FISCAL WELFARE, PROGRESSIVITY, AND DECENTRALIZATION: THE CASE OF SPAIN <PROVISIONAL DRAFT>

Transcription

1 XIII CONGRESO DE LA F.E.S. Gijón, 3 de Junio-2 de Julio de 216 FISCAL WELFARE, PROGRESSIVITY, AND DECENTRALIZATION: THE CASE OF SPAIN <PROVISIONAL DRAFT> José A. Noguera jose.noguera@uab.cat Paula Hermida paula.hermida@uab.cat Group of Analytical Sociology and Institutional Design (GSADI) Universitat Autònoma de Barcelona (UAB) Abstract Fiscal welfare has often been used as a political tool in order to deliver benefits which have regressive distributional impact. The reason for using the tax system in order to do so may be this: if a government, for whatever reasons, wants to benefit the non-poor (and even the rich), then tax expenditure might be the preferred option because of its low visibility and the opacity of its distributional impact. In this paper we analyze how the political decentralization of tax administration and regulation may create more political chances to choose that kind of strategy. We study fiscal welfare in the specific case of Spain, a state with a quite high degree of decentralization of its tax system: regional governments have full control of 5% of income tax (except for the Basque Country and Navarra, who control all taxes). At the national level, the general trend in the last decades has been the reduction and/or disappearance of many tax deductions linked to welfare (such as private health expenditure, private pension savings, or housing acquisition), but this evolution has paralleled a large spreading of welfare tax expenditures in virtually all the regional tax regulations. Regional tax expenditure has had most of the times a regressive distributional impact, but has proliferated in areas such as child allowances, private schooling expenditure, private care of dependant persons, and the like. Using the large income taxpayers sample of the Spanish Institute for Fiscal Studies (IEF), we present and analyze the main data that describe this dynamics, as well as the distributional income of tax expenditure and Spain at the national and regional levels. We conclude that decentralization of tax regulations in Spain has paved the way for spreading regressive fiscal welfare (sometimes linked to practices of political patronage by regional governments) at a low political cost because of its low visibility; this contrasts, in many cases, with the scarcity of the regional budgets allocated to antipoverty cash benefit programs (which are also fully decentralized). 1

2 Outline 1. Fiscal welfare, progressivity, and political visibility 2. Fiscal welfare and decentralization: the Spanish case 3. Hypothesis, data and methods 4. Analysis of fiscal welfare in Spain (I): expenditure 5. Analysis of fiscal welfare in Spain (II): types of fiscal benefits 6. Analysis of fiscal welfare in Spain (III): distributional impact 7. Fiscal welfare and minimum income programs: a comparative analysis 8. Conclusions and further research 2

3 1. Fiscal welfare, progressivity, and political visibility Monetary transfers between the public administration and citizens are mainly of two different types: social benefits, by which the state tranfers income to those citizens who, for any reason, cannot cover their basic needs by themselves, and taxes, through which those above a given income level transfer part of that income to the state. However, as Titmuss (1958) noted long ago, many transfers from the State to citizens are in fact paid through the tax system as reductions of their taxable income or direct tax deductions. When these tax allowances are linked to traditional social expenditure functions, they may be rightfully considered as hidden welfare benefits. In his pioneering essay on The Social Division of Welfare, Titmuss convincingly argued that this kind of fiscal benefits may reach an important level of aggregated expenditure and may have a regressive distributional impact which is not usually transparent for the public opinion; however, they attract scarce attention from scholars and researchers who study the welfare state, since they usually focus exclusively on the analysis of direct welfare benefits. In recent decades, a growing number of scholars have remarked this fact, and have developed several studies on these hidden benefits under labels such as fiscal welfare, social tax expenditures or hidden welfare state (references to be added). 1 The abovementioned contributions pave the way to study taxes and the welfare state as a system of communicating vessels, where an intervention in one of them has a corresponding economic equivalence in the other: lowering income tax for a given taxpayer equals to pay a benefit to him; totally or partially withdrawing a welfare benefit to a recipient equals to raising the income tax he is paying. However, the respective rationales and designs of both vessels are not usually well coordinated, and are often inconsistent from the standpoint of horizontal and vertical equity; for instance, it is often likely that many citizens are entitled for different social tax allowances while they would not be eligible for direct social benefits designed for similar purposes; conversely, they might be eligible for social benefits but may not be able to benefit from tax allowances with similar rationale because they do not have enough income to pay taxes. As a result, it is to be expected that many fiscal benefits have a regressive impact on income distribution which may counteract the progressive impact of many social benefits (Ganghof, 26; Goodin y Le Grand, 1987; Hills, 2; Parker, 1989; Vanderborght, 26). Spain has been traditionally a very good example of this trend (Antón, 27; Calero, 22; Sevilla, 1999; Sevilla, 23). But distributional inconsistency is not the only problem associated with fiscal welfare. Other studies suggest that fiscal welfare often works in a less 1 In this paper, for simplicity, we will use expressions such as fiscal welfare, social tax expenditures, or tax benefits as equivalent, although we are aware that there are some nuances and different connotations related with each one of them. 3

