protection money, tax compliance and colonial legacy in africa
TRANSCRIPT
Protection money, tax compliance and colonial legacy inAfrica
MERIMA ALI(CMI, Norway)
ODD-HELGE FJELDSTAD(CMI, Norway)
BOQIAN JIANG(Syracuse University, USA)
ABDULAZIZ SHIFA(Syracuse University, USA)
Arusha, 8 December 2014
(Foo and Bar) Arusha, 8 December 2014 1 / 26
Fiscal capacity
A broad consensus:
Fiscal capacity is important for development (Bird, 1980; Tanzi andZee, 2000).Fiscal capacity one aspect of state power (Herbest 2000; Acemoglu,2005; Besley and Persson, 2011).
Origin of state capacity?
Inter-state wars.Population density and size.Ethnic fragmentation.
Our question: Is there a systematic difference between Francophoneand Anglophone countries?
(Foo and Bar) Arusha, 8 December 2014 2 / 26
Fiscal capacity
A broad consensus:
Fiscal capacity is important for development (Bird, 1980; Tanzi andZee, 2000).Fiscal capacity one aspect of state power (Herbest 2000; Acemoglu,2005; Besley and Persson, 2011).
Origin of state capacity?
Inter-state wars.Population density and size.Ethnic fragmentation.
Our question: Is there a systematic difference between Francophoneand Anglophone countries?
(Foo and Bar) Arusha, 8 December 2014 2 / 26
Colonial legacy: Indirect vs direct rule
Britain fostered “competition between ethnic groups... practiced adivide and rule strategy....purposely maintained opposing traditionalstructures of control in order to keep the different ethnic populationswithin a colony from forming a coalition to challenge Britishhegemony.... Thus the British structured ethnic conflict into theirsystem” (Blanton et al 2001).
France pushed “for a centralized state, which incorporated individualsfrom different regions and ethnic groups into a single social system,all under the control of an administrative state modeled after andcontrolled by the French state” (Blanton et al 2001).
(Foo and Bar) Arusha, 8 December 2014 3 / 26
Outcome variable
We compare differences in three outcome variables b/n Francophoneand Anglophone countries:
1 The relative importance of ethnic versus national identity.2 Citizens attitude towards tax obligation.3 Protection against extortion (by non-state actors, e.g. gangs).
RESULT: In all of the above measures, the outcomes are consistentwith the hypothesis of “direct” and “indirect” rule.
(Foo and Bar) Arusha, 8 December 2014 4 / 26
Outcome variable
We compare differences in three outcome variables b/n Francophoneand Anglophone countries:
1 The relative importance of ethnic versus national identity.2 Citizens attitude towards tax obligation.3 Protection against extortion (by non-state actors, e.g. gangs).
RESULT: In all of the above measures, the outcomes are consistentwith the hypothesis of “direct” and “indirect” rule.
(Foo and Bar) Arusha, 8 December 2014 4 / 26
Outline
1 Introduction
2 Data
3 Descriptive Figure
4 Regression results: OLS
5 Regression results: IV
6 Conclusion
(Foo and Bar) Arusha, 8 December 2014 5 / 26
Figure: Francohpone (yellow) and Anglophone (red) countries in the sample.
(Foo and Bar) Arusha, 8 December 2014 6 / 26
Surevey question on ethnic vs national identity:
Let us suppose that you had to choose between being a Kenyan andbeing a ———–[R’s ETHNIC GROUP]. Which of the followingstatements best expresses your feelings?
1 I feel only Kenyan.2 I feel more Kenyan than ————— [insert R’s ethnic group].3 I feel equally Kenyan and ————— [insert R’s ethnic group].4 I feel only ———– [insert R’s ethnic group].
(Foo and Bar) Arusha, 8 December 2014 7 / 26
Survey question on attitudes towards tax obligation:
Please tell me whether you think that not paying the taxes peopleowe on their income is:
1 not wrong at all.2 wrong but understandable.3 or wrong and punishable.
