Cooperating Estonians and Exiting Lithuanians: Trust in Times of Crisis

Transcription

1 Cooperating Estonians and Exiting Lithuanians: Trust in Times of Crisis This paper argues that substantial differences in political legitimacy can help explain why Estonia dealt with the recent economic crisis more successfully than Lithuania. In the last crisis, Lithuania saw its budget deficit expand substantially, Estonia managed to keep the deficit under 3 percent of GDP and consequently was invited to join the eurozone as of The experience of these countries presents an interesting puzzle, as the divergent fiscal performance cannot be attributed to purely economic factors. Both countries have similar economic structure, and both were similarly affected by the crisis. Furthermore, both pursued similar expenditure and tax policies during the crisis. Based on quantitative and qualitative evidence, it is argued that higher tax compliance and subsequently higher tax revenues can explain the difference. In turn, this compliance gap can be attributed to different level of trust in political institutions in Estonia and Lithuania. As both countries implemented harsh austerity measures, Lithuanian society did not present open contestation (in terms of protests or electoral backlash), but resisted (or at least did not support the consolidation policy) by turning to shadow economy (as well as emigrating in large numbers). This also explains how Lithuanian state can at the same time be strong (in terms of reacting fast and possibly radically) but at the same time its efforts are often undermined in areas that require good state-society relationships. The discussion has several important implications. First, shadow economy, lack of tax compliance and emigration can be important indicators of legitimacy (as a form of exit ). Second, the importance of legitimacy might actually increase in hard times. Third, times of crisis might encourage the development of different equilibria (high trust-high performance virtuous cycle as well as low trust-low low performance one). 1

2 Introduction You only find out who is swimming naked when the tide goes out Warren Buffett (Norris, 2008) Taxation has been at the core of establishing modern states and maintaining their successful operation. The ability to extract revenue from a population has been used as an indicator of a state s capacity. This capacity is not a direct function of coercive capacities, however, but has been found to rely heavily on voluntary compliance on the part of population. In this light, the importance of trust and legitimacy arises. The quality of the relationship between the state and its citizens (as reflected in political trust) as well as among citizens themselves ( generalized trust) are obviously important aspects per se, but there are also important reasons to treat them as preconditions for successful economic performance in general and good fiscal results in particular. The motivation for this paper comes from the diverging experience among the Baltic countries in dealing with the financial crisis of The conventional wisdom attributes Estonia s success to two factors: the fiscal reserve accumulated before the crisis and more successful restraint in expenditure during the downturn. Either way, the primary importance is attributed to better decisions on the part of Estonian government before and during the crisis. While these factors undoubtedly contain important elements of truth, they also provide incomplete explanations for the Baltic divergence. This paper argues that Estonia s better fiscal results during the crisis can be primarily attributed to differences in government revenue, in particular tax revenue. Subsequently, the case is made that these differences are in turn explained to a significant extent by differences in the level of trust in the Baltic societies. The primary importance is attributed to political trust, i.e. the extent to which citizens trust national political institutions. It is argued that Estonia had accumulated higher resources of political trust and has managed to build a social contract of higher quality between the state and society, which enabled the pursuit of more responsible fiscal policy before the crisis as well as during it. Hence, the words by Warren Buffet on the tide revealing the extent of swimmers clothing: while precrisis boom years masked differences between the Baltic countries, the subsequent hard times have separated the clothed from the naked. The paper is organized as follows. First, the divergence in Baltic countries performance is described. It is also argued that this divergence cannot be attributed to structural-economic factors. Subsequently, a theoretical case is made for the importance of trust ( political and generalized ) for fiscal performance. The next section presents comparative evidence on the levels of trust in Estonia and Lithuania. This evidence is supplemented by quantitative analysis relating trust and tax morale. The next section discusses Estonian state s better performance in terms of providing public goods, fighting corruption and ensuring fiscal 2

