Tax evasion and trust: A comparative analysis
Ronald Wintrobe
Department of Economics, University of Western Ontario, London, Ontario, Canada and
Klarita Gërxhani
∗
Amsterdam Institute for Advanced Labor Studies;
Amsterdam School for Social science Research; and Faculty of Economics and Econometrics, University of Amsterdam, The Netherlands
Abstract
The standard theory of tax evasion cannot adequately explain differences in tax evasion across countries. Public choice theory provides an important step forward, but is not completely satisfactory either. We complement these theories by incorporating insights from institutional economics. By specifically focusing on the informal institution ‘trust’, we argue that tax evasion is higher in some countries (i.e., transition countries) than in others (i.e., institutionally advanced countries) because of differences in the level of trust. This is tested using two hypotheses, relating individual tax evasion to trust in the government and in others’ willingness to pay taxes, respectively. The data support our hypotheses.
JEL codes: H26, H11, O17, O5
Key words: Tax evasion, public choice, institutions, transition
∗ Corresponding author: Amsterdam Institute for Advanced Labor Studies (AIAS), University of Amsterdam,
1. Introduction
In the standard theory of tax evasion, individuals and corporations pay taxes only because they are forced to (i.e., because they believe that if they did not, they would be liable to prosecution by the state). If this were the case, it would be essential that the probability of being discovered for tax evasion, and the size of the penalty if caught and convicted are sufficiently large to deter evasion. One problem with the standard view is that for some taxes such as self-reported income taxes, it is hard to believe that the probability of being caught for evasion is very large. In fact, all countries do encounter tax evasion, even those with the most sophisticated systems for gaining compliance. To illustrate, the United States Internal Revenue Service estimates that the proportion of all individual tax returns that are audited was 0.8% in 1990 (down from 4.75% in 1965). Civil penalties range from 20% of the portion of the underpayment resulting from a specific misconduct such as negligence or substantial understatement to 75% if there is evidence of substantial intentional wrongdoing. In very serious cases, criminal penalties may be applied. However, in 1995, only 4.1% of all U.S. taxpayers who were reassessed following an audit received any penalty at all. Yet, the IRS estimates that, for tax year 1992, 91.7% of income that should have been reported was in fact reported.1 What explains this high degree of compliance in the face of low penalties and probabilities of detection?
A second and related question is, why does the level of tax evasion differ so much across countries? Table 1 provides a comparison of the size of the shadow economy across developed, developing and transition countries. The table shows that tax evasion is particularly severe in developing countries. In Europe, transition –especially former Soviet Union- countries appear to have the highest levels. Why is compliance lower in these countries? And why don’ t these countries simply raise the penalties for non-compliance and solve the problem that way?
TABLE 1 HERE
We believe the answer to these questions lies in the level of ‘trust’ in these countries. Trust is often considered to be an important aspect of informal institutions (i.e., rules of behavior, norms of trust, culture) or social capital characterizing a country. In this respect, our view can also be linked to new institutional economics, arguing that the informal institution of trust co-determines the extent of tax compliance. The importance of informal institutions is stressed by Chhibber, for example: “Even when formal rules are similar, informal rules or social capital can in some situations explain a significant part of the reason why some societies progress faster than others” (2000: 297). We support this view by combining the public choice approach to the problem of tax evasion with insights from new institutional economics and social capital. The theoretical predictions we derive will be tested with a unique database for Albania where tax evasion and some informal institutions with respect to taxes are measured. By doing so, this paper contributes to the scarce literature on the micro-foundations of tax evasion and how it varies in size from one country to another.
The paper is organized as follows. The next section describes the standard theory of tax evasion. Section 3 describes the public choice approach. Section 4 takes us a step further in understanding the issue of tax evasion by introducing the concept of trust. Section 5 tests two theoretical predictions established in the previous section. Concluding remarks and some policy implications follow in section 6.
2. The standard theory of tax evasion
The standard view of tax compliance in tax theory is that taxes are a ‘burden’ or windfall harm. Individuals do not consider taxes in relation to the other side of the government ledger - expenditures. The chief problem in normative taxation theory is to devise taxes which minimize the ‘excess burden’ , i.e, how to minimize the total burden of taxation.
