determinants of tax evasion: a cross-country investigation

Journal of International Accounting, Auditing and Taxation15 (2006) 150–169

Determinants of tax evasion:A cross-country investigation

Grant Richardson ∗Department of Accountancy, Faculty of Business, City University of Hong Kong,

83 Tat Chee Avenue, Kowloon Tong, Hong Kong

Abstract

The purpose of this study is to expand on the work of Riahi-Belkaoiu [Riahi-Belkaoiu, A. (2004). Relation-ship between tax compliance internationally and selected determinants of tax morale. Journal of InternationalAccounting, Auditing and Taxation, 13, 135–143] and systematically investigate, on a cross-country basis,many of the key determinants of tax evasion identified by Jackson and Milliron [Jackson, B. R., & Milliron,V. C. (1986). Tax compliance research: findings, problems and prospects. Journal of Accounting Litera-ture, 5, 125–165]. Based on data for 45 countries, the results of the OLS regression analysis show thatnon-economic determinants have the strongest impact on tax evasion. Specifically, complexity is the mostimportant determinant of tax evasion. Other important determinants of tax evasion are education, incomesource, fairness and tax morale. Overall, the regression results indicate that the lower the level of complexityand the higher the level of general education, services income source, fairness and tax morale, the loweris the level of tax evasion across countries. These findings remain robust to a broad range of cross-countrycontrol variables, an alternative tax evasion measure and various interactions.© 2006 Elsevier Inc. All rights reserved.

Keywords: Tax evasion; Complexity; Education; Income source; Fairness; Tax morale

1. Introduction

Tax evasion has been the subject of a great deal of academic research in most developedcountries over a long period of time (Andreoni, Erard, & Feinstein, 1998; Cuccia, 1994; Jackson& Milliron, 1986; Kinsey, 1986; Long & Swingen, 1991; Richardson & Sawyer, 2001). Even so,little research has focused on the underlying determinants of tax evasion on a cross-country basis.This is disappointing since Andreoni et al. (1998, p. 856) and Tan and Sawyer (2003, p. 454) haveargued there is a need for international and cross-country comparisons on this topic.

∗ Tel.: +852 2788 7923; fax: +852 2788 7944.E-mail address: [email protected].

1061-9518/$ – see front matter © 2006 Elsevier Inc. All rights reserved.doi:10.1016/j.intaccaudtax.2006.08.005

mailto:[email protected]

dx.doi.org/10.1016/j.intaccaudtax.2006.08.005

G. Richardson / Journal of International Accounting, Auditing and Taxation 15 (2006) 150–169 151

Riahi-Belkaoiu (2004) considered the relationship between selected determinants of tax moraleand tax evasion, employing data from 30 countries. He finds empirical evidence to show that taxevasion across countries is negatively related to the level of economic freedom, the level ofimportance of the equity market, the effectiveness of competition laws and high moral norms.Notwithstanding, he only explored the broad link between tax evasion and some selected deter-minants of tax morale across countries.

The first major tax evasion literature review by Jackson and Milliron (1986)1 established14 key determinants of tax evasion. These include: age, gender, education and occupa-tion status (‘demographic’ determinants), income level, income source, marginal tax rates,sanctions and probability of detection (‘economic’ determinants), and complexity, fairness,revenue authority initiated contact, compliant peers and ethics or tax morale (‘behavioral’determinants).

The purpose of this study is to expand on the work of Riahi-Belkaoiu (2004) and systematicallyinvestigate on a cross-country basis, many of the key determinants of tax evasion identified byJackson and Milliron (1986). When these determinants are clearly identified in a systematicway by empirical analysis, appropriate policy conclusions can then be drawn and governmentpolicymakers are then in a position to design and implement strategies to minimize the damagingeffects of tax evasion.

Based on data for 45 countries, the results of the OLS regression analysis for this study show thatnon-economic determinants have the strongest impact on tax evasion. In particular, complexityis the most important determinant of tax evasion. Other significant determinants of tax evasionare denoted by education, income source, fairness and tax morale. On the whole, the results ofthe regressions show that the lower the level of complexity and the higher the level of generaleducation, services income source, fairness and tax morale, the lower is the level of tax evasionacross countries. These findings remain robust to a broad range of cross-country control variables,an alternative tax evasion measure and various interactions.

This study contributes to the literature in four ways. First, it builds upon the pioneering work ofRiahi-Belkaoiu (2004) and investigates more systematically many of the key demographic, eco-nomic and behavioral determinants of tax evasion. It thereby fills a major gap in the tax literatureby exploring the major determinants of tax evasion across countries. Second, it demonstrates that‘mixed’ models of tax evasion that integrate demographic, economic and behavioral determinants,offer valuable insights into our understanding of tax evasion across countries. Third, it providesa sound empirical framework, which includes a substantial list of multiple cross-country controlvariables, for further international research on tax evasion. Fourth, it also presents a key summaryof multiple data sources for future international tax research.

The remainder of the paper is organized into the following sections. Section 2 reviews themajor determinants of tax evasion discussed in the literature and formulates research hypothe-ses. Section 3 describes the research design. Section 4 summarizes and analyzes the empir-ical results of this study. Section 5 presents the conclusions, limitations and future researchopportunities.

1 Other major tax evasion literature reviews have been carried out by Kinsey (1986), Cuccia (1994), Andreoni etal. (1998), Long and Swingen (1991) and Richardson and Sawyer (2001). However, the literature review by Jacksonand Milliron (1986) represents the first to systematically identify and discuss in comprehensive terms, all of the keydeterminants of tax evasion.

152 G. Richardson / Journal of International Accounting, Auditing and Taxation 15 (2006) 150–169

2. Major determinants of tax evasion: theory and hypotheses

Why do taxpayers in some countries evade paying income taxes more frequently than taxpayersin other countries? This question can be answered by considering the major determinants oftax evasion previously identified in the literature. Jackson and Milliron (1986) provide the firstdetailed review on this topic. This study considers the impact of 10 of Jackson and Milliron’s(1986) key variables: age, gender, education, income level, income source, marginal tax rates,fairness, complexity, revenue authority initiated contact and tax morale.2

The chronological age of taxpayers is one of the most important determinants of tax evasion(Jackson & Milliron, 1986, p. 130). Studies find that older taxpayers are generally more compliantthan younger taxpayers (Tittle, 1980; Witte & Woodbury, 1985; Dubin & Wilde, 1988; Feinstein,1991; Hanno & Violette, 1996). Tittle (1980) explains the relationship between age and taxdeviance as attributable to lifecycle variations and generational differences. Younger taxpayersare more risk-seeking, less sensitive to penalties (a lifecycle variation), and reflect the social andpsychological differences related to the period in which they are raised (a generational difference).

Gender of the taxpayer has been revealed to be significant in previous studies. For example,Vogel (1974) and Mason and Calvin (1978) show that the compliance levels of female taxpayersare normally higher than for male taxpayers. Jackson and Milliron (1986, p. 131) argue that thecompliance gap between females and males is shrinking over time as new generations of liberatedwomen emerge. However, studies of gender and tax evasion since Jackson and Milliron (1986)tend to show that the compliance gap among females and males has been maintained (e.g., Brooks& Doob, 1990; Collins, Milliron, & Toy, 1992).

Education attainment is another important determinant of tax evasion. It usually relates to ataxpayer’s ability to comprehend and comply or not comply with income tax laws (Jackson &Milliron, 1986, p. 132). It has been argued by Jackson and Milliron (1986, p. 132) that educationhas two elements: the general degree of fiscal knowledge and the specific degree of knowl-edge regarding tax evasion opportunities. They claim that by enhancing the level of generalfiscal knowledge tax compliance improves because of more positive perceptions about taxation.Increased knowledge of tax evasion opportunities has a negative influence on tax compliance as itassists non-compliance. However, the vast majority of studies examining the impact of educationon tax evasion use a taxpayer’s general education level as the approach to measure education(Richardson & Sawyer, 2001, p. 162). Research by Song and Yarbrough (1978), Wallschutzky(1984) and Witte and Woodbury (1985) find a negative association between the general educationlevel of taxpayers and tax evasion.

