ON ISLAM and INEQUALITY
Introduction

Fish (2011) in his book named “Are Muslims Distinctive: A Look at the Evidence.” discusses whether Muslims show any systematic differences on individual and country levels in various issues such as democracy, large scale violence, corruption, tolerance and social inequality. In terms of inequality, Fish compares Muslim and non- Muslim countries in relation to gender and income inequality. As a result of his data analyses he concludes: “On gender based inequality Muslims lag; on income inequality they lead” (Fish, 2011: 228). I consider the finding about income inequality as an unexpected one because in the income inequality literature, religion in general and Islam in specific are not commonly considered as potential determinants. Fish`s proposed causal mechanism makes his study even more suspicious because his references are anecdotal and his reasoning is culturalist. This is why, I decided to investigate his findings deeper and conduct a test on his claims that Muslims are culturally more egalitarian than Christians. The plan of this paper works in the following way: I will start with a literature review and continue with a brief summary of Fish`s data analysis and his major claims about the causal mechanism behind the relationship between Muslim culture and egalitarian state policies. Afterwards, I will continue with my alternative replication of Fish`s regression analysis with a different dataset which measures inequality in an alternative way. Then, I will introduce my test in relation to one part of his causal mechanism and attempt to answer the question whether Muslims are really more egalitarian or not. Finally, I will discuss my alternative causal mechanism idea which is religiosity should be added to the analysis as an interactive variable with the religion because religion variable alone without religiosity is not an ideal indicator of culture. What matters is not being Muslim, Protestant or Jewish only but a religious Muslim, Protestant or Jewish, because only after than religion should be expected to play a significant impact on the economic ideals of individuals.

