نظریه ای از تنوع قومی و توزیع درآمد: رویکرد چانه زنی قانونی
|کد مقاله||سال انتشار||مقاله انگلیسی||ترجمه فارسی||تعداد کلمات|
|11094||2013||16 صفحه PDF||سفارش دهید||محاسبه نشده|
Publisher : Elsevier - Science Direct (الزویر - ساینس دایرکت)
Journal : European Journal of Political Economy, Volume 34, June 2014, Pages 52–67
In this paper, I examine how the two dimensions of income inequality and ethnic diversity affect the formation of political coalitions and the reallocation of income. I construct a legislative bargaining model to analyze when coalitions form along the income dimension and when it forms along the ethnic dimension. The results of the model suggest that the agenda setter selects a partner with a lower aggregate income: a smaller population size and a lower income level. Further, I extend the model from a one-round to a two-round game. The extended model shows that forming an oversized coalition is a possible equilibrium outcome and is consistent with the empirical findings.
Research suggests that ethnic diversity has a negative impact on an economy because it creates political conflict.1 In this regard, Alesina and Glaeser (2004) argue that ethnic diversity is not necessarily equivalent to ethnic conflict because the negative effect from diversity does not appear if the ethnic minority is rich. However, their argument does not quite hold for the case of Rwanda where the Tutsi minority is also relatively rich but still fights for redistribution. Here lies the starting point of this paper: If the ethnic minority is rich, does it always decrease the ethnic political conflict? This paper partially answers this question by analyzing the formations of political coalitions along two dimensions: income inequality and ethnic diversity. A gap in the literature exists based on the assumption that ethnic diversity leads to ethnic political conflict. Consequently, very few studies study the effect of diversity when the diversity is multidimensional. This paper tries to fill this gap by examining under what conditions ethnic diversity leads to political coalitions based on ethnicity. In this respect, Przeworski (2005) claims that the relative size of an ethnic group does not necessarily correlate with its share of the vote; that is, the population size of an ethnic group does not have to coincide with that of the political party that represents the group. This finding suggests that the political outcome does not always reflect the ethnic diversity in the population. In this regard, Posner (2005) also argues that people usually have several attributes such as linguistic, income level, or ethnicity along which they vote. The following is a modified example from Przeworski in which everybody can stand as a candidate as in the citizen–candidate model.2 There are two ethnic groups, A and B, and some of the members are rich and some are poor. Thus, we can classify people into four groups: A and rich, B and rich, A and poor, and B and poor. If each group has a population size smaller than one-half, then in order to get a majority each group has to form a coalition with another. They can form a coalition either along ethnic lines or according to their income class. Because they are dividing a fixed pie if they win, they want to form a coalition that is a majority but with as small a population size as possible. The table below presents two such cases. Full-size image (37 K) Figure options The numbers in each cell express the population of each group respectively. In both tables, the ethnic minority B is relatively rich (the proportion of the rich is larger than the poor). In the left table above, the rich group, A and B, forms a coalition because its population size of the rich (55%) is smaller than that of the ethnic majority group A (60%).3 On the other hand, in the right table, the ethnic group A can win with a smaller portion of votes (55%) than the coalition of the income class group, the poor (60%); thus, group A forms a coalition as in the case of Rwanda. In both cases, the ethnic minority B is relatively rich; however, the political result is different. Political ethnic conflict is more likely to occur in the case of the right table, where the coalition is formed along ethnicity. Although this paper is similar to that of Przeworski (2005), two main points exist that makes this paper significantly different. First, Przeworski's example mainly deals with voting; however, in this paper, I consider the formation of a government one stage after voting.4 Second, Przeworski only considers the population size of the group, whereas this paper also takes into account income distribution. In fact, adding income inequality to ethnic diversity leads to interesting findings: The agenda setter chooses to form a coalition with the group whose population size is smaller, and the income level is relatively lower. This result occurs because this type of group has less coalition options, and thus, subsidization is cheaper for this type of coalition partner. This finding might provide a key insight into the question of why ethnic-based coalitions exist in most of the countries in Africa where each ethnic group is small in population size and their income level is low. In Table B.1, I summarize the coalition types across world regions for the period 1975 to 2006. I define a noneconomic coalition as either (1) a political coalition in the government where the right- and left-wing parties coexist or (2) the largest party in the government where there is neither left nor right wing. As Table B.1 shows, compared to other regions, the share of noneconomic coalitions over all of the types of coalitions is much higher in Africa. Another theoretical finding of this paper is that when the poor are in the majority, and the income difference between the rich and the poor is greater, political coalitions form along class lines. In fact, as Table B.3 shows, this finding partially explains the political situation in Latin America where the income inequality is very high and coalitions based on class are more likely to be formed. When the rich are in the majority, this relation between the income inequality and the formation of coalitions reverses: If the