Income level represents an additional key determinant. It typically refers to the adjusted grossincome or total positive income of a taxpayer (Jackson & Milliron, 1986, p. 133). Mason andLowry (1981) and Witte and Woodbury (1983) find that middle income taxpayers are generallycompliant with tax laws, while low income level taxpayers and high income level taxpayers arerelatively non-compliant with tax laws. Richardson and Sawyer (2001) show however that overallempirical findings remain mixed.

Income source usually refers to the type or nature of the taxpayer’s income (Jackson & Milliron,1986, p. 134). Schmolder’s (1970) shows that when a large part of a country’s labor force is engagedin agriculture and small trading, income and profit taxation is unsuccessful. Wallschutzky’s (1984)

2 Occupation status, sanctions, probability of detection and compliant peers are not considered in this study because ofthe lack of available cross-country data for these determinants.


survey research of tax evaders and non-tax evaders finds that the greatest opportunity to evadeincome tax exists from those who derive their income from agriculture, independent trades orself-employment. The least opportunity exists for those taxpayers whose source of income isdependent on wages or salaries subject to withholding, such as from the services sector.

Marginal tax rates is another important tax evasion determinant, but empirical results are mixed.Clotfelter (1983) and Mason and Calvin (1984) find a positive association between marginal taxrates and tax evasion, while Feinstein (1991) and Christian and Gupta (1993) show a negativeassociation between them. Richardson and Sawyer (2001, pp. 200–201) argue that not controllingfor the correlation between marginal tax rates and income level may cause this inconsistency.They cite the work of Feinstein (1991) who tests an economic model of tax evasion using pooleddata. By pooling data from years in which different tax schedules were operating in the U.S.,Feinstein (1991) is able to separate out the effects of marginal tax rates and income level. Theresults show that higher marginal tax rates lead to reduced tax evasion (Feinstein, 1991, p. 24). Bycombining data from various countries with different tax schedules, this study is able to separateout the effects of marginal tax rates and income level.

It is generally accepted that perceptions about fairness and tax evasion are related (Jackson &Milliron, 1986, pp. 127–129). The importance of taxpayers ‘perceptions’ of fairness should notbe underestimated (Richardson & Sawyer, 2001, pp. 291–295). Spicer (1974) finds a significantnegative association between fairness and tax evasion generally. Song and Yarbrough (1978) alsodetect a significant negative association between these variables, with 75% of taxpayer subjectsstating that the fairness concept of ‘ability to pay’ is more significant for tax evasion than is the‘benefits’ concept. Moreover, Hite and Roberts (1992) find that fairness is significantly associatedwith perceptions of an improved tax system, and that fairness and tax evasion are negatively related.

As taxation systems have become increasingly complex over time in many developed countriesaround the world, complexity has become an important tax evasion determinant (Jackson &Milliron, 1986, p. 138; Richardson & Sawyer, 2001, p. 184). Prior research, utilizing archivaldata (Clotfelter, 1983; Long & Swingen, 1988) and survey data (Collins et al., 1992; Milliron &Toy, 1988; Vogel, 1974) methodologies provide strong empirical evidence which demonstratesthat complexity has a positive association with tax evasion.

Revenue authority initiated contact is another important determinant. In a major study ofEuropean tax structures by Strumpel (1969, pp. 27–30), he observes that while southern Europeancountries have a long history of trying to improve tax compliance by strengthening enforcement,these same countries have the lowest tax compliance rates in Europe (Jackson & Milliron, 1986,p. 139). Consistent with this observation, Spicer and Lundstedt’s (1976) survey of U.S. taxpayersfinds that taxpayers’ direct experience with the revenue authority is positively associated withincreased tax resistance and admitted tax evasion. Research by Wallschutzky (1984), Klepper andNagin (1989a, 1989b) and Brooks and Doob (1990) also supports this view.3 Conversely, reducingthe level of contact between public tax officials and taxpayers by means of a self-assessment taxsystem decreases the possibility of widespread tax resistance and tax evasion (Tanzi, 2000; Sarker,2003; Torgler & Murphy, 2004).

While tax morale is a rather nebulous concept (Jackson & Milliron, 1986, p. 136), it describesthe moral principles or values individuals hold about paying taxes (Torgler & Murphy, 2004,p. 301). Early survey research by Spicer (1974), Spicer and Lundstedt (1976) and Tittle (1980)

3 A few studies (e.g., Spicer & Hero, 1985; Witte & Woodbury, 1983) find that increased revenue authority initiatedcontact leads to reduced tax evasion. However, it appears that this particular association is infrequent (Richardson &Sawyer, 2001, p. 189).


finds that the tax morale of individuals has a negative association with tax evasion. More recently,Torgler (2003a) shows that tax morale and tax evasion are negatively correlated. In addition,Riahi-Belkaoiu (2004) finds empirical evidence which indicates that tax evasion across countriesis negatively related to selected determinants of tax morale.

Following from the above discussion, it is hypothesized that:

H1. All else equal, there is a significant negative association between older taxpayers and taxevasion in a country.

H2. All else equal, there is a significant negative association between female taxpayers and taxevasion in a country.

H3. All else equal, there is a significant negative association between the general education levelof taxpayers and tax evasion in a country.

H4. All else equal, there is a significant positive association between low income level/highincome level taxpayers and tax evasion in a country.

H5. All else equal, there is a significant positive (negative) association between income derivedfrom agriculture (services) and tax evasion in a country.

H6. All else equal, there is a significant negative association between high marginal tax ratesand tax evasion in a country.

H7. All else equal, there is a significant negative association between perceptions of fairnessand tax evasion in a country.

H8. All else equal, there is a significant positive association between complexity and tax evasionin a country.

H9. All else equal, there is a significant negative association between self-assessment and taxevasion in a country.

H10. All else equal, there is a significant negative association between tax morale and tax evasionin a country.

3. Research design

3.1. Sample

Determining the sample for this study was based solely on obtaining the requisite data on thevariables of interest as specified in the hypotheses. A total of 45 countries met this particularrequirement. They are reported in Table 1. The sample comprises both developed and developingcountries, and a mixture of countries differentiated by language, culture and geography. Thecountries included in the sample are diverse.

3.2. Data description

Data for this study are collected from a wide range of sources. Appendix A presents a com-prehensive description of data employed to measure the different variables used and their varioussources. To achieve robustness, both objective and survey measures of the variables are utilized.La Porta, Lopez-de-Silanes, Shleifer, and Vishny (1999, p. 234) argue that this is important for


Table 1List of countries

Argentina Greece PortugalAustralia Hungary RussiaAustria Iceland SingaporeBelgium India Slovak RepublicBrazil Indonesia SloveniaCanada Ireland South AfricaChile Italy SpainChina (PRC) Japan SwedenColombia Korea (South) SwitzerlandCzech Republic Mexico TaiwanDenmark Netherlands ThailandEstonia New Zealand TurkeyFinland Norway United KingdomFrance Philippines United StatesGermany Poland Venezuela

N = 45.

subjective assessments of variables, since “within the same survey responses” to different ques-tions may simply reflect some general underlying sentiment toward a country. When differentsurveys use different respondents, this potential risk is reduced. A brief discussion of the relevantvariables used now follows.

3.3. Dependent variable

The dependent variable in this study is represented by tax evasion (TEVA). Its measure isbased on a country survey rating of tax evasion collected by the World Economic Forum (WEF)and published in the Global Competitiveness Report (WEF, 2002, 2003, 2004). While the GlobalCompetitiveness Report is a valuable source of cross-country tax evasion data, using one questionin this study to measure tax evasion raises concerns about reliability due to measurement error.However, measurement error can be minimized by using average data for several years (Fisman& Gatti, 2002, p. 331; You & Khagram, 2005, p. 142).4 Therefore, averaged WEF tax evasiondata for several years (from 2002 to 2004) are used as the dependent variable instead of data fora single year.

