Elite Capture of Foreign Aid: Evidence from Offshore Bank ...
Transcript of Elite Capture of Foreign Aid: Evidence from Offshore Bank ...
![Page 1: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/1.jpg)
Elite Capture of Foreign Aid:Evidence from Offshore Bank Accounts
Jørgen Juel Andersen1 Niels Johannesen2 Bob Rijkers3
1BI Norwegian Business School, Oslo
2University of Copenhagen
3World Bank
February 28, 2020
![Page 2: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/2.jpg)
Introduction
I Classic concern: elite capture of aid
I Consistent withI theories of rent seeking (e.g. Svensson, 2000)I corruption in aid-receiving countries (Alesina and Weder,
2002)I Difficult to measure
I Do World Bank aid disbursements trigger increases inforeign deposits in bank accounts in havens?
![Page 3: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/3.jpg)
Introduction
I Classic concern: elite capture of aidI Consistent with
I theories of rent seeking (e.g. Svensson, 2000)
I corruption in aid-receiving countries (Alesina and Weder,2002)
I Difficult to measure
I Do World Bank aid disbursements trigger increases inforeign deposits in bank accounts in havens?
![Page 4: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/4.jpg)
Introduction
I Classic concern: elite capture of aidI Consistent with
I theories of rent seeking (e.g. Svensson, 2000)I corruption in aid-receiving countries (Alesina and Weder,
2002)
I Difficult to measure
I Do World Bank aid disbursements trigger increases inforeign deposits in bank accounts in havens?
![Page 5: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/5.jpg)
Introduction
I Classic concern: elite capture of aidI Consistent with
I theories of rent seeking (e.g. Svensson, 2000)I corruption in aid-receiving countries (Alesina and Weder,
2002)I Difficult to measure
I Do World Bank aid disbursements trigger increases inforeign deposits in bank accounts in havens?
![Page 6: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/6.jpg)
Introduction
I Classic concern: elite capture of aidI Consistent with
I theories of rent seeking (e.g. Svensson, 2000)I corruption in aid-receiving countries (Alesina and Weder,
2002)I Difficult to measure
I Do World Bank aid disbursements trigger increases inforeign deposits in bank accounts in havens?
![Page 7: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/7.jpg)
Introduction (continued)
I Havens: offshore banking centers known for secrecyand asset protection havens
I Main sample: 22 highly aid dependent countries,1990-2010 highly aid dependent countries
I annual WBG aid>2% GDPI absorb 10% of all WB lendingI Extensions
I period since 2009I less aid dependent countries
![Page 8: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/8.jpg)
Introduction (continued)
I Havens: offshore banking centers known for secrecyand asset protection havens
I Main sample: 22 highly aid dependent countries,1990-2010 highly aid dependent countries
I annual WBG aid>2% GDPI absorb 10% of all WB lendingI Extensions
I period since 2009I less aid dependent countries
![Page 9: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/9.jpg)
Premise: Haven deposits belong to elites
I Offshore wealth is extremely concentratedI Scandinavia: Top 0.01% (0.1%) richest households own
50% (80%) of wealth on offshore accounts (Alstadsaeter etal., 2017)
I By contrast, the poor are mostly unbankedI 9% of the bottom 40% households had a (domestic) bank
account in 2011 (Demirguc-Kunt et al, 2011)
![Page 10: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/10.jpg)
Premise: Haven deposits belong to elites
I Offshore wealth is extremely concentratedI Scandinavia: Top 0.01% (0.1%) richest households own
50% (80%) of wealth on offshore accounts (Alstadsaeter etal., 2017)
I By contrast, the poor are mostly unbankedI 9% of the bottom 40% households had a (domestic) bank
account in 2011 (Demirguc-Kunt et al, 2011)
![Page 11: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/11.jpg)
Main findings
I WB aid disbursements to highly aid dependentcountries coincide with significant increases in havendeposits within the same quarter
I Aid-induced increase in haven deposits ("Leakage rate")I ≈ 7.5% aid inflows for the average countryI ≈ 5% of aggregate flows to these countriesI tends to increase with ratio of aid/GDP
![Page 12: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/12.jpg)
Main findings
I WB aid disbursements to highly aid dependentcountries coincide with significant increases in havendeposits within the same quarterI Aid-induced increase in haven deposits ("Leakage rate")
I ≈ 7.5% aid inflows for the average countryI ≈ 5% of aggregate flows to these countriesI tends to increase with ratio of aid/GDP
![Page 13: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/13.jpg)
Additional findings
I Disbursements do not trigger flows to non-havens
I Improved financial transparency since 2009 has nothelped
![Page 14: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/14.jpg)
Additional findings
I Disbursements do not trigger flows to non-havens
I Improved financial transparency since 2009 has nothelped
![Page 15: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/15.jpg)
Interpretation
I Isolating the underlying mechanism is challenging
I we don’t know who owns the deposits=> cannot attribute agency
I Elite capture is a salient and plausible explanation
I Other explanations are possible
We would love to hear your thoughts!
