University of Groningen On Taxes and Taxpayers ten Kate, Fabian

(1)

University of Groningen

On Taxes and Taxpayers

ten Kate, Fabian

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below.

Document Version

Publisher's PDF, also known as Version of record

Publication date: 2019

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

ten Kate, F. (2019). On Taxes and Taxpayers: understanding the heterogeneous effects of taxation. University of Groningen, SOM research school.

Copyright

Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

Take-down policy

If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

(2)

531634-L-sub01-bw-tenKate 531634-L-sub01-bw-tenKate 531634-L-sub01-bw-tenKate 531634-L-sub01-bw-tenKate Processed on: 23-5-2019 Processed on: 23-5-2019 Processed on: 23-5-2019

Processed on: 23-5-2019 PDF page: 13PDF page: 13PDF page: 13PDF page: 13

Chapter 1 Introduction

Every tax ought to be so contrived as to take out and to keep out of the pockets of the people as little as possible, over and above what it brings into the public treasury of the state.

Adam Smith (1991/1776, p.499)

S

incethe time of Adam Smith, economists have been concerned with questions

about the optimality of different tax instruments. In the Wealth of Nations, Smith (1991/1776) suggests four maxims of taxation that are still relevant in the present day. The first maxim is often used to justify progressive tax policy and holds that each individual should be taxed in proportion to the benefits enjoyed from the protection of the state. The second and third maxims hold that taxes should be certain and not arbitrary, and levied at such a time as to be convenient for the taxpayer. Modern tax systems are reasonably in line with these three maxims. The fourth (see opening quote), however, is more complicated, and lies at the heart of optimal taxation.

Optimal taxation has long been a key area of research in the public finance literature. It is based on the notion that the choices of economic agents should be distorted only to the extent required to achieve the goal at hand. Crucially, the goal is generally not to raise a certain amount of revenue as efficiently as possible —in which case the solution would be trivial and consist of a lump sum tax—but rather to also take into account equity concerns. This efficiency-equity trade-off has been formalized in the work of Mirrlees (1971). Since then the field has produced a number of key insights, as summarized by Mankiw, Weinzierl, and Yagan (2009). Broadly speaking, these fall into two categories. First, there is the manner in which tax schedules should be constructed, which is especially relevant for the taxation of labor. Second, there are lessons concerning what should and should not be taxed. This relates to value added taxes, but also to the taxation of production inputs, particularly capital.

(3)

Processed on: 23-5-2019 PDF page: 14PDF page: 14PDF page: 14PDF page: 14 The problem of constructing an optimal tax schedule originates from there being

unobserved differences in ability between individuals in an economy (Mankiw et al., 2009). High-ability individuals should generally not be taxed at such a rate as to find it beneficial to pretend to be a low-ability individual. Understanding how ability is distributed across individuals, moreover, is crucial for the design of optimal tax policy. Depending on the distributional assumptions made, optimal marginal rates may be increasing (Saez, 2001) or decreasing with income (Tuomala, 1990). In part due to such ambiguities, a flat tax may be desirable (Caminada & Goudswaard, 2001). Such a tax, combined with a universal basic income, could even be optimal (Mankiw et al., 2009; Mirrlees, 1971).

Before the question of how to tax something is arrived at, one should first con-sider what should and what should not be taxed in the first place. On this topic there are three particularly important results. First, Diamond and Mirrlees (1971) demonstrate that intermediate goods should not be taxed. Intuitively, taxes that affect different industries, sectors or time periods differently should not be used, as they would move the economy away from the production frontier (Mankiw et al., 2009). Second, Atkinson and Stiglitz (1976) show that all final consumption goods should generally be taxed at the same rate. This result follows if an individual’s consumption choices do not provide any additional information about his level of ability that is not already provided by his income level. By implication, the pre-ferred method of redistributing income would therefore be to adjust income taxes accordingly, rather than to set, for example, high tax rates on luxury goods. Third, the work of Chamley (1986) and Judd (1985) demonstrates that it is generally unde-sirable to tax capital, as this reduces the capital stock in the economy and thereby the level of output. This effect is so large that even an individual who does not himself own any capital, would still be better off if the revenue was raised using a tax on labor rather than on capital (Mankiw, 2000)