3.4. Independent variables

The independent variables are represented in the study by age (AGE), gender (GEND), edu-cation (EDUC), income level (ILEVEL): low income level (LILEVEL) and high income level(HILEVEL), income source (ISOURCE): agriculture income source (AISOURCE) and servicesincome source (SISOURCE), marginal tax rates (MTR), fairness (FAIR), complexity (COMP),self-assessment (SELFA) and tax morale (MORALE). Where possible, data for these independentvariables are computed as 3-year averages, covering 2002–2004 so as to be consistent with themeasurement of the dependent variable, and to reduce the possibility of measurement error.

4 Assuming that measurement error has a normal distribution with a mean of zero and a variance of σ2, averaging of Nobservations will decrease the variance to σ2/N.


AGE (percentage of the population greater than 65)5 and GEND (percentage of the populationthat is female) are both measured from data collected by the World Bank and published in the2005 World Development Indicators (World Bank, 2005a). EDUC is measured on the basis of acountry survey rating of the quality of a country’s general education system. Data are gatheredby the Institute of Management Development (IMD) and published in the World CompetitivenessYear Book (IMD, 2002, 2003, 2004). LILEVEL is measured on the basis of the proportion ofhousehold income going to the lowest 20% of households, while HILEVEL is measured as theproportion of household income going to the highest 20% of households. Data for each of thesevariables are taken from the World Competitiveness Year Book (IMD, 2002, 2003, 2004). AISOUis measured on the basis of the percentage of employment in the agricultural sector, while SISOUis measured as the percentage of employment in the services sector. Data for these variablesare collected from the World Competitiveness Year Book (IMD, 2002, 2003, 2004). MTR ismeasured on the basis of the top marginal income tax rate for individuals. Data for this variableare taken from the 2005 World Development Indicators (World Bank, 2005a). FAIR is measuredin terms of a country survey rating of the fairness of tax policy. Data for this variable are gatheredby the Institute of Industrial Policy Studies (IPS) and published in the National CompetitivenessReport (IPS, 2002). COMP is measured in terms of a country survey rating of complexity inthe tax system. Data for this variable are collected from the Global Competitiveness Report(WEF, 2003, 2004). SELFA is measured by means of a dummy variable (1 if a country has aself-assessment tax system, 0 otherwise) based on information provided by the OECD (2004),PricewaterhouseCoopers (2004) and KPMG (2003). Finally, consistent with Torgler (2003a,2003b, 2005) and Torgler and Murphy (2004), MORALE is measured in terms of a countrysurvey rating of tax cheating. Data for this variable are gathered from the World Values Survey(Inglehart, 2003; Inglehart, Basanez, Diez-Medrano, Halman, & Luijkx, 2004).

3.5. Control variables

Since this study is undertaken at the country level of analysis, it is important to control forpotential cross-country effects. Therefore, several control variables relating to economic, politicaland cultural factors are included in the empirical analysis. The level of economic development(EDEV) can have a major impact on tax evasion across countries (Alm & Martinez-Vazquez,2003; Bird, 1992; de Soto, 2000; Quirk, 1997). Quirk (1997) asserts that countries in the earlystages of economic development are especially prone to tax evasion. Major studies of tax evasionundertaken in developing countries show that it is not uncommon for 50% or more of potentialincome tax to remain uncollected due to tax evasion (Bird, 1992; Gillis, 1989; Richupan, 1984).Das-Gupta, Lahiri, and Mookherjee (1995) observe that in India, the amount of income notsubject to taxation is estimated to be more than 200% of the assessed income. EDEV is measuredin this study based on the natural log of GDP per capita, which is collected from the 2005 WorldDevelopment Indicators (World Bank, 2005a).

Political institutions based on notions of democracy (DEMOC) can also influence the levelof tax evasion across countries (Alm, McClelland, & Schulze, 1999; Pommerehne & Weck-Hannemann, 1996). Specifically, because the taxpaying public is allowed to directly participatein the democratic political process through the right to vote on taxation issues, politicians become

5 The age of ‘65’ has been used as the cut-off point to represent older tax payers in tax evasion research by Clotfelter(1983) and Witte and Woodberry (1985), for example.


more accountable and transparent such that they must take taxpayer preferences into account.This improves taxpayer confidence and can reduce the level of tax evasion in a country (Feld& Tyran, 2002; Torgler, 2003b; Torgler, Schaltegger, & Schaffner, 2003). DEMOC is measuredbased on the political rights index developed by Freedom House (2005).

Culture (CULT) and religion (RELIG) can also influence tax evasion across countries. Forexample, survey research by Tittle (1980) in the U.S. finds that cultural and religious backgroundhas a relationship with the deviant propensity of taxpayers. Focus group research by Colemanand Freeman (1997) in Australia also shows that cultural and religious background affects taxcompliance. Finally, a cross-country survey study of tax evasion by Chan, Troutman, and O’Bryan(2000) in Hong Kong and the U.S. indicates that cultural background influences tax evasion. CULTis measured in this study on the basis of ethnolinguistic fractionalization, collected from Mauro(1995), while RELIG is measured by the percentages of Protestants (PROT),6 Catholics (CATH),Muslims (MUSL) and other denominations (OTHRD), gathered from La Porta et al. (1999).

It is also important in cross-country tax research (Brunetti & Weader, 2003; Treisman, 2000) toinclude additional controls for legal system (LEGAL), colonial heritage (COLONY) and regionaldeveloping countries (REGION). These additional controls consider whether tax evasion is drivenby differences between common law system countries versus code law system countries, colo-nial countries versus non-colonial countries and regional developing countries versus developedcountries. Dummy variables for LEGAL, COLONY and REGION are also included in the empir-ical analysis. LEGAL is measured in this study on the basis of the common law system countryclassification of La Porta et al. (1999). COLONY is measured by the colonial heritage countryclassification of Barro and Lee (1994). REGION is measured by the developing country group-ing classifications of: East Asia and Pacific region (EAPR), Europe and Central Asia region(EUCAR) and Latin America and the Caribbean region (LACR), gathered from the World BankGroup—Data and Statistics (World Bank, 2005b).

3.6. Base regression model

To investigate the determinants of tax evasion, the following base regression model is estimated:

TEVAi = α0 + β1 AGEi + β2 GENDi + β3 EDUCi + β4 LILEVELi + β5 HILEVELi

+ β6 AISOURCEi + β7 SISOURCEi + β8 MTRi + β9 FAIRi + β10 COMPi

+ β11 SELFAi + β12 MORALEi + εi (1)

where TEVAi is the tax evasion score for country i, AGEi the percentage of the population greaterthan 65 for country i, GENDi the percentage of the population that is female for country i, EDUCi

the general education score for country i, LILEVELi the proportion of household income goingto the lowest 20% of households for country i, HILEVELi the proportion of household incomegoing to the highest 20% of households for country i, AISOURCEi the percentage of employmentin the agricultural sector for country i, SISOURCEi the percentage of employment in the servicessector for country i, MTRi the top marginal income tax rate for individuals of country i, FAIRi

the fairness score for country i, COMPi the complexity score for country i, SELFAi the dummyvariable represented by 1 if country i has a self-assessment tax system, 0 otherwise, MORALEi

the tax morale score for country i and εi is the error term for country i.

6 Reference category only.


4. Empirical results and analysis

4.1. Descriptive statistics and correlations

Table 2 reports descriptive statistics for the variables used in this study from a cross-section of45 countries. Moreover, the Pearson pairwise correlation coefficients for this study’s dependentand independent variables are summarized in Table 3.