(Foo and Bar) Arusha, 8 December 2014 8 / 26
Surevey question on extortion by non-state actors:
In the last year, how often have powerful people or groups other thangovernment, such as criminals or gangs, made people in yourcommunity or neighbourhood pay them money in return forprotecting them, their property or their businesses?
1 Never.2 Only once.3 A few times.4 Often.
(Foo and Bar) Arusha, 8 December 2014 9 / 26
Ethnic vs. national identity
BENBFACIVGINMDG
MLI
NERSENTGO
BENBFA
CIVGINMDGMLINERSENTGO
BEN
BFA
CIV
GIN
MDG
MLI
NER
SENTGO
BENBFACIVGIN
MDG
MLI
NER
SENTGO
BEN
BFACIV
GIN
MDG
MLI
NER
SENTGO
BWA
GHAKENLSO
MWI
NAMNGASLETZAUGAZMBZWE BWA
GHAKENLSOMWINAM
NGA
SLETZA
UGAZMBZWE
BWA
GHA
KENLSO
MWINAM
NGA
SLE
TZA
UGA
ZMB
ZWE
BWA
GHAKEN
LSOMWI
NAMNGA
SLE
TZAUGAZMB
ZWEBWA
GHA
KEN
LSO
MWINAM
NGA
SLE
TZA
UGA
ZMB
ZWE
0.2
.4.6
.8
Only ethnic More ethnic Equal More national Only national
Francophone Anglophone
(Foo and Bar) Arusha, 8 December 2014 10 / 26
Attitude toward tax compliance
BEN
BFA
CIV
GINMDG
MLI
NER
SEN
TGO
BENBFA
CIV
GIN
MDG
MLI
NER
SEN
TGO
BEN
BFA
CIVGIN
MDG
MLINER
SEN
TGO
BWAGHA
KEN
LSO
MWI
NAMNGASLESWZ
TZAUGA
ZMBZWE
BWAGHA
KEN
LSO
MWINAM
NGA
SLESWZTZA
UGAZMBZWE
BWA
GHA
KEN
LSO
MWI
NAM
NGA
SLE
SWZ
TZA
UGA
ZMB
ZWE
0.2
.4.6
.8
Acceptable Understandable Not Acceptable
Francophone Anglophone
(Foo and Bar) Arusha, 8 December 2014 11 / 26
Protection money: I
BENBFA
CIV
GIN
MDGMLINER
SEN
TGO BWAGHA
KEN
LSO
MWI
NAM
NGA
SLE
SWZ
TZA
UGA
ZMB
ZWE
.7.8
.91
Francophone Anglophone
People never paying protection money
(Foo and Bar) Arusha, 8 December 2014 12 / 26
Protection money: II
BENBFA
CIV
GIN
MDGMLINERSENTGO BEN
BFA
CIV
GIN
MDGMLINER
SEN
TGO
BEN
BFA
CIV
GINMDGMLINER
SENTGO
BWAGHA
KEN
LSOMWI
NAM
NGA
SLE
SWZ
TZA
UGA
ZMB
ZWE
BWAGHA
KEN
LSO
MWI
NAM
NGA
SLE
SWZ
TZA
UGA
ZMB
ZWE
BWAGHA
KEN
LSO
MWI
NAM
NGA
SLESWZ
TZAUGAZMBZWE
0.0
5.1
.15
Only one A few times Often
Francophone Anglophone
(Foo and Bar) Arusha, 8 December 2014 13 / 26
Table: Regression results: Dependent variables rank level of national identity, taxcompliance and frequency of payment to non-state actors
Protection money Tax Compliance National Identity
(1) (2) (3) (4) (5) (6)
Anglophone 0.129∗∗∗
0.147∗∗∗
-0.100∗∗∗
-0.188∗∗∗
-0.240∗∗∗
-0.278∗∗∗
(0.007) (0.009) (0.010) (0.012) (0.019) (0.021)
Former German colony -0.035∗∗∗
-0.053∗∗∗
0.401∗∗∗
(0.011) (0.017) (0.029)
Log GDP per capita -0.033∗∗∗
0.106∗∗∗
-0.005(0.005) (0.006) (0.013)
Age -0.001∗∗∗
0.002∗∗∗
0.003∗∗∗
(0.000) (0.000) (0.000)
