Immigrants Rights and Benefits. A public opinion analysis for Spain Robert Duval-Hernández División de Economía, CIDE Ferran Martínez i Coma División de Estudios Internacionales, CIDE December 10, 2010 We would like to thank Gordon Hanson, Luciana Moscoso, Covadonga Messeguer, Joshua Dyck, and seminar participants at CIDE and MPSA 2010 for their valuable comments to this work. Needless to say, the remaining errors are our sole responsibility.
Abstract We present a theoretical model to explain the preferences of natives for granting voting rights and welfare benefits to immigrants. Our model predicts that in a country where immigrants are relatively more unskilled than natives, unskilled natives oppose granting health and education to immigrants because of the competition for resources implied. On the other hand, skilled natives oppose granting voting rights because of fear for costly redistributive fiscal policies. The predictions of the model are tested with a dataset of public opinion on immigration for Spain. The data supports these hypotheses in high migration regions. We also find support for other non-economic hypotheses posited in the literature. JEL Classification: D72, F22, H2, J61 Keywords: Public Opinion, Immigration, Spain, Political Economy Contact Author: Robert Duval-Hernández División de Economía Centro de Investigación y Docencia Económicas, A.C. Carretera México-Toluca 3655 Col. Lomas de Santa Fe, 01210 México, D.F. MEXICO Phone: (52-55) 5727-9800, ext. 2719 Fax: 5727-9878 robert.duval@cide.edu 1
1 Introduction A significant number of contributions in the literature on public opinion and immigration have focused on just one aspect of this relation, namely whether the amount of immigrants in a given country is about right or not. These studies rely on survey questions that ask respondents if they would like to see the levels of immigration increased, decreased, or stabilized. Although these are very valuable contributions, they are limited to analyzing the desired size of immigration. However, the relationship between public opinion and immigration is more complex than just this dimension. One important aspect to explore is the opinion about the rights and benefits that should be granted to immigrants. Some interesting questions on this regard include: What rights and benefits do natives want to offer to immigrants? What are the determinants of the political rights that the public would like to grant? What are the determinants of the willingness to grant economically costly social benefits, like education or health? Do the stated hypotheses that explain attitudes toward the level of immigration in the literature still explain the granting of benefits and rights? This paper explores some of these questions by formulating a simple political economy model that is empirically tested using a rich dataset of public opinion on immigration in Spain. By studying the rights and benefits that natives want to grant, we obtain a more complete view on the relationship between immigration and public opinion. Furthermore, this paper provides evidence on an interesting yet understudied case in the literature of immigration and public opinion, namely the Spanish case. 1 The Spanish case is interesting on its own, since Spain has shifted from being a country sending migrants to becoming a net receiver of immigrants in the last decades. In fact, beyond the North African migration, Spain has received the largest Latin American migration of any country in Europe, partly explained by historic ties and common language. Furthermore, the immigration from Eastern European countries has recently grown at extremely fast rates. This transformation has occurred in a relatively short period of time, and now Spain deals with an immigration rate of about twelve percent of its population. The paper is organized as follows. In section 2 we present a brief overview of the literature related to this study. Section 3 presents the theoretical model and discusses additional hypotheses that can be tested with the data. In section 4 we discuss the status quo regarding voting rights and welfare benefits to immigrants in Spain. Section 5 presents our data and empirical strategy, and section 6 presents the main results. We include our conclusions in 1 The works of Escandell and Ceobanu (2008), Escandell and Ceobanu (2010), and Martínez i Coma and Duval-Hernández (2009) are exceptions to this claim. 2
section 7. 2 Literature There is now a growing literature studying the characteristics and determinants of public opinion on immigration. The economic literature has focused mainly in analyzing two channels by which immigration affects public opinion, namely the labor market and the fiscal channel. Regarding the first channel, the main question is whether labor market competition between immigrants and natives of similar skill generates opposition to immigration, and whether natives whose skill is complementary to those of immigrants tend to support higher immigration flows. Evidence in this direction has been found for the United States in Scheve and Slaughter (2001); Hanson et al. (2007), and for a set of countries by Mayda (2006), to cite a few studies. The importance of a fiscal channel has also been explored, and evidence shows the relevance of this additional factor in explaining these attitudes (Dustmann and Preston, 2004; Hanson et al., 2007; Facchini and Mayda, 2009). Research has also shown that attitudes are also affected by whether the welfare state adjusts to immigration by modifying tax rates or the benefits provided (see for instance Facchini and Mayda, 2009). Our theoretical model on the preferences of natives for granting public benefits like education and health, and for granting the right to vote to foreigners builds upon the models of Dustmann and Preston (2004), Facchini and Mayda (2009), and Razin et al. (forthcoming). 2 The paper of Dustmann and Preston (2004) builds a model that incorporates labor market competition between natives and immigrants, together with considerations about impacts through the welfare state, mainly through tax adjustments after the entrance of immigrants. The paper also considers preferences over efficiency gains due to immigration. They test their model with data from the European Social Survey 2002 and find a strong role for fiscal factors in driving preferences toward immigration. The paper of Facchini and Mayda (2009) builds on this model and extends it to consider different types of adjustments of the welfare state to the entrance of immigrants. 3 In particular, it explores in more detail whether the welfare state adjusts to the aforementioned flows by altering the tax rate or the per-capita benefits offered to the population. The paper also extends the previous literature by exploring the cross-country heterogeneity in the impact of individual-level variables. 2 A related model is the one of Mariani (2007). 3 Hanson et al. (2007) present a model that captures similar features through a different approach. 3
