Tax Morale and Trust in Public Institutions

Transcription

1 Tax Morale and Trust in Public Institutions Preliminary version. Please do not cite and distribute Wilfried Anicet Kouamé Université de Sherbrooke July 30, 2015 Abstract One signicant puzzle in economics is to explain why people pay their taxes and why there are so many dierences in tax compliance across countries. Much recent research has mainly identied social and religious interactions such as norms, values, culture and history as factors that could justify tax compliance. This paper aims is to provide, both theoretically and empirically a new explanation to taxpayer' willingness to comply by highlighting trust in public institutions The theoretical framework goes beyond the standard model of tax evasion by allowing social and institutional interactions. For empirical evidence, i use the World Values Survey to estimate the causal impact of trust in public institutions on tax morale. The ndings suggest that in emerging and developing countries, social norms play a great role on tax morale, whereas in advanced countries institutional environment seems to be one of the most important factors. The results remain robust after exploiting and conducting several sensitivity analysis. JEL codes: D700; H260; H310; K420 Keywords: behaviour. Tax morale, tax compliance, trust, tax evasion, institutions, opportunistic Corresponding author: Wilfried Anicet Kouamé; Adress: 2500 Boulevard l'université, Sherbrooke QC, J1K2R1, (Canada); tel: ext.65060; wilfried.kouame@usherbrooke.ca 1

2 1 Introduction In developed and developing countries alike, the answer to the question why do people pay tax?" remains a key challenge for both economic theory and policy makers. Allingham and Sandmo (1972) were the rst to attempt to answer this question by developing the standard model of tax evasion. In this model, deterrence policies, i.e. probability of detection and degree of punishment, determine the extent of tax evasion. However, this model has been widely criticized because of the incompatibility between its predictions and the actual tax evasion. Indeed in many countries, the actual level of deterrence is too low to explain the high degree of tax compliance (Alm et al. (1992) ; Clotfelter (1983); Skinner and Slemrod (1985)). The literature highlights two explanations to this inconsistency. The rst argument strongly validated by theoretical and experimental studies 1, is related to the suggestive nature of the likelihood to be detected. Indeed, taxpayers have a tendency to overestimate the probability of being detected (Erard and Feinstein (1994) ). A misperception of this probability would explain why tax evasion is, in reality, lower compared to the predictions of the standard model of tax evasion. The second explanation, in which this paper is inscribed, highlights tax morale as a key factor that justifyies the gap between the predictions of Allingham and Sandmo (1972) models and facts (Alm et al. (1992); Cummings et al. (2009); Torgler (2006); Frey and Torgler (2007)). Tax morale is dened in this literature as "the intrinsic motivation of an individual to pay his taxes" (Torgler (2007); Feld and Frey (2007)). A large part of the existing literature identies social interaction variables (horizontal relationship) as determinants of tax morale. Besley et al. (2015), Bénabou and Tirole (2011) and Myles and Naylor (1996) show for example that the intrinsic motivation of taxpayers to pay their taxes is aected by social norms. Using the WVS, Alm and Torgler (2006) show that cultural dierences between the United States and Europe explain the dierence in tax 1 Alm et al. (1992), Bernasconi (1998), Erard and Feinstein (1994) 2

3 morale. Through their analysis of the tax reform under Margaret Thatcher's government in the UK in the early 1990s, Besley et al. (2015) show the evidence that temporary shocks aect the willingness to comply. Fischbacher et al. (2001), Fortin et al. (2007), and Frey and Torgler (2007) found evidences for conditional cooperation that taxpayers contribute to public goods according to others' contributions. Finally, Torgler (2006) shows strong evidence that religiosity aects tax morale. The eects of social interactions on tax morale have thus been widely discussed in the literature in terms of conditional cooperation, norms, culture, and social conformity. However, evidence from vertical relationship (interactions with public institutions) is rather sparse 2. Smith and Stalans (1991) were one of the rst to show that positive actions of states tend to improve attitudes and commitments of taxpayer against tax system. This result was conrmed by Pommerehne et al. (1994) who used a simulation study design to analyze the impact of scal exchange on tax compliance. Pommerehne et al. (1994) show that the more there are discrepancies between the optimal choice of public good, the level and the real quantity supplied, the more tax evasion is high. Based on survey data, Scholz and Lubell (1998) show the evidence that trust in government increases tax morale. Also, in comparative framework between West Germany and United-States, Slemrod (2002) shows that trust in government decreases the acceptability of tax evasion. Moreover, Ahmed and Braithwaite (2004) show that tax evasion is signicantly higher in the group of taxpayers who have lost condence in government. Finally at the macro-level, Frey and Torgler (2007) nd that quality of institutions increases tax morale. Papers on social interactions eects highlight that taxpayers' willingness to comply depends on the quality of their relationship with public authorities. Thus, as Feld and Frey (2007) and Akerlof (1982), i argue that the relationship between the taxpayers and public authorities can be perceived as a psychological or implicit contract. This contract involves a rewards system, punishments, loyalty and mutual expectation. Opportunistic behaviour like cheating 2 To our knowledge, the only papers that analyze vertical relationship are Ahmed and Braithwaite (2004), Slemrod (2002), Scholz and Lubell (1998), Pommerehne et al. (1994), and Smith and Stalans (1991). I'am going to discuss these papers later. 3