3 discipline. Following that, other indicators of support (or lack of it) are provided including electoral outcomes, emigration and protest activitiy. The paper ends with discussion and conclusions. Baltic divergence The three Baltic countries faced enormous economic challenges during the last global economic downturn. As a result of huge macroeconomic imbalances prior to the crisis, they suffered double-digit declines in economic output. In 2009, the countries that registered the biggest economic contractions in the world were Latvia, Estonia, Ukraine and Lithuania respectively. The three Baltic countries also adopted the same anti-crisis strategy called internal devaluation. This refers to an economic adjustment policy which aims to safeguard fixed exchange regimes while pursuing fiscal consolidation in order to achieve fiscal sustainability but also to bring down costs and wages within the economy (Purfield and Rosenberg, 2010). Also, all three countries experience is puzzling in the sense that they managed to implement internal devaluation strategies despite opposite predictions of numerous foreign analysts and economists, including Paul Krugman, Nouriel Roubini and Kenneth Rogoff (Kuokštis and Vilpišauskas, 2010). At the same time, their performance diverged in an important respect. Specifically, Estonia was the first one to bring its budgetary situation under control. As a result of successful deficit reduction, it was the first one to restore investor confidence and was also able to satisfy Maastricht criteria, thus being able to join the eurozone from Remarkably, in 2009 facing an economic decline of around 15 percent it managed to reduce its budget deficit compared to In contrast, both Latvia and Lithuania registered significant increases in budgetary deficits in 2009 (see Figure 1). 3

4 0-2 Estonia Latvia Lithuania Figure 1. Public balance (budgetary deficits) in the Baltic countries in in 2008 (blue) and 2009 (red). Source: Eurostat. The next question to ask is whether these trends can be attributed to expenditure or income changes. Perhaps Estonians were able to cut budgetary expenditure more forcefully? Figure 2 rejects this explanation. In fact, Latvia reduced its expenditure much more, while Lithuania cut its expenditure slightly more. Therefore, we can attribute the Baltic divergence to changes in budgetary revenue. Indeed, Figure 2 reveals important differences in how tax revenue declined within the Baltic countries: tax revenue fell by 4.2, and 19.7 percent in Estonia, Latvia and Lithuania respectively. 0-0,01-0,02 Estonia Latvia Lithuania 0-0,05 Estonia Latvia Lithuania -0,03-0,1-0,04-0,05-0,06-0,15-0,2-0,07-0,08-0,09-0,25-0,3 4

5 Tax revenue growth Figure 2. Budget expenditure (left) and revenue (right) changes in percentage in 2009 compared to Source: Eurostat; author s calculations. What makes these differences particularly interesting are the overall similarities in economic vulnerabilities and trends within the Baltic countries during the last crisis. Latvia was clearly the most overheated economy and also the only one to suffer a banking crisis and thus suffered the biggest economic contraction (Purfield and Rosenberg, 2010). However, Estonia and Lithuania found themselves in a very similar situation, as revealed by almost identical real GDP (as well as domestic demand) developments during 2009: real output in Estonia declined by 14.3 percent, and the corresponding figure for Lithuania was 14.7 percent. An even more accurate illustration for the puzzle emerges after considering the general trends in EU member countries. Figure 3 shows the relationship between economic activity change (nominal GDP) and tax revenue change in EU member states. It can be seen that neither of the Baltic countries is on the regression line. Latvia and Lithuania are both below the line, while Estonia is above it. In other words, Estonia managed to extract more revenue from the economy than predicted by the general model, while Lithuania and Latvia underperformed. Yet in other words, all Baltic countries are deviant (Estonia positively, Lithuania and Latvia negatively) cases, which makes them especially interesting cases for comparative analysis y = 0,3488x - 7,3791 R² = 0, EE LT -20 LV Nominal GDP growth Figure 3. Nominal GDP change (X axis, in percentage) and tax revenue change (Y axis, in percentage) in 2009 compared to 2008 in EU member countries. Source: Eurostat, author s calculations. The last possible explanation based on economic-structural conditions to consider is related to tax system changes in Perhaps Estonia increased taxes more than Lithuania did? It did not: both Estonia and Lithuania increased tax rates in a similar fashion during 5

6 Adjusted total tax revenue growth the crisis. Both countries raised their VAT, excise as well as social security contribution rates. Furthermore, Picture 4 takes into account discretionary tax changes and therefore arguably presents a better picture of the relationship between GDP decline and tax revenue. Indeed, it can be seen that the predictive power of the model increases (R-squared jumps from 6 percent to 34 percent), while the conclusions regarding the Baltic countries prevail. As before, Estonia is above the line, while Lithuania and Latvia below it ,00% -20,00% -15,00% -10,00% -5,00% 0,00% 5,00% -5 EE LT y = 66,727x - 1,7279 R² = 0, LV Nominal GDP growth -35 Figure 5. Nominal GDP change (X axis, in percentage) and adjusted tax revenue change (Y axis, in percentage) in 2009 compared to 2008 in EU member countries. Source: Eurostat; Taxation trends in the EU; author s calculations. In summary, it has been so far established that the big difference in tax revenue (and in turn budget deficit developments) between Estonia and Lithuania cannot be explained by economic-structural conditions, i.e. economic trends or discretionary tax system changes. 6