![Page 16: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/16.jpg)
Interpretation
I Isolating the underlying mechanism is challengingI we don’t know who owns the deposits
=> cannot attribute agency
I Elite capture is a salient and plausible explanation
I Other explanations are possible
We would love to hear your thoughts!
![Page 17: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/17.jpg)
Interpretation
I Isolating the underlying mechanism is challengingI we don’t know who owns the deposits
=> cannot attribute agency
I Elite capture is a salient and plausible explanation
I Other explanations are possible
We would love to hear your thoughts!
![Page 18: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/18.jpg)
Interpretation
I Isolating the underlying mechanism is challengingI we don’t know who owns the deposits
=> cannot attribute agency
I Elite capture is a salient and plausible explanation
I Other explanations are possible
We would love to hear your thoughts!
![Page 19: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/19.jpg)
Key Caveats & Common (Media) misperceptions (1/2)
I Correlation does not imply causationI money is fungible
![Page 20: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/20.jpg)
Key Caveats & Common (Media) misperceptions (2/2)
I our main estimates only apply to 22 highly aiddependent countries
I these account for 10% of all WB lendingI aggregate leakage in these countries = 5%
![Page 21: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/21.jpg)
Outline
Introduction
Data
Empirical Strategy
ResultsMain ResultsRobustnessHeterogeneityCalculating Leakage
Interpretation
Conclusions
Supplemental Slides
![Page 22: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/22.jpg)
Cross border bank deposits
I Quarterly data on bilateral non-bank deposits, i.e.privately held accountsI only observe direct ownership
I Source: Locational Banking Statistics of the Bank forInternational Settlements (BIS)I covers value of bank deposits in 43 financial centers owned
by residents of around 200 countriesI confidential data: 1990-2010I public data with partial coverage: 1990-2018
I Havenit : sum of all deposits owned by residents ofcountry i in the 17 havens in quarter t
![Page 23: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/23.jpg)
Foreign aid
I The World Bank Project Database (WBPD)I Approval dateI Commitment amountI SectorI Lending instrument type (e.g., investment projects or
development policy loans)I Timing of each disbursement
I Aidit : Sum of all IDA and IBRD project disbursementsin country i at time t (normalized by GDP)
![Page 24: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/24.jpg)
Other variables
I Wars - PRIOI Coups - Powell and Thyne (2011)I Financial Crises - Laeven and Valencia (2012)I Natural Disasters - International Disaster DatabaseI Petroleum Rents - WDII Financial Sector Development - WDII Capital Account Openness - Chinn and Ito (2006)I Disclosure Requirements - Djankov et al. (2010)I Democracy - Polity IVI Corruption - WGII Procurement - World Bank Procurement Database
![Page 25: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/25.jpg)
Aid and haven deposits growth distributions
050
100
150
200
250
Freq
uenc
y
0 .01 .02 .03 .04World Bank aid (relative to GDP)
Figure: Aid (% of GDP)0
100
200
300
400
Freq
uenc
y
-1 -.5 0 .5 1 1.5Change in haven deposits (in log-levels)
Figure: Haven deposits Growth
![Page 26: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/26.jpg)
Descriptive Statistics
Variable Mean Std. Dev. Min MaxAidWB disbursements (% GDP) 0.027 0.006 0.020 0.039Total foreign aid (%GDP) 0.101 0.078 0.001 0.647Foreign depositsHaven deposits (million USD) 61 68 0 621Non-haven deposits (million USD) 136 143 0 964Quarterly growth ratesHaven deposits growth 0.020 0.272 -0.875 1.240Non-haven deposits growth 0.015 0.209 -0.693 0.795GDP growth 0.017 0.040 -0.240 0.178
