ANALYZING FACTORS AFFECTING TAX MORALE LEVELS IN TURKEY USING ALTERNATIVE ORDERED RESPONSE MODELS.

• Author: Ozdemir, Dilek

1. Introduction

   Tax morale is commonly described as taxpayers' intrinsic motivation to pay taxes (Frey and Torgler, 2007; Lubian and Zarri, 2011; Lago-Penas and Lago-Penas, 2010). Therefore, the degree of tax morale is commonly considered to explain the reason of paying taxes by individuals (Russo, 2013; Slemrod, 2007). Similarly, tax morale is also regarded as an important determinant of explaining the high degree of tax compliance in many country cases with low deterrence level (Lisi, 2015), while some earlier research (Baldry, 1986; Cummings et al., 2009; Blaufus et al., 2015) highlights the relationship between the concept of tax morale and its significant role in explaining observed tax evasion. A recent study (Lubian and Zarri, 2011) indicates that individuals pay taxes due to their willingness, namely the sense of moral obligation, and the corresponding evidence is discussed under positive hedonic effects. However, as an individual attitude, the high degree of tax morale does not always guarantee the high degree of tax compliance as an individual behavior (Lisi, 2015). Frey and Torgler (2007) argue that since the behavior of other taxpayers may be crucial to better understanding tax compliance, pro-social behavior theories may be valuable to explain the corresponding behavior. In fact, as individuals' willingness to pay taxes increases, the other taxpayers' behavior as being honest also increases. Torgler and Schneider (2009) find strong evidence that higher degrees of tax morale and institutional quality provide a smaller shadow economy.

   Prior studies find evidence that tax morale is not dependent on an individual's economic situation. Particularly, Blaufus et al. (2015) concentrate on whether tax morale is directly associated with individuals' economic situation or there exists a self-serving bias and their empirical analysis concludes that tax morale is subject to a self-serving bias. Their results also reveal that individuals consider tax evasion as a less unethical behavior when they have an opportunity to evade. On the contrary, Lago-Penas and Lago-Penas (2010) find that individuals' income and the degree of tax morale are positively correlated. They also indicate a negative correlation between self-employment and the degree of tax morale in line with other empirical research (Torgler, 2004; Frey and Torgler, 2007; Kountouris and Remoundou, 2013). Filippin, Fiorio and Viviano (2013) explore the association between tax enforcement and tax morale with a motivation to analyze the relationship between formal and informal institutions. The empirical findings of a relatively large microdata conducted in 2004 by the Bank of Italy put forward that there exists a positive correlation between tax enforcement and tax morale. Besides, the impact of tax enforcement and social environment appears to be more influential in the case of low tax morale.

   The degree of tax morale is considered to be significantly associated with individuals' several demographic characteristics including age-group, religiosity, gender, income and educational level, occupational and marital status (Lago-Penas and Lago-Penas, 2010; Kountouris and Remoundou, 2013; Filippin, Fiorio and Viviano, 2013). For instance, a country-comparison study (Sipos, 2015) carried out on Hungarian and Romanian samples indicates the significant role of the age-group of the respondents since younger respondents are less likely to have a high degree of tax morale. The low degree of tax morale among younger individuals were also found in other empirical research (Torgler, 2004; Torgler, 2005b; Alm and Torgler, 2006; Frey and Torgler, 2007; Cummings et al., 2009; Martinez-Vazquez and Torgler, 2009; Lago-Penas and Lago-Penas, 2010; Kountouris and Remoundou, 2013; Russo, 2013). The results of a recent study (Gungor Goksu and Izgi Sahpaz, 2015) that compares the tax morale attitudes between university students in Turkey and Spain reveal that religious belief is an important factor that separates the respondents' perceptions on tax evasion between these two countries. Other studies (Torgler, 2005a, 2005b; Alm and Torgler, 2006; Torgler, 2006; Martinez-Vazquez and Torgler, 2009; Lago-Penas and Lago-Penas, 2010; Kaynar Bilgin, 2011; Kountouris and Remoundou, 2013) also highlight the significant role of religious belief on tax morale and they were found as positively correlated. Particularly, Torgler and Schneider (2007) find the statistically significant positive impact of religiosity on the higher degree of tax morale only in Belgium and Switzerland, for a three-country comparison.