In order to know more about this theory, let us have a look at the standard theoretical model (Allingham and Sandmo, 1972; Yitzhaki, 1974) in more detail. As is now common in the literature on tax evasion, the model visualizes an individual taxpayer facing a tax rate t on own income Y. If she chooses to evade taxes, she faces a punishment ftE where E is the amount of unreported income and f is the size of the punishment (the fine rate) if caught. In
one sense, the model adapts the standard crime model of Becker (1968) to the taxation case. In another sense, tax evasion is part of optimal portfolio choice: the individual who chooses to evade taxes in effect makes a risky bet that she will not be caught and convicted. However, the Yitzhaki (1974) model makes an odd prediction –namely that an increase in the tax rate t actually leads to less tax evasion. This result holds in the model as long as individual absolute risk aversion decreases as income increases. This prediction is at variance with empirical evidence (e.g., Clotfelter, 1983), the results of laboratory experiments (e.g., Friedland, Maital and Rosenberg, 1978)2 and, it would seem, even with common sense.
However, the logic is simple once one realizes that in these models, tax evasion is treated as a risky gamble or a problem in optimal portfolio choice. The penalty if an individual is caught, ftE, is simply a constant multiple of the amount of tax evaded tE. Thus, if the tax rate rises, both the gains from evasion and the penalty rise by the same proportion, and there is no substitution effect for or away from evasion. There is an income effect, however: the individual is poorer as a result of the possibility of paying a higher penalty. This will make her take less risk, hence evade less at higher tax rates. Of course this relationship is derived from individual behavior and only holds at the individual level. The aggregate level of evasion may well move in a different direction as the level of tax affects the number of taxpayers who choose to evade.
3. The public choice approach
The basic hypothesis of the field of economics known as ‘public choice’ is that the citizens of democratic political jurisdictions perceive a connection between the taxes they pay and the government services they receive. In other words, citizens elect governments to provide them with goods and services and there is a certain sense in which every citizen must be aware that taxes must be paid to finance public services, whether they think their own burden is too high or low. This implies that every citizen knows that, if taxes are reduced, a reduction in public services will follow.
2 In these experiments, the single most important factor resulting in evasion was the tax rate. On the other hand, raising the size of the penalty, even to exorbitant levels (e.g., from 3 to 15 times the amount evaded) lowered the
One version of this approach is used by Cowell and Gordon (1988), who introduce public goods into the Yitzhaki model of tax evasion.3 Their result is that, if individuals display
decreasing absolute risk aversion, the effect on tax evasion of an increase in the tax rate is positive or negative as public goods are under- or over-provided. Thus, if public goods are under-provided, an increase in tax rates means an increase in public goods as well. Individuals feel wealthier, and they wish to take more risk. Hence, they evade tax more when the increase in public goods and associated increase in tax rates makes them better off, and
less when it makes them worse off. However this result remains at variance with the
empirical evidence. The authors themselves find the result a bit counterintuitive and relate it to the fact that the relationship between government and taxpayer has more dimensions than just the provision of public goods, something that their model does not capture (Gërxhani, 2004a).
Note that in the public choice approach, which sees the government as a rational utility maximizer, one still has to assume that citizens trust their government to deliver the services it has promised in order to explain why they would ‘voluntarily’ pay their taxes. It is still rational for each citizen to free ride, since whether she pays the taxes or not has little to do with the level of public services she receives. For example, suppose there are 1,000 citizens in a jurisdiction, and each one is supposed to pay taxes of $1,000. Each citizen will reason that if she does not pay the taxes, but everyone else does, then her level of services will fall, but only by a tiny amount. This will hold if public services are constant cost and they are pure public goods, so that the non-tax paying citizen cannot be excluded from receiving services. Assuming services are shared equally, while own tax bill falls by 100%, own level of services will fall by only 1/1000 = 0.1%. Consequently, it is rational for everyone to free ride and not pay the taxes independently of whether government services are delivered or not, and whether other individuals pay or not. 4
amount evaded and the probability of an under-declaration, but only marginally.