crowding in
![Page 27: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/27.jpg)
Empirical Strategy
∆ log(Havenit ) = βAidit + γXit + µi + τt + εit
I ∆ log(Havenit ) growth rate in haven deposits owned bycountry i in quarter t
I Aidit quarterly WB aid disbursements/GDPI Xit : ∆ log GDPit (in robustness: foreign exchange rate; oil
wealth shocks)I µi : country fixed effectsI τt : time fixed effectsI Standard errors clustered at the country levelI Deposit and aid variables winsorized at the 1%/99% level.I Alternative specifications/LHS-variables:
∆ log (nonhavenit ) and ∆ log (havenit )−∆ log (nonhavenit )
![Page 28: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/28.jpg)
Empirical Strategy
∆ log(Havenit ) = βAidit + γXit + µi + τt + εit
I ∆ log(Havenit ) growth rate in haven deposits owned bycountry i in quarter t
I Aidit quarterly WB aid disbursements/GDP
I Xit : ∆ log GDPit (in robustness: foreign exchange rate; oilwealth shocks)
I µi : country fixed effectsI τt : time fixed effectsI Standard errors clustered at the country levelI Deposit and aid variables winsorized at the 1%/99% level.I Alternative specifications/LHS-variables:
∆ log (nonhavenit ) and ∆ log (havenit )−∆ log (nonhavenit )
![Page 29: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/29.jpg)
Empirical Strategy
∆ log(Havenit ) = βAidit + γXit + µi + τt + εit
I ∆ log(Havenit ) growth rate in haven deposits owned bycountry i in quarter t
I Aidit quarterly WB aid disbursements/GDPI Xit : ∆ log GDPit (in robustness: foreign exchange rate; oil
wealth shocks)
I µi : country fixed effectsI τt : time fixed effectsI Standard errors clustered at the country levelI Deposit and aid variables winsorized at the 1%/99% level.I Alternative specifications/LHS-variables:
∆ log (nonhavenit ) and ∆ log (havenit )−∆ log (nonhavenit )
![Page 30: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/30.jpg)
Empirical Strategy
∆ log(Havenit ) = βAidit + γXit + µi + τt + εit
I ∆ log(Havenit ) growth rate in haven deposits owned bycountry i in quarter t
I Aidit quarterly WB aid disbursements/GDPI Xit : ∆ log GDPit (in robustness: foreign exchange rate; oil
wealth shocks)I µi : country fixed effectsI τt : time fixed effects
I Standard errors clustered at the country levelI Deposit and aid variables winsorized at the 1%/99% level.I Alternative specifications/LHS-variables:
∆ log (nonhavenit ) and ∆ log (havenit )−∆ log (nonhavenit )
![Page 31: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/31.jpg)
Empirical Strategy
∆ log(Havenit ) = βAidit + γXit + µi + τt + εit
I ∆ log(Havenit ) growth rate in haven deposits owned bycountry i in quarter t
I Aidit quarterly WB aid disbursements/GDPI Xit : ∆ log GDPit (in robustness: foreign exchange rate; oil
wealth shocks)I µi : country fixed effectsI τt : time fixed effectsI Standard errors clustered at the country level
I Deposit and aid variables winsorized at the 1%/99% level.I Alternative specifications/LHS-variables:
∆ log (nonhavenit ) and ∆ log (havenit )−∆ log (nonhavenit )
![Page 32: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/32.jpg)
Empirical Strategy
∆ log(Havenit ) = βAidit + γXit + µi + τt + εit
I ∆ log(Havenit ) growth rate in haven deposits owned bycountry i in quarter t
I Aidit quarterly WB aid disbursements/GDPI Xit : ∆ log GDPit (in robustness: foreign exchange rate; oil
wealth shocks)I µi : country fixed effectsI τt : time fixed effectsI Standard errors clustered at the country levelI Deposit and aid variables winsorized at the 1%/99% level.