7 Estonia Lithuania Figure 6. VAT gaps in Estonia and Lithuania, as percentage of GDP. Source: (Center for Social and Economic Research, 2013). There are good reasons to suggest that a significant part of reduction in tax income was due to lower compliance and increase in shadow economy. Although far from a perfect measure, VAT gaps are often used to gauge the extent of (non)compliance. They are defined as the difference between the theoretical VAT liability and the collections of VAT (Center for Social and Economic Research, 2013, p. 10). VAT gaps are also a good way to compare countries because VAT tax systems are usually similar. Also, comparative data on VAT gaps exists. Furthermore, VAT tax revenue comprises a high share of budget revenue in all three Baltic countries. Finally, tax changes in Lithuania and Estonia were similar during the crisis, as both countries increased VAT rates and eliminated exemptions (Center for Social and Economic Research, 2013, p. 14). Figure 6 above reveals the developments of VAT gaps in the two Baltic countries prior to and during the Great Recession. Even prior to the crisis, Estonia displayed significantly lower level of VAT gap than Lithuania. Estonia belongs to the group of countries with the lowest VAT gap, and Lithuania has one of the highest ones in the EU. As can be seen from the figure, this divergence between Estonia and Lithuania grew even further during the downturn. In Lithuania, VAT gap jumped from 3.2 percent of GDP in 2007 to 4.9 percent in 2009, and then declined somewhat to 4.4 percent in Conversely, in Estonia VAT gap only increased from 0.9 percent in 2007 to 1.3 percent in 2009, and then up to 1.9 percent in In fact, the gap even decreased from 2008 to

8 Trust and fiscal performance: theory Why do people pay taxes? There are two main perspectives explaining tax compliance have emerged. The first one, which can be referred to as the deterrence theory, stems from a larger paradigm of an economic approach to human behavior (Becker, 1978). It explains tax compliance as an act based on an evaluation of expected benefits and costs of tax compliance versus tax evasion. Expected non-compliance costs are calculated as the probability of getting caught multiplied by the potential loss from non-compliance (a fine or a sentence). At the same time, this perspective runs into a number of important problems. First and foremost, the level of tax compliance observed in many countries is way too high to sit well with the theory s predictions. As was put by Joel Slemrod, the puzzle is not to explain why people evade, but rather why people pay taxes in the context of the standard economic model, people who voluntarily comply are exhibiting nothing short of pathological honesty (Slemrod, 1998, p. 485). In other words, the probability of getting caught in reality is much lower than what would be necessary to justify the levels of tax compliance that can actually be observed in many countries. Besides, researchers have found that countries with similar levels of tax enforcement efforts and compliance costs manage to collect different amounts of tax revenue (Cummings et al., 2006). This has led researchers to consider other important factors that may come into play while explaining tax compliance, in particular social norms related to trust. One important perspective relates the level of tax compliance with trustworthiness of government. It views the act of paying taxes as a form of social contract, whereby citizens comply providing funds for the state, and the latter provides public goods and services for the citizens. Based on this perspective, one could expect to observe variations in tax compliance with the quality of public goods, fairness of government and tax collection procedures and the associated notion of trust in political institutions ( political trust). Importantly, in view of this perspective, more active tax enforcement measures can actually backfire if they reduce government s legitimacy. In line with this expectation, some studies have actually found a negative relationship between tax enforcement measures and tax compliance (Frey and Torgler, 2007). Furthermore, while citizens might trust the government to provide public goods and deliver on its promises, they might still not expect other citizens to comply. Tax collection potentially poses huge collective action problems. A person, upon successful evasion of taxes, might still enjoy the benefits of a public good and thus free-ride. This observation gives birth to the third perspective that aims to explain tax compliance by trust in others ( generalized trust). This perspective could also be called conditional cooperation citizens cooperate (pay taxes) if they expect others to do the same. It must be noted that empirical findings in the literature generally support the claims of the sociological perspective, although evidence for political trust is stronger than for generalized trust 8