   Earlier studies (Torgler, 2005b; Alm and Torgler, 2006; Frey and Torgler, 2007; Torgler and Schneider, 2007) indicate that women tend to have a higher degree of tax morale than men. On the contrary, Lago-Penas and Lago-Penas (2010) unexpectedly reveal that male respondents have a relatively higher tax compliance than female counterparts. Previous studies (Torgler, 2005b; Alm and Torgler, 2006; Kountouris and Remoundou, 2013) find evidence that married people tend to have a higher degree of tax morale. Particularly, Frey and Torgler (2007) indicate that divorced and separated persons have the lowest tax morale in Western and Eastern European countries. The positive correlation between the degree of tax morale and respondents' educational level is also experienced in previous studies (Torgler, 2005a; Kountouris and Remoundou, 2013). On the contrary, Frey and Torgler (2007) find evidence that educational level and tax morale is negatively correlated. Along with commonly used determinants of tax morale in the existing literature, Russo (2013) seeks the relationship between tax morale and immigration and finds evidence that tax morale and immigration are positively correlated. In addition, the results of the corresponding study indicate that other determinants of tax morale are social capital, political participation, unemployment, and essential public services. Kountouris and Remoundou (2013) find evidence that if immigrants originate from countries with higher degrees of tax morale, they tend to have a higher moral attitude to pay taxes in a European sample. Moreover, their results reveal a statistically significant relationship between tax morale and cultural background. Among other significant variables, trust in government (Kaynar Bilgin, 2011; Torgler, 2012), politicians (Lago-Penas and Lago-Penas, 2010), the justice system (Torgler, 2012) or European Union (Torgler, 2012) were also found as statistically significant determinants of tax morale. Many earlier studies have performed standard or weighted OPROBIT (Alm and Torgler, 2004; Frey and Torgler, 2007; Kaynar Bilgin, 2011; Kountouris and Remoundou, 2013), standard or multi-level OLOGIT (Lago-Penas and Lago-Penas, 2010) models for the analysis of datasets, although as far as is known, no earlier studies consider alternative ordered response models as their estimation.

   As of December 31, 2017, the number of tax subjects has reached approximately 10.6 million people in Turkey (Turkish Revenue Administration, 2018). As a developing country, successful return of tax payments as an increasing tax revenue plays a crucial role for the Turkish economy to maintain its economic development and to achieve further economic goals. In that context, a better understanding of the behavior of taxpayers may give valuable information in terms of increasing tax revenues. The main objective of this paper is to examine the behavior of subjects in Turkey by concentrating on a specific sample area, namely, northern Turkey. The determinants of the degree of tax morale were considered as the basis that best explains the behavior of Turkish taxpayers. This paper performs OPROBIT and other alternative ordered response models to analyze the obtained data. The following sections introduce the sample area, present the estimation results, and deal with the discussion of the empirical findings with substantial recommendations for future tax policies and future research.

2. Results

   2.1. Data description and pre-estimation tests

   This paper aims to examine factors influencing the tax morale attitude of subjects in Turkey. Another purpose of this study is to determine the best model fit that explains the determinants of tax morale. For this purpose, a well-established written questionnaire was administered to 943 subjects in three northeastern cities of Turkey, namely, Erzurum, Erzincan, and Bayburt in July 2015. The corresponding questionnaire mainly follows earlier research (Blaufus et al., 2015; Gungor Goksu and Izgi Sahpaz, 2015) in the tax morale literature on selecting distinguishing questions. The questionnaire included three main sections. The first section involves questions about subjects' opinions about the current condition of the taxation system in Turkey. The second section involves detailed questions about subjects' opinions on tax morale, tax compliance, and tax evasion. The third section involves questions about subjects' non-confidential demographic characteristics. A simple random sampling methodology was followed in regard to the total population of three selected cities. According to the Nomenclature of Territorial Units for Statistics (NUTS) 2 classification of Turkey, the three selected cities are located in TRA1 sub-region of Turkey. Erzurum is the most populous city of this sub-region with almost 761 thousand inhabitants, while Erzincan city hosts approximately 232 thousand people. Bayburt is the least populous city of Turkey with almost 80 thousand inhabitants (TurkStat, 2018). To the best of our knowledge, no earlier studies have been made in this specific sub-region, which makes it one of the main reasons for selecting this sub-region as target sample. A...