Table 3 shows that there are a number of significant correlations between tax evasion and theindependent variables. For example, there are fairly high correlations (p < .01) between TEVA andFAIR (r = −.72), TEVA and COMP (r = .68), TEVA and SISOURCE (r = −.61), TEVA and EDUC(r = −.53) and TEVA and AISOURCE (r = .51). Correlations are also found (p < .05) between,TEVA and AGE (r = −.29), TEVA and SELFA (r = −.27) and TEVA and MORALE (−.24).

Table 2Descriptive statisticsa

Variable Number Mean S.D. Minimum Maximum

TEVA 45 4.33 1.13 1.80 6.10AGE 45 11.48 4.58 3.93 18.37GEND 45 .51 .01 .48 .53EDUC 45 5.20 1.50 2.30 8.41ILEVEL: LILEVEL 45 6.84 2.23 2.33 10.87ILEVEL: HILEVEL 45 43.92 8.26 32.53 65.37ISOURCE: AISOURCE 45 11.94 14.31 .30 62.45ISOURCE: SISOURCE 45 59.62 14.99 23.10 84.27MTR 45 .41 .10 .13 .59FAIR 45 5.43 1.56 1.30 8.00COMP 45 4.72 1.11 1.55 6.35SELFA 45 .47 .50 0 1MORALE 45 8.31 .88 4.82 9.82EDEV 45 9.19 1.20 6.14 10.80DEMOC 45 1.82 1.43 1.00 7.00CULT 44 .21 .22 .00 .83RELIG: PROTb 45 19.11 29.24 0 97.80RELIG: CATH 45 40.44 39.24 0 96.90RELIG: MUSL 45 5.43 16.71 0 99.20RELIG: OTHRD 45 35.02 34.13 .70 98.50LEGAL 45 .27 .45 0 1COLONY 45 .24 .43 0 1REGION: EAPR 45 .10 .31 0 1REGION: EUCAR 45 .12 .33 0 1REGION: LACR 45 .12 .33 0 1

a Variable definitions: TEVA, tax evasion; AGE, age; GEND, gender; EDUC, education; ILEVEL, income level(LILEVEL, low income level; HILEVEL, high income level); ISOURCE, income source (AISOURCE, agricultureincome source; SISOURCE, services income source); MTR, marginal tax rates; FAIR, fairness; COMP, complexity;SELFA, revenue authority initiated contact; MORALE, morale; EDEV, economic development; DEMOC, democracy;CULT, culture; RELIG, religion (PROT, Protestant; CATH, Catholic; MUSL, Muslim; OTHRD, other denominations);LEGAL, legal system; COLONY, colonial heritage; REGION, regional developing countries (EAPR, East Asia and Pacificregion; EUCAR, Europe and Central Asia region; LACR, Latin America and Caribbean region).

b Reference category only.

G.R

ichardson/JournalofInternationalA

ccounting,Auditing

andTaxation

15(2006)

150–169159

Table 3Pearson correlations for dependent and independent variables

1 2 3 4 5 6 7 8 9 10 11 12 13

1. TEVA 12. AGE −.29** 13. GEND .15 .44*** 14. EDUC −.53*** .26** −.16 15. LILEVEL .01 .62*** .07 .27** 16. HILEVEL .16 −.72*** −.06 −.38*** −.82*** 17. AISOURCE .51*** −.60*** −.38*** −.33*** −.15 .30** 18. SISOURCE −.61*** .49*** .07 .34*** .15 −.36*** −.73*** 19. MTR −.16 .51*** −.08 .23* .52*** −.58*** −.29** .26** 1

10. FAIR −.72*** .45*** −.03 .55*** .21* −.34** −.54*** .49*** .33** 111. COMP .68*** .08 .21* −.38*** .16 −.10 .13 −.20* .26** −.49*** 112. SELFA −.27** −.28** −.06 −.16 −.29** .33** .18* −.24** −.22* −.25** .10 113. MORALE −.24** −.08 −.38*** −.06 .22** −.31** .01 .18 .34*** .06 −.04 .03 1

See Table 1 for variable definitions. N = 45 for all variables.* Significant at .10 level.

** Significant at .05 level.*** Significant at .01 level.


However, insignificant correlations are found between TEVA and GEND, TEVA and LILEVEL,TEVA and HILEVEL and TEVA and MTR.7

These univariate results provide some preliminary support for H1, H3, H5, H7, H8, H9 andH10 of this study. In addition, these results show that behavioral and demographic variables havethe strongest influence on tax evasion as compared to economic variables. This represents aninteresting empirical finding which indicates that non-economic variables are fundamental andshould be investigated along with economic variables in ‘mixed models’ of tax evasion acrosscountries.

4.2. OLS regressions

Table 4 summarizes the results of the OLS regression analysis for the base OLS regressionmodel (Column 1), and includes a number of control variables (Columns 2–8) to consider potentialcross-country effects.8

Table 4 (Column 1) shows that the base OLS regression model is significant at the p < .01 level(F statistic = 13.20), while the adjusted R2 for this regression model is .80. Concerning the signifi-cance of the regression coefficients for the independent variables summarized in Table 4 (Column1), the results show that COMP is the most important determinant of tax evasion across countries.The association between COMP and TEVA is positive and significant (p < .01), therefore, H8 issupported by the results. Where a country’s tax system is highly complex, this can increase theincidence of tax evasion.

EDUC, SISOURCE, FAIR and MORALE represent the next most important determinants oftax evasion across countries reported in Table 4 (Column 1). Specifically for EDUC, its asso-ciation with TEVA is negative and significant (p < .05), consequently, H3 is supported by theresults. Where the general education level of taxpayers in a country is high, tax evasion canbe reduced. Concerning ISOURCE, only the SISOURCE association with TEVA is significant(p < .05), accordingly, the results partially confirm H5. Where a country has income that is subjectto withholding (e.g., services employment income), this can decrease the level of tax evasion.Concerning FAIR, its association with TEVA is negative and significant (p < .05), so H7 is con-firmed by the results. Where taxpayers perceive that their country’s tax policy is fair, this canreduce the level of tax evasion. For MORALE, its association with TEVA is negative and signif-icant (p < .05), hence, H10 is also corroborated by the results. Where tax morale in a country ishigh, it can decrease the level of tax evasion. Finally, for AGE, GEND, LILEVEL, HILEVEL,MTR and SELFA, no significant associations are found with tax evasion. Therefore, H1, H2, H4,H6 and H9 are not supported by the multivariate results.

Consistent with the univariate findings reported in Table 3, the behavioral and demographicvariables have the strongest impact on tax evasion in the base OLS regression model. These results

7 Significant correlations are also found between tax evasion and some of the control variables. For instance, there arereasonably high correlations (p < .01) between TEVA and EDEV (r = −.63), TEVA and COLONY (r = −.48) and TEVAand LEGAL (r = −.44). Correlations are also observed (p < .05) between TEVA and PROT (r = −.29), TEVA and EUCAR(r = .29) and TEVA and LACR (r = .28). Finally, some marginal correlations are also detected (p < .10) between TEVAand EAPR (r = .22) and TEVA and CATH (r = .21). No significant correlations are found between TEVA and DEMOC,CULT, MUSL or OTHRD.

8 The t-statistics in Table 4 are shown in parentheses, and significance of the estimates is based on White-correctedstandard errors. Moreover, the variance inflation factors (VIFs) for all of this study’s regression models indicate that noneof the independent and control variables have a multicollinearity problem. Specifically, none of the VIFs exceed 4, whichare below the commonly accepted VIF cut-off threshold of 10 (Hair, Anderson, Tatham, & Black, 1998, p. 193).