4 on taxes can therefore be a way for taxpayers to express their satisfaction or discontent with the way public resources are managed. This paper aims to analyze theoretically and empirically the vertical relationship between taxpayers and public institutions through trust. To achieve this goal, this paper defends that the proper functioning of the implicit contract between taxpayers and public authorities depends on trust. Indeed, the taxpayer's trust in public institutions is determined by the comparison between the benets received from public institutions (in the form of public goods or services) and the price paid for them (taxes). Thus, the more taxpayers trust public institutions, the higher the chances that they will have an incentive to pay their taxes. This paper brings three major contributions to the literature. The rst one is to integrate the relationship between taxpayer and public institutions (vertical relationship) in the standard model of tax evasion. The second contribution is to consider trust in public institutions instead of condence in government. The third contribution is to deal with endogeneity issues between tax morale and trust in public institutions. There are three papers close to this one in the literature. Frey and Torgler (2007) briey analyze at macro level the relationship between tax morale and quality of institutions. But they evoke neither condence in public institutions, neither the potential endogenous issue between tax morale and the quality of institutions. With survey data, Scholz and Lubell (1998) analyze the relationship between tax evasion and trust in government and trust in co-citizens. However, they are conned only to the condence in government and to the United-States framework. I think that outside condence in government, other factors like condence in civil service or the court aect the intrinsic motivation of taxpayers to comply. Finally, in a comparative framework between West Germany and United-States, Slemrod (2002) shows that in both countries those who have condence in government are less likely to nd tax evasion acceptable. However, the analysis of Slemrod (2002) was limited to a simple correlation between condence in government and tax evasion acceptance. This paper goes further than Scholz and Lubell (1998) and Slemrod (2002) consider condence in public institutions (instead of condence in government) and deal with endogeneity issues. 4

5 The next section discusses the previous work and ndings on the relationship between the taxpayers and public institutions. Section 3 presents the theoretical model. Section 4 empirically analyzes the relationship between trust in public institutions and tax morale. Section 5 provides some concluding remarks. 2 Some previous work on trust and vertical relationship It could be that tax evasion is the only channel through which taxpayers can express their antipathy (... ) we can be condent in our general prediction that if tax attitudes become worse, tax evasion will increase. (Lewis, 1982, p ) As noted previously, recent literature on tax morale highlights norms, culture, religion and history as determinants of the taxpayers' intrinsic motivation to pay their taxes. However, trust appears as the key factor to develop favorable values to tax morale and foster collective actions. Indeed, as Scholz and Lubell (1998) noted, without trust the potential benets of collective actions will depend solely on altruism and enforcement mechanisms that ensure that the prots of opportunistic behaviours are lower than the costs associated with deviation. Trust therefore plays an important role in maintaining credibility and enhances social capital by reducing monitoring and costly sanctions (Coleman (1994); Ostrom (1990)). Condence also allows for the reduction of transaction costs (Kreps (1990), North (1990)) and improves the social well-being in private market ( Fukuyama (1995), North (1990) ) and in public management (Putnam et al. (1994), Ostrom (1990)). Similarly, Bayer (2006) and Bayer and Sutter (2009) show that lack of condence leads welfare's loss for the community because of the waste of resources associated with concealment and detection costs. Thus, any policy and action aiming to increase condence in public institutions will raise taxpayers' intrinsic motivation to pay taxes. The evidences discussed in Güth et al. (2005), Pommerehne and Weck-Hannemann (1996), Pommerehne (1978), Torgler (2005) and Torgler and Werner (2005) 5

6 support this idea. Indeed, by analysing the relationship between tax morale and decentralization or local autonomy, Güth et al. (2005) nd that the latter increases the former. Based on Swiss cantons data, Pommerehne and Weck-Hannemann (1996), Pommerehne (1978) and Torgler (2005) show that direct democratic signicantly increases tax morale. Torgler and Werner (2005) use German data to show that scal autonomy strongly increases tax morale. In all these papers, decentralisation, local autonomy, direct democratic right and scal autonomy improve trust in public institutions because these policies or systems are strongly in line with the taxpayers' preferences, hence a willingness to comply. Beside public goods provision, taxpayers' condence in public institutions could be aected by the way they are treated by the tax authorities. Indeed, as argued by Tyler et al. (1989), the taxpayers' willingness to cooperate, as well as their assessment of tax authorities is affected by the way they are treated by the latter. Fehr and Schmidt (1999) also argue that fairness, justice or reciprocity feelings can increase the taxpayer intrinsic motivation to pay his taxes. The evidence discussed in Hofmann et al. (2008) supports the idea that perceptions of fairness of the tax system shape attitudes towards tax compliance. Thus, astorgler (2005) argues, unequal tax system could enhance the incentives to rationalize opportunistic behaviour. Trust therefore plays an important role in the taxpayers' willingness to pay their taxes whether at the national or local level. As Lewis (1982), this paper supports that pay taxes can be a way for the taxpayer to express his satisfaction about public policies or services implemented. If a taxpayer does not trust public institutions or has a lower perception of their quality, his intrinsic motivation to pay taxes will be low. The next section will show how trust in public institutions aects tax evasion. 6

