detecting corruption and state capture - world...

Detecting Corruption and State Capture

Bob Rijkers

Development Research GroupWorld Bank

October 28, 2019Policy Research Talk

Introduction

I Corruption is a development challenge

I Difficult to detectI especially at the highest levels of government

I Rigorous evidence is a prerequisite for interventionI false accusations are harmfulI identifying mechanisms =⇒ reform priorities

This talk: review methods to detect corruption

I How do dictators get rich?I does regime change curb capture?

I How do bureaucrats collude?I can technology help?

I Do elites capture aid?I can transparency initiatives help?

Evidence from our operational engagement

I State capture in Tunisia and IndonesiaI (ab)use of entry regulation & tax evasion

I Collusion in Madagascar customsI manipulation of allocation of declarationsI differential treatment by inspectors

I Elite capture of World Bank lendingI disbursements trigger increases in deposits in tax havens

Main messages

I Grand corruption is prevalent and costlyI undermines World Bank effectiveness

I Systematic corruption is difficult to dislodgeI regime change alone does not curb captureI IT can help, but is not a panaceaI transparency alone is not sufficient

I Combating corruption requires changing incentives

Mechanisms of State Capture: Evidence from Tunisia

What do we know about political connections?

I Politically connections are prevalent and valuable...I 2% of publicly listed firmsI 8% of market capitalization

I .... especially in countries with high levels ofcorruption, FDI restrictions, and lax conflict of interestlaws (Faccio, 2006)

Why are connections valuable?

FirmsI Privileged access to inputs (Johnson and Mitton, 2003)I Erection of regulatory barriers (Eibl and Malik, 2016)I Selective enforcement of regulation (Fisman and Wang, 2016)I Privileged access to import licenses (Mobarak and Purbasari,

2006)I Government contracts (Cingano and Pinotti, 2013)

PoliticiansI job creation - firms overhire during election years (Bertrand et

al., 2018)I campaign contributions (Claessens and Feijen, 2008)

Why Tunisia?

I Celebrated development model:I (Pre-) Arab Spring “Success” Story

I Data availabilityI 652 firms confiscated after the Arab SpringI tax, social security, and customs transaction data

Connected firms are especially profitable in regulatedsectors

0.2% 0.9% 0.2%

1.7%

10.5%

1.0%

5.3%

42.6%

1.2%

15.8%

54.5%

3.3%

ENTIRE ECONOMY HIGHLY REGULATED SECTORS OTHER SECTORS

Contributions of Ben Ali Firms

Firms Jobs Output ProfitsSource: Rijkers, Freund and Nucifora (2017)

There is a premium on being connected

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

p10 p50 p90 p95 p99

The Premium on Being Connected (Market Share)

Ben Ali premium Additional premium in highly regulated sectors

Source: Rijkers, Freund and Nucifora (2017)

The premium on being connected is

I larger for big firms

I higher in regulated sectors

Impacts on competition in highly regulated sectors

I Ben Ali MS ↑ =⇒ exit ↓ and concentration ↑

Connected firms are more likely to evade taxes

9.0%

15.5%

10.9%

21.1%

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

Missing tax declaration Inconsistent reporting

Prevalence of tax evasion

Other firms Ben Ali firmSource: Rijkers, Arouri and Baghdadi (2017)

Detecting tariff evasion

I Compare Tunisia’s import records with sourcecountries export’ records (Fisman and Wei, 2004)I Calculate Evasion Gaps for product p at time t

Gappt = Exportspt(partners)− Importspt(Tunisia)

I Large discrepancies are associated with evasion

I which firms evade?

Connected firms are more likely to evade tariifs

-0.10

-0.40

-0.20

0.00

0.20

0.40

0.60

0.80

1.00

All Offshore Onshore Public Ben Ali

Log

Evas

ion

Gap

Mean Evasion Gaps by Dominant ImporterEvasion Gap=Log Exports reported by partner-Log Imports reported in Tunisia

Low Tariff High TariffSource: Rijkers, Baghdadi and Raballand (2017)

Connected firms are more likely to evade tariifs

-0.10

-0.40

-0.20

0.00

0.20

0.40

0.60

0.80

1.00

All Offshore Onshore Public Ben Ali

Log

Evas

ion

Gap

Mean Evasion Gaps by Dominant ImporterEvasion Gap=Log Exports reported by partner-Log Imports reported in Tunisia

Low Tariff High Tariff Source: Rijkers, Baghdadi and Raballand (2017)

Ben Ali firms evaded 1.2 billion USD in tariffs between 2000 and 2009

What happened after the Revolution?

-16.30%

6.20%

-9.40%

5.80%

-20.00%

-15.00%

-10.00%

-5.00%

0.00%

5.00%

10.00%

Previously Ben Ali dominated Other products

Post Revolution Change in Evasion Gaps

High tariffs Low tariffsSource: Arouri, Baghdadi, and Rijkers (2019)

I Limited economic reformI limited reform of the investment code

I Corruption escalated and democratizedI tariff evasion democratized and escalated

Regime change does not curb capture: evidence fromIndonesia

Suharto connected firms lost market share

-.03

-.02

-.01

0.0

1.0

2β P

Cye

ar

Suharto era Post Suharto era'93 '94 '95 '96 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09

Source: Hallward-Driemeier, Kochanova and Rijkers (2019)

The Premium on Being Connected (Market Share)

Competition did not improve durably

-20

24

6β P

Cye

ar

Suharto era Post Suharto era'93 '94 '95 '96 '00 '01 '02 '03 '04 '05 '06 '07 '08 '09

Source: Hallward-Driemeier, Kochanova and Rijkers (2019)

Evolution of competition index

I Districts in which Suharto appointed majors remained in powerlonger have worse development outcomes (Martinez-Bravo et al, 2017)

How does evasion happen in practice? Collusion inMadagascar customs

Why Madagascar?