3 Bordignon (1993) develops a ‘fairness’ approach in which public goods are introduced as well. In this model, an increase in tax rates yields more evasion, in accordance with the empirical evidence.
4. Introducing trust
It has become increasingly common to emphasize that social capital in general and trust in particular play an important role in the performance of an economy.5
Several ‘new institutional economists’ emphasize the relationship between informal institutions like trust (or rules of behavior, cultural norms) and formal institutions (i.e., laws, regulations). For example Feige (1997) hypothesizes that more tax evasion will be observed when the two types of institutions are in conflict with each other. This hypothesis has recently found empirical support (Gërxhani, 2004b).
In this paper we first suggested that formal institutions, as specified in the standard theory of tax evasion (i.e., tax and fine rate), cannot adequately explain the distinct levels of tax evasion in different countries. In addition, in the standard framework, wherein taxes are a windfall burden, it should not matter to a citizen whether the government delivers the services promised or not, or whether or not other people pay. We then moved a step further to the public choice approach which introduces public goods as another aspect of formal institutions. The outcome is, however, that it is generally still rational for a citizen to completely free ride and not pay taxes, no matter what the government and other citizens do. As a result, the public choice approach does not solve the puzzle either. Here we broaden the analysis by introducing the level of trust, both between citizens and the government, and between the citizens themselves as variables to explain tax evasion.6
We break the logic of the standard free riding problem by assuming that each citizen will pay more of her tax bill (1) if she can trust everyone else to pay more; and (2) if she can trust the government to provide more of the services promised in return. As the level of trust
5 See Breton and Wintrobe (1982), Coleman (1990), North (1990), Knack and Keefer (1997), Putnam (1993, 2000), Chhibber (2000), Paldam and Svendsen (2000), Frey (2002).
6 Trust here is used in the same way as in Breton and Wintrobe (1982). A related concept is tax morality i.e., the public attitude towards the state (see Frey and Weck-Hannemann, 1984 and Schneider and Enste, 2003). Still another way to think about the trust variables used is in terms of social interactions. Benjamini and Maital (1985) introduce stigma (withdrawal from future interaction or exchanges, presumably because tax evasion can be interpreted as a violation of the implicit contract) as a form of social interaction into the standard model. In their model, an increase in stigma costs does not affect the level of evasion, but affects the decision whether to evade or not. The normal ‘entry’ condition that the net financial gain from evasion be positive is modified: the net financial gain must now be large enough to outweigh the expected penalty and the stigma cost. Hence, a number of individuals will decide not to evade if the tax rate is fairly low. However, as the tax rate rises, the financial gain from evasion may rise above the stigma cost, pushing previously honest taxpayers into becoming tax evaders.
by a citizen in the government’ s honesty or in the civic spirit of her fellow citizens falls, her own willingness to pay decreases. At some point, her own willingness to pay falls to zero, and she pays only to the extent that she believes that she will be caught, convicted and forced to pay a penalty if she does not pay. To see how our ideas on trust fit into the standard models, we will first proceed with a more theoretical discussion of these models.
4.1. A theoretical introduction
To introduce trust into the public choice approach, in figure 1, we assume one public good S, for simplicity, provided to the citizens of some jurisdiction. Since the good is public, all of the citizens in the jurisdiction must consume the same amount, whatever level the government desires to provide. Each citizen is assumed to be able to correctly calculate that the tax price to her of a unit of the public good is p. D is the individual’ s demand or marginal valuation curve for the public good. With that demand and at that price, the individual desires a level of public goods equal to S1. Suppose, however, that the government provides S0 instead of S1. As a consequence, this citizen loses the amount of surplus from the good
indicated by the shaded area L, which in turn will give her incentives to evade on tax p. FIGURE 1 HERE
Now, let us look at the level of tax evasion if tax rates rise. On this model, a rise in tax rates (e.g. to p1) with no change in the level of government services would indeed
increase the propensity of a citizen to evade taxes since it lowers the surplus of that citizen
from the government, as would normally be the case (the loss is given in area B). Presumably this is the case which is relevant to the experimental and other evidence discussed above.
