Anticorruption reforms, tax evasion, and the role of harassment

• Author: Dyuti S. Banerjee,Samarth Vaidya
• DOI: http://doi.org/10.1111/jpet.12304
• Published date: 01 February 2019
• Date: 01 February 2019

B

62 © 2018 Wiley Periodicals, Inc. wileyonlinelibrary.com/journal/jpet Journal of Public Economic Theory. 2019;21:62–80.

Received: 9 June 2017 Revised: 4 April 2018Accepted: 4 April 2018

DOI: 10.1111/jpet.12304

ARTICLE

Anticorruption reforms, tax evasion, and the role

of harassment

Dyuti S. Banerjee1Samarth Vaidya2

1Departmentof Economics, Monash University,

Clayton,VIC 3800, Australia

2Departmentof Economics, Deakin Business

School,Deakin University, Geelong, VIC 3220,

Australia

DyutiS. Banerjee, Department of Economics,

MonashUniversity, Clayton, VIC 3800, Australia

(dyuti.banerjee@monash.edu).

SamarthVaidya, Department of Eco-

nomics,Deakin Business School, Deakin

University,Geelong, VIC 3220, Australia

(samarth.vaidya@deakin.edu.au).

We examine the impact of anticorruption reforms on tax evasion

when corruption and potentially harassment are endemic among tax

auditors. We find that the threat of harassment may counterintu-

itively boost the impact of such anticorruption reforms on tax eva-

sion and also eliminate corruption. Specifically, a moderate anticor-

ruption policy can discontinuously reduce tax evasion to a leveleven

below that under no corruption. Furtherstrengthening of such policy

can nonetheless prove counterproductive and increase tax evasion.

On the contrary, in the absence of harassment,a moderate anticor-

ruption reform induces higher tax evasionand sustenance of bribery.

In this case, only a large reform can reduce tax evasion and eliminate

corruption.

1INTRODUCTION

The dual problems of tax evasion and corruption are pervasive in developing countries where “the payment of taxes

often tends to be the exception rather than the rule …. Moreover, evasion tends to be disproportionately severe for

direct (income) taxes …. Ultimately, corruption becomes pervasive…” (Das-Gupta & Mookherjee, 1998). Empirical

evidence from 27 transitional economies shows that business tax evasionthrives when bribes to tax officials are com-

monplace (Joulfaian, 2009). The extentof tax evasion can be considerable: Fjeldstad and Tungodden (2003) show that,

in developing countries, over 50% of tax revenues do not reach government treasuries. In the contextof Ghana, the

ratioof tax revenues to gross domestic product (GDP) was as low as 4.5% in 1983, before reforms were initiated (Chand

& Moene, 1999).

Further, in addition to bribery and tax evasion, the practice of harassment, by which tax officers threaten to

overassess tax liabilities, can be quite entrenched in tax administrations in countries as diverse as the United States,

India, and the Philippines.1Similarly, Safavian, Graham, and Gonzalez-Vega (2001) provide evidence of extortionary

behavior in Russia, where tax officials have sufficient discretion to arbitrarily change the tax liabilities of microenter-

prises. As vividly described in Hindriks, Keen, and Muthoo (1999), such overassessmentcan be achieved in many ways,

such as disallowing or challenging legitimate deductions and misreporting taxpayer characteristics.

Consequently, a literature has emerged which studies the various facets of the interplaybetween corruption and

harassment. Both the consequences of such illicit behavior and the role of penalties and rewards as deterrence

1SeeHindriks, Keen, and Muthoo (1999), Klitgaard (1988), and Marjit, Mukherjee, and Mukherjee (2000).

m

c

a

c

s

p

t

f

b

m

m

i

n

e

m

o

a

a

o

t

e

e

a

l

a

r

t

q

c

c

C

o

t

e

2

3

M

4

a

c

5

BANERJEE ANDVA IDYA 3

63

0.

s

r

a

l

-

t

,

e

d

o

s

,

y

r

-

s

,

d

e

measures have been examined. Polinsky and Shavell (2001) consider a general setting of law enforcement, where

corruptible officers can also leverage their ability to frame innocent civilians in order to impose fraudulent charges

and extract extortionary payments. Their paper makesa case in favor of an appropriate reward for law enforcers in

conjunction with maximal penalties for corruption and framing but no sanction for extortion. This is because such a

sanction may lead to either higher bribes or framing, thereby victimizing the innocent and reducing deterrence. The

paper also points out that while higher rewards to enforcers may reduce corruption, it may also lead to more framing.