These various lessons of optimal taxation seem to have been taken to heart by policymakers to a varying degree (Mankiw et al., 2009). First, it seems to be the case that tax systems are moving in the direction of flatter tax schedules, in line with the suggestions made by Mirrlees (1971). Similarly, there is a positive relationship between the extent of pre-tax income inequality and the amount of redistribution that takes place. On the taxation of final goods, on the other hand, the recommendations seem to be largely ignored. Value added taxes are employed almost everywhere, but typically consist of various different rates. Such a policy generally aims at reducing the tax burden of the poor by taxing necessities at a lower rate, but, as noted above, Atkinson and Stiglitz (1976) suggest that this is better done through income taxation. Similarly, taxes on capital are well above zero in almost any country. In light of these differences, Mankiw et al. note: “Where large

(4)

Processed on: 23-5-2019 PDF page: 15PDF page: 15PDF page: 15PDF page: 15 gaps between theory and policy remain, the harder question is whether policymakers

need to learn more from theorists, or the other way around” (2009, p.148).

One explanation for such gaps may be that policy makers recognize to a greater extent than theorists that these lessons are context-dependent. For one, some of the results on optimal income tax schedules depend strongly on the distribution of ability in the economy. Similar qualifications exist in the other areas. Taxing different commodities at different rates may be desirable if preferences for particular goods are related to an individual’s level of ability (Saez, 2002). Another well-known exception occurs when the consumption of certain goods comes with negative externalities, making it also desirable to tax such goods at a higher rate (Pigou, 1920). The optimality of a zero tax on capital furthermore depends on the time horizon of individuals (Conesa, Kitao, & Krueger, 2009), as well as on the nature of the growth process (Aghion, Akcigit, & Fernández-Villaverde, 2013). In all of these examples the standard recommendations of the optimal taxation literature may be seen as a baseline or benchmark case, yet its exact applicability will depend on the particular scenario at hand.

While other forms of government policy are recognized as having potentially different effects in different contexts (Rodrik, 2006), this has not been thoroughly considered for tax policy. The optimal taxation literature has identified various empirically plausible scenarios in which certain policies are optimal. Insufficiently recognized, however, is that one country may find itself in one particular scenario, while another may find itself in an entirely different one. For example, there is some evidence that the accumulation of capital depends on the formal and informal institutions that are in place (Méon & Sekkat, 2015). It may therefore not be such a stretch of the imagination that the taxation of capital would have different effects in different countries. This thesis investigates the potential heterogeneity in the effects that taxation has on economic outcomes. Each of the three subsequent chapters considers the role of heterogeneity in a particular dimension of taxation.

Chapter 2 of this thesis investigates the relationship between capital taxation and economic growth at the national level. As noted above, the conventional wisdom

among economists suggests that the taxation of capital is undesirable. Though

powerful and intuitive, there is reason to believe that the Chamley-Judd result may not apply universally. In various contexts, diﬀerent theoretical models suggest that it may not be optimal not to tax capital. In general, other forms of taxation also create distortions, which may make it desirable to tax capital and use the revenue to reduce those distortions (Jacobs, 2013). For one, taxes on labor may cause distortions in labor supply (Aghion et al., 2013) or in the accumulation of human capital (Jacobs & Bovenberg, 2009) so that a tax on capital may be growth enhancing if used to reduce a tax on labor. Moreover, some models with

(5)

Processed on: 23-5-2019 PDF page: 16PDF page: 16PDF page: 16PDF page: 16 based growth make the point that at least some forms of capital taxation may be

growth enhancing (Anagnostopoulos, Carceles-Poveda, & Lin, 2012; Peretto, 2007). Similar conclusions have been reached for some models with heterogeneous agents (Aiyagari, 1995; Golosov, Troshkin, Tsyvinski, & Weinzierl, 2013; Saez, 2013) and for models in which the government is politically constrained (Acemoglu, Golosov, & Tsyvinski, 2011; Cozzi, 2004; Gordon & Li, 2009).

Furthermore, the existing empirical literature on the effects of capital taxation on economic growth has produced mixed results. Some authors find a negative effect (Djankov, Ganser, McLiesh, Ramalho, & Shleifer, 2010; Lee & Gordon, 2005), whereas others fail to do so (Arachi, Bucci, & Casarico, 2015; Easterly & Rebelo, 1993; Mendoza, Milesi-Ferretti, & Asea, 1997; Widmalm, 2001). A careful reading of this literature, however, suggests that the explanation may lie in the selection of countries included in the analysis. Papers which did not find a significant effect are generally based on a sample of highly developed countries. Conversely, the papers that did find a negative effect have a wider country coverage. This pattern of results could be obtained if the effect of capital taxation on economic growth depends on the level of economic development, i.e. if the negative effect is driven by the developing countries.