4 transparent and visible way than direct welfare benefits programs; this is an incentive for governments to implement monetary benefits while avoiding harsh parliamentary, political or media control and accountability. This lack of transparency is many times reflected in the difficulty to obtain statistical data on fiscal welfare to make reliable estimations on its distributional impact. Governments who, for whatever reasons, have a preference for non-progressive redistribution may be tempted to use tax benefits in order to channel public resources towards middle and high income taxpayers at a low political cost, and even with electoral returns since these benefits are only clearly visible for those who benefit from them. Since fiscal welfare works by moving effective tax rates away from marginal tax rates, they may easily produce tax illusions which are politically useful. It is true that not all kinds of fiscal welfare follow a regressive distributional pattern. As always with monetary benefits, the distributional impact heavily depends on the design of the benefit, and the devil uses to lie behind the details of that design. Good examples of progressive tax benefits may be found in the different tax credits programs that have been introduced in the last 15 years by several UE countries, following the road opened by EITC in USA back in the 197s. Tax credits for low-income working families, which sometimes get close to the idea of a negative income tax, explicitly pursue a greater integration and consistency between personal tax and direct welfare benefits. Governments of different political orientation in countries with quite different tax and welfare systems have implemented such programs (Adam y Browne, 26; Atkinson, 1995; Immervoll y Pearson, 29; Parker, 1989). The UK is probably the country in the UE that went further on that road since the 199s (Brewer et al, 212), but also Germany, Belgium, Denmark, Finland, France, the Netherlands, Hungary, Ireland, Sweden and some Baltic countries have adopted important measures in the direction of tax-benefit integration. The European Comission (23) and the OCDE (23, 25, 28) showed also early interest about the issue. An interesting fact to note is that, in these cases, and contrary to more traditional forms of fiscal welfare, the progressive impact of most of the programs has worked as an incentive for enhancing their political visibility: governments have strongly publicized tax credits as a new and progressive social policy which (supposedly unlike traditional welfare) is work-friendly (expressions such as work pays or in-work benefits have been widely used to remark it). This seems to reinforce the plausibility of the hypothesis that the default for fiscal welfare is non-visibility in the political arena; hence fiscal welfare is in principle more attractive than direct payments when a government seeks to benefit the better-off. 4