9 (Uslaner, 2003). Generalized trust differences Do Estonians trust others more than Lithuanians do? The evidence is mixed. Based on data from , Estonians did not trust others more than Lithuanians. However, certain surveys from 2004 and later reveal that Estonians do trust (or have started to trust) others more than Lithuanians 2. According to European Values Survey (EVS) of 2008, 32.6 percent of Estonians answered that most people can be trusted and 67.4 percent declared that one cannot be too careful when dealing with other people. The corresponding figures for Lithuania were 29.9 percent and 70.1 percent. Nevertheless, differences between Lithuanian and Estonian population means is statistically insignificant. At the same time, Estonians are more trusting of others with regard to two other questions included in the EVS that can be related to general trust, and these differences are statistically significant. To the question whether others try to take advantage of others most of the time, 9.9 percent of Lithuanians answered positively and only 3.6 percent of Estonians did. Besides, as many as 19.4 percent of Lithuanians declared that most people tend to look after themselves and only 8.4 percent of Estonians reached this conclusion. Overall, however, it does not seem that general trust difference between Estonia and Lithuania are large. Political trust differences There are significant differences in terms of institutional quality and political trust between Estonia and Lithuania. Table 1 below provides measures of political trust for the three Baltic countries. As can be seen from the Table, there are important differences among the Baltic States. In Estonia, trust in government actually surpasses the average rate for the EU, while Lithuania (and Latvia) society s attitude towards the government could be described as one of prevailing mistrust. Table 1. Trust in national institutions in the Baltic countries Legal Autumn 07 Spring 08 Autumn 08 Spring 09 Autumn 09 Spring 10 Autumn 10 1 European Values Survey Special Eurobarometer

10 system Estonia Latvia Lithuania Government Estonia Latvia Lithuania Parliament Estonia Latvia Lithuania Political parties Estonia Latvia Lithuania Respondents who tend to trust a given national political institution, in percentage. Source: Eurobarometer. A look at another source Global Competitiveness Report published by World Economic Forum, confirms Eurobarometer's data. Apart from trust in politicians indicator, other aspects are included that are closely related to society's evaluation of government's competence and fairness. In every relevant category related to citizens' trust in government and evaluation of governmental procedures for decision-making as well as spending budgetary funds, there are major differences between Estonia on the one hand and Lithuania (as well as Latvia) on the other. Remarkably, Lithuania according to wastefulness of government spending managed to get the 117th place out of 139 countries surveyed, while Latvia took 120th place (2010 report). Estonia was way higher in the 49th place. This means that Lithuania and Latvia were surpassed by countries such as Angola, Zimbabwe, Russian Federation, and Azerbaijan. This might point to Lithuanian and Latvian societies negative bias in evaluating the performance of their government. In other words, perception of the level of government incompetence and corruption might actually be worse than the real situation. In terms of government s (lack of) legitimacy, perceptions matter more than objective facts. Table 2. Quality of institutions according to Global Competitiveness Report Estonia Latvia Lithuania Property rights

11 Diversion of public funds Public trust of politicians Favoritism in decisions of government officials Wastefulness of government spending Transparency of government policymaking Source: World Economic Forum, Global Competitiveness Report (2009). The numbers in the table indicate the place of the country among 133 countries covered in the Report. EVS also includes data on political trust. In particular, it includes the question on citizens trust in parliament, civil service, social security system, justice system, political parties, government, as well as the question on overall satisfaction with democracy and the level of pride in one s country (pride has been found to be a significant predictor in tax compliance studies) (Alm et al., 2006). On all of these measures, Estonia scores higher than Lithuania. Estonian government s better performance People s willingness to trust the government with their money depends on their judgement on how well this money would be spent. People are more likely to pay taxes if they think the government is competent, responsible and sticks to its promises. In Estonia, the society has more ground for trusting the government than in Lithuania (and Latvia). One reason for this is the difference in fiscal management. All three Baltic countries witnessed very high economic growth rates prior to the crisis. Despite these favourable circumstances, only Estonia managed to keep fiscal deficits under control. In fact, it achieved fiscal surpluses. To the contrary, both Latvia and Lithuania did not have a single year with a budget surplus (see Figure 7). From a public choice perspective, Estonian case is more surprising given the universal expectation that politicians will by prone to induce political business cycles. 11