The monograph by Razin et al. (forthcoming) analyzes several aspects of the relationship between the welfare state and immigration, in particular it studies how flows of immigrants of a specific skill group affect the welfare system, and hence the well-being of different types of natives. The book also studies the impact of the generosity of the welfare state on the type of immigrant that seeks to enter the economy, and it also analyzes the aforementioned impacts in a dynamic inter-generational setting. A related model analyzing the fiscal impact of immigration on the well-being of natives and that incorporates scenarios where migrants vote is Razin et al. (2002). This work shares several implications with our paper. Our paper starts from a framework similar to the one of Facchini and Mayda (2009). This allows us to capture labor market as well as welfare state effects, but instead of focusing on the two polar cases of tax vs. benefits adjustments to the entrance of immigrants, we consider the optimal choice of a publicly provided good through majority voting in the spirit of Razin et al. (forthcoming). In particular, we compare the welfare of natives when publicly provided benefits are granted to immigrants, and we study the equilibrium arising when the latter are allowed to vote over the optimal amount of taxes and of the publicly provided good. This framework allows us to focus on three different scenarios: one where immigrants do not have access to welfare benefits nor to vote, another one in which they have access to benefits but not to vote, and finally one where they are both fully participant in the welfare state and in the public decision making process. 4 It is important to remark that there is a large literature studying non-economic determinants of attitudes on immigration. In particular, the studies of Espenshade and Hempstead (1996); Dustmann and Preston (2007); Citrin and Sides (2006); Sides and Citrin (2007) to cite just a few, provide evidence on the importance of non-economic factors in explaining such attitudes. In particular, the importance of identitary factors, contact with immigrants, the perceived number of immigrants, among others are stressed in the literature in addition to traditional economic factors. Although we don t incorporate them explicitly in our theoretical model, we will control for several of these factors in the empirical specifications. Finally, one growing strand of studies closely related to ours appears in the sociology and social psychology literature. In particular, the papers of Scheepers et al. (2002), Raijman et al. (2003), Raijman et al. (2008), and Raijman (2010) study the preferences for granting civil rights to immigrants in European countries, Israel and the U.S. Their approach relies on arguments of competition (between natives and migrants), threat of immigration, and 4 The case where they are allowed to vote but not to access welfare benefits is not observed in practice in any country. Furthermore, as it will be seen below, virtually none of the respondents in the survey analyzed supported such a combination of rights. 4
identitary factors. Our approach shares several features of their competition arguments, since our theoretical model accounts for competition in the economic arena (mainly for welfare benefits and in the labor market). Also, as mentioned before, by incorporating variables capturing perceived threat of immigration, contextual variables, as well as identitary factors (like racial preferences) in the empirical analysis, we can test the importance of these reasons in shaping the respondents preferences on rights. 3 Theoretical Framework In this section we present a simple political economy model to shed light on what determines the preferences of natives for granting certain types of rights and benefits to immigrants. We start with a two-good, two-factor HO model extended with a welfare state similar to the models of Dustmann and Preston (2004) and Facchini and Mayda (2009). The productive factors are unskilled labor L u and skilled labor L s, which are combined in a Constant Returns to Scale production function q i = f i (L u, L s ) for i 1, 2. For simplicity, good 1 will be the numeraire good, and p will be the (relative) price of good 2. These prices are fixed, since we are assuming a small open economy. We will consider both the case of diversification (where the economy produces both goods) and the case of specialization in good 1. 5 The native population is normalized to 1, and the immigrant population is of size π < 1. The native population has a share ϕ of unskilled workers, while for the immigrant population this share equals ψ. For the rest of the paper we will assume that ϕ < ψ, i.e. immigrants are relatively more unskilled than natives, since that is the case for Spain, and of many developed economies today. 6 The implications of the opposite case can be easily derived from our model. Labor supply is inelastic for each group and equal to L u = ϕ + πψ L s = 1 ϕ + π(1 ψ) Wages are w s, w u for skilled and unskilled workers respectively, and we assume w s > w u. The equilibrium of the goods and the labor market in this economy has been presented elsewhere in the literature (see for instance Dustmann and Preston, 2004; Facchini and Mayda, 2009) and we include its crucial features in Appendix C. The main implication that follows from this equilibrium is that when the economy is diversified, the entrance of immigrants does not lead to wage changes, instead there is a 5 We assume no joint production between goods. 6 See for instance Instituto Nacional de Estadística (2007). 5