7 3 Theoretical Approach 3.1 Standard model of tax evasion Tax evasion standard model is based on Allingham and Sandmo (1972) and Yitzhaki (1974) papers. In this model, the taxpayer i has at each period an exogenous income Y i, that is not known by tax administration. The taxpayer declares (to the tax administration) an income D on which a tax τ is levied. I assume in this paper that the taxpayer chooses a share ϑ of his income Y i that he escapes i.e D i = (1 ϑ i )Y i. Let f be the ne paid by the taxpayer if he is discovered. In this case, his income will be: Y ai = Y i τ(1 ϑ i )Y fτϑ i Y i However, the taxpayer's income if he is not discovered will be: Y ni = Y i τ(1 ϑ i )Y i At each period, the taxpayer chooses ϑ that maximizes his expected utility : Where, p is the detection probability. max ϑ i Eu i (Y i ) = pu i (Y ai ) + (1 p)u i (Y ni ) (1) max ϑ i Eu i (Y i ) = pu i [Y i τ(1 ϑ i )Y i fτϑ i Y i ] + (1 p)u i [Y i τ(1 ϑ i )Y i ] (2) I assume that the taxpayer is risk averse, hence his private utility u i () is increasing and concave in consumption. The rst order will be Eu i (Y i ) ϑ i = 0 p(1 f)τy i u i(y ai ) + (1 p)y i τu i(y ni ) = 0 7

8 The optimal level of tax evasion can be obtained from Kuhn-Tucker conditions and be written as : ϑ i = ϑ i (τ c, f c, p c ) (3) Where, c is the country index. Thus, according to Allingham and Sandmo (1972)'s model the proportion of evaded income depends on the extent of deterrence policies like detection probabilities and nes. This means that an increase in detection probabilities p or on the ne for evasion f would reduces tax evasion. However, considering the weakness of actual detection policies (low detection probabilities and weak ne), we should observe a higher tax evasion than there really is (Alm et al. (1992)). Indeed, this standard model takes into account neither taxpayers perception about public services i.e. transparency and fairness (Cummings et al. (2009)), nor the way in which public expenditures are determined (Alm et al. (1993)) 3, nor the way in which application rules are determined (Alm et al. (1999)), nor the social norms (Besley et al. (2015); Bénabou and Tirole (2011); Myles and Naylor (1996)) and the collective actions (Naylor (1989)). The next section extends the standard model of tax evasion by taking into account the eect of condence in public institutions. The idea here is that the taxpayer condence in public institutions aects his tax morale. 3.2 Extension: inclusion of condence in public institutions Based on Brock and Durlauf (2001) 4 and Fortin et al. (2007) 5, i include an individual institutional utility function in the taxpayer private utility function. More specically, this model 3 Alm et al. (1993) nd for example that tax morale is higher when public goods are voted rather than when they are imposed. 4 First paper which includes an individual social utility function in the private utility function. 5 Fortin et al. (2007) apply Brock and Durlauf (2001)'s framework to the standard model of tax evasion by adding a social interactions function. 8

9 presents a general framework in which there are both social interaction function S i (ϑ i, Z i ) and institutional interaction function A i (ϑ i, X i ). The latter depends on the share of income evaded ϑ i and respectively on a set of variables Z i and X i. I assume that individual social and institutional interactions functions are linear in ϑ i6. This assumption entails that the marginal social and institutional functions depends only on variables X i and Z i. Taxpayers' expected utility will be : EV i (Y i ) = Eu i (Y i ) + S i (ϑ i, Z i ) + A i (ϑ i, X i ) The rst component is the private expected utility associated with tax morale, corresponding to expected utility in the standard model of tax evasion. The second and the third components are respectively social and institutional utility associated to tax morale. S i (ϑ i, Z i ) = s i (Z i )(1 ϑ i )Y i A i (ϑ i, X i ) = a i (X i )(1 ϑ i )Y i s i (Z i ) and a i (X i ) are respectively social and institutional marginal utility associated to tax morale. The expected utility function can be written as EV i (Y i ) = pu i [Y i τ(1 ϑ i )Y i fτϑ i Y i ] + (1 p)u i [Y i τ(1 ϑ i )Y i ] + [s i (Z i ) + a i (X i )](1 ϑ i )Y i (4) Where, a i (X i ) = a(t i, ɛ i ) s i (Z i ) = s i ( ϑ i e, K i, K i, ɛ i ) (5) I assume that the marginal institutional utility of tax morale depends on trust in public institutions T i and a random term ɛ i that captures unobservable individual characteristics and attributes that are common to all individuals in the same community. The idea here is that the relationship between the taxpayer and the state can be assimilated in terms of a psychological or implicit contract. This contract involves a rewards system, punishments, 6 Fortin et al. (2007) make the same assumption. 9

10 loyalty and mutual expectation. Thus, if the taxpayer is satised with this relationship, his condence in public institutions increases. In this case, the taxpayer intrinsic motivation to pay his taxes will be high. Thus, condence in public institutions have a positive eect on tax morale i.e. the marginal institutional utility of tax morale is positive. Concerning the marginal social utility of tax morale, i assume that it depends on individual subjective expectation of the mean of tax morale of co-citizens ϑ e i. As in Fortin et al. (2007), Myles and Naylor (1996) and Gordon (1989) a positive eect of ϑ i on tax morale corresponds to a social conformity eect. A negative eect corresponds to a social anti-conformity eect. In the latter case, the taxpayer's preferences indicate a pure free-riding behaviour. I assume also that the marginal social utility of tax morale depends on the mean observable attributes of the taxpayer community, Ki. Unlike Fortin et al. (2007), i suppose that public goods funded through taxes are fully take into account in individual utility 7. The slackness of this assumption is justied by the inclusion of condence in public institutions in tax evasion model. Indeed, the implicit contract between the taxpayer and the state requires the provision of desirable level of public good (quality and quantity). Thus, dissatisfaction with public services and expenditures causes deterioration of taxpayer's condence in public institutions. Also, the signals sent by the public institutions (failure, wastage, ineciency, corruption, abuse of power) aect the taxpayer condence in public institutions. Assuming that preferences satisfy Von Neumann-Morgenstern axioms, the taxpayer problem can be write as: 7 Many papers like Spicer and Lundstedt (1976), Alm et al. (1992) make implicitly the same assumption. 10