G.R


ccounting,Auditing

andTaxation

15(2006)

150–169161

Table 4OLS regression results (dependent variable: tax evasion, WEF)

OLS (1) OLS (2) OLS (3) OLS (4) OLS (5) OLS (6) OLS (7) OLS (8)

CONSTANT 10.263 (1.185) 13.321 (1.379) 8.833 (1.001) 13.037 (1.390) 14.063 (1.433) 14.995 (1.664) 14.995 (1.664) 13.482 (1.395)AGE −.256 (−1.331) −.201 (−.967) −.270 (−1.397) −.313 (−1.500)* −.181 (−.795) −.237 (−1.263) −.237 (−1.263) −.185 (−.877)GEND −.008 (−.065) −.031 (−.234) −.010 (−.076) −.024 (−.179) −.071 (−.488) −.075 (−.562) −.075 (−.562) −.056 (−.388)EDUC −.205 (−1.877)** −.192 (−1.723)** −.214 (−1.945)** −.206 (−1.721)** −.217 (−1.967)** −.142 (−1.248)* −.142 (−1.248)* −.190 (−1.637)**

LILEVEL .047 (.212) .009 (.039) .078 (.349) .025 (.100) .106 (.460) .041 (.185) .041 (.185) .018 (.075)HIILEVEL −.248 (−.859) −.292 (−.982) −.242 (−.837) −.295 (−.952) −.229 (−.785) −.305 (−1.074) −.305 (−1.074) −.284 (−.831)AISOURCE .012 (.079) .112 (.549) .042 (.269) .004 (.022) .014 (.090) .074 (.480) .074 (.480) .069 (.396)SISOURCE −.301 (−2.429)** −.265 (−1.978)** −.285 (−2.372)** −.351 (−2.435)** −.319 (−2.318)** −.337 (−2.732)** −.337 (−2.732)** −.330 (−2.343)**

MTR −.047 (−.416) −.023 (−.198) −.008 (−.069) −.017 (−.132) −.114 (−.744) −.078 (−.698) −.078 (−.698) −.026 (−.194)FAIR −.212 (−1.820)** −.202 (−1.708)** −.203 (−1.736)** −.235 (−1.876)** −.246 (−2.035)** −.219 (−1.924)** −.219 (−1.924)** −.218 (−1.772)**

COMP .438 (4.057)*** .438 (4.017)*** .442 (4.078)*** .405 (3.363)*** .416 (3.751)*** .427 (4.046)*** .427 (4.046)*** .418 (3.344)***

SELFA −.106 (−1.286) −.108 (−1.294) −.100 (−1.198) −.093 (−1.044) −.079 (−.857) −.042 (−.466) −.042 (−.466) −.096 (−1.102)MORALE −.245 (−2.318)** −.241 (−2.259)** −.269 (−2.424)** −.251 (−2.378)** −.222 (−1.861)** −.228 (−2.199)** −.228 (−2.199)** −.258 (−2.316)**

EDEV −.203 (−.744)DEMOC −.097 (−.925)CULT −.134 (−1.146)

RELIGPROTa

CATH .077 (.521)MUSL −.024 (−.208)OTHRD −.085 (−.481)

LEGAL −.162 (−1.509)COLONY −.162 (−1.509)

REGIONEAPR .088 (.761)EUCAR .077 (.700)LACR .103 (.747)

N 45 45 44 45 45 45 45 45R2 (adjusted) .80 .80 .80 .78 .80 .81 .81 .78F statistic 13.20*** 12.01*** 12.18*** 10.74*** 10.60*** 12.98*** 12.98*** 9.84***

See Table 1 for variable definitions. t-Statistics are in parentheses. Standard errors are corrected for heteroschedasticity.a Reference category only.* Significant at .10 level.



show that by incorporating not only economic variables, but also behavioral and demographicvariables into mixed models of tax evasion, more compelling results are found. The findingssupport the views of Cuccia (1994) and Cummings, Martinez-Vazquez, and McKee (2001) whoargue that by combining economic and non-economic perspectives of tax evasion, leads to a betterunderstanding of the subject.

Table 4 (Columns 2–8) summarizes the results of supplementary regression models, whichincorporate several cross-country control variables. The results show that these models are allsignificant at the p < .01 level (F statistics ranging from 9.84 to 12.98), while the adjusted R2’s forthese regression models (i.e. approximately .80) are quite consistent with the adjusted R2 for thebase OLS regression model (Column 1). Therefore, the explanatory power of the base regressionmodel is not improved by the inclusion of cross-country control variables.

In the supplementary regression models, COMP (p < .01) remains the most important determi-nant of tax evasion, and its regression coefficient is relatively stable across all of the supplementaryregression models. SISOURCE, FAIR, MORALE and EDUC remain the next most significantdeterminants of tax evasion and have reasonably stable regression coefficients (p < .05; EDUCfalls to p < .10 in some regressions). Non-economic variables are again shown to have the strongestimpact on tax evasion compared to economic variables across the supplementary regressions.

Concerning the significance of the regression coefficients of the cross-country control variablesin Table 4 (Columns 2–8), none of the control variables pertaining to EDEV, DEMOC, CULT,RELIG, LEGAL, COLONY and REGION are found to be significantly related to tax evasion.These results demonstrate that the base OLS regression model results in Table 4 (Column 1)described above remain robust and are not driven by cross-country differences.

Overall, the supplementary regression model results are comparable to the base OLS regressionmodel results presented earlier. Accordingly, this study’s findings, after controlling for a broadrange of cross-country variables, continue to be robust.

4.3. Sensitivity analysis

The regression analysis thus far has made use of the WEF’s measure of tax evasion (WEF, 2002,2003, 2004). This analysis raises the question of whether the results only reflect characteristics ofthe underlying data. For this reason, the robustness of the results is tested by using an alterativemeasure of tax evasion based on a country survey rating collected by the IMD (2002, 2003, 2004).Data for this measure of tax evasion are computed as 3-year averages, covering the 2002–2004years, to reduce the possibility of measurement error. The IMD tax evasion measure has a highcorrelation (r = .95; p < .01) with its WEF complement, suggesting that this particular measure oftax evasion is relatively sound. The results of the sensitivity analysis are reported in Table 5.

The results of the sensitivity analysis using the IMD measure of tax evasion show that the regres-sion coefficients of COMP, EDUC, SISOURCE, FAIR and MORALE are significant and remainquite stable across the different regression model specifications reported in Table 5 (Columns1–8). This finding is consistent with the earlier regression results using the WEF measure oftax evasion. However, the level of significance of some of these independent variables changesslightly when using the IMD tax evasion measure.

COMP remains the most significant determinant of tax evasion across countries (p < .01).EDUC (p < .05 or greater) and MORALE (usually p < .05), SISOURCE and FAIR also have signif-icant relationships with tax evasion (p < .10 or greater). For AGE, GEND, LILEVEL, HILEVEL,AISOURCE, MTR and SELFA, no significant association is found with tax evasion. Again, non-economic variables have the strongest impact on tax evasion. Finally, none of the cross-country

G.R


ccounting,Auditing

andTaxation

15(2006)

150–169163

Table 5OLS regression results (dependent variable: tax evasion, IMD)

OLS (1) OLS (2) OLS (3) OLS (4) OLS (5) OLS (6) OLS (7) OLS (8)

CONSTANT 15.354 (1.108) 17.775 (1.140) 15.039 (1.047) 18.546 (1.287) 19.966 (1.373) 18.511 (1.236) 18.511 (1.236) 20.949 (1.366)AGE −.154 (−.737) −.125 (−.547) −.157 (−.730) −.203 (−.944) −.122 (−.531) −.146 (−.687) −.146 (−.687) −.116 (−.510)GEND −.002 (−.011) −.011 (−.073) −.004 (−.030) −.013 (−.097) −.059 (−.397) −.029 (−.190) −.029 (−.190) −.029 (−.185)EDUC −.309 (−2.593)** −.302 (−2.461)** −.310 (−2.542)** −.290 (−2.345)** −.322 (−2.881)*** −.280 (−2.167)** −.280 (−2.167)** −.321 (−2.558)***