change in the output mix toward sectors intensive in the abundant input. This has been called factor price insensitivity in the trade literature, and is directly related to the Rybczinski s theorem (Leamer, 1995; Rybczynski, 1955). However, when the economy specializes in the production of a single good the entrance of immigrants will lead to a change in wages. In particular, if immigrants are relatively more unskilled than natives the wages of unskilled workers will fall, and the ones of skilled workers will rise. In this paper, we will develop the implications of our model under both a scenario of fixed wages and a scenario where wages change. There is a welfare system in the economy that levies a flat tax rate τ and redistributes uniformly a per-capita demogrant b among the participants in the welfare system. This demogrant will capture the public provision of health services and education. Denoting by the subscripts u, s the unskilled and skilled individuals, and by n, m the natives and migrants respectively, we express the (total) income of natives and immigrants as y n = w u ϕ + (1 ϕ)w s y m = w u ψ + (1 ψ)w s hence, under the assumption of a balanced budget, the budget constraint of the welfare system is τ(y n + πy m ) = b(1 + π). (1) In other words, the per-capita demogrant b equals the collected tax on the per-capita income of the economy, namely (y n + πy m )/(1 + π). An individual with skill j has a utility of u j = (1 τ)w j + γ j ln(b) j {u, s}. (2) In this quasilinear utility function the logarithmic term implies that the demogrant is an imperfect substitute of private consumtion. 7 The parameter γ j reflects the valuation of this good. We will assume that there is individual heterogeneity in this parameter, and that unskilled individuals value the demogrant more. In particular, we assume that γ j is distributed according to the density f j, i.e. γ j f j over the interval [0, γ s ] for skilled individuals and [ γ s, γ] for unskilled individuals. 8 We also assume γ < w u, a condition whose purpose will become clearer next. 7 Furthermore, this term satisfies an Inada condition that ensures the positivity of the b selected. 8 In other words the distribution of γ u First-Order Stocastically Dominates the one of γ s. 6
We assume the political equilibrium in this model is found by majority voting. In other words, the choice (b, τ ) of equilibrium will be the one that maximizes the utility of the median voter, i.e. (b, τ ) = arg max (b,τ) The solution to this problem is given by (1 τ)w µ + γ µ ln(b) subject to τ(y n + πy m ) = b(1 + π) b = γ µ y n + πy m w µ 1 + π τ = γ µ w µ (3) Our baseline scenario will be one where immigrants have access to the welfare system (by paying taxes and receiving the demogrant), but do not have the right to vote. As we will see below this situation describes the current status quo in Spain. Also note that at higher levels of γ µ there will be a more redistributive welfare system. 9 With this framework in mind, we now present the welfare analysis under different policy scenarios. 3.1 Immigration Without Wage Effects In this section we consider the simpler case where immigration does not have an impact on wages. As explained before, this corresponds to the case where our economy diversifies and produces both goods (q 1, q 2 ). 3.1.1 No Benefits, No Vote We begin with the case where immigrants are excluded from the welfare system (i.e. they receive nothing from the system, nor contribute anything to it either) and they have no voting rights. 10 The solution of the problem of the median voter in this case is obtained from (3) by setting π to zero, i.e. b = γ µ w µ y n τ = γ µ w µ (4) 9 The condition that γ < w u that we imposed earlier ensures we get tax rates strictly less than one. 10 Throughout the model we assume immigration as exogenously given, so that a change in b and τ will not impact the location decisions of potential immigrants. For research on this topic see for instance Borjas (1999) and Razin et al. (forthcoming). 7
In other words, the median voter doesn t change because, as in our baseline scenario, immigrants still do not vote. However, by excluding them from the welfare system there is less income to tax from, but at the same time there are less recipients of the demogrant among whom to share the fiscal revenues. From these results we derive our first proposition. The proof of this statement, and of all the other results in the paper, is included in Appendix A. Proposition 1. Excluding immigrants from the welfare system raises the per-capita benefits b accruing to natives. The reduction in b when (relatively unskilled) immigrants participate in the welfare system arises from a fiscal leakage effect (Razin et al., 2002). In other words, (relatively unskilled) immigrants are net beneficiaries of the system. While the above analysis models the economy as a single entity, in practice different regions within a country face very different amounts of immigration. Therefore, if a country has some degree of fiscal decentralization (as Spain does), the impact of immigration will vary according to regions. To capture this we will present a series of corollaries describing what happens to the derived results as the immigration rate rises. In practice, this is equivalent to treating each region as fully decentralized, having its own independent fiscal system. While this does not do justice to the intricacies of a mixed system where some degree of centralization/decentralization exists, it gives simple testable implications of what we should observe in high migration regions. Our first result in this spirit is the next corollary. Corollary 1. The gains in benefits from excluding immigrants from the welfare system will be greater in high immigration regions. This result is expected since in high immigration regions the fiscal leakage is larger, hence there are larger potential gains from preventing immigrants to accessing the welfare system. With these results in hand we can evaluate which natives will lose more from allowing immigrants to participate in the welfare system. This result constitutes our next proposition. Proposition 2. Unskilled natives will benefit more than the rest of natives from restricting the access of immigrants to welfare benefits. Since unskilled natives value more b, i.e. they have a larger γ, their losses are accentuated when this demogrant falls as a result of the fiscal competition with (relatively) unskilled immigrants. Furthermore, these utility changes will be accentuated in high immigration regions, as our next corollary shows. 8