11 max EV i (Y i ) = pu i [Y i τ(1 ϑ i )Y i fτϑ i Y i ] + (1 p)u i [Y i τ(1 ϑ i )Y i ] + [s(z i ) + a(x i )](1 ϑ i )Y i ϑ s/c 0 ϑ i 1 a i (X i ) = a i (T i, ɛ i ) s i (Z i ) = s i ( ϑ i e, Ki, K i, ɛ i ) First-order conditions are: (6) p(1 f)τy u i[y i τ(1 ϑ i )Y fτϑ i Y i ] + (1 p)y i τu i[y i τ(1 ϑ i )Y i ] [a i (T i, ɛ i ) + s( ϑ i e, Ki, K i, ɛ i )]Y = 0 Using the Kuhn-Tucker conditions, the optimal level of tax evasion will be: ϑ i = ϑ i (τ c, f c, p c, T i, ϑ i e, K i, K i, ɛ i ) Where, i and c are respectively individual and countries index. I assume that the taxpayer's expectations are self-consistent i.e. they are based on the contemporaneous mean. In this case, the Nash equilibrium is characterized by ϑ ie = ϑi. Thus, the optimal level of tax evasion can be write as, ϑ i = ϑ i (τ c, f c, p c, T i, ϑ i, K i, K i, ɛ i ) (7) By taking the total derivatives of tax evasion relative to condence in public institutions, mean of tax evasion, ne, detection probability, we have: dϑ i dt i = a i (T i,ɛ i ) T i (τy ) 2 [p(1 f) 2 u i (Y ai) + (1 p)u i (Y ni) (8) 11

12 dϑ i d ϑ i = s i ( ϑ i,k i, K i,ɛ i ) ϑ i (τy ) 2 [p(1 f) 2 u i (Y ai) + (1 p)u i (Y ni) (9) dϑ i df c = dϑ i dp c = p(1 f)τy u (Y a ) (τy ) 2 [p(1 f) 2 u (Y a ) + (1 p)u (Y n ) (f 1)τu (Y a ) + τy u (Y n ) (τy ) 2 [p(1 f) 2 u i (Y ai) + (1 p)u i (Y ni) (10) (11) From this model and expression above, we can draw four major predictions regarding tax evasion: dϑ i dt i 0 dϑ i d ϑ i 0 3. dϑ dp 0 4. dϑ df 0 Prediction 1 indicates that an increase of condence in public institutions would reduce tax evasion. This means that the more taxpayers' condence in public institutions is high, the higher their intrinsic motivation to pay taxes will be. This result can be explained by the fact that condence in public institutions reects taxpayer's satisfaction about the implicit contract between him and the state. In other words, tax resources collected to fund public goods would be used optimally, hence an incentive to comply. Prediction 2 states that an increase in the mean of tax evasion of the community raises individual tax evasion. I assume that the mean of community's tax evasion reduces the taxpayers' private social marginal utility 8. In this case, the taxpayer has an incentive to comply if paying taxes is a social norm. If cheating on tax is a social norm, the taxpayer will cheat on tax too. Gordon (1989), Fortin et al. (2007) and Myles and Naylor (1996) call this eect "`social conformity eect"'. However, we cannot exclude a specic case in which the 8 This mean that s( ϑ i,k i, K i,ɛ i) ϑ i 0 12

13 mean of community's tax evasion increases taxpayer private social marginal utility. In this case, taxpayer adopts a strategic substitutability behaviour (anti-social conformity eect). Predictions 3 and 4 state respectively that an increase in both detection probability and nes reduces tax evasion. The last two predictions are exactly the same as in Allingham and Sandmo (1972) model. 4 Empirical Approach The empirical assessment of our theoretical model is based on equation (7). I use the data from the sixth wave of the World Value Survey (WVS), which cover 52 countries from 2010 to In this survey data, respondents are interviewed on their attitudes concerning economic and social contemporary issues, family, religion and work. Although one can be critical about the cross-sectional nature of this survey, the WVS are widely used in political economics, sociology and in economy in many papers whose Knack and Keefer (1997); Slemrod (2002); Zak and Knack (2001); and Nunn and Wantchecon (2011). Moreover, using the experimental data of the Reader's Digest 9, Knack and Keefer (1997) provide an empirical proof of the validity of these data. Indeed, they show the existence of a strong correlation between the condence measure in the WVS and the number of wallets lost and returned with their content intact in many countries. For the purpose of simplication, the estimating equation is obtained by linearizing the expression of latent tax evasion (equation (7)). Moreover, from theoretical model to the empirical approach, i assimilate tax evasion variable to the tax morale. This choice can be justied by the fact that in WVS, i have a measure of tax morale 10. In addition, poor tax morale involves tax evasion and vice versa. Thus, the estimated equation will be: 9 This experimental study was conducted in several European countries and the United States. It consisted of making lose in several cities, the wallets containing 50$ each with the owners address and phone number. See The Economist of 22 June Frey and Torgler (2007), Torgler (2006) and Torgler (2005) use this variable as a measure of tax morale, while Slemrod (2002), uses it as a measure of tax evasion. 13