I Alternative specifications/LHS-variables:
∆ log (nonhavenit ) and ∆ log (havenit )−∆ log (nonhavenit )
![Page 33: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/33.jpg)
Empirical Strategy
∆ log(Havenit ) = βAidit + γXit + µi + τt + εit
I ∆ log(Havenit ) growth rate in haven deposits owned bycountry i in quarter t
I Aidit quarterly WB aid disbursements/GDPI Xit : ∆ log GDPit (in robustness: foreign exchange rate; oil
wealth shocks)I µi : country fixed effectsI τt : time fixed effectsI Standard errors clustered at the country levelI Deposit and aid variables winsorized at the 1%/99% level.I Alternative specifications/LHS-variables:
∆ log (nonhavenit ) and ∆ log (havenit )−∆ log (nonhavenit )
![Page 34: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/34.jpg)
Endogeneity
I Problem: Aid disbursements are endogenousI Natural disasters, financial crises and conflict can trigger
simultaneous surge in both aid and offshore deposits
I SolutionsI Use lags and leadsI Use an IV: Predicted disbursementsI Exclude periods of turmoil: war, coups, financial crises
![Page 35: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/35.jpg)
Endogeneity
I Problem: Aid disbursements are endogenousI Natural disasters, financial crises and conflict can trigger
simultaneous surge in both aid and offshore deposits
I SolutionsI Use lags and leadsI Use an IV: Predicted disbursementsI Exclude periods of turmoil: war, coups, financial crises
![Page 36: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/36.jpg)
Instrumenting Aid (Kraay, 2012, 2014)
-.01
0.0
1.0
2.0
3ai
d (re
sidu
aliz
ed)
-.01 0 .01 .02instrument (residualized)
coef = .87, se = .08
-.01
0.0
1.0
2.0
3ai
d (re
sidu
aliz
ed)
-.01 -.005 0 .005 .01instrument (residualized)
coef = .77, se = .09
I Predict quarterly disbursements for each project usingI initial commitments
I average disbursement schedule in the same sector andregion disbursementprofiles
I aggregate across all projects (excluding initial quarters)
![Page 37: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/37.jpg)
Instrumenting Aid (Kraay, 2012, 2014)
-.01
0.0
1.0
2.0
3ai
d (re
sidu
aliz
ed)
-.01 0 .01 .02instrument (residualized)
coef = .87, se = .08
-.01
0.0
1.0
2.0
3ai
d (re
sidu
aliz
ed)
-.01 -.005 0 .005 .01instrument (residualized)
coef = .77, se = .09
I Predict quarterly disbursements for each project usingI initial commitmentsI average disbursement schedule in the same sector and
region disbursementprofiles
I aggregate across all projects (excluding initial quarters)
![Page 38: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/38.jpg)
Instrumenting Aid (Kraay, 2012, 2014)
-.01
0.0
1.0
2.0
3ai
d (re
sidu
aliz
ed)
-.01 0 .01 .02instrument (residualized)
coef = .87, se = .08
-.01
0.0
1.0
2.0
3ai
d (re
sidu
aliz
ed)
-.01 -.005 0 .005 .01instrument (residualized)
coef = .77, se = .09
I Predict quarterly disbursements for each project usingI initial commitmentsI average disbursement schedule in the same sector and
region disbursementprofiles
I aggregate across all projects (excluding initial quarters)
![Page 39: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/39.jpg)
Baseline Results
(1) (2) (3)haven non-haven ratio
(% change) (% change) (% change)
Aid (%GDP) 3.391*** -1.511 4.973***(1.154) (1.035) (1.715)
GDP (%growth) 0.122 0.147** -0.001(0.141) (0.063) (0.150)
country FE Yes Yes Yestime FE Yes Yes YesObservations 1,648 1,652 1,645R-squared 0.101 0.092 0.076Standard errors are clustered by country
![Page 40: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/40.jpg)
Haven deposits increase when aid flows in
-50
5Pe
rcen
tage
cha
nge
in h
aven
dep
osits
-4 -2 0 2 4Quarter relative to disbursement
![Page 41: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/41.jpg)
Non-haven deposits do not respond to disbursements
-50
5Pe
rcen
tage
cha
nge
in n
on-h
aven
dep
osits
-4 -2 0 2 4Quarter relative to disbursement
![Page 42: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/42.jpg)
Aid differentially impacts haven vs non-haven deposits
-50
5Pe
rcen
tage
cha
nge
in h
aven
to n
on-h
aven
ratio
-4 -2 0 2 4Quarter relative to disbursement
![Page 43: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/43.jpg)
Instrumenting Aid
