.gov)
DHS (www.measuredhs.com)
Ethn
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
.
Council of Europe report
Census (www.census.gov)
DHS (www.measuredhs.com)
Census (www.ipums.umn.edu)
DHS (www.measuredhs.com)
DHS (www.measuredhs.com)
Note: “NSO” – National Statistical Office; “DHS” – Demographics and Health Survey. “Lang” means that language
data were used for ethnicity and “Enth” means that ethnicity data were used for language. This happens when there are no
independent sources of regional-level data for language and ethnicity and, at the same time, national-level statistics on ethnic
and linguistic diversity coincide. More detailed information about the data sources is available from the authors upon request.
57
58
Ethnicity
seg
inst
.373
.152
.012
.141
.004
.000
.067
.000
.011
.000
.189
.000
.005
.012
.047
.111
.203
.279
.066
.005
.210
.092
.075
.108
.047
.149
.092
.453
.235
.094
.009
.001
.042
.007
.006
.000
.106
.000
.022
.006
.085
.001
.280
.027
.044
.007
.147
.303
.008
.051
.004
.002
.024
.000
.
.
.322
.244
.
.
.018
.003
.214
.079
.064
.000
.020
.000
.026
.000
.001
.017
.112
.035
.002
.007
.384
.205
.242
.090
.
.
.215
.036
.003
.026
.004
.000
.090
.090
.179
.003
.
.
.003
.000
.191
.435
.006
.062
.002
.000
.019
.030
.028
.043
.235
.008
.002
.000
Language
seg
inst
.373
.401
.
.
.002
.000
.067
.000
.047
.392
.
.
.005
.000
.047
.103
.203
.327
.082
.014
.210
.037
.093
.118
.006
.000
.103
.141
.028
.019
.006
.001
.215
.072
.017
.000
.106
.000
.022
.001
.085
.000
.280
.033
.014
.000
.147
.109
.035
.229
.004
.002
.024
.000
.
.
.104
.001
.
.
.025
.088
.214
.021
.064
.070
.
.
.026
.015
.
.
.053
.047
.
.
.495
.021
.224
.055
.001
.000
.215
.000
.002
.016
.004
.000
.146
.004
.129
.001
.
.
.
.
.
.
.141
.001
.002
.000
.
.
.028
.036
.068
.002
.002
.000
Religion
seg
inst
.
.
.
.
.002
.002
.003
.000
.006
.011
.
.
.008
.063
.
.
.
.
.002
.004
.103
.111
.
.
.019
.001
.229
.630
.093
.167
.148
.008
.007
.010
.025
.000
.014
.146
.000
.000
.
.
.
.
.
.
.053
.160
.026
.071
.013
.011
.
.
.005
.000
.
.
.047
.031
.010
.071
.119
.360
.
.
.
.
.010
.047
.
.
.057
.036
.
.
.016
.024
.125
.173
.015
.000
.
.
.067
.317
.
.
.186
.127
.271
.225
.001
.006
.004
.000
.041
.179
.
.
.031
.000
.
.
.098
.752
.151
.246
.009
.000
Kyrgyzstan
Latvia
Lesotho
Lithuania
Macedonia
Madagascar
Malawi
Mali
Mauritius
Mexico
Morocco
Mozambique
Namibia
Nepal
Netherlands
New Zealand
Nicaragua
Niger
Nigeria
Norway
Pakistan
Panama
Paraguay
Peru
Philippines
Portugal
Qatar
Romania
Russia
Rwanda
Sao Tome
Saudi Arabia
Senegal
Slovakia
Slovenia
South Africa
Spain
Sri Lanka
Sweden
Switzerland
Taiwan
Tajikistan
Tanzania
Thailand
Togo
Turkey
Uganda
Ukraine
United Kingdom
USA
Uzbekistan
Vietnam
Zambia
Zimbabwe
Country
Ethnicity
seg
inst
.032
.050
.014
.022
.011
.000
.052
.057
.095
.044
.
.
.023
.014
.094
.155
.
.
.160
.078
.253
.142
.
.
.
.
.026
.004
.001
.000
.042
.000
.
.
.118
.180
.
.
.022
.000
.316
.204
.186
.000
.002
.002
.169
.048
.119
.000
.015
.000
.025
.000
.018
.007
.155
.012
.011
.025
.
.
.041
.000
.085
.097
.004
.009
.003
.004
.247
.096
.244
.005
.021
.000
.001
.000
.088
.069
.044
.000
.011
.197
.141
.008
.
.
.202
.083
.357
.468
.371
.024
.025
.016
.026
.000
.011
.012
.058
.121
.037
.003
.055
.031
.394
.406
Language
seg
inst
.273
.425
.153
.190
.011
.000
.052
.057
.095
.062
.
.
.019
.011
.094
.121
.021
.000
.142
.061
.253
.275
.166
.014
.318
.038
.042
.010
.
.
.003
.000
.116
.000
.118
.216
.052
.039
.022
.036
.316
.124
.186
.000
.022
.002
.254
.189
.107
.000
.015
.000
.
.
.016
.030
.299
.314
.048
.286
.
.
.041
.018
.085
.034
.004
.009
.007
.007
.247
.185
.244
.000
.