14 ϑ ilc = αt ilc + β ϑ ilc + γ K ilc + ηk ilc + ηf E c + ɛ ilc (12) Where, i, l and c are respectively individual, localisation (city) and country index. ϑ ilc is tax morale, T ilc trust in public institutions, ϑilc mean of tax morale in the taxpayer's community, K ilc mean of observable attributes of the taxpayer community. K ic is the taxpayers' individual characteristics and F E c is the country xed eects which capture country characteristics that do not vary from those taxpayers. In this model, i assume that nes and detection probability are the same for all taxpayers. These variables are therefore captured by the country xed eects. The general question that assesses the degree of tax morale is Please tell me whether you think that cheating on tax if you have the chance can always be justied, never be justied, or something in between. (scale from 1=never justied to 10=always justied) Our dependent variable, tax morale, is obtained by recoding the question above. Here 10 corresponds to never justied and 1 for always justied. Trust in public institutions variable is obtained by taking the weighted average of condence in government, parliament, legal system, civil service, education system 11. The question in order to obtain this variable is: I am going to name a number of organizations. For each one (the legal system, the police, the government in your nation's capital, parliament, the civil service, the education system), could you tell me how much condence you have in them: is it a great deal of condence, quite a lot of condence, not very much condence or none at all? (scale from 1=A great deal to 4=None at all). 11 In a simple comparison between West of Germany and the United-States of America, Slemrod (2002) uses a similar variable. However, Slemrod (2002) is limited to four indicators (condence in education and in legal system, condence in police and in civil service). I go further in this paper to include condence in the parliament and government. 14

15 Concerning taxpayers' individual characteristics, i consider age, education, religion, employment status and type of employment, marital status, gender, economic class and nancial satisfaction. These variables are dened below. The variable age is included to reect the fact that individual tax morale may depend on age. According to the literature, i'am expected a positive or a non-signicant eect of age on tax morale. Indeed, Tittle (1980) and Friedman et al. (2000) show the evidence of positive eect of age on tax morale, while other papers like Spicer and Becker (1980) and Mason and Calvin (1984) nd a non-signicant eect of age on tax morale. The variable religious controls for the taxpayers' beliefs. Like Torgler (2006), i'am expected that being religious increases tax morale. According to Spicer and Becker (1980), Tittle (1980) and social psychological literature, women are more compliant than men. To take into account the impact of gender on tax morale, i include a gender variable, gender, which is equal to 1 if taxpayer is a man and 0 otherwise. I control for employment status by including dummy variables capturing whether the taxpayer is a full time employee, part time employee, self-employed, retired, housewife not otherwise employed, student or unemployed (reference group : other). Also, working in the public sector or in the private sector may aect individual tax morale. Thus, i include two dummy variables public employee and private employee that indicate if the taxpayer works in the public or private sector. I control also for marital status by including dummy variables (single, living together, married, divorced, separated) that indicate if the taxpayer is single, married, divorced, separated or living together (reference group: widow). Indeed, Tittle (1980) shows that the tax morale of married persons is higher than that of single persons. The explanation of this evidence is that married persons are more constrained by their social network than single persons. Based on the theory of aspiration or reference point 12, i introduce a nancial satisfaction variable (nancial satisfaction) to take into account the eect of nancial satisfaction on tax morale. The intuition here is that the taxpayers who are least satised nancially are more likely to engage in tax evasion because they aspire for a better nancial situation. Thus, 12 See Simon (1955) and Kahneman and Tversky (1979) 15

16 the least nancially satised taxpayers will have the weakest tax morale. In similar ways as nancial satisfaction, the taxpayer's economic class can aect his motivation to pay taxes. For this purpose i introduce dummy variables, upper class, uppermiddle class lower class and working class (reference group: lowermiddle class). The latter captures the impact of economic class on tax morale. Finally, I control for the highest level of education attained by the taxpayer ( education). The intuition here is that the more the taxpayer is educated, the better he knows tax system (strengths and weaknesses) and the uses of collected resources through taxes. Thus, on one hand the taxpayer can use his knowledge to detect aws and make tax evasion, on the other hand, the taxpayer can have a higher motivation to pay his taxes because he better evaluates the benets of collective action. As Lewis (1982) and Torgler (2006), i do not expect education to have a specic eect on tax morale, it can increase or decrease a taxpayer's tax morale. All questions associated with these individual characteristics are presented in appendix. The mean observable attributes of the taxpayer community variables and the mean of tax morale in the taxpayer community are obtained by computing the weighted average of tax morale, economic class, education, nancial satisfaction, age and religion by locality for each taxpayer. These weighted average are specic to each taxpayer because they are obtained by excluding the observation of the taxpayer for which i calculate the average. The same approach is adopted for the mean of tax morale ( ϑ ilc ). Because the taxpayer is small relative to his locality/community, there are therefore no endogeneity issue between his tax morale and that of his locality/community. 4.1 Baseline estimations As the dependent variable (tax morale) is ranking information's between 1 and 10, ordered probit model is more relevant for estimations. In order to, to correct the sample and to reect national distribution, i use weighted variable and so, i use weighted ordered probit model 16