Dependent variable: % Change in haven deposits(1) (2) (3)
OLS IV IVBaseline exclude 1 quart exclude 3 quart
Aid (%GDP) 3.391*** 2.595*** 2.848*(1.154) (1.004) (1.564)
country FE Yes Yes Yestime FE Yes Yes YesKleibergen-Paap rk Wald-F 182.624 159.650Observations 1,648 1,648 1,648R-squared 0.101 0.005 0.005Standard errors are clustered by country
![Page 44: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/44.jpg)
Robustness
Dependent variable: % Change in haven depositsOmit periods with other shocks
No Wars No Coups No Disasters No Fin Crisis(1) (2) (3) (4)
Aid (%GDP) 2.754** 3.446*** 3.601*** 3.559***(1.296) (1.160) (1.237) (1.190)
Observations 1,386 1,618 1,586 1,511R-squared 0.107 0.100 0.104 0.102
Alternative specificationscountry-year FE Exchange controls Oil*time FE No winsorization
(5) (6) (7) (8)
Aid (%GDP) 3.290* 3.310*** 3.535** 3.591**(1.895) (1.126) (1.296) (1.500)
Observations 1,648 1,648 1,648 1,648R-squared 0.242 0.103 0.142 0.100
Standard errors are clustered by country. Country and time FE included
![Page 45: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/45.jpg)
Heterogeneity C
orru
ptio
n C
redi
t D
emoc
racy
Dis
clos
ure
Pro
cure
men
t
low corruption controlhigh corruption control
disclosure requiredno disclosure required
less democraticmore democratic
low credithigh credit
high local procurementlow local procurement
-5 0 5 10
Estimated impact of aid on haven deposits
![Page 46: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/46.jpg)
Does Type of Aid Matter?
(1) (2) (3)haven non-haven ratio
(% change) (% change) (% change)
IPF Aid (% of GDP) 5.122*** -0.435 5.787**(1.708) (1.563) (2.256)
DPF Aid (% of GDP) 2.697* -1.973 4.704**(1.348) (1.323) (2.060)
Observations 1,648 1,652 1,645R-squared 0.100 0.092 0.076
All specifications control for GDP growth and country and time FE.Standard errors are clustered by country. Country and time FE are included.
![Page 47: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/47.jpg)
Transparency (post 2008) has not helped
-20
24
68
β (Ai
d/G
DP)
1990-2009 2010-2019
Source: Andersen, Johannesen, and Rijkers (2019)
Impact of aid on growth in offshore deposits (public data)
![Page 48: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/48.jpg)
Quantifying the impact of aid on offshore depositsLeakage rate ≡ Aid-induced increase in offshore deposits
Calculating average leakageI βAid = 3.4
I => 1% Aid(%GDP) ↑ => 3.4% haven deposits ↑
I mean of haven/GDP: 0.022 (in the high aid sample)I haven/GDP increases by (3.4× 0.022 =) 0.075 when
Aid/GDP increases by 1=> leakage rate is 7.5%
Calculating aggregate leakageI Assume βAid = 3.4 applies uniformlyI mean of haven/GDP weighted by the share of average aid
received: 0.014I haven/GDP increases by (3.4× 0.014 =) 0.05 when
Aid/GDP increases by 1.=> leakage rate is 5%
![Page 49: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/49.jpg)
Quantifying the impact of aid on offshore depositsLeakage rate ≡ Aid-induced increase in offshore deposits
Calculating average leakageI βAid = 3.4
I => 1% Aid(%GDP) ↑ => 3.4% haven deposits ↑I mean of haven/GDP: 0.022 (in the high aid sample)
I haven/GDP increases by (3.4× 0.022 =) 0.075 whenAid/GDP increases by 1=> leakage rate is 7.5%
Calculating aggregate leakageI Assume βAid = 3.4 applies uniformlyI mean of haven/GDP weighted by the share of average aid
received: 0.014I haven/GDP increases by (3.4× 0.014 =) 0.05 when
Aid/GDP increases by 1.=> leakage rate is 5%
![Page 50: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/50.jpg)
Quantifying the impact of aid on offshore depositsLeakage rate ≡ Aid-induced increase in offshore deposits
Calculating average leakageI βAid = 3.4
I => 1% Aid(%GDP) ↑ => 3.4% haven deposits ↑I mean of haven/GDP: 0.022 (in the high aid sample)I haven/GDP increases by (3.4× 0.022 =) 0.075 when
Aid/GDP increases by 1=> leakage rate is 7.5%
Calculating aggregate leakageI Assume βAid = 3.4 applies uniformlyI mean of haven/GDP weighted by the share of average aid
received: 0.014I haven/GDP increases by (3.4× 0.014 =) 0.05 when
Aid/GDP increases by 1.=> leakage rate is 5%
![Page 51: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/51.jpg)