LILEVEL .184 (.763) .205 (.813) .180 (.721) .217 (.855) .065 (.281) .225 (.886) .225 (.886) .255 (.996)HIILEVEL −.328 (−1.159) −.352 (−1.181) −.327 (−1.129) −.309 (−1.193) −.301 (−1.123) −.354 (−1.110) −.354 (−1.110) −.386 (−1.188)AISOURCE .165 (.987) .110 (.489) .160 (.921) .218 (1.204) .165 (1.034) .136 (.778) .136 (.778) .181 (.960)SISOURCE −.212 (−1.566)* −.193 (−1.304)* −.211 (−1.527)* −.242 (−1.632)* −.195 (−1.401)* −.228 (−1.633)* −.228 (−1.633)* −.266 (−1.744)**

MTR −.085 (−.691) −.072 (−.556) −.079 (−.592) −.102 (−.749) −.114 (−.735) −.099 (−.782) −.099 (−.782) −.050 (−.346)FAIR −.186 (−1.461)* −.180 (−1.385)* −.185 (−1.418)* −.199 (−1.538)* −.222 (−1.817)** −.189 (−1.466)* −.189 (−1.466)* −.209 (−1.570)*

COMP .375 (3.180)*** .375 (3.123)*** .375 (3.119)*** .345 (2.786)*** .331 (2.954)*** .370 (3.091)*** .370 (3.091)*** .303 (2.242)**

SELFA −.109 (−1.210) −.110 (−1.198) −.108 (−1.171) −.071 (−.781) −.089 (−.946) −.080 (−.776) −.080 (−.776) −.109 (−1.160)MORALE −.199 (−1.723)** −.197 (−1.673)** −.201 (−1.689)** −.269 (−2.479)** −.196 (−1.628)** −.191 (−1.625)* −.191 (−1.625)* −.224 (−1.852)**

EDEV −.110 (−.365)DEMOC −.015 (−.125)CULT −.159 (−1.328)

RELIGPROTa

CATH .127 (.755)MUSL .010 (.081)OTHRD .002 (.009)

LEGAL −.074 (−.606)COLONY −.074 (−.606)

REGIONEAPR −.036 (−.290)EUCAR .064 (.536)LACR .146 (.979)

N 45 45 45 44 45 45 45 45R2 (adjusted) .76 .75 .75 .77 .79 .75 .75 .74F statistic 10.75*** 9.59*** 9.54*** 10.03*** 10.36*** 9.70*** 9.70*** 8.19***

See Table 1 for variable definitions. t-Statistics are in parentheses. Standard errors are corrected for heteroschedasticity.a Reference category only.* Significant at .10 level.



control variables of EDEV, DEMOC, CULT, RELIG, LEGAL, COLONY and REGION have asignificant relationship with tax evasion.

On the whole, the results of the sensitivity analysis indicate that the associations identified inthe earlier regressions are robust to an alternative measure of tax evasion.

4.4. Additional analyses

A review of the tax evasion literature by Jackson and Milliron (1986) and Richardson andSawyer (2001) suggest that major interactions between the key tax evasion determinants shouldalso be considered. They argue that a potential reason for some of the inconsistent findings inprevious tax evasion studies is that researchers are not taking into account interactions betweenthe key tax evasion determinants.

Table 3 above reports several interesting correlations between a number of the key tax evasiondeterminants. For instance, correlations (p < .01) are found between AGE and GEND (r = .44),AGE and ILEVEL (AGE and LILEVEL, r = .62; AGE and HILEVEL, r = −.72), AGE andISOURCE (AGE and AISOURCE, r = −.60; AGE and SISOURCE, r = .49), AGE and MTR(r = .51), AGE and FAIR (r = .45), EDUC and FAIR (r = .55), ILEVEL and MTR (LILEVEL andMTR, r = .52; HILEVEL and MTR, r = −.58), ISOURCE and FAIR (AISOURCE and FAIR,r = −.54; SISOURCE and FAIR, r = .49) and FAIR and COMP (r = −.49).

Accordingly, the study explored interaction terms for AGE*GEND, AGE*ILEVEL,AGE*ISOURCE, AGE*MTR, AGE*FAIR, EDUC*FAIR, ILEVEL*MTR, ISOURCE*FAIR andFAIR*COMP in separate regression models. The additional analyses considered whether theseinteractions are significant and/or change the base regression model findings summarized inTable 4 (Column 1). The regression results indicate that none of these interactions are signifi-cant predictors of tax evasion (p < .10).9 Moreover, the base regression model findings remainunchanged (with no changes in sign) after the inclusion of these interactions. The additional anal-yses demonstrate that the associations identified in the earlier regressions are robust to interactionsbetween the independent variables.

5. Conclusions, limitations and future research

While tax evasion has long been a popular academic research topic in most developed countries,there has not been detailed consideration of the major determinants of tax evasion on a cross-country basis. This study expands Riahi-Belkaoui’s (2004) pioneering work of the relationshipbetween selected determinants of tax morale and tax evasion and systematically investigates manyof the key determinants of tax evasion on a cross-country basis.

OLS regression analysis shows that non-economic determinants have the strongest impact ontax evasion in comparison with economic determinants. By integrating these various determinantsinto mixed models of tax evasion, our understanding is enhanced about tax evasion across coun-tries. Complexity is found to be the most important determinant of tax evasion. Other significantdeterminants are education, income source, fairness and tax morale. The results of the regressionsshow that the lower the level of complexity and the higher the level of general education, servicesincome source, fairness and tax morale, the lower is the level of tax evasion across countries.These results remain robust to different cross-country control variables, an alternative measure oftax evasion and several interactions.

9 These additional regression results are available from the author upon request.


The results have implications for governments that seek to reduce the level of tax evasionin society. A more simple tax system can reduce tax evasion. Moreover, general education isnegatively associated with tax evasion. Wage and salary income subject to withholding (e.g.,services employment income) is another important curb on tax evasion. In addition, perceptionsthat tax policy is fair are associated with reduced levels of tax evasion. Finally, where tax morale ishigh, lower levels of tax evasion can be expected. These specific insights should allow governmentpolicy-makers to gain a better understanding of the key variables that are significantly associatedwith tax evasion internationally, and design and implement appropriate strategies to minimize itsdamaging effects. This should lead to improvements in tax revenue collection by governments.

This study is subject to several limitations. First, because of data unavailability, other potentialtax evasion determinants such as occupation status, sanctions, probability of detection and com-pliant peers are not included, so the results might reflect omitted variable bias. Second, the samplesize of 45 countries is relatively small, which means that the findings may not be generalizable.However, this is a common problem of cross-country research. Third, tax evasion is measuredusing subjective survey ratings, which raises concerns about its reliability because it could beprone to measurement error. Data are averaged over several years to minimize the possibility ofmeasurement error. Fourth, using survey data measures for some of the independent and controlvariables raises additional concerns about measurement error. Data are collected from reputablesources and averaged data measures are used for the independent and control variables.

Future international research on tax evasion might consider four matters. First, variables relat-ing to sanctions, probability of detection and compliant peers might be analyzed, subject to theavailability of reliable cross-country data. Second, increased country sample sizes might enhancecross-country comparisons and the generalizability of findings. Third, improved survey measuresof tax evasion and various explanatory variables (e.g., complexity, education and fairness) couldimprove the reliability of the empirical results and reduce further the risk of measurement error.Finally, a greater longitudinal emphasis could be undertaken to examine the impact of changes inthe key determinants and other important variables on changes in tax evasion levels.

Appendix A. Data description and sources

Variable Description Source

Tax evasion (TEVA) • Country rating that tax evasion is minimal (on ascale from 1—strongly disagree to 7—stronglyagree) averaged for 2002–2004. This variable wastransformed to obtain an increasing scale of taxevasion.

Global Competitiveness Report(WEF, 2002, 2003, 2004).

• Country rating of tax evasion (on a scale from0—common to 10—not common) averaged for2002–2004. This variable was transformed to obtainan increasing scale of tax evasion.