17 in the baseline estimations. However, an ordered probit estimation has a non-linear form, which implies that estimated coecients give the right sign but not the right quantitative eect. Thus, we need to calculate marginal eects to assess the quantitative eect that a variable has on tax morale.in line with the literature 13, i present only the marginal eects for the highest value of tax morale i.e. cheating on taxes is never justied. This marginal eect indicates the change in the share of taxpayers belonging to a specic tax morale level when the independent variable increases by one unit. Columns (2) of table 1 below presents the results 14. The latter shows a strong positive correlation between condence in public institutions and tax morale. If condence in public institutions increases by one unit, the probability of arguing that cheating on tax is never justiable rises by 2.7 percentage points. This result is consistent with our theoretical model and our expectations. The ndings show also that the mean of tax morale, being religious, work in the public sector, be nancially satised and be part of working class increase tax morale. For example, if the mean of tax morale rises by one unit, the probability of stating that tax evasion is never justiable increases by 13 percentage points. The ndings shows also a strong conditional cooperation eect (or social conformity eect) as Frey and Torgler (2007) and Fortin et al. (2007). I notice also that being religious increases the probability of stating that cheating on taxes is never justiable by 1.51 percentage points. Finally, the ndings show that one's work and family status - a full time employee, self-employed, unemployed, student, single, divorced, living together belonging to the upper and the upper middle class - decrease tax morale. For example, belonging to the upper class reduces the probability of stating that tax morale is never justiable by 5.4 percentage points. All these results are consistent with those obtained in the literature. 13 See for example Torgler (2006) 14 Columns (1) is a OLS estimate use only for robustness. 17

18 Table 1: Trust in public institutions and tax morale Weighted ordored probit Weighted 2SLS Weighted OLS Weighted Ordored probit 2SLS (4) (5) (6) (7) Trust in public institutions (0.016)*** (0.010)*** (0.026)* (0.022)*** (0.011)*** (0.050)*** (0.030) [0.027] [0.047 ] [0.021] Social capital and religiosity Religious (0.024)** (0.015)*** (0.024)* (0.025)*** (0.019) (0.035)*** (0.033) [0.015] [0.040 ] [-0.001] Mean by locality Tax morale (0.024)*** (0.012)*** (0.024)*** (0.040)*** (0.012)*** (0.057)*** (0.026)*** [0.130] [0.10 ] [0.132] Education (0.024)*** (0.015)** (0.024)*** (0.044) (0.016)*** (0.056) (0.026)*** [-0.011] [ ] [-0.011] Age (0.027) (0.017) (0.027) (0.021) (0.020) (0.032) (0.025) [-0.006] [ ] [-0.003] Religious (0.027) (0.017) (0.027) (0.020) (0.019) (0.030) (0.025) [0.004] [0.002 ] [0.001] Financial satisfaction (0.026) (0.015) (0.026) (0.046) (0.016) (0.062) (0.029) [0.002] [0.009] [0.004] Economic class (0.057) (0.034) (0.058) (0.109) (0.036) (0.140) (0.062) [0.012] [0.018] [0.012] Employment status Full time employee (0.075) (0.057)* (0.075) (0.099) (0.068) (0.098) (0.115) [-0.036] [-0.051] [-0.022] Partial time employee (0.078) (0.059) (0.078) (0.105) (0.070) (0.107) (0.118) [-0.032] [-0.031] [-0.028] Self-employed (0.078)* (0.058)* (0.079)* (0.108)** (0.069) (0.115)** (0.117) [-0.039] [-0.080] [-0.021] Unemployed (0.085) (0.061)* (0.085) (0.110) (0.074) (0.117) (0.125) [-0.039] [-0.027] [-0.034] Retired

19 (0.077)* (0.058) (0.077)* (0.102) (0.071) (0.102) (0.117)* [0.027] [0.013] [0.040] Student (0.097)*** (0.065)*** (0.097)*** (0.116)*** (0.078) (0.143)*** (0.138) [-0.075] [-0.116] [-0.043] Home (0.082) (0.061)* (0.083) (0.110) (0.073) (0.112) (0.124) [-0.036] [-0.012] [-0.036] Public employee (0.043)** (0.027)** (0.044)** (0.055)** (0.030)*** (0.063)* (0.057)*** [0.021] [-0.037] [0.046] Private employee (0.041) (0.025) (0.041) (0.052)*** (0.028)** (0.058)*** (0.053)* [0.005] [-0.067] [0.022] Marital status Single (0.048)*** (0.032)*** (0.048)*** (0.064)** (0.037)*** (0.079)*** (0.060)*** [-0.048] [-0.051] [-0.047] Living together (0.056)*** (0.036)*** (0.056)*** (0.069) (0.043)*** (0.089)* (0.070)*** [-0.045] [-0.036] [-0.045] Married (0.042) (0.029) (0.043) (0.058) (0.034) (0.066) (0.054) [-0.001] [-0.000] [-0.002] Divorced (0.061)** (0.040)** (0.062)** (0.073) (0.049)* (0.090) (0.081)* [-0.030] [-0.026] [-0.032] Separated (0.086)*** (0.052)*** (0.087)*** (0.109)** (0.059)* (0.185)** (0.098)* [-0.052] [-0.085] [-0.038] Demographic factors Age (0.001) (0.000) (0.001) (0.001) (0.001) (0.001) (0.001) [-0.000] [0.000] [-0.000] Gender (0.021)*** (0.013)*** (0.021)*** (0.025)*** (0.016)*** (0.034)*** (0.027)*** [-0.032] [-0.062] [-0.016] Education (0.005)** (0.003) (0.004)* (0.007) (0.006)* (0.010) (0.006) [0.001] [-0.000] [0.002] Economic situation Financial satisfaction (0.005)** (0.003)** (0.005)*** (0.006)*** (0.004) (0.009)*** (0.006) [0.003] [0.009] [0.000] Upper class

