Compliance with Endogenous Audit Probabilities

• Author: Kai A. Konrad,Salmai Qari,Tim Lohse
• Date: 01 July 2017
• DOI: http://doi.org/10.1111/sjoe.12182
• Published date: 01 July 2017

Scand. J. of Economics 119(3), 821–850, 2017

DOI: 10.1111/sjoe.12182

Compliance with Endogenous Audit

Probabilities∗

Kai A. Konrad

Max Planck Institute for Tax Law and Public Finance, DE-80539 Munich, Germany

kai.konrad@tax.mpg.de

Tim Lohse

Berlin School of Economics and Law, DE-10825 Berlin, Germany

tim.lohse@hwr-berlin.de

Salmai Qari

Berlin School of Economics and Law, DE-10825 Berlin, Germany

salmai.qari@hwr-berlin.de

Abstract

In this paper, we study the effect of endogenous audit probabilities on reporting behavior

in a face-to-face compliance situation, such as at customs. In an experimental setting in

which underreporting has a higher expected payoff than truthful reporting, we find an

increase in compliance of about 80 percent if subjects have reason to believe that their

behavior towards an officer influences their endogenous audit probability. Higher compliance

is driven by considerations about how their own appearance and performance affect their

audit probability, rather than by the social and psychological effects of face-to-face contact.

Keywords: Audit probability; compliance; customs; face value; tax evasion

JEL classification:JEL classification:C91; H26; H31; K42

I. Introduction

The audit probability shows up in the theory of crime (Becker, 1968)

as a key determinant of deceptive behavior. This probability can be ex-

ogenously given; however, in many real life instances, it is not. When

individuals communicate face-to-face, for example, with tax authorities,

∗We kindly thank the Munich Experimental Laboratory for Economic and Social Sciences

(MELESSA), University of Munich, for providing laboratory resources. We thank Hans

Mueller for developing the web-based environment. We thank Ralph Bayer, Nadja Dwenger,

Wer n e r G ¨

uth, Changxia Ke, Vilen Lipatov, Elena Manzoni, Florian Morath, Hans-Werner

Sinn, Fangfang Tan, and audiences at various conferences and seminars for helpful comments.

We also thank two anonymous referees for their helpful reviews and suggestions. We are very

grateful to customs officers from the airports of Berlin and Munich for giving us valuable

insights into their work.

CThe editors of The Scandinavian Journal of Economics 2016.

822 Compliance with endogenous audit probabilities

with their boss about the budgeting costs of a business project they are

responsible for, or with their professor about the potential plagiarism of

their thesis, etc., they have discretion as regards their behavior. In these

situations, individuals need to conjecture how their appearance and per-

formance affect their endogenous probability of receiving an audit. Some

individuals might be self-confident in a face-to-face situation because they

believe that they are highly capable of deception. Therefore, they assume a

low probability for themselves. Others might believe that they are prone to

bad luck, and that Fortuna will always pick on them whenever they attempt

to cheat. In this paper, we make a first attempt to understand the effect

of endogenous audit probabilities on compliance behavior in a face-to-face

situation.

The experimental tax compliance framework we use resembles the sit-

uation at customs. We analyze three treatments to study whether or not

an individual’s compliance decision is affected (at all) by the fact that

their audit probability is endogenous (rather than exogenous), and therefore

depends on their behavior. In our main treatment (later labeled T3) indi-

viduals meet face-to-face with an officer to whom they must make an oral

compliance declaration and who assesses the honesty of their declaration.

This assessment is influential for the question whether or not the individual

receives an audit. This implies that the audit probability is endogenous. We

compare this treatment T3 with a scenario in which there is as little room

as possible for subjective deviations from an exogenously imposed audit

probability. In that treatment (labeled T1), the audit probability is hard

wired in a computerized audit system, and declarations occur directly via

a computer. Hence, there are no officers the individuals meet or could in-

teract with. This benchmark treatment T1 resembles the prototypical set-up

of a large set of computerized tax compliance experiments used throughout

the literature.

However, a whole set of aspects of the compliance situation changes be-

tween a purely computerized audit system with an exogenous audit prob-

ability as in treatment T1 and a compliance situation with face-to-face

contact with an officer and an endogenous audit probability as in treatment

T3 (e.g., Holm and Kawagoe, 2010). Personal contact with a customs offi-

cer (rather than declaring on a computer screen) can, for instance, cause a

higher mental cost of lying (Vanberg, 2008; Lundquist et al., 2009), might

invoke shame (Coricelli et al., 2010) or guilt aversion (see Charness and

Dufwenberg, 2006, among others), or trigger other psychological effects on

self-image, etc. These are relevant dimensions for deception decisions, and

it is evident from some of these studies that the strength of the psycho-

logical motivations depends on the specific environment of the compliance

CThe editors of The Scandinavian Journal of Economics 2016.