5 2. Fiscal welfare and decentralization: the Spanish case The rise of tax credits has only reached Spain in a very timid fashion. No comprehensive program of tax credits for low-income working families has been approved so far. The only existing tax credit was introduced in 23 and consists on a 1 /month tax deduction for working women who have children from to 3 years old; this amount is payable directly through the wage pack. However, its effect is clearly not progressive but equally distributed among deciles of income tax payers; this is to be expected, since there is no income targeting, and in fact, mothers must show a minimum contribution to social security in order to be eligible, something that typically leaves out many mothers working part-time or with low wages. Recently, a similar tax credit has been introduced for large families, mono-parental families with 2 children, and households with disabled children or elders, but data on its effects are not yet available. However, the variety and budget of tax expenditures linked to welfare in Spain is anything but low: millions were spent in 211 in dozens of tax reductions and deductions in income tax; that figure amounts to 18% of the total income tax revenue for that year, and leaves aside the tax reductions known as life minimum and reduction for income from work which are generally applied to all taxpayers, and almost triplicate the expenditure in fiscal welfare. It is true, as we will detail in section 4, that budget constraints have led the Spanish government to cut or eliminate some tax benefits at the national level: for example, tax reductions linked to private health expenditures were long ago suppressed, deductions for private pension plans have been reduced, and those linked to housing property are only kept with retroactive effects. Anyway, the expenditure remains high at the national level and even grows at the regional level. This fact, together with the virtual absence of tax credits for low income families, may raise the suspicion that in Spain tax expenditures have been mainly used to redistribute income in favour of the better-off, and that regional governments have resorted to them to protect some middle and high income groups from the effects of budgetary and fiscal cuts coming from the central tax authorities. The Spanish income tax regulations are in this sense a breeding ground for regional political patronage. Since 26, fifteen Spanish regional governments control 5% of the income tax (among other taxes), and have the capacity to regulate tax rates, income bands, and tax expenditures affecting half of the income base. The two remaining regions (Navarra and the Basque Country) have always controlled 1% of their income tax, for historical reasons recognized in the Spanish Constitution. This capacity has been used scarcely to alter marginal tax rates but much more often to introduce tax deductions for specific purposes, many of them linked to social welfare such as childbirth, childbearing, adoption, family care, disability, housing, private school expenditure, private health expenditure, job training, self-employment. Hence, the decentralized structure of income tax has produced a more fragmented system of fiscal welfare, and most likely a less visible 5

6 and transparent one, since it is difficult for the public opinion, the media and the analysts to monitor seventeen different fiscal welfare policies with different designs and distributional effects. 3. Hypothesis, data and methods The main hypothesis of this paper is that fiscal welfare has been mostly used in Spain in order to hide monetary transfers with regressive distributional impact, which would not have been politically acceptable had they been channelled directly as social benefits. The low visibility of tax expenditures and the opacity of its distributional impact are even greater in a decentralized tax system such as the Spanish one; all of this makes the use of fiscal welfare very attractive in political terms for governments with regressive distributional agendas. It is then to be expected that, despite the budgetary constraints imposed by the economic crisis and the European Union, expenditure in fiscal welfare in Spain has been kept at a relatively high level, that its distributional impact has been mostly regressive, and that different social tax expenditures have proliferated at the regional level even in a period of budgetary cuts. In order to test the hypothesis, we have analyzed the comprehensive information on tax expenditures included in the micro-data of the Income Taxpayers Panel created by the Institute of Fiscal Studies (IEF, an official organism which depends on the Ministry of Finance). We have selected the data for fiscal years 28 (when the economic crisis started to have a tangible effect) and 211 (the most recent year with data available when we started our study). The Panel contains a sample of near two million income tax declarations ( for 28, and for 211), which are statistically representative of the whole income tax declarations; it does not contain declarations for Navarra or the Basque Country, due to their specific tax status. The micro-data provided by the IEF Panel are very useful for several reasons: first, they have never been used to date in order to analyze fiscal welfare in Spain. Second, they contain real data on virtually all income tax declarations: the sample has a raising factor which allow to work with nearly 2 million cases ( for 28 and for 211), with a sampling error lower than 1,1% and confidence level of 99,7%. Third, it allows for estimating the distributional impact of all tax magnitudes, including deductions and reductions in the taxable income. No other data source allows for performing this kind of analysis with similar detail. For example, the OCDE database on taxes and social benefits includes mostly aggregated data and often makes difficult or unfeasible to identify clearly social tax expenditures. Official public reports on tax expenditure budgets show the first problem, if not also the second; additionally, analysis based on them cannot account for the fact that tax expenditures budgets may not in the end match 6