12 3,0 2,0 1,0 0,0-1, Estonia Latvia Lithuania -2,0-3,0-4,0 Figure 7. Public balance (surplus/deficit) in the Baltic countries, in percentage of GDP. Source: Eurostat. This meant that Estonia came into the crisis more prepared. It had accumulated a sizable fiscal reserve, and was thus not forced to borrow from foreign creditors (like Latvia) or from the bond market (like Lithuania). Equally or perhaps even more importantly, Estonian government could expect higher support from its population during the crisis because it had delivered prior to it. In contrast, people in Latvia and Lithuania largely attributed the crisis to the recklessness of their governments, and thus were less likely to comply with austerity policies. On a broader level, the Estonian state has delivered better results than its Lithuanian (and Latvian) counterpart. Importantly, all three started at a similar level of economic and institutional development in the early 1990s. Nevertheless, Estonia implemented structural reforms earlier and more consistently. In consequence, it emerged as the leader of the pack. It was the first country to start negotiation talks with the EU, to be invited to the OECD, and finally join the eurozone. On many economic development indicators, such as real wages and FDI, Estonia is doing better. Finally, it scores significantly higher on fighting corruption. As Figure 8 reveals, Estonia along with Slovenia have the lowest corruption numbers among the new member states. Lower level of perceived corruption in Estonia is yet another reason for people to be more trusting of the government and more willing to part with their money in paying taxes. 12

13 Slovenia Estonia Czech Republic Hungary Latvia Slovakia Lithuania Poland Figure 8. Corruption perception index in 2008, on scale from 1 to 10, 10 being the lowest corruption. Source: Transparency International, The relationship between trust and tax morale: individual level data The hypothesis that fiscal performance is driven by differences in political trust is based on the argument that trust affects tax morale i.e. citizens willingness to comply with the tax code. Indeed, data reveals that Estonians have higher tax morale. 13

14 Estonia 30 Latvia Lithuania never always Picture 5. Tax morale in the Baltic countries, attitudes towards cheating on tax. Source: European Values Suvery EVS includes the question do you justify teaching on tax?. Based on data from 2008, in Estonia people were least tolerant of tax cheating (tax morale was highest) among the Baltic States. The proportion of respondents that chose the answer never was 53 percent in Estonia, just above 40 percent in Latvia and 35 percent Lithuania. The differences are statistically significant: at 0.01 significant level, the Mann-Whitney test rejects the hypothesis that there is no difference between Estonian and Lithuanian populations regarding tax morale. As has already been mentioned, EVS includes questions answers to which can be used as measures of political trust, general trust as well as tax morale. This provides opportunities to check whether there is a significant relationship between trust (both political and general ) and tax morale. Table 3 provides correlation measures for different items on political and generalized trust. Since the data is ordinal and does not satisfy normal distribution requirements, Spearman s and Kendal tau s correlation tests were used instead of Pearson s correlation coefficient. Table 3. Correlation coefficients between tax morale and different dimensions of trust 14

15 Spearman s correlation Kendall tau correlation coefficient with tax morale coefficient with tax morale Political trust Confidence in parliament 0.092** 0.081** Confidence in civil service 0.134** 0.119** Confidence in social security system 0.124** 0.111** Confidence in justice system 0.115** 0.101** Confidence in political parties Confidence in government 0.120** 0.106** Satisfaction with democracy 0.105** 0.092** Pride in one s country 0.216** 0.189** Generalized trust Other people can be trusted Most people try to be fair 0.135** 0.110** Most people try to be helpful 0.108** 0.089** Source: European Values Survey, 2008; author s calculations, ** refers to significance at 0.01 level. Concerning political trust, results are overall consistent with theoretical expectations. All correlation coefficients are of the expected sign and all but one confidence in political parties are significant. The latter observation is not particularly surprising: trust in political parties can be expected to be heavily driven by various motivations, such as ideological considerations, that are not expected to correlate significantly with evaluations on the quality of public goods provided. As for generalized trust, evidence is mixed. On the one hand, the most direct measurement of generalized trust opinion on whether other people can be generally trusted does not correlate significantly with tax morale. At the same time, however, the two other measures do have significant correlation coefficients with tax morale and are of the expected sign. Since it was established that Estonia does score higher on trust measures (statistical tests also confirm the significance of these differences), and since these measures correlate positively and significantly with tax morale, one can expect that at least part of tax morale gap between Estonia and Lithuania can be explained by the gap in the trust capital. More robust answers can be provided by regression analysis. 15