On the other hand, if the rise in tax rates raised the level of public goods provided, and they had been under-provided previously, so that the net effect of the tax increase is to
increase citizens’ surplus, then her propensity to evade taxes would decrease. This case is
illustrated in figure 2. In figure 2, a rise in the tax rate from p to p1is accompanied by an increase in the public good (from S0 to S2). Provided that the area A (the increased surplus
from government increases, and therefore so does her propensity to pay taxes. If the surplus decreases, the opposite is true. It is worth emphasizing that these results are the reverse of those of Cowell and Gordon (1988) discussed earlier.
FIGURE 2 HERE
4.2. Mutual trust
One way trust fits into this model is by affecting the perceived or actual surplus depicted in figures 1 and 2. Why should there be a connection between the level of trust and the surplus from government goods and services? The most obvious reasons why the government might provide less than the desired quantity of public goods are: a) the government does not know what the citizen wants; b) it has to provide all citizens with the same amount, and most others want a different amount; or c) the government is corrupt, and appropriates some of the revenue which could be used to produce S for itself. In all these cases, if there were more trust the government could rectify the problem. If trust increased between the citizen and the government, first, it is easier for the citizen to signal own preferences to the government and for the government to receive them, and second, if the government has to provide others with
S0 and thus fails to satisfy this particular citizen, it could try to satisfy her in some other way. For example, it could provide her with a private good, which is a substitute for S, thus lowering her demand for S, and decreasing her loss in L (cf. figure 1). The more trust between the citizens and the government, the easier it is for the citizens and the government to communicate with each other, and the more responsive the government would be to the citizens’ preferences. And the more trust there is by citizens in the government, the more the government has to lose by being corrupt.
On this formulation, there is an exchange or an implicit contract between the citizenry and the government: the government provides goods and services to citizens in exchange for their support. (In some versions of this type of model, the government tries to maximize this support, as in probabilistic voting models of the government sector.7) In the aggregate, the government tries to maximize the sum of citizens’ surpluses -value of public goods and
services minus taxes- from the public sector. Each citizen, in turn, is more likely to support the government and, ceteris paribus, less likely to evade taxes, the greater the surplus she receives from the public sector. On the other hand, the smaller this surplus, the less likely individuals are to support the government, and the less likely they are to pay their taxes.
Although for illustrative purposes we are collapsing the two dimensions of trust into one, it is important here to emphasize that there are at least two directions of trust: (1) trust between the citizen and the government; and (2) trust between citizens themselves. This second variable operates in the same way as the first. The more citizens can trust their fellow citizens to pay taxes, the greater the surplus they expect to receive from the public sector, and the more willing they are themselves to do so and vice versa.
This is similar to the ideas in Benjamini and Maital (1985) or Myles and Naylor (1996). In these papers, the utility of evasion to a taxpayer is positively related to the number of others who evade. This dependence of individual decision-making on the decisions of others leads to multiple equilibria, which can be broadly classified into two: one in which people assume that others are paying and so most of them also pay; and in the other equilibrium, it is assumed that people do not pay their taxes and do everything they can to evade. The theoretical analysis in Benjamini and Maital (1985) or Myles and Naylor (1996) shows that there is a tipping point, as is common in the analysis of group interdependencies: when the number of tax evaders reaches a certain level, everyone is better off evading and evasion becomes epidemic. Consequently a small change in exogenous variables, e.g., the tax rate or other variables that precipitate a change in the number of evaders, can produce an epidemic of evasion.