7 real expenditure, due to lack of prevision, changes in the distribution of income, changes in the behaviour of taxpayers, or non take-up. Finally, estimations models such as EUROMOD do not work with real data on distributional impact, but estimate this impact by simulating the application of tax-benefit rules to the distribution of income according to the Income and Living Conditions surveys (EU- SILC). With the IEF micro-data at hand, we analyze tax expenditures which may reasonably qualify as fiscal welfare or social tax expenditures. We consider as such all tax deductions and reductions whose declared rationale is supporting taxpayers who face situations like childbirth, adoption, childcare, family care, disability, housing costs, private health costs, private school expenditures, job training and other expenditures linked to employment, pension plans, and similar expenditures related with traditional welfare benefits. For simplicity, we (provisionally) group all these tax benefits into seven categories: disability and old-age (including pension plans), childbirth/adoption, caring of dependant persons (including childcare), aid to work-family conciliation, schooling costs, housing costs, and aid to employment. We quantify the volume of all those tax expenditures in the national income tax tier as well as in the regional one, and we calculate their distributional impact by income deciles in both cases. Gross income deciles have been calculated using an income variable specifically built for this study, which aggregates the taxable income from any source before any personal reductions have been applied. The resulting income thresholds for deciles are shown in Table 1 (the source for all tables is the Income Taxpayers Panel of the IEF for Spain, 28 and 211). Table 1 Income thresholds by income tax deciles. Spain, 28 and 211 Percentiles Income thresholds (up to) 28 Percentiles Income thresholds (up to) , , , , , , , , , , , , , , , , , ,51 7

8 4. Analysis of fiscal welfare in Spain (I): expenditure Table 2 shows the aggregated expenditure in fiscal welfare in Spain for the two years we compare. The total expenditure in 211 amounts to more than 18% of total income tax revenue. We decompose the total expenditure in two tiers: the national income tax tier and the regional one. In the first case, we also distinguish between reductions in taxable income and direct deductions in the tax to pay (at the regional level all fiscal benefits are channelled as tax deductions). Table 2. Aggregated expenditure in fiscal welfare in by government level Rate of variation (%) National government , ,14-17,45 Reductions in taxable income , ,8-8,1 Tax deductions , ,34-25,53 Regional governments , ,39 134,38 TOTAL , ,53-15,46 The table clearly shows that the volume of social tax expenditure in the national tier is much higher than in the regional one. This is mainly due to the nature of the tax benefits that are included in the national tier, which reach a very high number of taxpayers: the tax credit for working mothers with children from to 3 years old, the deduction for home purchase, and the contributions to pension plans. The variation between 28 and 211 shows a negative trend: the total expenditure has been reduced by 15%. But this hides an opposite evolution in the expenditure at the regional level, which has been more than doubled during the same period. Therefore, the aggregated trend is concealing a very differentiated evolution between different levels of government, as expected. Additionally, it can be observed that the reduction of expenditure in fiscal welfare at the national level has been much more intense in tax deductions than in reductions of taxable income. Since the income tax rate is progressive, deductions have usually a less regressive impact than reductions of taxable income, so it is to be expected that this different variation rate has had a regressive net distributional effect. Table 3 shows the data on expenditure in fiscal welfare by regional government. It is clear that the growth of expenditure is a general pattern in all of them, with the only exceptions of Asturias and La Rioja. Generally speaking, this growth has more to do with a higher number of eligible taxpayers and to higher deduction amounts than to the passing of new types of deductions. However, most 8

9 DISABILITY AND OLD AGE CHILDBIRTH / ADOPTION CARING OF DEPENDANT PERSONS WORK-FAMILY CONCILIATION SCHOOLING COSTS HOUSING COSTS JOB TRAINING AND EMPLOYMENT TOTAL regional governments (12 out of 15) have in fact approved new tax deductions during this period, thus confirming the expected evolution. Table 3. Aggregated expenditure in fiscal welfare in by regional government Rate of variation (%) ANDALUCÍA , ,8 75,29 ARAGÓN , ,5-2,86 ASTURIAS , ,73 184,94 BALEARES , ,98 141,11 CANARIAS , ,35 22,99 CANTABRIA , ,39 123,3 C. LA MANCHA , ,63 119,29 C. LEÓN , ,13 33,79 CATALUÑA , ,48 58,15 C.VALENCIANA , ,8 133,7 EXTREMADURA , ,74 78,66 GALICIA , ,74 474,72 MADRID , ,74 35,24 MURCIA 55.83, ,5 961,81 LA RIOJA , ,5-3,12 TOTAL , ,39 134,38 5. Analysis of fiscal welfare in Spain (II): types of fiscal benefits Table 4 shows the distribution of the expenditure in fiscal welfare by type, both at the national and regional level. As expected, there is a wider variety of fiscal benefits in the regional tier than in the national one. Table 4. Expenditure in fiscal welfare in million by type and government level NATIONAL 6.21, , , ,7 REGIONAL 18,6 17,7 42,8 7,3 4,2 8,3 1,5 172,8 TOTAL 6.4, ,9 42,8 7,3 4, ,9 1, ,6 NATIONAL 5.534,2 78, 4.412, ,8 REGIONAL 31,9 75,5 37,4 25,3 33, 197,9 3,7 45,1 TOTAL 5.566,2 855,6 37,4 25,3 33, 4.61,5 3, , 9