I WB Operational Engagement: Public SectorPerformance Project (P150116) (since 2014)

I Customs account for 48% of government revenue

I Data availability: rich customs transaction data withinformation on inspector and broker

Do allocations of declarations follow official rules?

I Formal rules minimum workload principle (randomassignment)I Probability of a broker matching with a given inspector is

equal to her “productivity”

I Reality excess interaction (manipulation)I certain inspectors and brokers interact excessivelyI allocation of 10% of declarations is manipulated by the IT

system and the port manager

Allocations of declarations do not follow official rules

0.0

5.1

.15

.2Fr

eque

ncy

0 5 10 15 20 25Inspector share of a given broker's declarations (%)

Predicted ObservedSource: Rijkers, Fernandes, Raballand, Chalendar, and Mattoo (2019)

Shares by brokerDistribution of Declarations Across Inspectors

Deviant declarations are risky

66.

57

7.5

8ris

k sc

ore

-.1 0 .1 .2 .3excess interaction between broker and inspector

Fitted 95% CISource: Rijkers, Fernandes, Raballand, Chalendard, and Mattoo (2019)

Deviant declarations are subject to higher taxes

.3.3

5.4

tax

rate



Deviant declarations are undervalued

-1-.8

-.6-.4

-.2lo

g (u

nit p

rice/

med

ian

unit

pric

e)



Inspectors treat deviant declarations differently

-3-2

-10

1βe

xces

s in

spec

tion

shar

e

outcomes

time (log hours) fraud recordedchange in log tax log potential revenue loss

Rijkers, Fernandes, Raballand, Chalendard, and Mattoo (2019)

How (long) was collusion curbed?

05

1015

20%

of c

ollu

sive

dec

lara

tions

-35 -30 -25 -20 -15 -10 -5 0 5 10Month

excess interaction manipulation of re-randomizationExternal re-randomization of allocation of declarations starts at month 0.

before and after introduction of re-randomizationEvolution of non-random assignment

I 5 inspectors were sanctioned (or volunteered to retire)I A 3rd party re-randomized the allocation of

declarations to inspectorsI temporarily eliminated non-random assignment

I IT manipulation resurfacedI IT staff enabled “bypassing” of re-randomization

Is World Bank aid diverted?

Elite Capture of Aid

I Do World Bank aid disbursements trigger increases inforeign deposits in havens?I Havens: offshore banking centers known for secrecy and

asset protectionI e.g. Switzerland, Luxembourg, Cayman Islands

I Premise: money flowing to havens belongs to elitesI 50% (80%) of Scandinavian assets in havens belong to the

0.01% (0.1%) wealthiest households (Alstadsaeter et al., 2017)

Data

I Haven deposits: quarterly bilateral non-bank depositsfrom Locational Banking Statistics of the Bank forInternational Settlements (BIS)I confidential: 1990-2010I publicly available - selected banking centers only:

1990-2018I Aid: quarterly IDA plus IBRD disbursements

I Sample: highly aid dependent countriesI average annual WB aid > 2% of GDPI absorb 10% of IDA & IBRD disbursements

We cannot identify the mechanism(s)

Caveats:I we don’t know who owns the depositsI aid is fungibleI elite capture of aid is a possible explanation, but there are

others

Your feedback is appreciated!

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

Source: Andersen, Johannesen and Rijkers (2019)

Non haven deposits do not respond to aid inflows

-50

5Pe

rcen

tage

cha

nge

in n

on-h

aven

dep

osits



Aid differentially impacts haven vs non-haven deposits

-50

5Pe

rcen

tage

cha

nge

in h

aven

to n

on-h

aven

ratio



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

Transparency (post 2009) 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)

Conclusion

I Corruption is widespread, systemic, and costlyI undermines aid effectiveness

I Difficult, but not impossible, to dislodgeI regime change alone does not curb captureI IT solutions can help - but are also susceptible to captureI transparency alone does not suffice

I Altering incentives is crucial

References

I Arouri, Hasen, Leila Baghdadi and Bob Rijkers. How do Dictators Get Rich? State Capture in Ben Ali’sTunisia, Ch 6. in Crony Capitalism in the Middle East Business and Politics from Liberalization to the ArabSpring (editors: Ishac Diwan, Adeel Malik, and Izak Atiyas) Oxford University Press 2019.

I Andersen, Jorgen, Niels Johannesen and Bob Rijkers. Elite Capture of Aid: Evidence from Offshore BankAccounts, 2019.

I Hallward-Driemeier, Mary, Ana Kochanova and Bob Rijkers. Cronyism and Competition in IndonesianManufacturing Before and After Suharto, 2019. (R&R at EJ)

I Rijkers, Bob, Hassen Arouri, and Leila Baghdadi and Gael Raballand. Are Politically Connected Firms MoreLikely to Evade Taxes?, World Bank Economic Review ABCDE Supplement, 2017.

I Rijkers, Bob, Ana Fernandes, Gael Raballand, Cyril Chalendard, and Aaditya Mattoo. Collusion in Customs:Evidence from Madagascar, 2019.

I Rijkers, Bob, Caroline Freund and Antonio Nucifora. All in the Family: State Capture in Tunisia, Journal ofDevelopment Economics, 2017, Vol 124:41-59

I Rijkers, Bob, Leila Baghdadi and Gael Raballand. Political Connections and Tariff Evasion: Evidence fromTunisia, World Bank Economic Review, 2017, Vol 31(2):459-482.

detecting corruption and state capture - world...

Documents