.
.001
.000
.088
.140
.
.
.011
.257
.141
.006
.020
.072
.312
.222
.357
.865
.371
.008
.012
.007
.170
.000
.070
.604
.055
.163
.037
.002
.030
.019
.394
.539
.022
.048
.030
.059
.005
.012
.010
.196
.
.018
.
.001
.007
.010
.000
.020
.028
.158
.001
.010
.
.017
.011
.004
.009
.
.088
.
.007
.
.
.220
.017
.048
.001
.033
.
.017
.021
.054
.014
.023
.047
.022
.134
.116
.058
.000
.007
.007
.083
.
.046
.
.001
.031
.000
.000
.007
.104
.132
.003
.000
.
.037
.145
.006
.048
.
.000
.
.022
.
.
.390
.032
.067
.003
.324
.
.000
.022
.044
.053
.036
.244
Religion
seg
inst
.095
.308
.
.
.
.
.013
.270
.
.
.056
.000
.014
.026
.021
.090
.062
.000
.022
.005
Note: “seg” indices the value of the index of segregation, while “inst” indicates the value of the instrument for segregation.
Afghanistan
Argentina
Armenia
Australia
Austria
Bahrain
Bangladesh
Belarus
Belgium
Belize
Benin
Bolivia
Brazil
Bulgaria
Burkina Faso
Cambodia
Cameroon
Canada
Central African Rep.
Chile
China
Colombia
Costa Rica
Cote D’Ivoire
Croatia
Czech Rep.
Denmark
Dominican Rep.
Ecuador
Egypt
Estonia
Ethiopia
Finland
France
Gabon
Germany
Ghana
Greece
Guatemala
Guinea
Haiti
Honduras
Hungary
Iceland
India
Indonesia
Iran
Ireland
Israel
Italy
Japan
Jordan
Kazakhstan
Kenya
Korea
Country
Table A.2: Segregation indices Sb and their corresponding instruments
59
Ethnicity
seg
inst
.412
.169
.012
.141
.006
.000
.067
.000
.014
.000
.189
.000
.005
.012
.054
.127
.203
.279
.098
.008
.214
.095
.075
.108
.047
.149
.100
.495
.320
.129
.021
.005
.071
.013
.007
.000
.116
.000
.022
.006
.104
.001
.299
.029
.044
.007
.153
.323
.019
.151
.021
.013
.024
.000
.
.
.322
.244
.
.
.027
.004
.367
.146
.088
.001
.017
.001
.080
.000
.005
.140
.175
.058
.003
.013
.384
.205
.244
.091
.
.
.215
.036
.003
.026
.004
.000
.090
.090
.357
.006
.
.
.003
.000
.191
.435
.006
.062
.003
.000
.019
.030
.040
.063
.273
.010
.002
.000
Language
seg
inst
.412
.446
.
.
.009
.000
.067
.000
.047
.392
.
.
.005
.000
.054
.118
.203
.327
.173
.030
.214
.038
.156
.204
.003
.000
.103
.141
.291
.262
.021
.006
.407
.140
.073
.000
.116
.000
.022
.001
.104
.000
.299
.036
.014
.000
.153
.117
.035
.229
.021
.013
.024
.000
.
.
.107
.001
.
.
.056
.196
.367
.040
.088
.103
.
.
.080
.048
.
.
.267
.293
.
.
.398
.022
.244
.061
.001
.000
.215
.000
.002
.016
.004
.000
.373
.010
.447
.003
.
.
.
.
.
.
.141
.001
.003
.000
.
.
.040
.052
.288
.009
.002
.000
Religion
seg
inst
.
.
.
.
.007
.012
.005
.000
.016
.027
.
.
.010
.081
.
.
.
.
.014
.025
.105
.114
.
.
.018
.001
.231
.639
.093
.167
.018
.014
.072
.094
.029
.000
.016
.173
.005
.000
.
.
.
.
.
.
.054
.166
.032
.089
.025
.023
.
.
.020
.000
.
.
.047
.031
.013
.096
.138
.427
.
.
.
.
.012
.066
.
.
.057
.036
.
.
.031
.053
.125
.173
.015
.000
.
.
.067
.326
.
.
.280
.202
.282
.237
.002
.016
.004
.001
.072
.322
.
.
.087
.000
.
.
.096
.787
.154
.253
.015
.000
Kyrgyzstan
Latvia
Lesotho
Lithuania
Macedonia
Madagascar
Malawi
Mali
Mauritius
Mexico
Morocco
Mozambique
Namibia
Nepal
Netherlands
New Zealand
Nicaragua
Niger
Nigeria
Norway
Pakistan
Panama
Paraguay
Peru
Philippines
Portugal
Qatar
Romania
Russia
Rwanda
Sao Tome
Saudi Arabia
Senegal
Slovakia
Slovenia
South Africa
Spain
Sri Lanka
Sweden
Switzerland
Taiwan
Tajikistan
Tanzania
Thailand
Togo
Turkey
Uganda
Ukraine
United Kingdom
USA
Uzbekistan
Vietnam
Zambia
Zimbabwe
Country
Ethnicity
seg
inst
.040
.065
.031
.052
.002
.000
.066
.072
.128
.060
.
