Business Tax Evasion in Transition Economies: A Cross-Country Panel Investigation.

• Author: Abdixhiku, Lumir
• Position: Report

1. Introduction

   With taxes comes evasion (Cowell, 1990). Tax evasion imposes economic costs: it slows down economic growth by weakening the government's ability to provide adequate public goods (Johnson et al., 2000); it diverts resources to unproductive activities such as establishing financial subsidiaries to cover-up evasion (Slemrod, 2007); it provides an incentive for firms to remain small and invisible to facilitate evasion, thereby missing opportunities from the formal economy (Nur-tegin, 2008); and it generates inequity between the evaders and the honest taxpayers by shifting the burden to the latter group, thereby creating an incentive for further evasion (Feinstein, 1991).

   Tax evasion is one of the major problems facing developing (Fuest and Riedel, 2009) and transition economies (Pirttila, 1999). The literature on the factors shaping tax evasion is fairly well developed (reviews include: Jackson and Milliron, 1986; Cowell, 1990; Andreoni, et al., 1998; Franzoni, 2008; Torgler 2011). However, most of it relates to individuals.

   The lack of research on tax evasion by businesses is unfortunate, especially given the fact that in most countries the bulk of taxes is paid by firms and firms account for the bulk of tax evasion too (McCaffery and Slemrod, 2004; Crocker and Slemrod, 2005; Chang and Lai, 2004, Nur-tegin, 2008). Moreover, as suggested by Andreoni et al (1998) there is a huge gap and thus a permanent need for international and cross country research on tax evasion; while the work in the context of transition countries is still less developed. This paper aims to reduce this gap by introducing some empirical findings for businesses, cross-country and transition features of tax evasion.

   The starting assumption in our work is similar to the assumption made Generally in the current literature on the tax behaviour of businesses, which is that the behavior of businesses is similar to the behaviour of individuals, and that--as a corollary--the determinants of business tax evasion may be similar, at least qualitatively, to the determinants of tax evasion by individuals or households. (4) As Slemrod (2007, p.36) points out, the literature on business tax evasion "adapts the theory of tax evasion, which for the most part concerns individual decision makers, to the tax compliance decisions made by businesses'".

   So far, cross-country investigations on tax evasion are rare. Through this paper we want to build upon pioneering work of cross-country investigation in Riahl-Belkaoiu (2004) and Richardson (2006) who have analysed individual tax evasion in respectively 30 and 45 countries. Riahl-Belkaoiu (2004) examines the international differences in tax evasion and relates these differences to selected determinants of tax morale. His findings show that tax evasion is lowest in countries characterized by high economic freedom, a developed equity market, effective competition laws and a low serious crime rate.

   Richardson (2006) on the other hand advances cross-country investigation of individual tax evasion using a larger sample and finds that non-economic determinants have the strongest impact on tax evasion in comparison with economic determinants; most notably, the complexity of the tax system, education, income source, fairness and tax morale are highly correlated with tax compliant behaviour. We extend their (crosscountry) approach by focusing on business instead of individual tax evasion and by focusing only on transition countries. (5)

   Firm level analysis using BEEPS data have been also conducted previously By Nur-tegin (2008) Joulfaian (2009) and Abdixhiku et al. (2016). In addition to Riahl-Belkaoiu (2004) and Richardson (2006), we build upon the work of Nurtegin, (2008), Joulfaian (2009) and Alon and Hageman (2012) by first, introducing new-- and, so far, unobserved--determinants to the tax compliance model given the aggregation of micro data from various non-BEEPS sources; and second, by increasing the time period of dataset to three years (as compared to one in the past studies).

   In a very recent publication, Alm, Martinez-Vazquez and McClellan (2016) Also investigate the relationship between corruption and tax evasion--i.e. on 'how the potential for bribery of tax officials affects a firm's tax evasion decision'. They develop a new theoretical model of firms reporting when bribery is an option. Our cross-country panel investigation, instead, uses the traditional model--and focuses on a wider range of determinants of tax evasion that go beyond the causal relationship between tax evasion and corruption related to tax administration. In our work, we focus on macroeconomic factors (GDP per capita, unemployment, inflation, non-performing loans); institutions (corruption and institutional development); and socio-cultural differences (social norms and education). Furthermore, when treating corruption in tax administration as a control variable for tax evasion, (6)* we come to similar results as in Alm, Martinez Vazquez & McClellan (2016)- that the likelihood of the decision to evade increases in countries where there is corruption in tax administration and that the possibility of bribing the tax administration officials affects business decision to evade.