Corollary 2. The utility loss for unskilled natives by giving immigrants access to the welfare system will be greater in high immigration regions. Proposition and Corollary 2 give us our first empirical implication of the model, namely that ceteris paribus unskilled natives will be more likely to oppose granting welfare benefits to immigrants, especially in high immigration regions. Our model contrasts with the tax vs. benefits-adjustment models of Facchini and Mayda (2009). In their case the fiscal system can adjust to the entrance of unskilled immigrants by either adjusting downward benefits (as in our model) or by raising taxes to keep the demogrant fixed. In our model, this last situation does not arise because of the voting model we propose. In particular, since natives are relatively more skilled than immigrants, in a voting context they will not support tax increases that will benefit immigrants more than proportionately. 11 3.1.2 Welfare Benefits and Voting Rights We next consider the case where immigrants have full access to the welfare system and have the right to vote. The main implication of granting voting rights to immigrants is that now the median voter preferences are affected by the preferences of immigrants as well. In fact, as we will show next, since immigrants are relatively more unskilled, and unskilled individuals value more the demogrant, allowing immigrants to vote will lead to more redistributive policies. This constitutes our first result in this section. Proposition 3. By granting voting rights (in addition to welfare benefits) to immigrants, the selected b and τ will be higher in comparison with the baseline scenario. In particular, in the proof of this proposition it is shown that the new selected (b, τ) will be given by γ µ y n + πy m b = w µ 1 + π > b τ = γ µ w µ > τ where γ µ and w µ are respectively the preference parameter and wage of the new median voter. However, the increase in benefits is not necessarily larger in high immigration regions. In particular, 11 In our model we do not consider different levels of generosity of b depending on the national origin of the recipient (see for instance, Hanson, 2005). 9
Corollary 3. In high immigration regions the tax increases brought about by voting immigrants will be larger than in low immigration regions. At the same time, the increase in benefits might be greater, equal or smaller than in low immigration regions. The intuition for this result is that as immigrants are allowed to vote they support more redistributive policies, financed through higher taxes. In particular, in high immigration regions we can expect a higher increase in taxes. However, these higher taxes need not translate in larger increases in b than the ones observed at low immigration regions. The reason for this seemingly paradoxical result is that while a higher immigration rate leads to more redistribution for a given level of per-capita income, per-capita income per-se falls as relatively unskilled immigrants enter the economy, making the final effect over the per-capita demogrant ambiguous. With these results in hand we can now evaluate who wins and who loses from granting voting rights to immigrants. Proposition 4. If an individual has γ < γ µ then he will be worse-off when immigrants vote, otherwise he will be better-off. In other words individuals who valued the demogrant b less than the original median voter in the baseline scenario will lose when a more redistributive policy is implemented. Again, these effects will be accentuated in high immigration regions. Corollary 4. The gains and losses of natives brought about by giving immigrants the right to vote are accentuated in high immigration regions. These last two results give us our second empirical implication for the paper. Namely, that since unskilled natives value more the demogrant, they will be more likely to support granting the right to vote to immigrants, than skilled natives who will fear a move toward more redistributive policies. This effect should be particularly stark in high immigration regions. These results complete the analysis of preferences for welfare benefits and voting rights in the absence of labor market effects. The next section takes labor market effects into consideration. 3.2 Migration With Labor Market Effects In our setting labor market effects will be characterized by changes in factor prices. This case occurs in our model when the economy specializes in the production of one good and the skill mix of immigrants is different from the one of natives. The derivations of the labor market effects of immigration can be found in Appendix C. 10
In particular, if immigrants are relatively more unskilled than natives then the wages of unskilled workers will fall, and the ones of skilled workers will rise. The new wages and income after the immigration shock will be denoted by (ŵ j, ŷ k ) respectively, for j {s, u} and k {m, n}. Furthermore, we will assume an initial immigration level of zero, i.e. π = π π 0 = π. This implies ŵ j = w j + w j π π = w j + w j π π ŷ k = y k + y k π π = y k + y k π π j {s, u}, k {m, n}. Our first proposition compares the demogrant and tax rate in the baseline scenario with and without wage changes. Proposition 5. If (ˆb, ˆτ ) denote the equilibrium demogrant and tax rate in an scenario where wages change, then ˆb < b and ˆτ < τ if the median voter is skilled. The opposite will be true if the median voter is unskilled. 