.
.078
.048
.108
.182
.
.
.160
.078
.253
.142
.
.
.
.
.098
.014
.003
.000
.046
.000
.
.
.128
.196
.
.
.022
.000
.401
.265
.186
.000
.011
.009
.169
.048
.197
.000
.016
.000
.025
.000
.030
.012
.245
.019
.013
.031
.
.
.041
.000
.122
.144
.012
.026
.007
.011
.254
.099
.355
.008
.052
.000
.004
.000
.189
.149
.086
.000
.015
.356
.257
.014
.
.
.213
.090
.357
.468
.489
.032
.054
.035
.027
.000
.069
.087
.076
.161
.051
.005
.055
.031
.394
.406
.004
.189
.
.015
.257
.033
.380
.357
.489
.069
.170
.092
.098
.051
.242
.394
.000
.303
.
.465
.011
.124
.290
.865
.011
.037
.000
1
.604
.003
.157
.540
Language
seg
inst
.273
.425
.165
.214
.002
.000
.066
.072
.128
.084
.
.
.078
.046
.108
.142
.021
.000
.142
.061
.253
.275
.166
.014
.534
.065
.131
.034
.
.
.011
.000
.128
.000
.128
.235
.563
.497
.022
.036
.401
.161
.186
.000
.125
.011
.281
.224
.263
.000
.016
.000
.
.
.032
.062
.299
.314
.048
.286
.
.
.041
.018
.122
.050
.012
.026
.010
.010
.254
.191
.355
.000
Religion
seg
inst
.096
.315
.
.
.
.
.017
.402
.
.
.056
.000
.015
.027
.014
.110
.062
.000
.023
.005
.
.
.131
.136
.064
.190
.031
.117
.066
.066
.006
.000
.060
.033
.011
.012
.200
.085
.
.
.019
.051
.
.
.010
.014
.011
.058
.127
.000
.007
.000
.021
.020
.092
.354
.158
.132
.017
.046
.014
.000
.
.
.017
.037
.015
.203
.022
.033
.017
.090
.
.
.090
.000
.
.
.037
.126
.
.
.
.
.220
.390
.040
.078
.049
.068
.003
.073
.034
.396
.
.
.019
.000
.032
.038
.059
.049
.038
.142
.023
.058
.049
.316
Note: “seg” indices the value of the index of segregation, while “inst” indicates the value of the instrument for segregation.
Afghanistan
Argentina
Armenia
Australia
Austria
Bahrain
Bangladesh
Belarus
Belgium
Belize
Benin
Bolivia
Brazil
Bulgaria
Burkina Faso
Cambodia
Cameroon
Canada
Central African Rep.
Chile
China
Colombia
Costa Rica
Cote D’Ivoire
Croatia
Czech Rep.
Denmark
Dominican Rep.
Ecuador
Egypt
Estonia
Ethiopia
Finland
France
Gabon
Germany
Ghana
Greece
Guatemala
Guinea
Haiti
Honduras
Hungary
Iceland
India
Indonesia
Iran
Ireland
Israel
Italy
Japan
Jordan
Kazakhstan
Kenya
Korea
Country
Table A.3: Segregation indices Se and their corresponding instruments
Table A.4: Summary statistics
variable
Obs.
Mean
SD
Min
Panel A: Segregation and Fractionalization Indices
e
Segregation (ethnicity) S
97
0.12
0.12
0
b
Segregation (ethnicity) S
97
0.10
0.11
0
e
Segregation (language) S
92
0.16
0.14
0
b
Segregation (language) S
92
0.11
0.11
0
e
Segregation (religion) S
78
0.06
0.06
0
b
Segregation (religion) S
78
0.05
0.06
0
Fractionalization (ethnicity)
97
0.37
0.27
0
Fractionalization (language)
93
0.36
0.27
0
Fractionalization (religion)
78
0.43
0.24
0
Panel B: Dependent and control variables
Voice and accountability
109
0.07
0.93
-1.63
Political stability
109
-0.11
0.88
-2.26
Government effectiveness
109
0.13
1.00
-1.46
Regulatory quality
109
0.15
0.86
-2.12
Rule of law
109
0.05
1.00
-1.68
Control of corruption
109
0.06
1.06
-1.41
ln (population)
109
16.41
1.57
11.84
ln (GDP per capita)
109
8.53
1.20
6.27
Protestants share
109
12.96
22.38
0
Muslims share
109
19.61
33.04
0
Catholics share
109
34.09
36.76
0
Latitude
109
0.32
0.20
0
English legal origin
109
0.25
0.43
0
French legal origin
109
0.44
0.50
0
German legal origin
109
0.06
0.23
0
Socialist legal origin
109
0.21
0.41
0
Scandinavian legal origin
109
0.05
0.21
0
Democratic tradition
109
4.86
3.62
0
Mountains
109
0.28
0.26
0
Panel C: Instrumental variables
e (ethnicity)
Predicted S
97
0.08
0.11
0
b (ethnicity)
Predicted S
97
0.06
0.10
0
e (language)
Predicted S
92
0.12
0.19
0
b (language)
Predicted S
92
0.09
0.15
0
e (religion)
Predicted S
78
0.11
0.15
0
b (religion)
Predicted S
78
0.09
0.14
0
60
Max
0.49
0.39
0.56
0.49
0.28
0.27
0.92
0.89
0.83
1.54
1.48
2.29
1.67
2.07
2.47
20.95
10.41
97.80
99.40
96.90
0.72
1
1
1
1
1
10
0.94
0.49
0.47
1
0.86
0.79
0.75
61
e
S
b
S
e
S
b
S
F
e
S
b
S
e
S
b
S
F
e
S
b
S
e
S
b
S
F
seg
seg
inst
inst
frac
seg
seg
inst
inst
frac
seg
seg
inst
inst
frac
1
0.48
0.53
0.42
0.80