10 DISABILITY AND OLD AGE CHILDBIRTH / ADOPTION CARING OF DEPENDANT PERSONS WORK-FAMILY CONCILIATION SCHOOLING COSTS HOUSING COSTS JOB TRAINING AND EMPLOYMENT TOTAL In absolute terms, expenditure in all categories of national fiscal welfare has been reduced, while the opposite trend is observed at the regional level in all types of fiscal welfare with only one exception (a slight reduction in caring for dependant persons). But in relative terms (as we can see in Table 5), a more nuanced picture may be drawn: the loss in the relative weight of some tax benefits at the national level (childbirth and housing costs) has been accompanied by a higher relative weight of those categories at the regional level (strongly in the case of childbirth/adoption), while the opposite trend is observed in the disability/old age category. As for those categories of fiscal benefits that exist only at the regional level, some have dramatically fallen in relative terms (like aid for caring of dependant persons) while others have grown (such as support for schooling costs, mainly due to a new tax deduction approved by the regional government of Madrid). Some of these trends seem to suggest that many regional governments have intended to partially compensate tax benefit cuts (or absence of some specific types of fiscal welfare) at the national level. Table 5. Expenditure in fiscal welfare in % by type and government level NATIONAL 46,34 1,57 43,8 1, REGIONAL 1,82 1,28 24,78 4,28 2,45 46,48,92 1, TOTAL 45,88 1,57,33,6,3 43,13,1 1, NATIONAL 51,59 7,27 41,14, 1, REGIONAL 7,9 18,66 9,23 6,25 8,17 48,87,92 1, TOTAL 5, 7,69,34,23,3 41,42,3 1, 6. Analysis of fiscal welfare in Spain (III): distributional impact In this section we present the main results of our analysis of the distributional impact of fiscal welfare in Spain. We compare the distribution of social tax expenditures by income deciles in 28 and 211, both at the national and regional levels. The main result, again as expected, is that fiscal welfare has 1

11 % mainly a regressive impact, so the taxpayers at the lower income deciles not only do not get a higher portion of the deductions-reductions pie, but they do not even get a proportional part. The bulk of the resources spent is appropriated by the higher and middle income deciles (depending on the type of benefit and the specific government). If we observe Figure 1, we can easily see that the distribution of total fiscal welfare at the national tier clearly follows a regressive pattern, by benefitting taxpayers more the higher their income level is. This trend has even intensified from 28 to 211, as a result of the raise in the relative weight of reductions in taxable income with respect to deductions: Figures 2 and 3 show that the latter have a much more regressive distribution than the former; however, as one can see in Figure 3, deductions have also followed a regressive distributional pattern, and this pattern has even been intensified from 28. FIGURE 1. NATIONAL LEVEL DISTRIBUTION OF TOTAL FISCAL WELFARE BY , 35, 3, 25, 2, 15, 1, 5,, 11

12 % % FIGURE 2. NATIONAL LEVEL DISTRIBUTION OF TAX DEDUCTIONS BY , 15, 1, 5,, 6, 5, 4, 3, 2, 1, FIGURE 3. NATIONAL LEVEL DISTRIBUTION OF REDUCTIONS IN TAXABLE INCOME BY , 12