Male 0.009 0.017∗∗∗
0.040∗∗∗
(0.006) (0.006) (0.011)
Employment 0.008 -0.037∗∗∗
-0.026(0.008) (0.009) (0.016)
Urban 0.016 0.058∗∗∗
0.028(0.011) (0.011) (0.021)
Education -0.002 0.024∗∗∗
0.021∗∗∗
(0.002) (0.002) (0.004)
Wealth 0.099∗∗∗
0.027 0.012(0.017) (0.018) (0.034)
F 299.923 37.900 91.781 87.054 161.997 45.419R-squared 0.011 0.016 0.004 0.034 0.009 0.025Observations 35738 32167 35406 32021 35343 31672
(Foo and Bar) Arusha, 8 December 2014 14 / 26
Ordered logit
Protection money Tax Compliance National Identity
(1) (2) (3) (4) (5) (6)
Anglophone 1.154∗∗∗
1.232∗∗∗
-0.268∗∗∗
-0.515∗∗∗
-0.427∗∗∗
-0.490∗∗∗
(0.068) (0.077) (0.030) (0.034) (0.032) (0.036)
Former German colony -0.063 -0.132∗∗∗
0.711∗∗∗
(0.070) (0.049) (0.053)
Log GDP per capita -0.189∗∗∗
0.292∗∗∗
-0.012(0.028) (0.018) (0.021)
Age -0.008∗∗∗
0.006∗∗∗
0.004∗∗∗
(0.001) (0.001) (0.001)
Male 0.067∗∗
0.057∗∗∗
0.064∗∗∗
(0.033) (0.019) (0.018)
Employment 0.093∗∗
-0.098∗∗∗
-0.045∗
(0.045) (0.027) (0.026)
Urban 0.057 0.166∗∗∗
0.038(0.059) (0.032) (0.035)
Education -0.019 0.065∗∗∗
0.031∗∗∗
(0.013) (0.007) (0.007)
Wealth 0.627∗∗∗
0.050 0.018(0.102) (0.053) (0.056)
Observations 35738 32167 35406 32021 35343 31672
(Foo and Bar) Arusha, 8 December 2014 15 / 26
Dummy dependent variable OLS regression
Protection money Tax Compliance National Identity
(1) (2) (3) (4) (5) (6)
Anglophone 0.082∗∗∗
0.090∗∗∗
-0.060∗∗∗
-0.116∗∗∗
-0.141∗∗∗
-0.145∗∗∗
(0.004) (0.005) (0.008) (0.009) (0.008) (0.009)
Former German colony -0.005 -0.012 0.164∗∗∗
(0.007) (0.012) (0.012)
Log GDP per capita -0.019∗∗∗
0.068∗∗∗
-0.013∗∗
(0.003) (0.005) (0.006)
Age -0.001∗∗∗
0.002∗∗∗
0.001∗∗∗
(0.000) (0.000) (0.000)
Male 0.006∗
0.017∗∗∗
0.015∗∗∗
(0.003) (0.005) (0.005)
Employment 0.010∗∗
-0.018∗∗∗
-0.014∗∗
(0.004) (0.007) (0.007)
Urban 0.005 0.038∗∗∗
0.008(0.005) (0.008) (0.009)
Education -0.002 0.013∗∗∗
0.007∗∗∗
(0.001) (0.002) (0.002)
Wealth 0.054∗∗∗
-0.005 0.015(0.009) (0.014) (0.015)
F 456.943 54.192 60.591 63.305 308.272 54.239R-squared 0.015 0.020 0.003 0.024 0.017 0.030Observations 35738 32167 35406 32021 35343 31672
(Foo and Bar) Arusha, 8 December 2014 16 / 26
Dummy dependent variable Probit regression
Protection money Tax Compliance National Identity
(1) (2) (3) (4) (5) (6)
Anglophone 0.569∗∗∗
0.611∗∗∗
-0.151∗∗∗
-0.294∗∗∗
-0.357∗∗∗
-0.367∗∗∗
(0.031) (0.036) (0.019) (0.022) (0.021) (0.023)
Former German colony -0.017 -0.029 0.418∗∗∗
(0.037) (0.032) (0.032)
Log GDP per capita -0.102∗∗∗
0.173∗∗∗
-0.034∗∗
(0.015) (0.012) (0.015)