0.84
0.22
0.31
0.39
0.21
0.22
0.23
0.22
0.07
b seg
S
1
0.97
0.19
0.28
0.41
0.53
0.57
0.17
0.12
0.13
0.32
0.29
-0.15
1
0.18
0.31
0.46
0.47
0.56
0.15
0.11
0.13
0.31
0.27
-0.15
Ethnicity
e inst
b inst
S
S
1
0.55
0.32
0.10
0.05
0.89
0.29
0.23
0.21
0.23
0.31
F frac
1
0.80
0.28
0.25
0.54
0.46
0.41
0.20
0.19
0.10
e seg
S
1
0.29
0.43
0.36
0.16
0.13
0.18
0.16
0.03
b seg
S
1
0.89
0.05
-0.03
0.01
-0.02
-0.04
0.00
1
0.03
-0.10
-0.06
0.03
0.01
-0.10
Language
e inst
b inst
S
S
1
0.34
0.26
0.26
0.26
0.26
F frac
1
0.89
0.50
0.52
0.07
e seg
S
1
0.50
0.58
0.01
b seg
S
1
0.96
0.05
1
0.05
Religion
e inst S
b inst
S
1
F frac
Note: “seg” indicates the actual indices of segregation; “inst” indicates their corresponding instruments; “frac” indicates indices of fractionalization.
Ethnicity
Ethnicity
Ethnicity
Ethnicity
Ethnicity
Language
Language
Language
Language
Language
Religion
Religion
Religion
Religion
Religion
e seg
S
1
0.96
0.40
0.43
0.50
0.85
0.80
0.20
0.25
0.48
0.32
0.33
0.28
0.27
0.09
Table A.5: Pair-wise correlations: Indices of actual and predicted segregation and indices of fractionalization
Table A.6: Sources of control variables
Variable
ln (Population)
ln (GDP per capita)
Religion
Legal origin
Latitude
Democratic tradition
Fertility
Investment
Openness
Mountains
Colonial origin
Region
Island
OECD
Transition
Partitioned
Squiggliness
Elevation
Rivers
Definition
Natural log of population in the country. Average for the years 1995-2004. Source:
World Development Indicators 2006.
Natural log of GDP in constant 2000 international dollars per capita. Average for
the years 1995-2004. Source: World Development Indicators 2006. For initial value
of GDP per capita we use natural log of GDP in constant 2000 international dollars
per capita. Average for the years 1975-1980. Source: World Development Indicators
2006.
Identifies the percentage of the population of each country that belonged to the three
most widely spread religions in the world in 1980. For countries of recent formation,
the data is available for 1990-95. The numbers are in percent (scale from 0 to 100).
The three religions identified here are: (1) Romanic Catholic; (2) Protestant; and
(3) Muslim. Source: La Porta et. al. (1998). Original sources: World Christian
Encyclopedia 1982, Worldmark Encyclopedia of Nations 1995, Statistical Abstract of
the World 1995, Demographic Yearbook 1995, CIA World Factbook 1996
Identifies the legal origin of the Company Law or Commercial Code of each country.
There are five possible origins: (1) English Common Law; (2) French Commercial
Code; (3) German Commercial Code; (4) Scandinavian Commercial Code; and (5)
Socialist/Communist laws. Source: La Porta et. al. (1998). Original sources: CIA
World Factbook 1996.
The absolute value of the latitude of the country, scaled to take values between 0 and
1. Source: La Porta et. al. (1998). Original source: CIA World Factbook 1996
Democracy score index. Scale from 0 to 10, with lower values indicating a less democratic environment. Average for the years 1975-2004. Source: Polity IV Project:
Political Regime Characteristics and Transitions, 1800-2006.
Fertility rate (births per woman). Average for the years 1975-2004. Source: World
Development Indicators 2006.
Investment share as % of GDP. Average for the years 1975-2004. Source: Penn World
Table 6.2.
Export plus Import as % of GDP. Average for the years 1975-2004. Source: Penn
World Table 6.2.
Measure of mountains in the country. Source: William Easterly’s data.
Identifies countries that were colonized by a Western overseas colonial power since
1700 for at least 10 years. Source: Teorell and Hadenius (2005).
Identifies the region where the country is situated. There are six possible regions:
(1) East Asia and Pacific; (2) Europe and Central Asia; (3) Latin America and
Carribean; (3) Middle East and North Africa; (4) North America; (5) South Asia;
and (6) Sub-Saharan Africa. Source: World Bank.