3.2.1 No Benefits, no Vote From our analysis in section 3.1 we recall that if immigrants were relatively unskilled with respect to natives, there were utility losses for the latter from grating immigrants access to the welfare system, and these losses were larger for unskilled natives. The next two propositions compare the changes in benefits, taxes, and utilities in a scenario with wage changes versus another with fixed wages. Proposition 1.a. Excluding immigrants from the welfare system raises the per-capita benefits b accruing to natives, and the increases in b will be larger in the presence of labor market effects. Proposition 2.a. Unskilled natives will benefit more than the rest of natives from restricting the access of immigrants to welfare benefits, and their gains will be larger the presence of labor market effects. 3.2.2 Welfare Benefits and Voting Rights When immigrants are allowed to vote, and the economy is specialized in the production of one good the results from section 3.1.2 will still apply, although the changes brought about by the policy changes will differ in magnitude from the ones with fixed wages. In particular we have the following two propositions. 11
Proposition 3.a. By granting voting rights and welfare benefits to immigrants the welfare system will be more redistributive in comparison with the baseline scenario. Furthermore, the extent of the redistribution will be larger when wages are flexible. Proposition 4.a. If an individual has γ < γ µ then he will be worse-off when immigrants vote, otherwise he will be better-off. Furthermore, comparing the scenario with wage changes with the one of fixed wages we have that: If the median voter was unskilled before immigrants voted then skilled natives will have larger utility losses under flexible wages, while unskilled natives will have the same change in utility under both wage scenarios. If the median voter remains skilled after immigrants vote then unskilled natives will have larger utility gains under flexible wages, while skilled natives will have the same change in utility under both wage scenarios. If the median voter goes from being skilled to unskilled because of immigrants voting, then the utility gains and losses of natives will be larger under flexible wages. In summary, what we have learned from this section is that while the existence of labor market effects can alter the magnitude of the effects predicted in section 3.1, the direction of the effects is unaltered by changing wages in response to immigration. 3.3 Additional Economic and Non-economic factors Besides the individual economic factors identified in the model there are other potential causes to consider in the empirical analysis. An economic element affecting the willingness to provide welfare benefits and voting rights is determined by the perceived social impact of immigration. In other words, someone who considers that immigration economically hurts other members of society will be less likely to favor the provision of benefits that might function as welfare magnets to immigrants. Also non-economic factors like identitary feelings and contact with immigrants might shape the support for certain rights and benefits. 12 For starters the racial attitudes of natives are an important determinant of preferences toward immigration. In particular, individuals valuing racial diversity will be more likely to support granting voting rights and welfare benefits to immigrants, while the opposite 12 In an earlier version of the paper we also considered the perceived number of immigrants. However, since many respondents failed to answer such question we ended up with a reduced sample size that called into question the representativeness of our results. 12
will occur for those rejecting other races and seeing immigration as a threat to the core of national identity. Another factor to consider is that not all immigrant groups are the same. Some generate more acceptance while others more rejection. For instance, in the case of Spain according to the 2731 CIS survey in September 2007, 36.7% of the surveyed stated that there were some immigrants groups that inspire more trust than others. Concretely, 31.2% stated they trust Latin Americans more than any other group. Conversely, 51.2% affirmed that there were particular groups that did not inspire trust. For instance, almost 30% said that Romanians do not inspire trust; 13 and 18.9% stated they do not trust Moroccan, Maghrebis, and Algerians. 14 Finally, it is important to consider the type of contact natives have with immigrants. In particular, it is hypothesized that natives used to interacting with immigrants more often would be more inclined to expand rights and benefits than those who lack such contact. This is the contact hypothesis, first posed by Williams (1947) and later revised by Allport (1954) among others. However, the nature of contact must also be considered. In Burns and Gimpel (2000) words: presumably, interaction in the workplace is qualitatively different from interaction in other settings such as neighborhood, church or school (p. 209). The attitudes that any given individual has toward voting and social immigrants rights will vary depending on whether the contact with the immigrants is voluntary and if it leads to positive outcomes. 15 4 The current status of voting and welfare rights for immigrants in Spain The Spanish Constitution states that the welfare rights of foreign residents are the same as those of Spanish nationals (article, 13.1). Political rights though, are different since article 13.2 excludes the possibility of immigrants being elected and holding voting rights. There are three exceptions. First, since the adoption of Maastricht Treaty, EU citizens can vote in municipal and European elections. Second, immigrants may vote when so established by a treaty. And, third, when the principle of reciprocity applies -if a national from country A can vote in the elections of country B when residing there, nationals from country B can vote in the elections of country A, when residing there. 13 This percentage rises to 37.2% when adding the Eastern Europeans category 14 This figure raises up to 43.8% when including similar categories like Arabs (8.6%), Muslims (6.4%) or Moors (9.9%). 15 Those ideas were also posed by Sherif and Sherif (1953), and Allport (1954). 13