Our approach complements these in explaining why citizens might be expected to behave this way. When a citizen expects that the number of others evading is large, the citizen expects that the government will not receive much revenue, and therefore will not provide much public goods. As a result, citizens’ surplus from government goods and services will not be large. So why would she herself want to pay taxes? Under these circumstances the government is not going to honor its implicit contract to provide services in return. On the other hand, when the converse is true, and most people are expected to pay their taxes, the citizens’ surplus from government goods and services can be expected to be
large, and the citizen is willing to honor her own implicit contract to pay taxes.8
We can refer to this latter case as the institutionally advanced country equilibrium. We will refer to the other equilibrium, in which it is assumed that people do not pay their taxes, as the transition country equilibrium, due to the well-known problems these countries have with tax evasion.9 In the latter group, the Russian (Rose, 2000) or Albanian (Gërxhani, 2002) equilibrium can be considered amongst the worst.
4.3. Tax evasion and trust in transition countries
Given that the problem of tax evasion appears to be more substantial in institutionally less developed countries (i.e., transition countries), and since in this paper we intend to look at the role of informal institutions on the decision to evade taxes, transition countries provide an excellent test bed for our ideas. About a decade ago, these countries went through an institutional shock, caused by the collapse of former communist regime. The level of the institutional shock varied per country, depending on the type of regime. On one hand, the communist regime was over-organized, where bureaucratic orders and ideological repression determined what individuals had to do. On the other hand, it was characterized by organizational failure, which motivated individuals to create and rely on informal networks. “Such a ‘dual society’ of formal versus informal networks [institutions] was far more developed in the Soviet Union, where it had been in place for more than 70 years, than in the Czech Republic [for example]” (Rose, 2000: 166). In Eastern Europe, similar characteristics were observed in Albania, where the totalitarian regime lasted for more than 40 years. As a consequence, these societies experienced significant distrust in the government and formal institutions. The substitute was found in family-, friends- or local networks. After the collapse of communism, in countries where the ‘dual society’ was dominant, and where in addition the new governments did not manage to function properly, trust has eroded even further, forcing people to invest and rely more on networks.
Indeed, the level of trust in the Russian government appears to be extremely low
8 In addition, because trust between citizens needs to be mutual, it correlates their willingness to pay taxes. As a consequence, free riding has much higher costs than just losing a fraction of the public good provided. It leads to a loss in trust that will decrease contributions by others, thus eroding the provision by much more than just the own withdrawn contribution.
based on survey data used in international comparisons (Hjolland and Svendsen, 2001). A Russian scholar, Anton Oleinik (n.d.) reports that only 3.4% of the respondents think that they can trust the state. Indeed, Oleinik suggests that it was the “non-reciprocal behavior of the state confirmed during the August 1998 crisis [which] led to a dramatic decline of the citizens’ willingness to pay the taxes.” (Oleinik, n.d., p. 22).
Based on a survey run across several cities in Albania, De Soto et al. (2002) find that ‘people in all areas generally lack confidence in government’ . Only 25% of people appear to trust public institutions. The highest level of trust is expressed towards family members.
5. An empirical test
Following the logic of the previous section, we can hypothesize that a citizen will be more inclined to pay her tax bill: (1) the more she trusts the government will provide the services promised in return; and (2) the more she trusts everyone else is paying. In reversed form, as the level of trust in the government’ s honesty or in the civic spirit of the fellow citizens falls, own willingness to pay decreases.10
Below, we test these two hypotheses based on data collected from a field survey conducted by one of the authors (Gërxhani) in the urban area of Tirana (the capital of Albania) in 2000.11
The response rate was 89.3% out of a sample of 1,500 households. This data set contains information on the informal institutions with respect to taxes, as well as sufficient information about income and taxes to derive estimates (where applicable) of the extent of personal income tax evasion per respondent. This gives us a unique opportunity to explore the relationship between tax evasion and the informal institutions like trust.12
Due to sensitivity issues (i.e., respondents may be reluctant to confess noncompliance), various indirect questions were used to gather information and construct a variable measuring
10 Some empirical support of these hypotheses for an institutionally advanced country, the United States, can be found in Scholz and Lubell (1998). As argued in the previous section, the focus here is on transition countries. 11 See Gërxhani (2002) for more on the survey and the questionnaire.