Identifies countries that are situated on islands and therefore have no bordering countries. Source: CIA World Factbook 1996
Identifies countries that are currently members of OECD. These countries are Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Japan, Mexico, Netherlands, New
Zealand, Norway, Portugal, Slovak Republic, Slovenia, Spain, Sweden, Switzerland,
Turkey, United Kingdom, USA. Source: wikipedia.org.
Identifies transition countries. These countries are Armenia, Belarus, Bulgaria, Croatia, Czech Republic, Estonia, Hungary, Kazakhstan, Kyrgyzstan, Latvia, Lithuania, Macedonia, Romania, Russian Federation, Slovak Republic, Slovenia, Tajikistan,
Ukraine, Uzbekistan, Source: wikipedia.org.
Percent of the population of each country that belongs to groups partitioned by the
border. Source: Alesina et. al. (2006).
Log of basic fractal index based on World Vector Shoreline Dataset (GIS format).
This variable measures squiggliness of each country’s border. Source: Alesina et. al.
(2006).
Standard deviation of elevation of each country in meters. Source: GIS dataset.
Share of area of the country covered by large perennial bodies (rivers, lakes, seas).
Source: GIS dataset.
62
63
Control
of corr.
0.11
[0.84]
0.52*
[0.28]
0.831
yes
full
78
-5.07***
[1.32]
0.94**
[0.44]
0.146
no
full
78
0.39
[0.72]
0.45
[0.30]
0.854
yes
democ
64
Rule of
law
0.03
[0.93]
0.36
[0.25]
0.839
yes
full
78
-4.53***
[1.33]
0.76*
[0.42]
0.118
no
full
78
0.06
[0.81]
0.36
[0.26]
0.861
yes
democ
64
Note: Robust standard errors adjusted for heteroscedasticity in brackets. * significant at 10%; ** significant at 5%; *** significant at 1%.
Voice
Religion
Political
Govern-t
Regul.
stability
effectiv.
quality
Panel A. Baseline: All controls and full sample
Segregation
0.51
-0.97
0.28
0.9
[0.91]
[1.20]
[0.88]
[0.85]
Fractionalization
0.13
0.59*
0.43
0.15
[0.25]
[0.35]
[0.27]
[0.23]
R-squared
0.814
0.679
0.838
0.762
Controls
yes
yes
yes
yes
Sample
full
full
full
full
Obs.
78
78
78
78
Panel B. No controls and full sample
Segregation
-3.71*** -4.72*** -4.29***
-3.53***
[1.40]
[1.39]
[1.29]
[1.25]
Fractionalization
0.78*
0.72**
0.75*
0.63*
[0.40]
[0.35]
[0.43]
[0.37]
R-squared
0.109
0.153
0.11
0.103
Controls
no
no
no
no
Sample
full
full
full
full
Obs.
78
78
78
78
Panel C: All controls; sample excludes dictatorships
Segregation
-0.23
-1.3
0.53
0.38
[0.92]
[1.11]
[0.78]
[0.85]
Fractionalization
0.49**
0.68*
0.45
0.06
[0.24]
[0.37]
[0.28]
[0.25]
R-squared
0.809
0.722
0.862
0.789
Controls
yes
yes
yes
yes
Sample
democ
democ
democ
democ
Obs.
64
64
64
64
Table A.7: Religious segregation and the quality of government, OLS
64
Control
of corr.
-1.11
[1.98]
0.58**
[0.28]
0.827
yes
full
78
16.08
29.49
-7.50**
[3.07]
0.94**
[0.45]
0.124
no
full
78
14.96
38.64
0.22
[1.64]
0.46
[0.30]
0.854
yes
democ
64
22.24
26.03
Rule of
law
-0.87
[1.98]
0.4
[0.25]
0.837
yes
full
78
16.08
29.49
-6.65**
[2.92]
0.76*
[0.43]
0.1
no
full
78
14.96
38.64
-0.04
[1.65]
0.36
[0.26]
0.861
yes
democ
64
22.24
26.03
Note: Robust standard errors adjusted for heteroscedasticity in brackets. * significant at 10%; ** significant at 5%; *** significant at 1%. “F-stat (het)” reports F-statistics for
the excluded instrument from the first stage under the assumption of heteroscedasticity; and “F-stat (hom)” – under the assumption of homoscedasticity.
Voice
Religion
Political
Govern-t
Regul.
stability
effectiv.
quality
Panel A. Baseline: All controls and full sample
Segregation
0.55
-2.13
-1.14
0.97
[2.09]
[2.34]
[2.03]
[2.03]
Fractionalization
0.12
0.65*
0.50*
0.14
[0.27]
[0.38]
[0.27]
[0.25]
R-squared
0.814
0.675
0.834
0.762
Controls
yes
yes
yes
yes
Sample
full
full
full
full
Obs.
78
78
78
78
F-stat (het)
16.08
16.08
16.08
16.08
F-stat (hom)
29.49
29.49
29.49
29.49
Panel B. No controls and full sample
Segregation
-5.72
-4.39
-6.56**
-4.46
[3.52]
[2.84]
[2.70]
[2.74]
Fractionalization
0.78*
0.72**
0.75*
0.63*
[0.41]
[0.35]
[0.43]
[0.38]
R-squared
0.09
0.152
0.089
0.099
Controls
no
no
no
no
Sample
full
full
full
full
Obs.
