People have different preferences and tastes about how much a government should redistribute between the rich and the poor. For example, Figure 1 shows that about 80% of Spaniards in a representative survey believe that taxes on low-income earners are currently too high. But where does this come from and what creates the differences in people’s preferences for redistributive polices? Why do some people believe taxes are too low for the poor, while the majority does not? Evidence has shown that individuals tend to be more generous towards one another within ethnic, linguistic, cultural, religious, and nationality groups (Alesina and Stantcheva, 2020). The willingness to redistribute income between the rich and the poor depends on diversity along those lines. But do people really have the information they need to evaluate the tax system? Are people even aware of how much of their income is taxed by the government? And do people know which government—central or autonomous—taxes their income? This project first assesses how informed citizens are with regard to taxation. Then a simple experiment is implemented: the relevant information is provided to one group of participants and their attitudes towards taxation, inequality, and tax evasion are compared to those of the group which has not received the information. 

It is particularly interesting to implement this experiment in Spain because of the high level of diversity across regions and the strong regional identity of citizens. The study combines this feature with the institutional setting of personal income tax in Spain, in particular the fact that it has been partially ceded to the regional governments. Half of the personal income tax base is taxed at the level of the autonomous communities, and revenues remain directly at that level of government. This feature, however, is not very salient to citizens. The survey shows that only approximately 10% of respondents are aware of this detail. The first treatment provides this information to participants and allows estimation of the extent to which the knowledge that a large part of the tax collected remains within their region influences individuals’ preferences for redistributive policies.

Research has also shown that the information people have about their own position in the income distribution shapes preferences for redistribution (Cruces et al., 2013; Fernández-Albertos and Kuo, 2018). However, given people’s position in the income distribution, an important element is the information they have about the level of tax they currently pay. Citizens oftentimes have a strong misperception about the real tax burden they face. Due to the complexity of the tax system many taxpayers are not able to accurately guess the share of taxes they pay. In other words, it is difficult for most citizens to assess how much they contribute to the government. The study simulates the true tax burden for the income reported by each participant. The simulation takes into account the tax rates set by the central and regional government and the number of children. Comparing this simulation to what participants guessed shows that individuals on average underestimate their marginal tax rates by approximately 10% but tend to overestimate the average tax rate. In the second step, the study removes this bias from individuals by informing them about their true marginal tax rate.

The results (see Foremny, 2020, for a detailed discussion) show that knowledge about taxation is very im-portant for citizens when it comes to different outcomes that affect everyday life and decision-making. The survey first asks participants about their preferred tax schedule across the income distribution, how they evaluate the level of taxation in Spain for low and high incomes, the trade-off between higher taxes and better public services, and their opinion about tax evasion. Results show that the tax information treatment changes the way respondents think about a tax reform. They implement more progressive tax reforms than the control group, and these reforms are characterized by higher marginal tax rates at the very top of the income distribution. Individuals who learn about the regional split of personal income tax regard inequalities as a bigger social problem relative to those who did not receive this information.

Citizens seem to have large misconceptions about the amount of tax they pay. The study documents this for the average and the marginal tax rate, which are two important features of the personal income tax system. The average tax rate is the share of the total amount of tax paid out of the income a person earns. This aver-age summarizes the total tax burden for an individual. The marginal tax rate is important when we make economic decisions, such as accepting extra work. The marginal tax rate is the percentage of tax to be paid out of an additional euro of income. For example, a person without children who earns 50,000 euros and lives in Catalonia would pay approximately 28% in tax—the average tax rate. If this person earned 100 euros more, he or she would keep 63 euros and would need to pay 37 euros in tax on those additional 100 euros. The marginal tax rate is 37% for people in that income group. In a progressive tax system, marginal tax rates are higher than average tax rates, because richer people pay more tax than poorer ones. 
 

Like in the example given above, the survey respondents were asked how much they thought they would need to pay on an extra income of 100 euros. Figure 2 illustrates the distribution of the results. The blue bar indicates that only approximately 1,000 respondents, or 16%, were able to predict correctly how much tax they needed to pay. The pink bars indicate the share of participants who underestimated their tax burden. This group of people thought they would need to pay less in taxes than they would actually have to pay. The orange bars, on the other hand, indicate responses which overestimated the tax burden. This group is particularly large, because many people (44% in total) believed they would have to pay taxes on their hypothetical extra income even though they would not have to pay anything—or very little—in reality. The study finds that this happens in particular for low-income earners, who have incomes below the minimum threshold (approximately 14,000 euros) under which income remains untaxed. 

Spanish personal income tax is shared between the central and each regional government. Half of the labor tax base is taxed at rates set by the central government in Madrid, and the other half at tax rates set by the autonomous communities. The results document that people have an average misperception of 28 points. The graph shows that on average only 25% of respondents know about this feature of the tax system.  
 

However, there are differences across participants. The data shows that people who paid personal income tax in the 2018 tax year are slightly more aware of this feature (the average in this group is 25%) compared to people who did not submit a tax declaration (19%). This bias is large and exists across all ages and education levels, even though it is slightly lower for highly educated and self-employed participants, while respondents who are not in the labor force show the largest bias here.

Given the large bias documented above, an important question is whether or not participants evaluate policies differently once they have the correct information at hand. Therefore, the study provides detailed personalized information about the marginal tax rates to a randomly selected group of respondents in the survey. Another group sees a short video which explains the important role of the autonomous communities in personal income taxation.

Participants were asked to what extent they were concerned about differences between the rich and the poor. Receiving information and thinking about tax policy makes an important difference here. Most people were already very concerned and saw inequality as a serious problem. Approximately 33% of respondents made this statement. Comparing them to those who received the information about their true tax payments, this fraction increased to 36%. A similar increase can be observed for those who receive the information that half of their income tax is received by their autonomous community. The differences from the control group are statistically significant.  

The study (Foremny, 2020) documents other outcomes. For example, respondents were asked to design their own tax system. They could decide tax rates for four different individuals, earning between 10,000 and 180,000 euros per year. Participants who received the information about their true tax burden were particularly different from the control group and set lower tax rates for the hypothetical low-income and middle-class taxpayers. They compensated this by setting significantly higher tax rates for the upper end of the income distribution. Other results show that information also helps to reduce the acceptance of tax evasion. People are more likely to respond that this cannot be tolerated under any circumstances once they have learned more about the taxes they themselves pay. 

Tax systems are complicated and difficult for citizens to understand. It is, however, important to resolve this information bias because the amount of information people have affects how people think about redistributing between the rich and the poor. People are more likely to be in favor of a more redistributive system after learning about the taxes they pay and that part of the redistribution takes place within their own autonomous community. Providing citizens with more information could therefore help to gain their political acceptance of a more inclusive tax system and their support for the weaker part of society. 

ALESINA, ALBERTO and STANTCHEVA, STEFANIE (2020), Diversity, Immigration, and Redistribution, NBER Working Paper 26620
 
CRUCES, GUILLERMO, PEREZ-TRUGLIA, RICARDO and TETAZ, MARTIN (2013), Biased perceptions of income distribution and preferences for redistribution: Evidence from a survey experiment, Journal of Public Economics, 98, pp. 100-112

FERNÁNDEZ-ALBERTOS, JOSÉ and KUO, ALEXANDER (2018), Income perception, information, and progressive taxation: Evidence from a survey experiment, Political Science Research and Methods, 6(1), pp. 83-110

FOREMNY, DIRK (2020), Information, Ethnic Diversity, and Preferences for Redistribution (manuscrito inédito)