Chapter 2specifically asks if capital taxation is generally harmful for economic growth, as the conventional wisdom goes, or whether its effect depends on the level of economic development. It presents a theoretical model in which capital taxation has different effects for countries at different levels of development. The model is a variant of a multi-country version of the Schumpeterian growth model of Aghion and Howitt (1992), modified to allow for capital accumulation and taxation. It thus features a link between capital taxation and innovation, which is the engine of growth in the model. This approach makes it possible to study the effects of capital taxation on growth for economies that vary in their proximity to the technological frontier.

Empirically, the primary method of analysis used is a series of growth regressions that account for both the standard growth determinants and unobserved differences between countries and time periods. With this specification, which is standard in the literature, it is straightforward to assess the effect of capital taxation. Capital and its taxation, however, are convenient concepts in theory, but do not translate one-on-one to the real world. As such, the analysis uses different measures of capital taxation, and includes both narrow and broad measures. By employing a large sample of over 70 countries at different levels of development, it is furthermore possible to examine whether the growth effect of capital taxation varies with the level of development.

The implications of a given tax policy may not only vary across countries, but

(6)

Processed on: 23-5-2019 PDF page: 17PDF page: 17PDF page: 17PDF page: 17 also across regions within the same country. While all regions within a country are

in principle subject to one common tax policy, they may nonetheless be affected differently by it. There are substantial income differences between regions in the same country (Gennaioli, La Porta, Lopez-de Silanes, & Shleifer, 2013). Combined with the fact that tax schedules are generally progressive, this implies that the average citizen in a high-income region is taxed at a higher rate than one in a low-income region. Chapter 3 asks to what extent differences in regional tax rates, as they apply to the average citizen, are economically meaningful.

The chapter presents a new data set of average and marginal rates of personal income taxation as they apply to citizens with an average income in different Euro-pean regions. These tax rates are calculated directly from the respective country’s tax code. As such, they do not only take into account the tax schedule, but also tax relief, i.e. various measures aimed at reducing the tax burdens of (groups of) individuals without altering the overall tax schedule. This relief may come in vari-ous forms, such as a reduction of the amount of income over which taxes have to be paid, or a tax credit that may be directly subtracted from the tax bill. Tax relief has the effect of making it substantially more difficult to assess how much an individual is actually taxed, as it lowers the average rate at which he is taxed while possibly increasing his marginal rate. The latter effect occurs if the tax relief is reduced as income rises, which is often the case.

Besides discussing the procedure used to create this data set, the chapter also describes and explores the resulting data. By means of a variance decomposition it assesses whether the variation in the calculated tax rates occurs primarily between regions or within regions over time. Moreover, it also examines the extent to which there are diﬀerences in the extent of regional dispersion of these tax rates between countries. The same approach is used to quantify how large the variation is between regions in the same country relative to the variation between countries.

Another interesting question that is explored is what regions with relatively high tax rates (compared to their country’s mean) have in common. In principle, a relatively high tax rate must imply that a region has a relatively high income level. However, other factors may account for the observed variation as well, such as common geographical (physical or economic) characteristics, and similar economic structures, culture or institutions.

The eﬀects of taxes on personal income have been extensively studied in the literature. At the individual level, taxes aﬀect the labor supply decision. Higher taxes may cause individuals to work fewer hours (Hayo & Uhl, 2015; Kessler & Nor-ton, 2016), or otherwise reduce their income (Blomquist & Selin, 2010; Heim, 2010; Kessler & Norton, 2016; Lehmann, Marical, & Rioux, 2013). At the national level, countries that have higher labor income tax rates tend to have higher

(7)

Processed on: 23-5-2019 PDF page: 18PDF page: 18PDF page: 18PDF page: 18 ment rates as well (Daveri & Tabellini, 2000; Hausman, 1981; Planas, Roeger, &

Rossi, 2007; Triest, 1990). The impact of income taxes on unemployment has not been assessed at the regional level, but may be particularly relevant as there are large diﬀerences in unemployment rates between regions (Taylor and Bradley 1997). As an application of this new data set presented in Chapter 3, the relationship between income taxes and regional unemployment is analysed.