Although there are ius soli elements, the main way to automatically obtain Spanish nationality is through ius sanguini. 16 The debate on immigrant voting rights is on two grounds, namely: i) determining the territorial level at which immigrants could vote, and ii) which legal criteria should be considered to granting voting rights to foreigners (Zapata-Barrero, 2004; Zapata-Barrero and Zaragoza, 2009). A Spanish national can vote in 4 elections: local, autonomic, national and European. The reason why EU and reciprocity countries immigrants have the right to vote in the local elections is because it is argued that such elections are administrative and do not affect national sovereignty. This interpretation implies also that expanding immigrants voting rights to other elections, would necessarily bring a constitutional reform. 5 Data and Methods 5.1 Data The data used in this study are the Barómetro de Noviembre provided by the Centro de Investigaciones Sociológicas for November of 2005. The survey is a nationwide sample of Spaniards 18 years of age and older. The sample covers 47 provinces, and the sampling procedure is multi-stage stratified by clusters. The questionnaires were applied through direct interviews at the home of the interviewees. We have two dependent variables. The first one takes value 1 if the respondent believes that immigrants should be granted the right to vote in both general and local elections, and is zero otherwise. The second one takes value 1 if the respondent believes immigrants should have access to public education (for them and their children) and to free public health, and is zero otherwise. 17 A summary of the interrelationship between the attitudes toward voting and welfare rights for immigrants is presented in Table 1. TABLE 1 ABOUT HERE This table shows that the vast majority of respondents (about 85%) support giving both 16 Also, automatic nationality is guaranteed if one of the parents was born in Spain or if the person would be stateless. 17 A detailed description of the variables used in the analysis is included in Appendix B. 14
health and education benefits to immigrants, and more than 60% support allowing them to vote in both general and local elections. In fact, about 56% of the respondents support granting both full welfare and voting rights. However, the fact that more than 30% of them refuses offering any voting rights to immigrants, while only 4% do so for welfare benefits indicates that there is something substantially different between these two rights. While these are high values in terms of the willingness to grant such benefits and rights, we have to consider that there might be non-economic factors behind them, and that in an interview respondents might overstate such willingness. However, we will show that even at these high levels of support, the economic hypothesis previously posited hold in the data. Before discussing the econometric model it is interesting to make a simple exploratory analysis of how health and education benefits change with immigration. In Figure 1 we show the per-capita public expenditure in Health and Education against the immigration rate at the Autonomous Community level. It is clear from this graph that there exists a negative relationship between these two variables, giving support to the idea that the provision of such benefits is adjusted downwards as immigrants enter the economy. FIGURE 1 ABOUT HERE 5.2 Empirical Methodology In this paper we model the decision to support granting voting rights and welfare benefits to immigrants through two equations relating each binary dependent variable to a set of sociodemographic characteristics that can help us test the hypotheses posited. In particular, we use a Bivariate Probit model to estimate these equations. In other words, we assume that individuals have an underlying propensity y1 to provide political rights and another one y2 regarding the provision of welfare benefits. These propensities are related to a set of sociodemographic characteristics X by the following equations: y 1 = Xβ 1 + u 1 y 2 = Xβ 2 + u 2 where u 1, and u 2 are a set of unobserved characteristics assumed to be jointly distributed with a bivariate standard Normal distribution, i.e. 15
( u1 u 2 ) N ( 0, 1 ρ 0, ρ 1 In practice we do not observe the underlying propensities, but indicators on whether the respondents want to provide rights and benefits, i.e. ) y 1 = 1 if y 1 0 y 2 = 1 if y 2 0 The set of explanatory variables include characteristics of the respondent like gender, age, schooling level, marital status, religion, and employment status. It also includes variables at the regional level like the province immigration rate, the shares of different immigrant groups, the regional GDP per capita, and a dummy variable indicating whether the province is a regionalist autonomous community. 18 We include interactions between the education level, the employment status of the individual and the immigration rate at the province level. The interactions between education levels and the province immigration rates serve to see how the natives preferences vary with skills at different levels of immigration. These effects will be crucial to test the hypotheses derived from the theoretical model. The further interaction of these effects with the labor market status of the respondent is done to isolate preferences driven by fiscal reasons from those driven by both fiscal and labor market impacts. In particular, we consider the effects obtained for individuals out of the labor force to be unaffected by labor market effects, and hence we will focus on them to test our fiscal hypotheses. A second set of estimations includes, in addition to the previous variables, a set of indicators on whether the respondent has a preference for living in a culturally diverse environment, prefers not having an immigrant as a boss or married to a close relative, (to test identitary hypotheses), has contact with immigrants through work, or as a neighbor, (to test the contact hypothesis ) and we also include variables on whether the respondent believes immigration: reduces natives wages, negatively affects the poor (as long as the respondent is non-poor) or helps fill vacant jobs (to test non-individualistic economic hypotheses). The reason for including this additional set of variables in a separate estimation is because some of them could be potentially endogenous to the stated preference toward rights and benefits. In spite of this potential endogeneity, we include them to explore the relationship between these factors and the dependent variables after controlling for socio-demographic 18 The provinces considered under this category are Catalonia, Galicia and País Vasco. 16