12 Note that our focus on informal institutions is possible because the data set has no cross-sectional variation in formal institutions (i.e., tax and fine rate, or public goods), which the standard theory of tax evasion and the
the evasion of personal income tax. For example, if the response to the question ´Does your employer (state or private) deduct your personal income tax from your monthly salary´ is ´No´, then this was considered as one indication of tax evasion; or if the response to the question ´Please indicate who pays your tax on personal income´ is ´Nobody´, this was another indication of tax evasion.13 There were five such questions representing five (indirect) indications of tax evasion. For the group of respondents liable to personal income tax, the five indications were summarized to obtain our main variable on tax evasion: ‘the extent of personal income tax evasion’ (PITE). The value of this variable gives the number of indications of tax evasion, with a minimum of 0 and maximum of 5. Table 2 summarizes the information.
TABLE 2 HERE
The informal institutions (individuals’ rules of behavior, trust norms, traditions) towards taxes were captured in the survey by including several attitudinal statements in the questionnaire. Respondents had to express their (dis)agreement with each statement on a 5-point scale. Given that in this paper we aim to specifically look at ‘trust norms’ , we test our hypotheses based on two statements of the survey: (1) “I do not feel like paying taxes as long as the government cannot be trusted”; and (2) “The majority of people in Albania do not pay taxes”. Table 3 shows the distribution of responses and the number of cases.
TABLE 3 HERE
Because a high (low) score on each statement reflects (dis)agreement, we can say that a high score on statement (1) reflects the respondents’ trust in government. A high score on statement (2) reflects trust in other individuals in Albania paying taxes. For the purpose of regression analysis below, we will refer to these statements as ‘trust in government’ and ‘trust in others’ .
different types of taxes is also taken care of since we focus on the extent of personal income tax only. 13 The Albanian tax law is similar to that in most western countries: individuals employed in the public or private sector are subject to tax on personal income.
Now that we have data on tax evasion and individuals’ trust in government and other individuals, we can directly test our two main hypotheses. We do this by running a logit regression, where the dependent variable -the extent of personal income tax evasion (PITE)- is transformed into a binary variable, taking the value 0 if there was no indication of tax evasion at all, and the value 1 if there was at least one indication of personal income tax evasion (cf. table 2). Table 4 gives the results of this logit regression.14 We try to explain variation in this dependent variable by including the two statements –TRUST IN GOVERNMENT and TRUST IN OTHERS- as independent variables. We also include gender (a
dummy with value 0 for males and 1 for females) and age (divided by 100) to control for individual differences.
TABLE 4 HERE
The results provide support for both hypotheses. They show that when respondents reveal more trust in government and other individuals, the probability that they will evade the personal income tax decreases significantly. The control variables also affect tax evasion, showing that the probability of older individuals and females to evade is lower. All these effects are significant at the 5% level, with the exception of age(at the 10% level).
6. Concluding remarks and policy implications
There is an abundant amount of research trying to understand the phenomenon of tax evasion. Nevertheless, it remains unclear why some societies experience substantially higher levels of tax evasion than others. In this paper we try to find the answer by first discussing the major theories and streams that have contributed to this field. It seems that the formal institutions put forward by the standard theory of tax evasion (i.e., tax and fine rate) and public choice (i.e., public goods) cannot fully explain differences across countries. In addition, according to both theories, it is rational for a citizen to free ride and not pay taxes
irrespective whether government provides the promised services and whether other citizens pay taxes. We suggest a complementary institutional approach by incorporating trust as one aspect of informal institutions. Trust can play an important role in the distinct levels of tax evasion in different countries. We specifically hypothesize that the level of (non)compliance is related to trust in the government fulfilling its promises and trust in others paying taxes. These hypotheses were tested using a unique data set from a household survey in Tirana, containing information on both, tax evasion and the informal institutions with respect to taxes. The test supports the hypotheses that when individuals trust the government and other individuals, they are more inclined to pay taxes. This result remains strong even after controlling for age and gender, which also affect individuals’ decision to evade taxes. Given the relevance of our findings to policy, we provide a few suggestions on what can be done. The framework outlined in this paper suggests that one important avenue of solution to the problem of tax evasion is to develop and emphasize the logic of democracy, which is that there is an exchange relationship between the citizen and the government. This relationship can only be based on mutual trust, since a government cannot be sued if it does not deliver on its promises. Building trust implies de-emphasizing the relationship of coercion, implicit in models of taxation, which neglect the expenditure side of government. One implication of this approach is that the problem of tax evasion will not be solved by punitive measures and may in fact be worsened that way. The empirical evidence on penalties for tax evasion (e.g., Andreoni et al. 1998) supports this point: the evidence that higher penalties are in fact an effective deterrent remains weak. This is in line with the work on criminal penalties which casts doubt on their effectiveness in solving the crime problem.15
One, obviously very difficult, line of reform is to take measures to increase trust in the government. Here there is a distinction to be made between the rule of law, and authoritarianism. Historically, the only strong governments Albania or Russia have had have been authoritarian governments. These societies have not had experience with a government which is strong and implements the rule of law, both for itself and for its citizenry, but which
were the same. The results are very similar.