20 (0.086)*** (0.045)*** (0.089)*** (0.105)*** (0.050) (0.204)*** (0.098)*** [-0.054] [-0.153] [-0.026] Uppermiddle class (0.027)* (0.017)* (0.028)* (0.031)** (0.021) (0.042)** (0.036) [-0.011] [-0.025] [-0.006] Lower class (0.040) (0.024) (0.040) (0.068) (0.029) (0.099) (0.045) [0.008] [0.010] [0.014] Working class (0.025)*** (0.016)*** (0.025)*** (0.030) (0.018)*** (0.042) (0.031)*** [0.029] [0.003] [0.042] F-stat of excl. IVs Hansen J statistic R2 N Dependent variable is tax morale. Columns (1) to (3) are the results estimation on the total sample; columns (4) and (6) results on advanced countries and columns (5) and (7) are estimations on less advanced countries. A column (1) is OLS estimate reported for robustness. All estimates include country xed eects. The marginal eects in brackets [] is calculated at the highest tax morale score. Robust standard errors in parenthesis. * Signicance at p < 0.1; ** idem at p < 0.05; *** idem at p <

21 4.2 Econometric issues: Causality The ndings above show evidence of positive correlation between condence in public institutions and tax morale. However, one might wonder about endogeneity issues (error measurement and reverse causality) between these variables. Indeed, individuals may not accurately report their true tax morale and their condence in public institutions. This could lead to potential measurement error problems. Concerning the reverse causality, we cannot exclude the fact that people's tax morale aects their condence in public institutions. These error measurement and reverse causality problems could bias the previous estimated results. Thus, to identify the causal eect of condence in public institutions on tax morale, i use the instrumental variables method which seeks to nd an instrument for condence in public institutions. The latter must satisfy both relevance and validity conditions. To nd any variables which satisfy these conditions, i wonder how condence in public's institutions is formed? The answer to this question brings us to identify two potential instrumental variables: condence in political parties and perception of human rights. For condence in political parties, i argue that the political environment sends a signal about the reliability of public institutions. Indeed, political scandals such as bribes and abuse of authority deteriorate the taxpayer's condence in public institutions as a result of governments' unaccountability. Thus, condence in political parties aects tax morale through condence in public institutions. I see no reason why condence in political parties could directly aect tax morale. Indeed, even in the more extreme case of partisan position, we cannot reject the pertinence of our instrument. Supposing for example that condence in political parties is aected after a change of government which pushed the taxpayer to change his intrinsic motivation to pay his taxes. In this case, change of condence in political parties aects condence in public institutions which in turn aects the tax morale. Thus, variation in tax morale would be purely due to the changes in condence in the new public institutions. Concerning perception of human rights, it obviously aects the taxpayers' condence in public institutions. The less human rights are respected, the lower the condence in public 21

22 institutions will be. Indeed, respect for human rights implies respect for property rights, as well as fairness and equal treatment of taxpayers. Outside the channel of condence in public institutions, i do not see how perception of human rights could directly aect tax morale. In line with the literature 15, i use 2SLS method in our identication strategy. Column (3) of table 1 shows that condence in public institutions signicantly increases tax morale. Weak identication and over identication tests presented in table 1 conrm that our instruments are valid and relevant. All other results are substantially the same excepted results on full time employee, unemployed, housewife not otherwise employed (home) and education variables. The rst three variables have no more eect on tax morale relative to the previous estimates, while education has now a positive and signicant eect on tax morale. This last result shows that in this paper, the level of education helps to encompass the benets of collective actions 16. Besides highlighting that condence in public institutions causes tax morale, our results provide a theoretical justication for papers which show that participative democracy, scal autonomy and decentralisation increases tax morale 17. Indeed, these policies or systems show the need for the direct implication of the taxpayer in the decision-making process. The direct result of this system is to strengthen the condence of taxpayers in public institutions because they are an integral part of decision-making process and can directly evaluate how tax revenues are used. 4.3 Heterogenous Eects In this section, i consider potential heterogeneous eects of the relationship between condence in public institutions and tax morale. Indeed, in an area with good institutions, the taxpayers may be more sensitive to the signal sent by these institutions. In order to assess 15 See Frey and Torgler (2007), Torgler and Schneider (2007) 16 Lewis (1982) et Torgler (2006) nd the same eect of education on tax morale. 17 Güth et al. (2005), Torgler and Werner (2005), Torgler (2005), Pommerehne and Weck-Hannemann (1996), Pommerehne (1978). 22

23 the existence of heterogeneous eects, i divide the sample into two sub-groups: advanced countries and non-advanced countries 18 and shows the same estimations as the identication strategy. The ndings presented in columns (4) to (7) of table 1 show the existence of heterogeneous eects. Indeed, i notice that the eect of condence in public institutions on tax morale is higher in advanced countries relative to emerging and developing countries. The estimates with weighted ordered probit indicate, for example, that in advanced countries, if condence in public institutions rises by one unit, the probability of stating that cheating on taxes is never justiable increases by 4.7 percentage points. Whereas in emerging and developing countries the same probability increases by 2.1 percentage points. However, in weighted 2SLS estimates condence in public institutions has no more eect on tax morale in emerging and developing countries, although this variable still has a positive eect on tax morale in advanced countries. This heterogeneous eect can be explained by the taxpayers' expectations. Indeed, in countries with better institutions, the taxpayers attribute a great importance to the latter. This therefore generates high sensitivity to the signal sent by these institutions among taxpayers. In such conditions, the implicit contract between taxpayer and public institutions will have more value in advanced countries in contrast to emerging and developing countries. The ndings show also that the conditional cooperation eect of tax morale is highest in emerging developing countries relative to the advanced countries. These results suggest that in emerging and developing countries, social norms play a great role on tax morale, whereas in advanced countries institutional environment seems to be the most important factor. The results on the other variables remain almost the same. 4.4 Sensitivity Analysis To check the robustness of our results, i use dierent specications. The rst consists in using the 2-stage conditional maximum likelihood (2SCML) procedure in order to test and con- 18 I use the World Economic Outlook Database, October This classication is presented in appendix. 23