Age -0.004∗∗∗
0.004∗∗∗
0.003∗∗∗
(0.001) (0.001) (0.001)
Male 0.036∗∗
0.043∗∗∗
0.038∗∗∗
(0.017) (0.012) (0.012)
Employment 0.052∗∗
-0.047∗∗∗
-0.035∗∗
(0.023) (0.017) (0.018)
Urban 0.030 0.098∗∗∗
0.021(0.030) (0.021) (0.024)
Education -0.008 0.033∗∗∗
0.018∗∗∗
(0.007) (0.005) (0.005)
Wealth 0.324∗∗∗
-0.012 0.037(0.051) (0.035) (0.038)
Observations 35738 32167 35406 32021 35343 31672
(Foo and Bar) Arusha, 8 December 2014 17 / 26
Country-level OLS
Protection Money Tax Compliance National Identity
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Anglophone 0.079∗∗∗
0.096∗∗∗
0.098∗∗∗
-0.056 -0.094∗∗
-0.103∗∗
-0.164∗∗
-0.106∗
-0.122∗
(0.024) (0.026) (0.020) (0.044) (0.044) (0.040) (0.063) (0.058) (0.058)
Log GDP per capita -0.024 -0.016 0.069∗∗
0.065∗∗
-0.097∗∗
-0.102∗∗
(0.017) (0.014) (0.029) (0.027) (0.041) (0.041)
Landlocked country -0.025 -0.008 -0.013 -0.015 -0.116∗
-0.127∗∗
(0.024) (0.019) (0.040) (0.039) (0.056) (0.057)
Ethnic fractionalization -0.001 -0.180∗
-0.291(0.051) (0.102) (0.181)
Pre-colonial state centralization -0.032∗∗∗
-0.040∗∗
-0.017(0.009) (0.018) (0.027)
F 10.900 4.778 7.711 1.667 2.615 3.200 6.755 5.918 4.195R-squared 0.353 0.443 0.707 0.077 0.304 0.500 0.262 0.511 0.583Observations 22 22 22 22 22 22 21 21 21
(Foo and Bar) Arusha, 8 December 2014 18 / 26
Relative Distance IV: Colonial expansion in Africa
Resource: Late Colonialism in Africa, Wilma A. Dunaway
(Foo and Bar) Arusha, 8 December 2014 19 / 26
Distance IV First Stage Regression
Relative distance Log relative distance
(1) (2) (3) (4) (5) (6) (7) (8) (9) (10)
Relative distance to British than to French -0.160∗∗∗
-0.163∗∗∗
-0.153∗∗∗
-0.157∗∗∗
-0.155∗∗∗
(0.048) (0.052) (0.049) (0.050) (0.052)
Log relative distance to British than to French -0.395∗∗∗
-0.403∗∗∗
-0.382∗∗∗
-0.392∗∗∗
-0.391∗∗∗
(0.052) (0.055) (0.055) (0.055) (0.057)
Landlocked country -0.049 -0.034 -0.065 -0.091 -0.074 -0.061 -0.102 -0.106(0.192) (0.182) (0.193) (0.203) (0.119) (0.116) (0.118) (0.126)
Log GDP per capita 0.214∗
0.200 0.179 0.117 0.096 0.092(0.118) (0.123) (0.131) (0.078) (0.079) (0.084)
Pre-colonial state centralization 0.049 0.034 0.066 0.063(0.085) (0.092) (0.053) (0.058)
Ethnic fractionalization -0.267 -0.053(0.508) (0.323)
Constant 0.786∗∗∗
0.811∗∗∗
-0.771 -0.605 -0.282 0.407∗∗∗
0.434∗∗∗
-0.417 -0.192 -0.130(0.106) (0.146) (0.880) (0.941) (1.141) (0.061) (0.076) (0.573) (0.592) (0.717)
F 10.991 5.271 5.043 3.728 2.911 56.735 27.690 20.412 16.171 12.202R-squared 0.355 0.357 0.457 0.467 0.476 0.739 0.745 0.773 0.792 0.792Observations 22 22 22 22 22 22 22 22 22 22