16 Because the dependent variable (tax morale) is ordinal and not dichotomous, ordered probit (or logit) estimation is necessary. The dependent variable (question on justification for tax compliance in European Values Survey from 2008) has been recoded to take on four values (1, 2, 3, 4), where 1 refers to highest score on tax morale (tax cheating is never justified), while value 4 incorporates all values from 10 to 4. The expectations before performing regression analysis are the following: 1. Being an Estonian has a positive and significant influence on tax morale. 2. Both political and generalized trust variables are expected to have positive signs and be statistically significant (except for variables people try to take advantage most of the time and people tend to look after themselves, which are expected to have negative influence on tax morale). 3. Upon including into the regression measures of political and/or generalized trust, the coefficient on the dummy variable representing countries gets lower and/or loses its significance. If this is the case, it would lend support to the interpretation that the difference between Estonians and Lithuanians in tax morale can be partly or totally explained by differences in existing political and generalized trust in the two countries. Table 4. Regression equations where the dependent variable is tax morale Model 1 Model 2 Model 3 Model 4 Country dummy (LT=0; EE=1) *** *** *** *** Confidence police Confidence parliament Confidence civil service in in in 0.142*** 0.141*** ** Confidence in social security system Confidence in healthcare system * Confidence justice system in

17 Confidence in political parties Confidence government in Satisfaction with democracy Pride in one s country *** 0.263*** Trust in people ** * People try to take advantage People look after their own interests *** *** * Chi-squared *** *** *** *** Pseudo R-squared (Nagelkerke value) Source: European Values Survey 2008; author s calculations. *** indicates significance at 0.01 level, ** at 0.05 level and * at 0.1 level. Table 4 presents four equations with different combinations of independent variables. The first one includes only country dummy variable and, as expected, it predicts statistically significant differences in tax morale for Estonians and Lithuanians (since this is ordered probit regression, no direct interpretation of the coefficient size (magnitude) is available). Equation 2 includes various measures of political trust. The only two significant political trust variables are confidence in police and pride in one s country with the expected positive signs. Other variables are insignificant and some of them have unexpected signs. However, it must be mentioned that all trust variables are highly correlated among themselves. Separate regression analyzes have been performed including each of these variables alone, which yielded positive and statistically significant coefficients (regressions are not reported for matters of convenience). As expected, country dummy in equation 2 is lower, although it is still negative and statistically significant. This means that some, but not all, of the difference in tax morale between Estonia and Lithuania can be attributed to differences in political trust. 17

18 The third equation includes country dummy variable and generalized trust variables. In contrast to political trust variables, they do not perform as well (pseudo-r squared does not increase as much, and larger part of the difference between Estonia and Lithuania is left unexplained). All variables are significant, and the coefficients on variables people try to take advantage and people look after their own interests have the expected negative signs. An unexpected finding is that trust in others is negatively related to tax morale. In the final equation, all variables are included. Confidence in police, civil service and social security system, pride in one s country, and opinion on whether others are trying to take advantage have the expected and statistically significant signs. Trust in others is still negative and statistically significant. Finally, the country dummy stays negative and statistically significant, although it is smllaer than in the first equation. This means that part of the difference in tax morale between the two countries can be explained by trust differences, but there is still a portion of difference left unexplained. One interpretation of this finding could be that Estonians are somehow culturally more inclined towards cleaner attitudes (cultural explanation). The trouble with this hypothesis is that culturally-determined attitudes could be expected not to vary a lot and to have also existed at the beginning of transition, but data from early and even late 1990s reveal no differences in tax morale between Lithuania and Estonia 3. Other indicators of support What are other indicators of differences in legitimacy during the economic adjustment period? First of all, one could look at protest activities. The first thing to note here is that in general protest activities during the economic downturn in all three Baltic countries were surprisingly low. There is a robust positive relationship between austerity policies (especially ones concentrating on expenditure cuts, which was exactly the case in the Baltics) on the one hand and various forms of social turmoil on the other (Ponticelli and Voth, 2011). Given a very high level of economic contraction in the Baltic and given sizeable expenditure-based fiscal consolidation, one could have predicted a significant rise in protest activities. However, this largely did not happen. Some authors (Aslund and Dombrovskis, 2011) have attributed this to the broad level of support for the government s policies. However, figures provided above about the level of trust in government call into question such an interpretation in the cases of Latvia and Lithuania. Arguably, a better explanation would be that Latvian and Lithuanian populations did not support the government and its policies during the crisis, but instead of providing open contestation ( voice ) towards it, they opted for exit. One form of exit, as mentioned, has been the turn to shadow economy and lack of tax compliance. 3 European Values Surveys 1998/