Quantifying the impact of aid on offshore depositsLeakage rate ≡ Aid-induced increase in offshore deposits
Calculating average leakageI βAid = 3.4
I => 1% Aid(%GDP) ↑ => 3.4% haven deposits ↑I mean of haven/GDP: 0.022 (in the high aid sample)I haven/GDP increases by (3.4× 0.022 =) 0.075 when
Aid/GDP increases by 1=> leakage rate is 7.5%
Calculating aggregate leakageI Assume βAid = 3.4 applies uniformly
I mean of haven/GDP weighted by the share of average aidreceived: 0.014
I haven/GDP increases by (3.4× 0.014 =) 0.05 whenAid/GDP increases by 1.=> leakage rate is 5%
![Page 52: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/52.jpg)
Quantifying the impact of aid on offshore depositsLeakage rate ≡ Aid-induced increase in offshore deposits
Calculating average leakageI βAid = 3.4
I => 1% Aid(%GDP) ↑ => 3.4% haven deposits ↑I mean of haven/GDP: 0.022 (in the high aid sample)I haven/GDP increases by (3.4× 0.022 =) 0.075 when
Aid/GDP increases by 1=> leakage rate is 7.5%
Calculating aggregate leakageI Assume βAid = 3.4 applies uniformlyI mean of haven/GDP weighted by the share of average aid
received: 0.014
I haven/GDP increases by (3.4× 0.014 =) 0.05 whenAid/GDP increases by 1.=> leakage rate is 5%
![Page 53: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/53.jpg)
Quantifying the impact of aid on offshore depositsLeakage rate ≡ Aid-induced increase in offshore deposits
Calculating average leakageI βAid = 3.4
I => 1% Aid(%GDP) ↑ => 3.4% haven deposits ↑I mean of haven/GDP: 0.022 (in the high aid sample)I haven/GDP increases by (3.4× 0.022 =) 0.075 when
Aid/GDP increases by 1=> leakage rate is 7.5%
Calculating aggregate leakageI Assume βAid = 3.4 applies uniformlyI mean of haven/GDP weighted by the share of average aid
received: 0.014I haven/GDP increases by (3.4× 0.014 =) 0.05 when
Aid/GDP increases by 1.=> leakage rate is 5%
![Page 54: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/54.jpg)
Leakage increases with aid dependence
0.0
5.1
.15
Impl
ied
leak
age
rate
02
46
810
Poin
t est
imat
e an
d co
nfid
ence
bou
nds
.01 .015 .02 .025 .03Sample threshold (Mean WB aid / GDP)
Model point estimate (left axis) Implied leakage rate (right axis)
I Aggregate leakage in aid-dependent countries: 5%.I Aid might be most susceptible to capture in countries that
need it most
![Page 55: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/55.jpg)
Which mechanisms are at play? (1/3)
Highly unlikely explanations
I Portfolio adjustments by commercial and centralbanksI data only cover private deposits
I Profit shifting by multinational firmsI data do not cover indirect ownership
I Economic stimulusI GDP growth is already controlled forI output response to stimulus is typically protractedI not clear why money does not flow to non-havens
![Page 56: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/56.jpg)
Which mechanisms are at play? (1/3)
Highly unlikely explanations
I Portfolio adjustments by commercial and centralbanksI data only cover private deposits
I Profit shifting by multinational firmsI data do not cover indirect ownership
I Economic stimulusI GDP growth is already controlled forI output response to stimulus is typically protractedI not clear why money does not flow to non-havens
![Page 57: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/57.jpg)
Which mechanisms are at play? (1/3)
Highly unlikely explanations
I Portfolio adjustments by commercial and centralbanksI data only cover private deposits
I Profit shifting by multinational firmsI data do not cover indirect ownership
I Economic stimulusI GDP growth is already controlled forI output response to stimulus is typically protractedI not clear why money does not flow to non-havens
![Page 58: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/58.jpg)
Which mechanisms are at play? (2/3)
Elite capture
I Consistent withI money flows to havens, but not to non-havensI concurrence aid inflows and capital outflows within the
same quarterI outflows ↑ when corruption ↑ and disclosure requirements ↑
![Page 59: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/59.jpg)
Which mechanisms are at play? (2/3)
Elite captureI Consistent with
I money flows to havens, but not to non-havens