World Competitiveness YearBook (IMD, 2002, 2003, 2004).

Age (AGE) Percentage of the population which is greater than65 years of age averaged for 2002–2004.

2005 World DevelopmentIndicators (World Bank, 2005a).

Gender (GEND) Percentage of the population which is femaleaveraged for 2002–2004.


Education (EDUC) Country rating of the quality of the generaleducation system for a competitive economy (on ascale from 1—low to 7—high) averaged for2002–2004.



Appendix A (Continued )


Income level (ILEVEL) • Proportion of household income going to thelowest 20% of households (LILEVEL) averaged for2002–2004.


• Proportion of household income going to thehighest 20% of households (HILEVEL) averagedfor 2002–2004.

Income source(ISOURCE)

• Employment by sector: agriculture percentage oftotal employment (AISOURCE) averaged for2002–2004.


• Employment by sector: services percentage of totalemployment (SISOURCE) averaged for 2002–2004.

Marginal tax rates (MTR) The top marginal income tax rate for individuals of acountry averaged for 2002–2004.


Fairness (FAIR) Country rating of the fairness of tax policy (on ascale from 0—low fairness to 8—high fairness) forthe 2002 year.

IPS National CompetitivenessReport (IPS, 2002).

Complexity (COMP) Country rating of tax system complexity (on a scalefrom 1—high complexity to 7—low complexity)averaged for 2003–2004. This variable wastransformed to obtain an increasing scale of taxsystem complexity.

Global Competitiveness Report(WEF, 2003, 2004).

Revenue authorityinitiated contact(SELFA)

Dummy variable of 1 if the country has aself-assessment tax system, 0 otherwise.

OECD (2004).http://www.oecd.org/dataoecd/28/2/33866659.pdf.PricewaterhouseCoopers (2004).KPMG (2003).http://www.kpmg.com.hk.

Tax morale (MORALE) Country rating of cheating on taxes if you have thechance (on a scale from 1—never justifiable to10—always justifiable) averaged for the 1981, 1990,1995 and 1999 World Value Surveys. This variablewas transformed to obtain an increasing scale of taxmorale.

Inglehart (2003).http://nds.umdl.umich.edu/cgi/s/sda/hsda?harcWEVS+wevs.

Inglehart et al. (2004).

Economic development(EDEV)

Natural log of GDP per capita averaged for2001–2003.


Democracy (DEMOC) Political rights index (on a scale from 1—highpolitical rights to 7—low political rights), averagedfor 2002–2004. This index was transformed toobtain an increasing scale of democracy.

Freedom House (2005).http://www.freedomhouse.org/ratings/index.htm.

Culture (CULT) Ethnolinguistic fractionalization index measures theprobability that two randomly selected individualswithin a country belong to the same ethnic group. Itis an index between 0 and 100, with 100 denotinglower fractionalization.

Mauro (1995).

Religion (RELIG) The percentages of Protestants (PROT), Catholics(CATH), Muslims (MUSL) and otherdenominations (OTHRD) in 1980 or 1990–1995 forcountries of recent formation.

La Porta et al. (1999).

Legal system (LEGAL) Dummy variable of 1 if the country is a commonlaw system country, 0 otherwise.

La Porta et al. (1999).

http://www.oecd.org/dataoecd/28/2/33866659.pdf


http://www.kpmg.com.hk/

http://nds.umdl.umich.edu/cgi/s/sda/hsda?harcWEVS+wevs


http://www.freedomhouse.org/ratings/index.htm



Appendix A (Continued )


Colonial heritage(COLONY)

Dummy variable of 1 if the country was ever acolony, 0 otherwise.

Barro and Lee (1994).

Regional developingcountries (REGION)

• Dummy variable of 1 if the developing country isin the East Asia and Pacific region (EAPR), 0otherwise.

World Bank Group—Data andStatistics (World Bank, 2005b).http://www.worldbank.org/data/countryclass/classgroups.htm.• Dummy variable of 1 if the developing country is

in the Europe and Central Asia region (EUCAR), 0otherwise.• Dummy variable of 1 if the developing country isin the Latin America and the Caribbean region(LACR), 0 otherwise.

References

Alm, J., & Martinez-Vazquez, J. (2003). Institutions, paradigms, and tax evasion in developing and transition countries.In J. Martinez-Vazquez, & J. Alm (Eds.), Public finance in developing and transitional countries: Essays in honor ofRichard Bird (pp. 146–178). Cheltenham, UK: Edward Elgar.

Alm, J., McClelland, G. H., & Schulze, W. D. (1999). Changing the social norm of tax compliance by voting. KYKLOS,48, 141–171.

Andreoni, J., Erard, B., & Feinstein, J. (1998). Tax compliance. Journal of Economic Literature, 36, 818–860.Barro, R., & Lee, J. (1994). Sources of economic growth. Carnegie Rochester Conference Series on Public Policy, 40,

1–46.Bird, R. (1992). Tax policy and economic development. Baltimore: The Johns Hopkins University Press.Brooks, N., & Doob, A. H. (1990). Tax evasion: Searching for a theory of compliant behavior. In M. L. Friedland (Ed.),

Securing compliance: Seven case studies (pp. 120–164). Toronto: University of Toronto Press.Brunetti, A., & Weader, B. (2003). A free press is bad news for corruption. Journal of Public Economics, 87, 1801–1824.Chan, C. W., Troutman, C. S., & O’Bryan, D. (2000). An expanded model of taxpayer compliance: Empirical evidence

from the United States and Hong Kong. Journal of International Accounting, Auditing and Taxation, 9, 83–103.Christian, C. W., & Gupta, S. (1993). New evidence on ‘secondary evasion’. Journal of the American Taxation Association,

16, 72–93.Clotfelter, C. T. (1983). Tax evasion and tax rates: An analysis of individual returns. The Review of Economics and

Statistics, 65, 363–373.Coleman, C., & Freeman, L. (1997). Cultural foundations of taxpayer attitudes to voluntary compliance. Australian Tax

Forum, 13, 311–336.Collins, J. H., Milliron, V. C., & Toy, D. R. (1992). Determinants of tax compliance: A contingency approach. Journal of

the American Taxation Association, 14, 1–29.Cuccia, A. D. (1994). The economics of tax compliance: What do we know and where do we go? Journal of Accounting

Literature, 13, 81–116.Cummings, R. G., Martinez-Vazquez, J., & McKee, M. (2001). Cross cultural comparisons of tax compliance behavior,

Working Paper No. 01-3, International Studies Program. Atlanta, Georgia: Andrew Young School of Policy Studies,Georgia State University.

Das-Gupta, A., Lahiri, R., & Mookherjee, D. (1995). Income tax compliance in India: An empirical analysis. WorldDevelopment, 23, 2051–2064.

de Soto, H. (2000). The mystery of capital: Why capitalism triumphs in the west and fails everywhere else. New York:Basic Books.

Dubin, J. A., & Wilde, L. L. (1988). An empirical analysis of federal income tax auditing and compliance. National TaxJournal, 41, 61–74.

Feinstein, J. S. (1991). An econometric analysis of income tax evasion and its detection. RAND Journal of Economics,22, 14–35.

Feld, L. P., & Tyran, J. R. (2002). Tax evasion and voting: An experimental analysis. KYKLOS, 55, 197–222.

http://www.worldbank.org/data/countryclass/classgroups.htm



Fisman, R., & Gatti, R. (2002). Decentralization and corruption: Evidence across countries. Journal of Public Economics,83, 325–345.

Freedom House. (2005). Freedom House Country Ratings. New York: Freedom House (http://www.freedomhouse.org/ratings/index.htm).

Gillis, M. (1989). Tax reform in developing countries. Durham, NC: Duke University Press.Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis (5th ed.). Upper Saddle

River, NJ: Prentice-Hall.Hanno, D. M., & Violette, G. R. (1996). An analysis of moral and social influences on tax behavior. Behavioral Research

in Accounting, 8, 57–75.Hite, A., & Roberts, M. L. (1992). An analysis of tax reform based on taxpayers’ perceptions of fairness and self-interest.