(Foo and Bar) Arusha, 8 December 2014 20 / 26
Relative distance IV correlation check
Relative distance Log relative distance
(1) (2) (3) (4) (5) (6)
Landlocked country -0.984 -1.191 -1.072 -0.422 -0.549 -0.420(0.778) (0.837) (0.882) (0.466) (0.498) (0.514)
Pre-colonial state centralization 0.201 0.280 0.089 0.176(0.378) (0.413) (0.225) (0.241)
Ethnic factionalization 1.181 1.285(2.204) (1.286)
Constant 1.565∗∗∗
1.880∗∗
1.106 -0.340 -0.165 -1.007(0.487) (0.658) (1.593) (0.292) (0.391) (0.929)
F 1.598 1.031 0.757 0.818 0.616 0.744R-squared 0.071 0.098 0.112 0.037 0.061 0.110Observations 23 22 22 23 22 22
(Foo and Bar) Arusha, 8 December 2014 21 / 26
Country-level 2SLS
Protection Money Tax Compliance National Identity
(1) (2) (3) (4) (5) (6) (7) (8) (9)
Anglophone 0.086∗∗∗
0.104∗∗∗
0.091∗∗∗
-0.065 -0.090∗
-0.115∗∗∗
-0.168∗∗
-0.101∗
-0.127∗∗
(0.027) (0.027) (0.020) (0.048) (0.046) (0.039) (0.070) (0.061) (0.057)
Log GDP per capita -0.026∗
-0.015 0.068∗∗
0.068∗∗∗
-0.098∗∗∗
-0.101∗∗∗
(0.016) (0.012) (0.027) (0.024) (0.038) (0.035)
Landlocked country -0.026 -0.007 -0.014 -0.014 -0.116∗∗
-0.127∗∗∗
(0.022) (0.017) (0.037) (0.033) (0.051) (0.048)
Ethnic fractionalization -0.003 -0.185∗∗
-0.294∗
(0.044) (0.087) (0.154)
Pre-colonial state centralization -0.032∗∗∗
-0.040∗∗∗
-0.017(0.008) (0.015) (0.022)
Shea Partial R2 0.739 0.729 0.746 0.739 0.729 0.746 0.743 0.736 0.751R-squared 0.350 0.440 0.705 0.075 0.303 0.497 0.262 0.511 0.583Observations 22 22 22 22 22 22 21 21 21
(Foo and Bar) Arusha, 8 December 2014 22 / 26
IV Predicted value: Protection Money
SEN
MDG
GINMLI
CIV
BEN
TGO
BFANER
GHA
MWI
SWZ
TZAZWE
NAMZMBBWA
UGA
KEN
NGA
SLE
LSO
0.1
.2.3
−.5 0 .5 1 1.5Fitted values
(Foo and Bar) Arusha, 8 December 2014 23 / 26
IV Predicted value: Tax Compliance
SEN
MDG
GIN
MLI
CIV
BEN
TGOBFA
NERGHA
MWI
SWZ
TZA
ZWE
NAM
ZMB
BWA
UGA
KEN
NGA
SLE
LSO
.3.4
.5.6
.7
−.5 0 .5 1 1.5Fitted values
(Foo and Bar) Arusha, 8 December 2014 24 / 26
IV Predicted value: National Identity
SEN
MDGGIN
MLICIV
BENTGO
BFA
NER
GHAMWI
TZA
ZWE
NAM
ZMB
BWAUGA
KEN
NGA
SLE
LSO
0.2
.4.6
.8
−.5 0 .5 1 1.5Fitted values
(Foo and Bar) Arusha, 8 December 2014 25 / 26
Conclusion
Both citizens’ attitude toward tax obligation and national identityappear to be weaker in Anglophone countries.
IV results reaffirm this pattern.
Caution: the exact mechanism is not identified.
(Foo and Bar) Arusha, 8 December 2014 26 / 26