is democratic. That is the only form of government which promotes voluntary tax compliance. The latter will occur because the rule of law will be enforced, including appropriate penalties for tax evasion, but within the parameters of democracy, where citizens feel their relationship to the state to be one of exchange and not coercion.
References
Allingham, M.G., and Sandmo, A., 1972. Income Tax Evasion: A Theoretical Analysis.
Journal of Public Economics 1, 323- 38.
Andreoni, James, Erard Brian and Feinstein Jonathan, 1998. Tax Compliance. Journal of
Economic Literature 36, 818-860.
Becker, Gary S., 1968. Crime and Punishment: An Economic Approach. Journal of Political
Economy 76, 169-217.
Benjamini, Y. and Maital, S., 1985. Optimal tax evasion and optimal tax evasion policy: behavioral aspects. In Gaertner, W., Wenig, A. (Eds.), The Economics of the Shadow
Economy. Berlin: Springer-Verlag, pp. 245-264.
Bordignon, Massimo, 1993. A Fairness Approach To Income Tax Evasion. Journal of Public
Economics 52, 345-62.
Breton, Albert and Wintrobe, Ronald, 1982. The Logic of Bureaucratic Conduct. New York: Cambridge University Press,.
Chhibber, A., 2000. Social capital, the state, and development outcomes. In: Dasgupta, P., Serageldin I. (Eds.), Social Capital: A Multifaceted Perspective. The World Bank, Washington. pp. 296-310,
Clotfelter, C.T., 1983. Tax Evasion and Tax Rates. Review of Economics and Statistics 65, 363-73.
Coleman, James, 1990. Foundations of Social Theory. Cambridge: Harvard University Press.
Cowell, F.A. and Gordon, J.P.F., 1988. Unwillingness to Pay. Journal of Public Economics 36, 305-21.
De Soto, Hermine, Gordon, Peter, Gedeshi, Ilir and Sinoimeri, Zamira, 2002. Poverty in Albania - A Qualitative Assessment. World Bank Technical Paper no. 520, The World Bank.
Feige, Edgar L., 1997. Underground Activity and Institutional Change: Productive, Protective and Predatory Behavior in Transition Economies. In Nelson, J.M., Tilley, C., Walker, L. (Eds.), Transforming Post-communist Political Economies, National
Academy Press, Washington, D.C., pp. 21-35.
Frey, Bruno S. and Weck-Hannemann, Hannelore, 1984. The hidden economy as an ‘unobserved’ variable. European Economic Review 26 (1), 33-53.
Frey, Bruno S., 2002. ‘Direct Democracy for Transition Economies’ , paper for the Collegium Budapest, Institute for Advanced Study.
Friedland, N., Maital, S., and Rutenberg, A., 1978. A Simulation Study of Income Tax Evasion. Journal of Public Economics 10, 107-16.
Gërxhani, Klarita, 2002. The Informal Sector in Transition: Tax Evasion in an Institutional Vacuum, Ph.D. thesis, University of Amsterdam. Tinbergen Institute Research Series, no. 265, Amsterdam: University of Amsterdam.