income inequality is greater, ethnic coalitions are likely to form. Kuznets (1955) shows that income inequality increases when countries are in transition from low- to middle-income. Afterwards, when countries become more economically developed, income inequalities tend to decrease. Therefore, coalitions along ethnicity form only in poor countries but not in the middle-income and rich countries. This finding might give some insight into why ethnic coalitions form in developing countries, but not often in developed countries, as Table B.1 indicates. These results occur only for a one-round bargaining game and are consistent with Riker's (1962) minimal winning coalition theory: A coalition forms with as small a group as possible as long as it is winning. However, in reality, this is not the case all of the time. In fact, as in Volden and Carrubba (2004), almost half of all coalitions are oversized rather than minimal. In this paper, an oversized coalition is defined as “the coalition which cannot be formed with a smaller population size by replacing the party inside of the coalition with the party outside of the coalition.”5 In Table B.2, I summarize the share of oversized coalitions over all of the types of coalitions across the different regions. In fact, in some regions, more than half of the coalitions are oversized. To explain the mechanism of oversized coalitions, I extend the model from a one-round to a two-round bargaining game. When the game has two rounds, the agenda setter chooses the group with the larger population size and higher income with which to form a coalition. This is because the group with the larger population size and higher income has a lower expected utility when the game proceeds to the second round, and thus, its reservation utility is lower in the first round. In turn, a lower reservation utility allows the agenda setter to more easily make an offer to the group. This finding is a new explanation in the field of oversized coalitions, which is another contribution of this paper. Most theoretical studies predict minimal winning coalitions,6 but there are some that predict oversized coalitions similar to this paper. But there are few theoretical papers that try to explain the mechanisms of oversized governments.7 For example, similar to this paper, Diermeier and Merlo (2000) also construct a legislative bargaining model to analyze why oversized or minority governments form. However, the main difference between this paper and theirs is that in their model, the oversized government is caused because the formateur could extract more from the other parties. But Bandyopadhyay and Oak (2008) show that an oversized coalition is the result of the formateur trying to balance the policy around his or her ideal points.8 In contrast, in this paper, the coalition with the larger population size can be preferred when the rich are the majority because the rich prefer lower taxes and less deadweight cost for redistribution. As argued in Volden and Carrubba (2004), there are others who try to investigate the mechanisms of an oversized coalition. However, none of the existing studies' arguments for oversized coalitions are similar to this paper. In this regard, this paper tries to give a new viewpoint on oversized coalitions vis-à-vis the reservation utilities of the partner groups. There are several empirical studies that report a negative relation between ethnic or religious diversity and economic development. The leading papers are by Easterly and Levine (1997), Alesina et al. (1999) and Alesina et al. (2003). Alesina et al. (2003) extend the measurement of diversity in society by adding linguistic and religious diversity. Further, Keely and Tan (2008) empirically show, using the General Social Survey data, that people believe that redistribution should be based on race, sex, and income class in the United States. Their results strongly encourage a theoretical analysis in which the political parties differ in multidimensions. In this respect, Roemer (2004) theoretically analyzes political equilibrium with multidimensional preferences. This paper is more specific and elaborates on the stage of the government's formation. There are only a few theoretical studies regarding ethnic diversity. The first approach introduces the externalities that arise from the diversity of people and examines how these externalities affect the welfare of society or economic growth. For example, Esteban and Ray (1999) introduce externalities and analyze the links between the level and pattern of conflict and the distribution of the groups of people with specific interests. The second approach analyzes political games among parties to implement their preferred policy, as in this paper. One of the studies categorized in the second approach, Fernandez and Levy (2008), investigates political equilibrium with diversity in the population and argues that there might be less redistribution to the poor with a higher degree of diversity. There are some limitations to their model. First, their result of a nonmonotonic relation between diversity and redistribution depends crucially on the assumption that an exogenous fixed cost exists for the provision of public goods. Second, they assume that rich people are ethnically homogeneous, and only the poor differ along ethnic lines. In my model, the rich also differ along ethnic lines, and whether the rich unite across different ethnic groups or not is endogenous to the political process of forming a coalition. Similarly Azzimonti-Renzo (2005) analyzes how a government policy results in an inefficient allocation when there are two opposing political groups that have a conflict over the redistribution allocation. She uses a recursive formulation with probabilistic voting and succeeds in showing the negative relation between ideological diversity in society and economic growth. In other words, the assumption is that once the party wins the election, the promised platform is implemented (direct democracy). This paper analyzes the stage after voting, that is, the formation of a political coalition within the government.9 The paper proceeds as follows. In the next section, I present the model and its findings.