Thus far, the discussion has been about the different effects particular tax in-struments may have in different contexts. Another important element of taxation is whether taxes are paid in the first place. There are large differences between coun-tries in the prevalence of tax evasion (Richardson, 2006). These differences can be largely explained by the extent to which people are willing to pay taxes (Cummings, Martinez-Vazquez, McKee, & Torgler, 2009), which is the topic of Chapter 4.

People may have different motivations to pay their taxes. First, there are eco-nomic considerations, as cheating on your taxes involves the risk of being caught and fined. In practice, however, the expected cost of tax evasion is rather small, as are both the probability of being caught and the corresponding fine (Alm, McClelland, & Schulze, 1992; Dell’Anno, 2009; Pyle, 1991). Instead, it seems that non-pecuniary motivations drive the tax evasion decision, which are referred to in the literature as tax morale (Luttmer & Singhal, 2014). Tax morale can explain to a large extent the differences in tax compliance between different countries (Alm et al., 1992; Lewis, 1982; Pommerehne, Hart, & Frey, 1994; Torgler, 2007).

Tax morale varies substantially between individuals and between geographical regions. For example, the most recent wave of the European Values Study indicates that 91% of individuals residing in Norway find it unacceptable to cheat on their taxes, compared to 28% in Belgium. Differences between individuals can to some extent be explained by various personal characteristics (Schwartz & Orleans, 1967). Specifically, age (older people find tax evasion less acceptable), gender (the tax morale of men is generally lower), marital status (married individuals tend to have higher tax morale), and employment status (being self-employed in particular is associated with lower tax morale) are all relevant for tax morale.

Moreover, the literature suggests that tax morale is strongly associated with various cultural values. For example, perceptions of fairness (Cummings, Martinez-Vazquez, McKee, & Torgler, 2005; Fortin, Lacroix, & Villeval, 2007), attitudes to-wards the government (Cummings et al., 2005; Feld & Frey, 2002; Scholz & Lubell, 1998), civic virtues (Orviska & Hudson, 2003), social capital (Alm & Gomez, 2008), and religiosity (Alm & Torgler, 2011), have all been linked to tax morale. These factors explain to a substantial extent the variation in tax morale across countries. It is important to note, however, that there are not only large diﬀerences between countries in terms of the prevalence of these values, but also in the extent to which

(8)

Processed on: 23-5-2019 PDF page: 19PDF page: 19PDF page: 19PDF page: 19 they are shared across individuals (Schwartz & Sagie, 2000).

From the perspective of an individual taxpayer, taxes may be seen as a contri-bution to the resources of a group. A large literature in psychology indicates that people exhibit group favoritism and are more willing to share resources with in-dividuals whom they see as being similar to themselves (Tajfel, Billig, Bundy, & Flament, 1971; J. C. Turner & Oakes, 1986). This has its origin in the notion of social identity, in which an individual’s concept of himself depends (at least par-tially) on his membership of social groups (Tajfel & Turner, 1986). Given this, it seems likely that individuals would be more willing to pay taxes if they perceive those taxes as a contribution to a group with which they identify themselves.

The analysis in this chapter considers whether tax morale is driven by the extent to which individuals subscribe to the same values as their fellow citizens. For this purpose it first quantifies the extent to which an individual is similar, in terms of cultural values, to a reference group (taken to be fellow citizens residing in the same region). It considers a variety of different measures that vary in how the similarity

of individuals is calculated, as well as the set of cultural values considered. It

then analyses for a large sample of over 50,000 individuals the relationship between the extent to which an individual conforms to the group and the individual’s tax morale. The effects of this individual value conformity are furthermore contrasted with those of regional value conformity, which captures the extent to which people in general are similar to each other. It also considers a variety of robustness checks by using different econometric estimation techniques and samples. In general, using this approach a cultural value that would normally be expected to have a positive effect on tax morale may instead have a negative effect, if this value is not shared widely within society.

Combined these chapters present the argument that there is heterogeneity in the effects that taxation has on various economic outcomes. A similar tax may have different effects in different contexts. This is largely recognized by the literature on optimal taxation, which has demonstrated that assumptions regarding e.g. the distribution of ability (Saez, 2001; Tuomala, 1990) or the nature of technological progress (Aghion et al., 2013) may have drastic effects on the conclusions. Some of these assumptions may better describe the economies of some countries rather than others. If this is the case, then one would expect there also to be heterogeneity in the effects that taxes have. For this reason it is not desirable to give general policy recommendations regarding taxes, but rather to craft the recommendations specifically for each context and country at hand. Adding this nuance is the main point and contribution of this thesis.

(9)