Literature on the Preferences for Redistribution

Fish is right in his claim that there is a scarcity of academic research when it comes to the relationship between religion and redistribution preferences of individuals. Still, the demand side of redistribution is largely covered in the comparative political economy literature, from which there are some important implications to take before digging deeper into Fish`s study. My goal with this short literature review is to see where the cumulative research efforts stand in approaching to religion in general and Islam specifically as a determinant of equality preferences of individuals. It will also guide me in clarifying what some key control variables should be and what alternative causal mechanisms are proposed by the leading academicians of the field. There are three main groups of approaches to determinants of redistribution: personal history, cultural factors and historical experiences in a larger scale. Redistribution preferences as a consequence of these three factors are investigated through a wide range of indicators (Piketty, 1995). History of misfortune in one`s life might have a negative effect on egalitarian ideas. Similarly, in the country scale, going through traumatic periods such as economic crises might turn people into less or more supportive individuals of redistribution (Alesina and Glaeser, 2004). Long term indoctrination such as in the case of communist regimes are known to be important in defining redistribution preferences also (Alesina and Schundeln, 2007). Again, as a part of life experience, perceptions of justice and effort in relation to success can have a decisive role in one`s life (Bonabou and Tirole, 2006). Structure of the family in terms of economic and social dependency on the other family members can be another indicator of the impact of personal history on egalitarianism (Todd, 1985 and Esping- Andersen, 1999). Some cultures might emphasize redistribution as a virtue more than others, as Alesina and Gleaser`s (2004) comparative analysis of Europe and the US shows. Finally, the desire to act in accordance with public values or the will to obtain social status might say something about people`s preferences for sharing through welfare (Corneo and Gruner, 2002). Among these studies, there are some common findings guiding researchers in terms of what variables matter to control for. First of all, individual income is expected to matter: richer people are expected to be less supportive of redistribution. Second, gender matters: women controls show consistent significance in terms of egalitarianism (Alesina and Giuliano, 2009:3). Education is an interesting variable because its significance becomes somewhat ambiguous once controlled for income. However, it is still worth to control for, in general. Race as the black people is another important variable in the context of the US politics, but there is no sufficient evidence to generalize that conclusion for a wider spectrum of countries. Accordingly, in my analysis I will be controlling for education, GDP per capita income in my analysis. I also want to add life expectancy measure because Fish does so and finds statistical significance in his analysis. When it comes to religion, there are some commonly reached results among academia: Protestants are tended to be found less egalitarian while being raised as Catholic or Jewish increases the tendency for support for equality (Alesina and Giuliano, 2009:17). When controlled for ideology, all religions including Protestants become more egalitarian. This finding suggests us to include ideology in the regression analysis also. Finally, regardless of religion, religiosity seems to be a factor increasing tastes of redistribution in general. This finding is an important one because religiosity, in addition to religion, is arguably an important variable in determining redistribution preferences of individuals. There are some studies supporting this theoretical direction that I want to briefly summarize below. Davis and Robinson (2007) have an interesting theory, which they name as the moral cosmology theory. According to this view, there are two types of theological approach to economic beliefs: modernist vs. religious orthodoxy. Modernism dictates individual responsibility and does not expect individuals to feel moral responsibility for the poor. On the other hand, religious orthodoxy is communitarian and views the community of people as a group of believers, who are responsible of each other. After testing this theory in the US, 21 European countries and Israel comparatively and finding supporting results, they continue with the Muslim world and search for conditions of support for economic justice in seven predominantly Muslim countries: Algeria, Bangladesh, Egypt, Jordan, Indonesia, Pakistan and Saudi Arabia. They test their theory, which they claim applies to all Abrahamic religions, for the Muslim countries via survey data analysis and find out that those who support Islamic rule of sharia also show communitarian tendencies within these seven Muslim countries. Guisa, Sapienza and Zingales` study (2003) is another importantly relevant one at this point because it combines religiosity and religion to see their effects on people`s preferences in various realms of life including egalitarianism. As I will be doing in this paper, they also majorly rely on World Values Survey in conducting their research. About the general effects of religiosity, they find that, on average, religious people trust others more, trust the government more, they are less willing to break the law, and believe more in the fairness of the market. However, they are more intolerant and they have less progressive attitudes towards women (Guisa, Sapienza and Zingales, 2003:25). The writers ask, whether these figures change once controlled for six major religious denominations: Catholics, Protestants, Jews, Muslims, Hindus, and Buddhists (Guisa, Sapienza and Zingales, 2003:27). Results show that religious upbringing has a negative and statistically significant effect on trust for Catholics, Muslims, and Hindus. Religious denominations also differ in their attitudes toward private ownership.
When it comes to ideas on redistribution, Protestants, Catholics, and Hindus want more private ownership, while Muslims want significantly less private ownership. Catholics are more in favor of competition than any other religious group, including Protestants, while religious Muslims and Hindus are strongly against competition. In general Muslims seem to have anti-market tendencies. Inspired by these research, I claim that religion`s effect on preferences of income inequality depends on the religiosity factor. As Guisa, Sapienza and Zingales` research (2003) shows, religiosity matters a lot in understanding the role of religion on issues from trust to economic preferences. Davis and Robinson (2007)`s study, on the other hand, shows us that religios communities of Abrahamic religions have common qualifications cutting through different cultures. That is why religiosity is a relatively more manageable measure which should not be neglected when the main independent variable is religion. Accordingly, I expect, what matters is not coming from a Muslim country or Christian alone but also coming from a religious background. Thus, I think, these two variables have an interactive, instead of additive, relationship and religiosity is the key factor in telling us a story about religion`s impact on redistribution preferences.

Fish: Predominantly Muslim Countries Are More Equal Fish asks whether countries with a majority of Muslim population are more egalitarian or not in terms of “class inequality,” which he operationalizes as income inequality. Fish compares Muslim countries with the Christian ones on the one hand and the non-Muslim ones on the other hand in terms of the Gini scores from UNDP data source. The 0 Gini score means perfect equality while 100 indicates perfect inequality. None of the countries are perfectly equal or unequal, the real range in practice is between the score 25 making Denmark, Sweden, Iceland, Japan the most egalitarian cases and the score 74, which declares Namibia as the most unequal country.
Fish`s comparative data shows that predominantly Muslim countries have moderately lower average Gini scores compared to the non- Muslim as well as Christian countries. Results show that in Muslim countries the UNDP Gini score ranges between 25 and 50, while for the non-Muslim countries the range is 25 to 74. The average score for Muslim countries of the sample size 26 for the years between1990 to 2006 is 38.0, while it is 41.1 for the 114 non-Muslim countries. On the other hand, for Muslim- Christian comparison, the 15 Muslim countries` average is 39.0 compared to 42.3 average of the 19 Christian countries (See Figure 1). The analysis is controlled for the variables of GDP per capita, life expectancy, level of democracy and post- Communist country. In four different models, the Muslim percentage explanatory variable seems to hold its statistical significance (See Figure 2).