13 % If we now look at Figure 4, we can see that the distribution of the aggregated fiscal benefits in the regional tier has mainly favoured taxpayers in the middle income deciles, so the regressive pattern is less intense than the one we observed at the national level. This differentiated distributional pattern at both levels of government is most likely due to two different factors: first, the fact that all fiscal welfare in the regional tier is channelled through deductions instead of reductions of taxable income; second, and probably more important, the fact that the specific design of many regional deductions establishes upper income thresholds for eligibility. Even so, it can still be observed that taxpayers in the lower deciles do not obtain a proportional share of tax benefits, and the share they obtain is far away from that going to the upper deciles. Notwithstanding this general regressive pattern, the distribution of fiscal welfare in 211 at the regional level is slightly less regressive than in 28, contrary to what could be observed at the national level. This fact, however, should be treated with caution, since this evolution may be to a great extent the effect of the generalized reduction in deciles thresholds from 28 to 211 (see above, Table 1), which is a consequence of the generalized income loss produced by the economic crisis. FIGURE 4. REGIONAL LEVEL DISTRIBUTION OF REGIONAL TAX DEDUCTIONS BY , 3, 25, 2, 15, 1, 5,, 13

14 % TAX DEDUCTION Finally, Figures 5 to 19 show the distribution of regional tax deductions in each region. Although none of them follow a progressive pattern, it is possible to distinguish broadly between a group of regions where the expenditure is more concentrated in the middle income deciles (Andalucía, Asturias, Baleares, Cantabria, Catalunya, Extremadura with one of the less regressive distributions-, La Rioja and Madrid), and another group where fiscal welfare is clearly directed towards upper income deciles (Aragón, Canarias, Castilla y León, Castilla-La Mancha, Galicia, Murcia and Comunidad Valenciana). This broad difference suggests that more case-to-case analysis on the structure and rationale of specific deductions and on the policy process that led to their implementation would be needed to fully understand the dynamics of fiscal welfare at the regional level. But it is remarkable that virtually all the regions in the second group have been governed by the right-wing Popular Party during most of the period under analysis. 35, FIGURE 5: ANDALUCÍA DISTRIBUTION OF TAX DEDUCTIONS BY , 25, 2, 15, 1, 5,, 14

15 % TAX DEDUCTION % TAX DDEDUCTION FIGURE 6: ARAGÓN DISTRIBUTIONOF TAX DEDUCTIONS BY FIGURE 7: ASTURIAS DISTRIBUTION OF TAX DEDUCTIONS BY , 3, 25, 2, 15, 1, 5,, 15

16 % TAX DEDUCTIONS % TAX DEDUCTIONS FIGURE 8: BALEARES DISTRIBUTION OF TAX DEDUCTIONS BY , 8, 6, 4, 2,, FIGURE 9: CANARIAS DISTRIBUTION OF TAX DEDUCTIONS BY

17 % TAX DEDUCTIONS % TAX DEDUCTIONS FIGURE 1: CANTABRIA DISTRIBUTION OF TAX DEDUCTIONS BY FIGURE 11: CASTILLA LEÓN DISTRIBUTION OF TAX DEDUCTIONS BY DECLIES

18 % TAX REDUCTIONS % TAX DEDUCTIONS FIGURE 12: CASTILLA LA MANCHA DISTRIBUTION OF TAX DEDUCTIONS BY , 4, 35, 3, 25, 2, 15, 1, 5,, FIGURE 13: CATALUÑA DISTRIBUTION OF TAX DEDUCTIONS BY , 3, 25, 2, 15, 1, 5,, 18

19 % TAX REDUCTIONS % TAX REDUCTIONS FIGURE 14: EXTREMADURA DISTRIBUTION OF TAX DEDUCTIONS BY FIGURE 15: GALICIA DISTRIBUTION OF TAX DEDUCTIONS BY , 3, 25, 2, 15, 1, 5,, 19

20 % TAX REDUCTIONS % TAX REDUCTIONS FIGURE 16: LA RIOJA DISTRIBUTION OF TAX DEDUCTIONS BY FIGURE 17: MADRID DISTRIBUTION OF TAX DEDUCTIONS BY

21 % TAX DEDUCTIONS % TAX DEDUCTIONS FIGURE 18: MURCIA DISTRIBUTION OF TAX DEDUCTIONS BY , 7, 6, 5, 4, 3, 2, 1,, FIGURE 19: COMUNIDAD VALENCIANA DISTRIBUTION OF TAX DEDUCTIONS BY