24 trol for endogeneity problems. As the dependent variable is ranking information, non-linear model is more appropriate. I check therefore robustness of previous results by using 2SCML in identication strategy. This procedure was developed by Alvarez (1994) and Rivers and Vuong (1988) and consists of using ordinary least squares regression to estimate, rst, the reduced form equations for endogenous variables in the structural equation. Then, i include the residuals from the latter regressions as additional variables in the weighted ordered probit estimate of the structural equation. According to Alvarez (1994), the 2SCML procedure is more ecient than the standard 2-stage probit procedure. The results reported in Table 7 (appendices 1) show that in advanced countries if trust in public institutions increases by one unit, the probability of stating that tax evasion is never justiable increases by 2.8 percentage points. These results also show that trust in public institutions has no eect in tax morale in emerging and developing countries and conditional cooperation eect is higher in these countries in contrast to advanced countries. These ndings conrm the robustness of the previous results. The second sensitivity test consists of creating various subcomponents of condence in public institutions. In our baseline model, condence in public institutions is the mix of condence in courts, police, parliament, the civil service, universities and the government. One would think that the above results are aected by this combination. To prove that this is not the case, i separate condence in public institutions into three subcomponents: condence in system of law (police and courts), legislative and executive power (parliament and government) and other public services (universities and civil service). Indeed, if condence in public institutions aects tax morale, the latter will also be sensitive to the dierent components of condence in public institutions. Table 2 indicates that there is a positive and signicant relationship between tax morale and the three components of condence in public institutions. As in the previous results, these variables have no eect on tax morale in emerging and developing countries in 2SLS estimations. In the next step of our sensitivity analysis, i check whether or not the results are sensitive 24

25 to the use of another wave of WVS. Thus, i use the last two wave of WVS 19 to estimate the relationship between condence in public institutions and tax morale. The results presented in Table 3 show that condence in public institutions still has a positive and strong eect on tax morale. As in the baseline results, this eect is weaker in emerging and developing countries. Finally, i try to better isolate the eect of condence in public institutions on tax morale and verify the results are sensitive to some variables. First, i include the taxpayer political position. The taxpayer political position could aect simultaneously his tax morale and his condence in public institutions. The failure to account for this variable could therefore lead to omission bias. Thus, i include the taxpayer's political position to check and ensure the robustness of our results. The results presented in the panel A of table 4 show that condence in public institutions still has a strong eect on tax morale. The results highlight also that condence in public institutions has no eect on tax morale in emerging and developing countries. Afterward, i test the sensitivity of our results to vertical condence (in addition to horizontal condence). I thus include condence in co-citizens as an independent variable, in order to check if our results are sensitive to taking into account both horizontal and vertical condence 20. The results reported in the panel B of table 4 show that condence in public institutions still has a positive and signicant eect on tax morale in total sample and in advanced countries but has no eect in emerging and developing countries. Trust in co-citizens has a negative and signicant eect on tax morale in weighted ordered probit estimations, but not in our identication strategy. Recent literature highlights norms and culture as one of the important determinants of tax morale. In our cross-sectional analysis, i implicitly take this into account by using country- xed eects. However, some norms and culture can be specic to the language and history. Thus, to better isolate the eect of condence in public institutions on tax morale and check 19 I'am focusing on the wave 4 and 5, because there are the only one with questions about selected variables. 20 Scholz and Lubell (1998) discussed the need to take account both aspects of condence. 25

26 its robustness, i include a dummy variable indicating if countries have romanic culture 21. The results reported in table 5 show that condence in public institutions always increases tax morale, with weaker or no eect in emerging and developing countries. These ndings also show that tax morale is smaller in romanic countries in contrast to other countries. The last thing i do in our sensitivity analysis is to take into account taxpayers' risk attitudes. Indeed, risk aversion can reduce the incentive to make illegal activities and according to Torgler (2006) individual tax compliance decision could be a function of taxpayers' risk attitudes. We thus include a proxy of taxpayers' risk aversion 22. The results reported in table 6 conrm that condence in public institutions increases tax morale. As in the previous estimates, condence in public institutions has a weaker or no eect in emerging and developing countries. As expected, risk aversion has a positive and signicant eect on tax morale. All these robustness checks conrm that condence in public institutions increases tax morale. This eect is greater in advanced countries than in emerging and developing countries. These results portrays also that social norms are more important in emerging and developing countries than in advanced countries. 21 The variable take 1 for the countries that have a romanic culture and 0 otherwise. A country has a romanic culture if latin language ( such as french, italian, spanish, portuguese or romanian) is spoken there as an ocial language and/or if the country was colonized by a country that has a romanic culture. 22 The risk aversion variable take 1 if the taxpayer is risk averse and 0 otherwise. This variable is described in Appendix. 26