Advances in Taxation, 4, 115–137.Inglehart, R. (2003). Human values and social change: Findings from the values surveys. Leiden, Boston: Brill

(http://nds.umdl.umich.edu/cgi/s/sda/hsda?harcWEVS+wevs).Inglehart, R., Basanez, M., Diez-Medrano, J., Halman, L., & Luijkx, R. (2004). Human beliefs and values: A cross-cultural

sourcebook based on the 1999–2002 value surveys. Mexico: Siglo XXI.Institute for Industrial Policy Studies. (2002). IPS national competitiveness report. Seoul, South Korea: IPS.Institute of Management Development. (2002). The world competitiveness yearbook. Lausanne, Switzerland: IMD.Institute of Management Development. (2003). The world competitiveness yearbook. Lausanne, Switzerland: IMD.Institute of Management Development. (2004). The world competitiveness yearbook. Lausanne, Switzerland: IMD.Jackson, B. R., & Milliron, V. C. (1986). Tax compliance research: Findings, problems and prospects. Journal of Accounting

Literature, 5, 125–165.Kinsey, K. (1986). Theories and models of tax cheating. Criminal Justice Abstracts, 2, 403–425.Klepper, S., & Nagin, D. (1989a). The anatomy of tax evasion. Journal of Law, Economics and Organization, 5, 1–24.Klepper, S., & Nagin, D. (1989b). The role of tax preparers in tax compliance. Policy Sciences, 22, 167–194.KPMG. (2003). Asia pacific taxation. Hong Kong: KPMG (http://www.kpmg.com.hk).La Porta, R., Lopez-de-Silanes, F., Shleifer, A., & Vishny, R. (1999). The quality of government. The Journal of Law,

Economics and Organization, 15, 222–279.Long, S., & Swingen, J. (1988). The role of legal complexity in shaping taxpayer compliance. In P. J. Van Koppen, P. J. D.

Hessing, & G. Van den Heuvel (Eds.), Lawyers on psychology and psychologists on law (pp. 127–146). Amsterdam,The Netherlands: Swets and Zeitlinger.

Long, S., & Swingen, J. (1991). Taxpayer compliance: Setting new agendas for research. Law and Society Review, 25,637–683.

Mason, R., & Calvin, L. (1978). A study of admitted income tax evasion. Law and Society Review, 12, 73–89.Mason, R., & Calvin, L. (1984). Public confidence and admitted tax evasion. National Tax Journal, 37, 489–496.Mason, R., & Lowry, H. (1981). An estimate of income tax evasion in Oregon, Working Paper. Corvallis, Oregon: Survey

Research Center, Oregon State University.Mauro, P. (1995). Corruption and growth. The Quarterly Journal of Economics, 110, 681–712.Milliron, V. C., & Toy, D. R. (1988). Tax compliance: An investigation of key features. Journal of the American Tax

Association, 10, 84–104.Organization for Economic Co-operation and Development. (2004). Tax Administration in OECD Countries. Paris: OECD

(http://www.oecd.org/dataoecd/28/2/33866659.pdf).Pommerehne, W. W., & Weck-Hannemann, H. (1996). Tax rates, tax administration and income tax evasion in Switzerland.

Public Choice, 88, 161–170.PricewaterhouseCoopers. (2004). Individual taxes 2004–2005: Worldwide summaries. New York: John Wiley.Quirk, P. J. (1997). Money laundering: Muddying the macro economy. Finance & Development, 34, 7–9.Riahi-Belkaoiu, A. (2004). Relationship between tax compliance internationally and selected determinants of tax morale.

Journal of International Accounting, Auditing and Taxation, 13, 135–143.Richardson, M., & Sawyer, A. J. (2001). A taxonomy of the tax compliance literature: Further findings, problems and

prospects. Australian Tax Forum, 16, 137–320.Richupan, S. (1984). Income tax evasion: A review of the measurement of techniques and some estimates for the developing

countries, Department Memorandum No. DM/84/46, The International Monetary Fund. Washington, DC: FiscalAffairs Department.

Sarker, T. K. (2003). Improving tax compliance in developing countries via self-assessment systems—What couldBangladesh learn from Japan? Asia-Pacific Tax Bulletin, 9, 170–176.

Schmolders, G. (1970). Survey research in public finance: A behavioral approach to fiscal theory. Public Finance, 25,300–306.




http://www.kpmg.com.hk/



Song, Y. D., & Yarbrough, T. E. (1978). Tax ethics and tax attitudes: A survey. Public Administration Review, 38, 442–452.Spicer, M. W. (1974). A behavioral model of income tax evasion. Unpublished Ph.D. Thesis. Ohio State University.Spicer, M. W., & Hero, R. (1985). Tax evasion and heuristics: A research note. Journal of Public Economics, 26, 263–267.Spicer, M. W., & Lundstedt, S. B. (1976). Audit probabilities and the tax evasion decision: An experimental approach.

Journal of Economic Psychology, 2, 241–245.Strumpel, B. (1969). The contribution of survey research to public finance. In A. T. Peacock (Ed.), Quantitative analysis

in public finance (pp. 13–38). New York: Praeger.Tan, L. M., & Sawyer, A. J. (2003). A synopsis of taxpayer compliance studies: Overseas vis-a-vis New Zealand. New

Zealand Journal of Taxation Law and Policy, 9, 431–454.Tanzi, V. (2000). Policies, institutions and the dark side of economics. Cheltenham, UK: Edward Elgar.Tittle, C. (1980). Sanctions and social deviance: The question of deterrence. New York: Praeger.Torgler, B. (2003a). Tax morale and tax evasion: Evidence from the United States, WWZ Discussion Paper 02/-7. Basel,

Switzerland.Torgler, B. (2003b). Tax morale and institutions, Working Paper No. 2003-09. Basel, Switzerland: Center for Research

in Economics, Management and the Arts.Torgler, B. (2005). Tax morale in Latin America. Public Choice, 122, 133–157.Torgler, B., & Murphy, K. (2004). Tax morale in Australia: What shapes it and has it changed over time? Journal of

Australian Taxation, 7, 298–335.Torgler, B., Schaltegger, C. A., & Schaffner, M. (2003). Is forgiveness divine? A cross-culture comparison of tax amnesties.

Swiss Journal of Economics and Statistics, 139, 375–396.Treisman, D. (2000). The causes of corruption: A cross-national study. Journal of Public Economics, 76, 399–457.Vogel, J. (1974). Taxation and public opinion in Sweden: An interpretation of recent survey data. National Tax Journal,

27, 499–513.Wallschutzky, I. G. (1984). Possible causes of tax evasion. Journal of Economic Psychology, 5, 371–384.Witte, A. D., & Woodbury, D. F. (1983). The effect of tax laws and tax administration on tax compliance, Working Paper

83-1. Chapel Hill, North Carolina: Department of Economics, University of North Carolina.Witte, A. D., & Woodbury, D. F. (1985). The effect of tax laws and tax administration on tax compliance: The case of the

U.S. individual income tax. National Tax Journal, 38, 1–13.World Bank. (2005a). 2005 world development indicators. Washington, DC: World Bank.World Bank. (2005b). 2005 world bank group—Data and statistics. Washington, DC: World Bank

(http://www.worldbank.org/data/countryclass/classgroups.htm).World Economic Forum. (2002). The global competitiveness report 2001–2002. Geneva: WEF.World Economic Forum. (2003). The global competitiveness report 2002–2003. Geneva: WEF.World Economic Forum. (2004). The global competitiveness report 2003–2004. Geneva: WEF.You, J. S., & Khagram, S. (2005). A comparative study of inequality and corruption. American Sociological Review, 70,

136–157.


determinants of tax evasion: a cross-country investigation

Documents