22 9th and 1th. TAX DEDUCTIONS. MEDIA = 1 A more refined picture of how the worst-off and the better-off are treated by the different regions tax deductions policy is offered in Figures 2 and 21, where regions are plotted according to which proportion of fiscal welfare goes to the lower income deciles (1 st and 2 nd, horizontal axis) and to the upper ones (9 th and 1 th, vertical axis). Some regions such as Catalunya, Madrid or Asturias are over the mean in the first and below the mean in the second, so in relative terms they treat better the worst-off than other regions, and treat worse the better-off (note that this does not exclude that middle and/or upper deciles eventually get the bulk of the expenditure, nor that the general pattern is regressive). Other regions such as Andalucía and Extremadura fare under the mean in both measures. Finally, a third group of regions (like Castilla-La Mancha, Castilla y León, Comunidad Valenciana or Galicia) treat relatively better the better-off, and worse the worst-off. Interestingly, and with the exception of Aragón in 28, there are no regions over the mean in both axis, suggesting that when social tax expenditure reaches a threshold, there is a distributional trade-off between benefitting the richer and the poorer. This pattern is quite similar in the two years under consideration. 3 FIGURE 2: % OF TAX DEDUCTIONS FOR THE TWO UPPER AND LOWER INCOME IN SPANISH REGIONS. 28 FIRST 2% OF INCOME MEDIA = 1 25 MURCIA GALICIA CANARIAS ARAGÓN C. LA MANCHA C. VALENCIANA C. LEÓN CANTABRIA LA RIOJA MADRID ANDALUCÍA CATALUÑA LAST 2% OF INCOME MEDIA= 1 ASTURIAS EXTREMADURA BALEARES, 1, 2, 3, 4, 5, 6, 7, 8, 1st and 2nd. TAX DEDUCTIONS. MEDIA = 1 22

23 9th and 1th. TAX REDUCTIONS. MEDIA = FIGURE 21: % OF TAX DEDUCTIONS FOR THE TOW UPPER AND LOWER INCOME IN SPANISH REGIONS. 211 C. LA MANCHA FIRST 2% OF INCOME. MEDIA = ARAGÓN C. LEÓN CANARIAS GALICIA C. VALENCIANA CANTABRIA MURCIA LA RIOJA ANDALUCÍA CATALUÑA LAST 2% OF INCOME. MEDIA = 1 MADRID EXTREMADURA BALEARES ASTURIAS st and 2nd. TAX REDUCTIONS. MEDIA = 1 It is worth to note that, similarly to what happens at the national aggregated level, the distributional pattern of aggregated fiscal welfare in each region often conceals particular regional tax deductions that have a strong regressive character, or the opposite. Disaggregating distributional impact at the level of each tax deduction (an operation allowed by our micro-data) will make possible to analyze case-by-case some interesting specificities. For example, it will make clear that some regional governments have used some tax deductions with a clear intention of benefitting upper income taxpayers, and of disguising those benefits under a general pattern which is apparently less regressive. An illustration may be found in Figure 22, which shows the distributional pattern of a tax deduction for private schooling costs in the region of Madrid in

24 % FIGURE 22: MADRID DISTRIBUTION OF THE SCHOOLING COSTS TAX REDUCTION BY, Fiscal welfare and minimum income programs: a comparative analysis Even if the general distributional pattern of regional tax deductions is far from being progressive, it should be acknowledged that the data presented in the previous sections do not necessarily show by themselves an explicit and deliberate political purpose to benefit the best-off, although they certainly raise a grounded suspicion in that sense. A way to reinforce this hypothesis may be to compare the regional expenditure in fiscal welfare with the resources devoted to minimum income programs targeted at individuals and families in situation of poverty and social exclusion. These programs are in Spain fully decentralized: the legal capacity to develop and implement them pertains to the regional governments, so there is no national minimum income program, but seventeen regional ones, with very different patterns as regards to coverage, eligibility conditions, amount of the benefits, and behavioural requirements. What they all have in common is that they are targeted at those households with income under the poverty threshold and not entitled to other benefits, so the benefits should have by definition a progressive distributional impact. Therefore, they may give a proxy indicator as to how committed with the welfare of the worst-off a regional government really is. By relating the expenditure on (regressive) fiscal welfare with that on (progressive) minimum income programs we could perhaps have a more complete approximation to this commitment. 24