   Specifically, this paper contributes to the literature by using the data for the years 1999, 2002 and 2005 to investigate business tax evasion in 24 transition economies. In our study, we incorporate institutional and macroeconomic indicators alongside tax rate and cultural influences on business evasion in transition economies.

   To analyse the data from 24 transition economies for the years 1999, 2002 and 2005, we employ a conventional fixed effects approach as well as a recent innovation in fixed effect panel analysis, known as fixed effect vector decomposition (FEVD), which hitherto has not been used in this context. The main benefit of this approach is that it enables us to model the effect of time-invariant (or, at least, "slow moving") variables, most notably proxies for institutional development. These determinants were not captured by previous studies (Riahl-Belkaoiu, 2004; Richardson, 2006; Nur-tegin, 2008; Joulfaian, 2009; and Alon and Hageman, 2012)

   The paper is organized as follows. In Section 2 we provide a brief overview of existing cross-country investigations. In Section 3 we describe the data used in our study and we review the major determinants of tax evasion. In Section 4 we outline the general form of the regression model. In Section 5 we focus on our approach to estimation and highlight the importance of diagnostic testing. Sections 6 and 7 report and discuss the empirical findings. The final section concludes.

2. Theory and determinants of tax evasion

   Over four decades of research on tax evasion have given rise to an Enormous amount of work (for reviews see Jackson and Milliron, 1986; Cowell, 1990; Andreoni, et.al, 1998; and Franzoni 2008 and Torgler 2011). The vast majority of this work, however, has neglected three important factors. The first factor relates to the importance of cross-country investigations. In one of the most insightful reviews of tax evasion, Andreoni et al. (1998, p.855), while concluding and providing directions for future research, argue that "... a broadening of the empirical database will improve the power of statistical tests of theoretical models, and spur comparative analysis across countries ". The second factor relates to the lack of studies on business tax evasion. As Torgler (2011) argues, "... business tax evasion in general, has received very little attention. Work in this area is therefore highly relevant (p.6)". Last, the context of transition economies in tax evasion studies has received limited attention (Pirttila, 1999).

   In this paper we attempt to fill these gaps by introducing all three components: business tax evasion; cross-country comparison; and the transition context. In order to do so, we start by assuming that the behaviour of businesses is similar to the behaviour of individuals, and that the determinants of business tax evasion may be similar, at least qualitatively, to the determinants of tax evasion by individuals or households. The decision on evasion, or compliance, is made by individual managers or entrepreneurs who, in essence, act as individuals (Arias, 2005). As Slemrod (2007, p.36) points out, the literature on business tax evasion "adapts the theory of tax evasion, which for the most part concerns individual decision makers, to the tax compliance decisions made by businesses". This is particularly true of small and medium sized enterprises (SMEs) where the decision making entrepreneur makes compliance decisions both as an individual and as a manager.

   Ever since Allingham and Sandmo's conventional model was introduced in 1972, theoretical and empirical literature on tax evasion has flourished. Advances incorporating interactions between institutions and taxpayers, cultural and behavioural differences as well as individual socio-demographic characteristics have also been made. These and conventional determinants of tax evasion, namely the tax rate, fine rate and audit rate, have contributed profoundly to modelling compliance decisions.

   In the traditional model the level of income tax evasion is negatively related to the level of punishment (7) imposed by law and the probability of audit by tax examiners. (8) When analysing the impact of tax rates on evasion, the model predicts an ambiguous effect with the occurrence of both an income effect (as tax rates rise, people become poorer and, in the presence of decreasing absolute risk aversion, they evade less) and a substitution effect (rising taxes means that the return from evasion is higher, thus the taxpayer prefers the risky choice to the safer one). However, Yitzhaki (1974) argued that the ambiguity was a result of an unrealistic assumption of the model that the penalty is imposed on the amount of income not...