19 Furthermore, although the general level of protest in the three Baltic countries was low, there was variation within them. In Estonia, there were almost no significant protests. In contrast, Latvia and Lithuania saw the biggest level of protests since regaining independence. In Latvia, around 1000 medical workers staged a protest in the Autumn of Furthermore, in January of 2009, a large protest rally occurred in the country and it eventually turned violent. Finally, a large, although nonviolent, protest happened in June of Following the January events in Latvia, a protest also occurred in Lithuania. The Lithuanian protest also included violent manifestations, as a group of people started throwing stones at the Parliament building. Finally, electoral results of the ruling parties and leaders support the interpretation as well. In Lithuania, the popularity of the political parties in the ruling coalition decreased enormously during the crisis. An interesting situation developed where outside analysts and politicians (for instance, EU and IMF representatives) were praising Lithuanian governments efforts in tackling the crisis but Lithuanians themselves turned against the ruling coalition (Vilpišauskas et al., 2014). In the end, the ruling parties lost the election in 2012 and the opposition came to power. At the same time, it must be noted that this pendulum pattern has been characteristic of Lithuanian political history, and is not peculiar to the crisis period. However, the Lithuanian situation still stands in stark contrast to the Estonian. In Estonia, government s and prime minister s popularity actually increased during the crisis, and the main party in the ruling coalition increased its number of seats (by two in a parliament of 101 members) in the national parliamentary elections in In the wake of Estonia s election, The Economist argued that the country managed to bust another stereotype (The Economist, 2011). These two stereotypes were the inability of a democratic country to achieve such dramatic fiscal consolidation and the inability of a government implementing such policies to get reelected. Last but not least, one could also take a look at emigration levels from Estonia and the other Baltic countries (see Table 5). One should always carefully treat migration numbers as they are notoriously unreliable. Furthermore, the Lithuanian figure or 2009 is artificually inflated due to the new requirement to pay compulsory health insurance contributions, which prompted many actual emigration to declare there emigration officially. Nevertheless, one could rather confidently conclude that the crisis sparked a subtantially higher emigration wave in Lithuania (as well as Latvia) than Estonia. Even more than tax non-compliance, emigration is often considered a form of exit (instead of voice ) in expressing one s dissatisfaction with the way things are going in one s environment. Table 5. Net migration from the Baltic countries Estonia Latvia Lithuania

20 Increase in VAT gap Crude rate of net migration plus statistical adjustment. Source: Eurostat. Disccusion and conclusions In this paper, it was argued that fiscal policy results are influenced by the relationship between the government and society, namely the level of trust in political institutions. The evidence provided weaker support for the argument that generalized trust has an important influence. It was shown that economic-structural factors (economic vulnerabilities, trends or economic policy changes) cannot adequately account for the divergence in government income between Estonia and Lithuania during the crisis. The conclusions of this paper therefore carry implications for the study of the relationship between trust and fiscal performance. More specifically, based on the argument presented in this paper, one could argue that trust s importance for fiscal performance actually increases during hard times. In other words, the relationship between trust and compliance (and thus fiscal situation) might be conditional on the economic environment. While further investigation of this novel hypothesis is both necessary and desirable, one could present some evidence in support of it. Figures 6 and 7 below depict the relationship between trust in parliament in 2008 and increase in VAT gap as well as tax morale and VAT gap during the crisis from 2008 to It shows that overall VAT gaps increased during the economic downturn, but this increase was smaller in countries with higher level of political trust and higher tax morale. 3,5 3 2,5 2 1,5 1 0,5 0-0,5-1 y = -0,0241x + 1,4383 R² = 0, Trust in parliament Figure 6. Relationship between trust in parliament (percentage of people who tend to trust national parliament, Spring of 2008) and increase in VAT gap between 2008 and Sources: Eurostat for trust figures and Center for Social and Economic Research for VAT gap estimates. 20