characteristics. This allows us to corroborate whether the sign predicted by some of the hypotheses listed above is confirmed by the data. 6 Results 6.1 Voting Rights and Welfare Benefits Table 2 includes the marginal effects of the explanatory variables on the probability of granting voting rights Pr(Right to Vote=1) and of granting welfare benefits Pr(Welfare Benefits =1) for the first set of estimations (the reduced model without the attitudinal controls). This table presents only the effects that serve to test the hypotheses previously mentioned. The full set of results can be found in Appendix B. TABLE 2 ABOUT HERE For individuals out of the labor force, having a high level of education and being in a region without immigrants is associated with a propensity almost 30% higher to offer voting rights relative to our reference group, the unskilled individuals in low migration regions. In principle this would seem to contradict our theoretical prediction that skilled natives oppose voting rights for immigrants. However, as immigration grows at the province level there is a decrease in the willingness to grant such rights, and this decline is steeper for the highly skilled. For this same group of natives out of the labor force, the highest opposition to giving welfare benefits is found among the unskilled in high immigration regions, as the model predicts. The individuals in the labor force have their preferences affected by the fiscal impacts of immigration as well as by the potential changes occurring in the labor market. For these individuals we still find a higher opposition to offering voting rights among the skilled in high immigration regions (e.g. a probability 2.2 percent smaller vs. one 1.7 percent smaller for the unskilled), although such difference is not statistically significant. For the case of health and education benefits, the effects are quite similar to the case of natives out of the labor force, namely there is a higher opposition among the unskilled, especially in high immigration regions. To visualize these relationships more clearly we present in Figure 2 the predicted differences in probability of support for each type of policy between the more skilled Spaniards (those with some College education or higher) and the less skilled (those with 6 years or less 17
of education), by immigration at the province level and labor force status. These predictions arise from the bivariate probit previously estimated. In this figure a positive number means that ceteris paribus skilled natives support a given policy more than the unskilled do. A 90 percent confidence interval is also presented to visualize the statistical significance of the predicted difference. FIGURE 2 ABOUT HERE In this graph we can observe that for natives out of the labor force the skilled living in low migration regions are more likely to favor voting rights for immigrants, yet such preferences are drastically reversed as we move to high immigration regions where the opposition is the highest among this group and statistically different from zero. For individuals in the labor force we don t find a statistically significant difference in support to voting rights between the skilled and the unskilled. This might reflect that skilled workers benefit from an influx of unskilled immigrants due to their complementarities in the labor market, 19 and hence they are not as opposed to granting them rights if that is going to foster migration flows that benefit them. Regarding the provision of health and education benefits the graphs show that the rejection to such provision occurs among unskilled natives in high immigration regions. Furthermore, this relative rejection becomes higher among unskilled natives in the labor force, something consistent with the effects of labor market competition between groups of similar skill. Regarding the relationship of the dependent variables with the penetration of specific groups of foreigners at the province level, we find a large negative relation between the provision of these rights and benefits and the share of Eastern European immigrants (Bulgarians and Rumanians) at the province level. Interestingly, a higher share of Moroccans increases the propensity to offer welfare benefits, something that would appear controversial given the negative attitudes of Spaniards toward Muslims. A possible explanation for these findings is that the rapidly growing migration of Eastern Europeans in recent years has generated a more negative attitude toward this group, while natives might be more prone to grant health and education to a group like Moroccans in order to speed-up their assimilation process. Table 3 presents the results for the model that includes attitudinal variables about immigrants as regressands in the model. 19 Amuedo-Dorantes and de la Rica (2008) present some evidence in this direction. 18
It is worth noting that in the Vote equation the parameter for skilled natives out of the labor force in low migration regions falls by 7% once we control for the attitudinal variables. This indicates that part of the high positive propensity found among such group is in part reflecting non-economic factors. Secondly, the parameters for the ethnic shares of different immigrant groups also change, mostly in magnitude rather than in direction. This indicates that the attitudes toward these groups are partly based on identitary and contact factors. Regarding the relation between identitary factors and the dependent variables, we find that having a preference for cultural diversity is positively associated with granting voting rights and welfare benefits to immigrants, while individuals who reject other races tend to have about a two percent smaller willingness to grant immigrants the right to vote as well as health and education. In other words, the racial attitudes of natives