Gërxhani, Klarita, 2004a. The informal sector in developed and less developed countries: A literature survey. Forthcoming in Public Choice.
Gërxhani, Klarita, 2004b. Tax Evasion in Transition: Outcome of an Institutional Clash? Testing Feige’ s Conjecture in Albania. Forthcoming in the European Economic
Review.
Hjolland, Lena and Tingaard Svendsen, Gert, 2001. “ Standard Measurement of Social Capital: Preliminary Results from Russia” , paper delivered at the European Public Choice Society Meetings, Paris, April 2001.
Knack, Stephen and Keefer, Philip, 1997. Does social have an economic payoff? A cross- country investigation. Quarterly Journal of Economics 112 (4), 1251-88.
Mueller, Dennis, 1989. Public Choice II. New York: Cambridge University Press.
Myles, G.D., and Naylor Robin, 1996. A Model of Tax Evasion With Group Conformity and Social Customs. Journal of Public Economics 12, 49-66.
North, Douglas C., 1990. Institutions, institutional change and economic performance. Cambridge: Cambridge University Press.
Oleinik, Anton, n.d., “ A Trustless Society: the influence of the August 1998 crisis on the institutional organization of the Russians’ everyday life” .
Paldam, Martin and Gert Tinggaard Svendsen, 2000. An Essay on Social Capital.
Looking for the fire behind the smoke. European Journal of Political Economy 16 (2), 339-366.
Putnam, Robert, 1993. Making Democracy Work. Princeton, NJ: Princeton University Press. ____________, 2000. Bowling Alone. New York: Simon and Schuster.
Rose, R., 2000. Getting things done in an anti-modern society: Social capital networks in Russia. In: Dasgupta, P., Serageldin I. (Eds.), Social Capital: A Multifaceted
Perspective. The World Bank, Washington. pp. 147-172.
Schneider, Friedrich and Enste, Dominik H., 2003. The shadow economy: an international
survey. Cambridge: Cambridge University Press.
Scholz, John T. and Lubell, Mark, 1998. Trust and taxpaying: Testing the heuristic approach to collective action. American Journal of Political Science 42 (2), 398-417.
Wintrobe, Ronald, 2001. “ The Economics of Group Behavior” . Paper delivered at the 2001 European Public Choice Society Meetings, Paris, April 2001.
Wittman, Donald, 1995. The Myth of Democratic Failure. Chicago: University of Chicago Press.
Yitzhaki, S., 1974. Income Tax Evasion: a Theoretical Analysis. Journal of Public
Table 1: The average size of the shadow economy in developed and less developed countries
Countries/continents Size as % of GNP*
Developed OECD countries
Transition Former Soviet Union
Middle and Eastern Europe
Developing Africa Latin America Asia 12 25 20 44 39 35
Source: Based on Schneider and Enste (2003, p.37, table 4.5). (*) derived from Physical input method, 1989-1993
Table 2: The extent of personal income tax evasion
PITE (personal)
Number of cases Percentages
No tax evasion at all 544 61.4
One indication of evasion 146 16.5
Two indications 76 8.6
Three indications 68 7.7
Four indications 43 4.9
Five indications 9 1.0
Total 886 100
Table 3: The distribution of responses to the statements
Strongly
Agree Mildly Agree Neither Agree nor Disagree
Mildly
Disagree Strongly Disagree Total cases I do not feel like paying taxes as long as the government cannot
be trusted
27.2 19.0 14.6 15.9 23.4 1254
Table 4: A direct test of the ‘trust’ hypotheses PITE TRUST IN GOVERNMENT -.239 (0.000)* TRUST IN OTHERS -.140 (0.045)* AGE -.015 (0.079)** GENDER -.460 (0.008)* # observations 886
(*) indicates statistical significance at the 5% level; (**) indicates statistical significance at the 10% level.
Note: Numbers represent the regression coefficient; p-values in parentheses.
Figure 1 B S S¹ Sº D p p¹ p L
S S² Sº D p p¹ p A B Figure 2