Figure 1: Largest Muslim vs. Christian Countries and Income Inequality

Figure 2: Regressions of Income Inequality on Hypothesized Regressors (Fish, 2011:219)

Fish: Muslims are more egalitarian than Christians.

Fish provides anecdotal evidence indicating the importance of informally institutionalized rule of justice in the Muslim thought. Fish claims, justice is commonly understood in socioeconomic terms in the Muslim world. He argues that socio-economic justice is a prioritized norm in the Muslim world. Its source is the Quran. Fish`s comparative textual interpretation shows that even though other holy books also refer to redistribution as a virtue, Quran is the most explicit one about its practice. He specifically focuses on the practice of Zakat in Islam, which is one of the five mandatory acts for Muslims in addition to the other four: being profession of faith, prayer five times a day, fasting during Ramadan month every year and visiting Mecca once in a life time In Islamic teaching, the poor is portrayed as a part of larger community, in relation to whom he has material and social rights. Accordingly, the poor is given the right to some specific percentage of resources of the other members of the community. The wealthy, on the other hand, is depicted as specifically vulnerable because they are in a bigger danger of forgetting about the non-material after world that they should live for. They are also responsible of the wellbeing of the poor in their close community. To Fish (2010: 220-221) this kind of explicit guidance for redistribution is unique to Islam. Justice, in the Muslim world, is understood specifically in reference to socio-economic equality and it is a source of pride for Muslims to be “just.” Accordingly, Fish states: “In broadest general terms, it (socio-economic justice) is the essence of Muslim ideal and message, much as the essence of the Christian ideal is love. The Buddhist equivalent is…oneness, or transcendental unity, of all things.”
As a result, this norm reflects to the political party discourses and policy options of the political parties in the Muslim world. To get the votes of ordinary citizens, political parties have to emphasize redistributive policies. This is how predominantly Muslim countries end up more equal than the non-Muslim ones. Fish compares the norm of justice in the so called “Muslim economy” with the individualist norm of “Christian economy.” He argues that Christian understanding of economy cherishes being rich and independent, that is why people do not understand justice as a norm of socio-political economy, but more as a norm of economic liberalism, property rights and economic mobility. Fish (2011:222-223) refers to the so called gospel of evangelical Protestanism as the opposite of Muslim economic philosophy. Evangelical Protestants spread the idea that being rich is a sign from God showing that He appreciates the efforts of the hardworking followers. In the Muslim understanding, there is no place for this kind of an approach, claims Fish.
In World Valus Survey, there are various questions related to perceptions and preferences of inequality asked to people from all around the world since 1981, when the first wave was conducted. The most direct question measuring this issue is the following one:

Now I'd like you to tell me your views on various issues. How would you place your views on this scale? 1 means you agree completely with the statement on the left; 10 means you agree completely with the statement on the right; and if your views fall somewhere in between, you can choose any number in between. Sentences: Incomes should be made more equal vs We need larger income differences as incentives. (WVS, Question E035, Aggregated Dataset, 1981- 2008) If Fish is right, we should find statistically significant evidence that individuals from predominantly Muslim countries are tended to agree more with the first statement while all the others and specifically Protestant Christians, should be tended to agree more with the second statement. I expect that coming from a predominantly Muslim country specifically, and religion in general, is not going to provide any statistical significance in determining the answer of individuals to this question. As mentioned before, the Muslim world is too large to act as one homogeneous entity. However, there are theories supporting the idea that religiosity factor for religions can draw a more manageable picture in terms of measuring the effect of religious culture on individual preferences. Accordingly, I think Fish`s argument is incomplete, if not wrong. I expect to see some effect of Islam on egalitarianism once religiosity is integrated to the picture as an interactive variable with religion. However, before continuing with my results of this test, I will share my alternative replications of Fish`s data analysis.

Are Predominantly Muslim Countries Really More Egalitarian?