78
78
78
78
F-stat (het)
14.96
14.96
14.96
14.96
F-stat (hom)
38.64
38.64
38.64
38.64
Panel C: All controls; sample excludes dictatorships
Segregation
0.1
-2.41
-0.61
-0.39
[1.99]
[2.79]
[1.42]
[2.21]
Fractionalization
0.48*
0.72*
0.49*
0.08
[0.25]
[0.40]
[0.28]
[0.26]
R-squared
0.808
0.718
0.859
0.788
Controls
yes
yes
yes
yes
Sample
democ
democ
democ
democ
Obs.
64
64
64
64
F-stat (het)
22.24
22.24
22.24
22.24
F-stat (hom)
26.03
26.03
26.03
26.03
Table A.8: Religious segregation and the quality of government, the second stage of 2SLS
Table A.9: Linguistic segregation and alternative measures of government quality
Dep. var.
b
LS
LF
Obs.
F-het
F-hom
R-sqrd
Specification
b
LS
LF
Obs.
F-het
F-hom
R-sqrd
Specification
Dep. var.
b
LS
LF
Obs.
F-het
F-hom
R-sqrd
Specification
b
LS
LF
Obs.
F-het
F-hom
R-sqrd
Specification
ICRG quality of
TI corruption
government
index
OLS
2SLS
OLS
2SLS
-0.36*
0.09
-3.08**
-0.52
[0.18]
[0.61]
[1.51]
[4.81]
-0.24***
-0.31***
-2.23***
-2.60***
[0.08]
[0.11]
[0.73]
[0.90]
79
79
91
91
16.41
15.52
12.99
19.65
0.16
0.11
0.12
0.11
Full sample, no additional controls
-0.34*
-0.49
-1.76
-2.04
[0.17]
[0.32]
[1.07]
[2.22]
0.11
0.13
0.31
0.35
[0.09]
[0.09]
[0.57]
[0.55]
66
66
75
75
4.54
7.04
8.30
14.01
0.84
0.83
0.87
0.87
Democracies sample, all controls
EF corruption
index
OLS
2SLS
-32.40**
-10.67
[15.61]
[47.03]
-23.85***
-27.02***
[7.39]
[8.78]
91
91
15.52
19.65
0.14
0.13
EF Property
EF Regulation
rights index
index
OLS
2SLS
OLS
2SLS
-50.92***
-11.5
-15.31
-10.1
[17.77]
[62.79]
[15.89]
[42.47]
-19.01**
-24.77**
-22.55***
-23.31***
[7.95]
[11.27]
[6.05]
[7.77]
91
91
91
91
15.52
15.52
19.65
19.65
0.13
0.10
0.16
0.16
Full sample, no additional controls
-53.12***
-66.84**
-4.9
-46.3
[19.20]
[29.89]
[19.57]
[33.35]
4.8
6.71
-1.09
4.67
[7.64]
[8.01]
[7.01]
[8.74]
75
75
75
75
7.04
7.04
14.01
14.01
0.80
0.80
0.69
0.64
Democracies sample, all controls
Tax evasion
index
OLS
-4.73***
[1.32]
-0.01
[0.67]
33
-25.71**
[10.75]
3.35
[5.48]
75
0.89
0.23
-1.65
[1.72]
-0.06
[0.78]
32
0.76
-48.49**
[21.33]
6.52
[6.15]
75
7.04
14.01
0.88
2SLS
-1.06
[3.62]
-0.61
[0.78]
33
9.93
6.78
0.11
-0.97
[2.98]
-0.13
[0.78]
32
3.83
3.92
0.76
Note: Robust standard errors adjusted for heteroscedasticity in brackets. * significant at 10%; ** significant at 5%; ***
significant at 1%. “F-stat(het)” reports F-statistics for the excluded instrument from the first stage under the assumption of
heteroscedasticity; and “F-stat(hom)” – under the assumption of homoscedasticity.
65
66
-1.30**
[0.53]
0.14
[0.14]
97
0.839
Segregation (ethnicity)
-1.27**
[0.54]
0.00
[0.00]
97
0.837
Political stability
Govern-t effectiv.
Regul. Quality
Rule of law
Control of corr.