25 TAX DEDUCTIONS (Euros per capita) This is done in Figures 23 and 24, where regions are plotted according to their per capita expenditure in minimum income programs (horizontal axis) and fiscal welfare (vertical axis). Since these expenditures have grown from 28 to 211, the dispersion is greater in Figure 24, but we can observe a similar pattern which takes the form of a U inclined to the right. What this suggests again is that there seems to be a trade-off between middle-high spending in regressive tax deductions and in progressive minimum incomes. Also, regions that treat very well the better-off through tax deductions tend to be those which spend less in minimum income programs than in fiscal welfare. Figures 24 and 25 show a very similar pattern when we consider recipients instead of per capita expenditure. This indirectly supports the suspicion that the structure of fiscal welfare at the regional level has a lot to do with a political commitment with regressive distributional policy options. 2 FIGURE 23: PER CAPITA EXPENDITURE IN TAX DEDUCTIONS AND MINIMUN INCOMES IN SPANISH REGIONS. 28 CANARIAS C.VALENCIANA C. LA MANCHA LA RIOJA BALEARES MADRID CASTILLA LEÓN EXTREMADURA CANTABRIA ANDALUCÍA ARAGÓN GALICIA CATALUÑA ASTURIAS MURCIA MIP(Euros per capita) 25

26 TAX DEDUCTIONS (% Recipients) TAX DEDUCTIONS (Euros per capita) FIGURE 24: PER CAPITA EXPENDITURE IN TAX DEDUCTIONS AND MINIMUN INCOMES IN SPANISH REGIONS CANARIAS 2 CASTILLA LEÓN 15 MADRID C.VALENCIANA BALEARES 1 CASTILLA LA MANCHA GALICIA ASTURIAS 5 EXTREMADURA LA RIOJA CANTABRIA ANDALUCÍA CATALUÑA MURCIA ARAGÓN MIP (Euros per capita) FIGURE 25: % OF POPULATION WHO IS RECOPIENT OF TAX DEDUCTIONS OR MINIMUN INCIMES IN SPANISH REGIONS CANARIAS CASTILLA LA MANCHA 6 LA RIOJA 4 BALEARES EXTREMADURA C.VALENCIANA ASTURIAS 2 CASTILLA LEÓN MADRID CANTABRIA ANDALUCÍA CATALUÑA MURCIA ARAGÓN GALICIA,2,4,6,8 1 1,2 1,4 MIP (% Recipients) 26

27 TAX DEDUCTIONS (% Recipients) FIGURE 26: % OF POPULATION WHO IS RECIPIEN OF TAX DEDUCTIONS OR MINIMUN INCOMES IN SPANISH REGIONS CANARIAS 12 CASTILLA LA MANCHA 1 8 C.VALENCIANA BALEARES 6 4 EXTREMADURA LA RIOJA MADRID CASTILLA LEÓN CANTABRIA ASTURIAS 2 MURCIA GALICIA CATALUÑA ANDALUCÍA ARAGÓN,2,4,6,8 1 1,2 1,4 1,6 1,8 2 MIP (% Recipients) 8. Conclusions and further research The analysis presented above show that the expenditure in fiscal welfare in Spain is relatively high and that, since it is mostly concentrated in the national income tax tier, its reduction from 28 to 211 conceals a remarkable raise in social tax expenditures in the regional tier, in spite of budget constraints. Generally speaking, tax benefits follow a regressive distributional pattern which benefits middle and upper income deciles (with some degree of variation depending on the regional government). Besides, the data suggest that the expenditure in regressive fiscal welfare in Spanish regions is in a relation of trade-off with the expenditure in minimum income programs targeted at poor households. This supports the hypothesis that regional fiscal welfare policy has been driven by a regressive distributional political commitment. Further research would no doubt be needed to reinforce the hypothesis and to refine the analysis. In successive stages of the study we plan to widen the temporal period under study, to include the data on the Basque Country and Navarra, and to analyze in more detail the policy process behind some key cases of regressive regional tax deductions, so more substantive evidence on the hypothesized political explanation may be raised. References <TO BE ADDED> 27