matter in determining the rights and benefits they want to offer foreigners, as identitary hypotheses posit. Regarding the contact hypotheses we find that contact with neighbors increases the willingness to grant health and education by about four percent We find some support for non-individualistic economic hypotheses given that individuals who think immigration lowers wages, and hurts the poor are less likely to grant voting rights and welfare benefits respectively, while those respondents who believe immigrants fill needed job vacancies are more likely to grant such rights. In figure 3 we reproduce again the predicted differences in support for rights and benefits between skilled and unskilled natives. This graph shows an even clearer pattern of opposition to immigrants voting among skilled natives out of the labor force, and a higher relative support for welfare benefits among skilled respondents. FIGURE 3 ABOUT HERE The fact that the opposition of skilled natives to voting rights in high immigration regions is clearer now, is an indication that their preferences are mediated by non-economic factors, that once controlled for, lead to a clearer test of the posited hypotheses. In summary, when looking at regions with a high influx of immigrants the empirical findings are in line with the predictions stemming from the theoretical model. While there might be non-economic factors making skilled natives more prone to supporting voting rights when they live in a region with few immigrants, these factors are outweighed in high immigration regions by the fear of potential redistributive policies brought about by voting immigrants. 19
7 Conclusions In this paper we present a model to explain why natives might be willing to grant voting rights and welfare benefits to immigrants and we test the predictions of this model using a rich dataset on public opinion about immigration for Spain. The theoretical model predicted that when immigrants are more unskilled than natives are, unskilled individuals reject the provision of health and education to immigrants, because of the competition for fiscal resources it implies to them. On the other hand, it predicted that skilled natives would oppose granting voting rights to immigrants, because that would lead to more redistributive fiscal policies to be financed through higher taxes. We find support for these hypotheses in the data at high immigration regions. We also find support for other hypotheses that have been used in the literature to explain a general attitude toward immigration. In particular, we find that respondents care not only about the direct impact of immigration on them, but also about the impact on the overall economy. We also find support for identitary hypotheses in the sense that the racial attitudes of respondents as well as the country of origin of immigrants matter in explaining the willingness to grant voting rights and health and education to immigrants. Contact with immigrants only matters to the extent that having immigrant neighbors leads to a higher willingness to granting welfare benefits to them. In general, this study confirms that the relationship between public opinion and immigration is more complex than simply a general attitude toward immigrants. This calls for conducting further studies focusing on several aspects of this relationship like the preferences on the type of admission policies to be implemented, and the mechanisms favored by the public to assimilate immigrants into the host society. Previous research of ours shows 20 that a vast majority of Spaniards want the current levels of immigration reduced. Yet in this paper we saw that at the same time the majority of them favors granting a wide range of rights and benefits to immigrants. It would be interesting to study the interrelationship between the desired number of immigrants and the types of rights and benefits that natives want to grant them. A better understanding of these issues not only provides a richer picture on the political preferences of natives in host countries, but is also relevant for public life, since it may define future political strategies regarding immigration rights, and the shaping of immigration policy in general. 20 See Martínez i Coma and Duval-Hernández (2009). 20
References Allport, G. (1954), The Nature of Prejudice, Cambridge, M.A.: Addison-Wesley. Amuedo-Dorantes, C. and S. de la Rica (2008), Does Immigration Raise Natives Income? National and Regional Evidence from Spain, IZA DP No. 3486. Borjas, G. J. (1995), The Economic Benefits from Immigration, Journal of Economic Perspectives, vol. 9(2): 3 22. Borjas, G. J. (1999), Immigration and Welfare Magnets, Journal of Labor Economics, vol. 17(4): 607 637. Burns, P. F. and J. G. Gimpel (2000), Economic Insecurity, Prejudicial Stereotypes, and Public Opinion on Immigration Policy, Political Science Quarterly, vol. 115(2): 201 225. Citrin, J. and J. Sides (2006), European immigration in the people s court, in Immigration and the Transformation of Europe, (eds.) C. A. Parsons and T. M. Smeeding, Cambridge: Cambridge University Press. Dustmann, C. and I. P. Preston (2004), Is Immigration Good or Bad for the Economy? Analysis of Attitudinal Responses, Research in Labor Economics, vol. 24(3). Dustmann, C. and I. P. Preston (2007), Racial and Economic Factors in Attitudes to Immigration, The B.E. Journal of Economic Analysis & Policy, vol. 7(1). Escandell, X. and A. M. Ceobanu (2008), When Contact with Immigrants Matter: Threat, Interethnic Attitudes and Foreigner Exclusionism in Spain s Comunidades Autónomas, Ethnic and Racial Studies, vol. 32(1): 44 69. Escandell, X. and A. M. Ceobanu (2010), Nationalisms and Anti-immigrant Sentiment in Spain, South European Society and Politics, forthcoming. Espenshade, T. J. and K. Hempstead (1996), Contemporary American Attitudes toward U.S. Immigration, International Migration Review, vol. 30(2): 535 570. Facchini, G. and A. M. Mayda (2009), Does the Welfare State Affect Individual Attitudes Toward Immigrants? Evidence Across Countries, The Review of Economics and Statistics, vol. 91(2): 295 314. Feenstra, R. (2003), Advanced International Trade: Princeton University Press. Theory and Evidence, New Jersey: 21