OLS
IV
OLS
IV
OLS
IV
OLS
IV
OLS
IV
Coefficient on segregation in baseline specification (presented for comparison)
-1.28 -1.98***
-3.65**
-0.45
-2.14***
-0.88
-2.10* -1.20** -2.47***
-0.57
-1.77***
[1.04]
[0.64]
[1.43]
[0.57]
[0.65]
[0.78]
[1.22]
[0.52]
[0.67]
[0.55]
[0.65]
Results with control for squiggliness from Alesina et al. (2010)
-1.34 -2.32*** -3.83***
-0.56
-2.06***
-1.06
-2.00 -1.35**
-2.40***
-0.62
-1.67**
[1.10]
[0.62]
[1.35]
[0.55]
[0.76]
[0.79]
[1.26]
[0.51]
[0.72]
[0.53]
[0.73]
-1.35
-5.01
-5.60
3.21
2.62
-0.87
-1.24
-0.35
-0.76
2.63
2.22
[2.52]
[3.04]
[3.38]
[2.88]
[3.13]
[2.81]
[2.85]
[2.95]
[3.15]
[3.62]
[3.69]
86
86
86
86
86
86
86
86
86
86
86
0.839
0.737
0.715
0.859
0.842
0.765
0.756
0.866
0.858
0.867
0.860
13.87
13.87
13.87
13.87
13.87
13.87
27.81
27.81
27.81
27.81
27.81
27.81
Results with control for partitioned dummy
-1.48 -2.18***
-4.10**
-0.55
-2.46***
-1.02
-2.42* -1.21** -2.60***
-0.59
-1.91***
[1.10]
[0.62]
[1.60]
[0.59]
[0.72]
[0.81]
[1.30]
[0.55]
[0.76]
[0.57]
[0.72]
0.15
0.23
0.35*
0.12
0.24
0.16
0.25
0.01
0.10
0.02
0.11
[0.15]
[0.17]
[0.19]
[0.16]
[0.18]
[0.14]
[0.16]
[0.12]
[0.14]
[0.15]
[0.17]
97
97
97
97
97
97
97
97
97
97
97
0.838
0.741
0.711
0.861
0.839
0.779
0.762
0.876
0.865
0.872
0.862
13.62
13.62
13.62
13.62
13.62
13.62
27.78
27.78
27.78
27.78
27.78
27.78
Results with control for the share of partitioned group from Alesina et al. (2010)
-1.03 -2.39***
-4.46**
-0.72
-2.88***
-0.91
-2.00 -1.32** -2.74***
-0.86
-2.39**
[1.54]
[0.69]
[2.12]
[0.61]
[0.95]
[0.77]
[1.51]
[0.55]
[0.94]
[0.59]
[0.99]
-0.00
-0.00
0.00
0.00
0.00*
-0.00
-0.00
0.00
0.00
0.00
0.00*
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
79
79
79
79
79
79
79
79
79
79
79
0.839
0.735
0.699
0.860
0.830
0.758
0.747
0.882
0.869
0.868
0.855
12.52
12.52
12.52
12.52
12.52
12.52
16.86
16.86
16.86
16.86
16.86
16.86
Results with control for the share of partitioned ethnic group constructed using our data
-1.57 -1.78***
-3.49**
-0.48
-2.67***
-0.98
-2.70* -1.19** -2.74***
-0.62
-2.23***
[1.06]
[0.61]
[1.47]
[0.60]
[0.84]
[0.82]
[1.46]
[0.55]
[0.85]
[0.58]
[0.78]
0.00
-0.00
-0.00
0.00
0.00
0.00
0.00
-0.00
0.00
0.00
0.00
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
[0.00]
97
97
97
97
97
97
97
97
97
97
97
0.837
0.739
0.716
0.860
0.831
0.777
0.752
0.876
0.862
0.872
0.858
12.19
12.19
12.19
12.19
12.19
12.19
23.35
23.35
23.35
23.35
23.35
23.35
IV
Note: Robust standard errors adjusted for heteroscedasticity in brackets. * significant at 10%; ** significant at 5%; *** significant at 1%. “F-stat (het)” reports F-statistics
for the excluded instrument from the first stage under the assumption of heteroscedasticity; and “F-stat (hom)” – under the assumption of homoscedasticity. Results are
similar for the linguistic diversity. All reported regressions contain the standard set of controls.
Observations
R-squared
F-het
F-hom
Partitioned (our data)
Segregation (ethnicity)
Observations
R-squared
F-het
F-hom
Partitioned (Alesina et al.)
Segregation (ethnicity)
Observations
R-squared
F-het
F-hom
Partitioned dummy
Observations
R-squared
F-het
F-hom
-1.18*
[0.61]
-0.00
[0.00]
79
0.839
-1.43***
[0.50]
-1.38
[2.54]
86
0.839
Segregation (ethnicity)
Squiggliness
-1.18**
[0.51]
Segregation (ethnicity)
Voice
OLS
Table A.10: Controls for artificial states
67
Political stability
Govern-t effectiv.
Regul. Quality
Rule of law
Control of corr.