21 Increase in VAT gap 3,5 3 2,5 2 1,5 1 0,5 0-0,5-1 y = -0,0337x + 2,4702 R² = 0, Tax morale Figure 7. Relationship between tax morale (percentage of people who say that to trust national parliament) and increase in VAT gap between 2008 and Sources: Eurostat for trust figures and Center for Social and Economic Research for VAT gap estimates. On another level, the paper s findings mean that explaining differences in economic performance among the Baltic States as well as improving future conditions in Lithuania (and Latvia) depends not only on economic-structural conditions and economic decisions. If important economic problems leaving social and political normative issues aside confronting Lithuania in terms of balooning public debt and other fiscal issues are to be resolved successfully, the quality of government-society relationship must improve. Divergence in informal institutional quality has led to the situation where Lithuania and Latvia, in contrast to Estonia, are increasingly relying on other sources of state revenue instead of taxation, be it in the form of borrowing from foreigners or using EU structural funds. Lower level of trust and associated poorer tax income inter alia speak of lower state capacity in Lithuania and Latvia. Furthermore, for both Latvia and Lithuania, there is a danger of a vicious cycle involving low trust (and low tax morale) along with worse fiscal performance, mutually feeding each other. At the same time, there is nothing inevitable about such a vicious cycle. Historical record shows that it is possible to break them and perhaps even build a virtuous cycle. For instance, Martin Daunton has demonstrated how the British state in the 19 th century was able to overcome a situation of deep mistrust and skepticism on the part of its population, effectively building legitimacy and consequently achieving high compliance form the taxpayers (Daunton, 2007, 1998). 21

22 References Alm, J., Martinez-Vazque, J., Torgler, B., Russian attitudes toward paying taxes before, during, and after the transition. Int. J. Soc. Econ. 33, Aslund, A., Dombrovskis, V., How Latvia came through the financial crisis. Peterson Institute, Washington, D.C. Becker, G.S., The Economic Approach to Human Behavior. University Of Chicago Press. Center for Social and Economic Research, Study to quantify and analyse the VAT Gap in the EU-27 Member States. Final report (Commissioned by the European Commission (DG TAXUD)). Cummings, R.G., Martinez-Vazquez, J., McKee, M., Torgler, B., Effects of tax morale on tax compliance: experimental and survey evidence. Daunton, M., Trusting Leviathan: British Fiscal Administration from the Napeolonic Wars to the Second World War, in: Levi, M., Braithwaite, V. (Eds.), Trust and Governance. Russell Sage Foundation, pp Daunton, M., Trusting Leviathan: The Politics of Taxation in Britain, Cambridge University Press. Frey, B.S., Torgler, B., Tax morale and conditional cooperation. J. Comp. Econ. 35, Kuokštis, V., Vilpišauskas, R., Economic Adjustment to the Crisis in the Baltic States in Comparative Perspective. Presented at the 7th Pan-European International Relations Conference, Stockholm, Sweden. Norris, F., Buffett s State of the World: There s Folly in Wonderland. N. Y. Times. Ponticelli, J., Voth, H.-J., Austerity and anarchy: Budget cuts and social unrest in Europe, CEPR Discuss. Pap. No. DP8513. Purfield, C., Rosenberg, C., Adjustment under a Currency Peg: Estonia, Latvia and Lithuania during the Global Finanacial Crisis Slemrod, J., On voluntary compliance, voluntary taxes, and social capital. Natl. Tax J. 51, The Economist, Estonia busts another stereotype. East. Approaches. Treisman, D., Russia s Tax Crisis: Explaining Falling Revenues in a Transitional Economy. Econ. Polit. 11, doi: / Uslaner, E.M., Tax evasion, trust, and the strong arm of the law, in: Trust, Institutions, and State Capacities: A Comparative Study. Presented at the Tax Evasion, Trust, and State Capacity, University of St. Gallen, St. Gallen, Switzerland, pp Vilpišauskas, R., Nakrošis, V., Kuokstis, V., Politics of reacting to the crisis in Lithuania in : exiting the crisis, entering politics as usual?, in: Bukovskis, K. (Ed.), Latvian Institute of International Affairs, Riga. 22