I concurrence aid inflows and capital outflows within thesame quarter
I outflows ↑ when corruption ↑ and disclosure requirements ↑
![Page 60: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/60.jpg)
Which mechanisms are at play? (2/3)
Elite captureI Consistent with
I money flows to havens, but not to non-havensI concurrence aid inflows and capital outflows within the
same quarter
I outflows ↑ when corruption ↑ and disclosure requirements ↑
![Page 61: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/61.jpg)
Which mechanisms are at play? (2/3)
Elite captureI Consistent with
I money flows to havens, but not to non-havensI concurrence aid inflows and capital outflows within the
same quarterI outflows ↑ when corruption ↑ and disclosure requirements ↑
![Page 62: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/62.jpg)
Which mechanisms are at play? (3/3)
Domestic aid-sponsored contractors depositing money inhavens
I consistent withI larger effects when the financial sector is small
I inconsistent withI larger effects when a greater share of procurement goes to
foreign firmsI money flowing to havens, but not to non-havens
![Page 63: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/63.jpg)
Which mechanisms are at play? (3/3)
Domestic aid-sponsored contractors depositing money inhavensI consistent with
I larger effects when the financial sector is small
I inconsistent withI larger effects when a greater share of procurement goes to
foreign firmsI money flowing to havens, but not to non-havens
![Page 64: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/64.jpg)
Which mechanisms are at play? (3/3)
Domestic aid-sponsored contractors depositing money inhavensI consistent with
I larger effects when the financial sector is smallI inconsistent with
I larger effects when a greater share of procurement goes toforeign firms
I money flowing to havens, but not to non-havens
![Page 65: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/65.jpg)
Conclusion
I WB aid disbursements coincide with increases inoffshore deposits held in havensI but only in highly aid dependent countries
I Consistent with elite capture
Any feedback you may have would be much appreciated!
![Page 66: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/66.jpg)
Crowding in of aid
[label=crowding in](1) (2)
Sample WB aid >2% WB aid >1%Annual non WB aid (% of GDP) Annual non WB aid (% of GDP)
Annual WB aid (% of GDP) -0.054 -0.128(0.385) (0.261)
Annual GDP (% growth) -0.048* -0.024(0.023) (0.019)
Observations 425 910R-squared 0.608 0.541
Standard errors are clustered by country. Country and time FE included.
![Page 67: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/67.jpg)
Results for Individual Havens
Dependent variable: % Change in haven depositsSwitzerland Luxembourg Belgium Jersey etc
Aid (%GDP) 2.872* 2.717** -1.035 -2.068(1.563) (1.197) (0.726) (2.026)
Observations 2,334 1,796 2,075 1,304R-squared 0.081 0.087 0.083 0.081Standard errors are clustered by country. Country and time FE are included.
![Page 68: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/68.jpg)
Disbursement Profiles
0.0
5.1
.15
.2Pr
edic
ted
Dis
burs
emen
t Sha
re
0 10 20 30 40 50Time (quarters) - (approval quarter=0)
average by sector-region overall average
Disbursement Profiles
IV
![Page 69: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/69.jpg)
Havens
I Based on Andersen et al. (2017)I financial centers identified by the OECD in 2008 as not
providing bank information to foreign governments plusMacao, SAR, China, and Hong Kong, SAR, China
I Bahamas, Bermuda, Cayman Islands, NetherlandsAntilles, Panama, Bahrain, Hong Kong, Macao, Singapore,Austria, Belgium, Guernsey, Isle of Man, Jersey,Liechtenstein, Luxembourg, and Switzerland
![Page 70: Elite Capture of Foreign Aid: Evidence from Offshore Bank ...](https://reader031.fdocuments.net/reader031/viewer/2022012519/61945084586d3f1c3c4524c9/html5/thumbnails/70.jpg)
Highly aid dependent countries
I Average annual WB aid exceeds 2% of GDP
I Afghanistan, Armenia, Burkina Faso, Burundi, Eritrea,Ethiopia, Ghana, Guinea-Bissau, Guyana, KyrgyzRepublic, Madagascar, Malawi, Mali, Mauritania,Mozambique, Niger, Rwanda, Sao Tome and Principe,Sierra Leone, Tanzania, Uganda, Zambia