OLS
IV
OLS
IV
OLS
IV
OLS
IV
OLS
IV
Coefficient on segregation in baseline specification (presented for comparison)
-1.28 -1.98***
-3.65**
-0.45 -2.14***
-0.88
-2.10*
-1.20**
-2.47***
-0.57
-1.77***
[1.04]
[0.64]
[1.43]
[0.57]
[0.65]
[0.78]
[1.22]
[0.52]
[0.67]
[0.55]
[0.65]
Results with control for mean of the respective quality of government indicator among neighbors
-1.38 -2.11***
-3.59**
-0.52
-1.97**
-0.92
-2.08
-1.23**
-2.22***
-0.58
-1.43*
[1.13]
[0.63]
[1.37]
[0.57]
[0.77]
[0.82]
[1.37]
[0.50]
[0.69]
[0.54]
[0.77]
0.22
0.13
0.33
0.09
0.29
0.21
0.37
0.11
0.25
-0.08
0.04
[0.27]
[0.29]
[0.34]
[0.19]
[0.25]
[0.27]
[0.34]
[0.20]
[0.25]
[0.21]
[0.25]
0.16
0.26**
0.25** 0.26***
0.26***
0.17*
0.17
0.31***
0.30***
0.40***
0.39***
[0.10]
[0.12]
[0.12]
[0.08]
[0.08]
[0.10]
[0.11]
[0.10]
[0.10]
[0.11]
[0.11]
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
90
90
90
90
90
90
90
90
90
90
90
0.846
0.744
0.724
0.873
0.857
0.772
0.759
0.887
0.880
0.897
0.892
14.26
13.47
13.62
13.77
13.60
13.62
26.93
26.86
26.94
27.02
26.90
26.91
Results with controls for means of segregation and fractionalization among neighbors
-1.16 -2.52*** -3.91***
-0.62 -2.19***
-1.06
-2.17* -1.39***
-2.61***
-0.77
-1.91***
[1.16]
[0.62]
[1.45]
[0.57]
[0.77]
[0.81]
[1.28]
[0.50]
[0.74]
[0.52]
[0.70]
0.20
0.30
0.52
0.08
0.32
0.22
0.39
0.08
0.27
-0.10
0.08
[0.28]
[0.30]
[0.37]
[0.25]
[0.30]
[0.28]
[0.35]
[0.25]
[0.30]
[0.27]
[0.31]
0.57
1.76**
2.20**
0.19
0.69
0.53
0.87
0.29
0.67
0.30
0.65
[0.69]
[0.83]
[1.00]
[0.68]
[0.73]
[0.73]
[0.82]
[0.61]
[0.70]
[0.66]
[0.74]
-0.16
-0.19
-0.25
0.00
-0.07
-0.30
-0.35
-0.21
-0.26
-0.03
-0.09
[0.30]
[0.42]
[0.45]
[0.35]
[0.35]
[0.40]
[0.38]
[0.32]
[0.32]
[0.33]
[0.33]
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
88
88
88
88
88
88
88
88
88
88
88
0.842
0.745
0.727
0.859
0.842
0.767
0.756
0.872
0.862
0.879
0.871
12.06
12.06
12.06
12.06
12.06
12.06
28.53
28.53
28.53
28.53
28.53
28.53
IV
Note: Robust standard errors adjusted for heteroscedasticity in brackets. * significant at 10%; ** significant at 5%; *** significant at 1%. “F-stat (het)” reports F-statistics
for the excluded instrument from the first stage under the assumption of heteroscedasticity; and “F-stat (hom)” – under the assumption of homoscedasticity. Results are
similar for the linguistic diversity. All reported regressions contain the standard set of controls. Similar results are for the linguistic diversity. Sample consists of states with at
least one neighboring country. Weighting by population size was used to calculate mean values of the quality of government, fractionalization, and segregation in neighboring
countries. Results are robust for using simple instead of population-weighted means.
Standard controls
Observations
R-squared
F-het
F-hom
Fractionalization in neighbors
Segregation in neighbors
Fractionalization (ethnicity)
Segregation (ethnicity)
Standard controls
Observations
R-squared
F-het
F-hom
Quality of government in neighbors
-1.41***
[0.50]
0.24
[0.22]
0.65
[0.67]
-0.17
[0.29]
Yes
88
0.843
-1.35**
[0.53]
0.21
[0.22]
0.16
[0.10]
Yes
90
0.846
Segregation (ethnicity)
Fractionalization (ethnicity)
-1.18**
[0.51]
Segregation (ethnicity)
Voice
OLS
Table A.11: Controls for average quality of government and average diversity of neighbor states
A.1
Discussion of Influential Observations
The results are robust to the exclusion of any one particular country from the sample. The two most
influential observations (which affect the results in favor of our story) are Chile (which has low ethnic
segregation and very high quality of government conditional on other covariates) and Zimbabwe (which has
very high ethnic segregation and low government quality). If we exclude both Chile and Zimbabwe from the
sample, the results become weaker. Nonetheless, in the sample that excludes dictatorships, the coefficient on
ethnic segregation remains statistically significant for government effectiveness, the rule of law, and control of
corruption in IV regressions and for voice, political stability, the rule of law and control of corruption in the
OLS regressions. Moreover, Chile and Zimbabwe have a countervailing force in the second stage regressions
for ethnic diversity: Bulgaria and Russia are very influential observations, but they work against our story.
Excluding Bulgaria (which has a relatively high quality of government and an extremely high predicted
ethnic segregation) and Russia (where both predicted ethnic segregation and the quality of government are
low) strengthens the negative effect of ethnic segregation on government quality.
Linguistic segregation also has a statistically significant negative effect on voice, political stability, the rule
of law and control of corruption in the OLS regressions without Chile and Zimbabwe. But the instruments
become weak in the second stage. Yet, once we exclude the USA—the most influential observation in the
first stage—in addition to Chile and Zimbabwe, the instrument for language becomes strong enough, and
then the statistically significant results are obtained in the second stage regressions for voice and political
stability. We conclude that the effect of segregation cannot be explained by the presence of outliers.
68