Mookherjee (1997) also studies the implications of reward and punishment schemes for deterring corruption and

tax evasion. Unlike Polinsky and Shavell (2001), Mookherjee (1997) endogenizes the auditor's monitoring effort and

finds that rewards work better for tax compliance whereas increased punishment can perversely increase tax evasion

by discouragingaudit effort. However, this paper also cautions that rewards can lead to credible threats of overassess-

ment and therefore increased extortion. Dhillon, Nicolo, and Xu (2017) also show that raisingmonetary compensation

mayperversely incentivize reputation-concerned public officials to behave corruptly. Marjit and Shi (1998) point to the

ineffectiveness of such bonuses when the tax officer's monitoring effort is endogenous.

Hindrikset al. (1999) instead focus on the equity and efficiency implications of taxation systems in which officers can

extract bribes and also threaten citizens with overassessment.In this context, they allow for general taxation, punish-

ment, and remuneration schemes but assume that monitoring is costless and thus abstract from theirincentive effects

on audit efforts. They show that unfettered corruption and harassmenthave regressive effects on the tax system. They

also examinethe characteristics of taxation schedules and reward and punishment schemes that are tax-evasion-proof

and corruption-proof. This analysis reveals that an honest implementation of progressive taxation requires that tax

officers receive rewards but only for high income reports. This needs to be financed through a higher averagetax rate,

thus bringing to light an additional trade-off between equity and efficiency due to the latent threat of corruption and

evasion. Marjit et al. (2000) also consider the equity implications of harassment in the context of corruption and tax

evasion and, likeHindriks et al. (1999), find that it has a regressive impact on income tax burdens.

The commonality in the findings of Mookherjee (1997), Polinskyand Shavell (2001), and Hindriks et al. (1999) is that

a policy mix aimed at inducing law-abiding behavior needs to be comprehensive,an integral part of which is rewards to

law enforcers. However,Das-Gupta and Mookherjee (1998) mention the difficulty in achieving such all-encompassing

reforms. For example, the tax administration is often not grantedautonomy over its personnel management and has

to comply with the general rules governingemployment of civil servants across all government agencies. The following

quote from TheJakarta Poston the recent stalling of tax reform in Indonesia illustrates this aptly:2

Critics have said that despite much-touted tax reform,inrealitylittlehas changed….Despite theremuneration

policy [applied in 2007] and theimplementation of the internalmonitoring system,taxmafia operations are still

rampant in tax offices.It's an indication that tax reformhas not progressedas expected.

While broad-based reforms are difficult to implement, public sentiment against corruption is typically high in many

countries,and governments find it politically expedient to be seen to be acting tough on the corrupt behavior of bureau-

crats by raising penalties on such offenses.3The Indian Parliament's recent approval of the anti-graftLokpal Bill and

Chinese president Xi Jinping's ambitious anticorruption drive, which snared dozens of businessmen and government

officials, are pertinent examples.4In light of this, instead of examining comprehensive reforms, as has been done in

the above literature, we focus only on the effectiveness of anticorruption penalties in alleviating corruption and tax

evasion.5

2See“Internal tax reform is stalling: Watchdogs,” TheJakarta Post, July 16, 2012.

3As the Ghana case reveals, the tax reforms of the Rawlings government started with severepunishments to corrupt officers. For details, see Chand and

Moene(1999).

4See TheTimes of India, December 18, 2013, http://timesofindia.indiatimes.com/india/Lok-Sabha-passes-amended-Lokpal-Bill-Anna-Hazare-ends-fast/

articleshow/27580904.cms. For China, see Bloomberg, March 4, 2014, http://www.bloomberg.com/news/print/2014-03-03/china-s-xi-broadens-graft-

crackdown-to-boost-influence.html.

5Strimbuand Gonzalez (2018) take the alternative route of examining the role of transparency in alleviating political corruption.