For all my data analyses I use the integrated dataset “Democracy Crossnational Data, Release 3.0 Spring 2009” merged and shared by Norris from Harvard University. This dataset is suitable for comparisons of 191 contemporary states. It contains data on the social, economic and political characteristics of 191 nations with over 1000 variables.
In my replication of Fish`s research, I first used the same dataset he used to measure the dependent variable: UNDP Gini index from 2004. My main independent variable is the dummy variable of being predominantly Muslim country from CIA Fact book. In my first model I control for GDP per capita (World Bank 2007), Freedom House rating of democracy (2008), life expectancy at birth in years (UNDP, 2008) and adult literacy rate (15+) (UNDP 2007). In terms of controls I tried to follow Fish`s choices to see whether I will receive the similar significance for the Muslim variable. Results show that (Figure 3) I do.

Figure 3 Replication with UNDP Gini Index This first replication confirmed Fish and I found statistical significance for the dummy variable Muslim (See Figure 3). No I want to add some new controls to the picture and see whether Fish`s results hold or not. We know that ideology is a potentially relevant factor in shaping inequality in both individual and country level (See Pontuson and Rueda, 2010; Alesina and Giuliano, 2009 among others). My ideology measure is a simple mean of left vs. right ideology from WVS. As the Figure 4 illustrates, the effect of the Muslim variable disappeared once controlled for ideology. This situation indicates potential spurious relationship between being a predominantly Muslim country and having low Gini score.

Figure 4 Replication with UNDP Gini Index Additional Control by Ideology

A more interesting result is received when the source and the measure of the dependent variable is changed from UNDP to World Bank, 2006. Regressing this variable on the same independent variable and control variables produced a very different result (See Figure 5 and 6). Test results with the World Bank Gini scores make the strong support for Fish`s argument from UNDP data disappear. Once the measure of the inequality changes, there is no argument left for Fish to formulate the causal mechanism for. Nevertheless, change in the R squared from 24% to 97% is a potentially problematic one, because going from 24% to 97% by only changing the dependent variable measure is unexpected.

Figure 5: Test with the UNDP datasets shows high significance for the Muslim variable with extra control for religiosity, GDP is significant democracy is almost significant and education shows some significance too.

Figure 6: Test with the World Bank dataset with ideology control shows no significance for the Muslim variable, GDP is very significant democracy is not significant any more

Interestingly, the two measures of the same inequality variable are negatively correlated with each other (See Figure 7). Their effect of the egalitarianism is also in opposite directions.

Figure 7: Difference direction and size of the correlations between the two measures of inequality with the egalitarian perceptions of individuals and the significantly negative correlation of the two measures with each other

Are Muslims Really More Egalitarian?

As mentioned before, the World Values Survey is a useful dataset in testing Fish`s causal mechanism argument. The income inequality question is coded as a continuous variable between the values of zero and ten. The answers given to that question are normally distributed: most of the individuals from 73 countries gave medium level answers to the question (See Figure 7 and 8). However, to be able to see the distribution of attitudes towards equality, I converted this continuous variable into a categorical one. I created five categories in the following way:

* [1,3) strongly pro Equality * [3,5) pro Equality * [5,7) balanced * [7,9) pro Incentives * [9,11)strongly pro Incentives

This format gave me the chance to form a cross- table of inequality perxceptions with the Muslim dummy variable. The results are in the direction that I expected:

Figure 7 Distribution of the Dependent Variable

Figure 8 Cross table showing the numbers and percentages of individuals from Muslim vs. non-Muslim majority countries in terms of their equality choices Opposite to Fish`s argument, half of the Muslims hold strongly inegalitarian views when it comes to economics. Out of 14 predominantly Muslim countries, seven of them take the strongly pro-incentives position, while most of the rest of the countries hold more balanced positions (about 80%). Non- Muslim coutnries` individuals seem to be relatively more egalitarian compared to the Muslims. Only one Muslim country, Tanzania, acts in the way Fish expected it to act and supported equality over incentives provided by inequality. These results confirm my expectations.
The results provided by the crosstab descriptive analysis need confirmation from the regression analysis with the same dependent variable of egalitarianism measure. As the Figure 9 below shows, Muslim variable is not significant in explaining variance in the perceptions of equality of individuals.

Figure 9 Regression of inequality perception measure on the Muslim dummy variable

If Not Muslims, Then Who; If Not Religion, Then What?

Since Fish was making various arguments about Christians and their approach to inequality in explaining the Muslim distinctiveness, it is useful to check with the Protestant and Catholic dummy variables to see whether being a Christian shows any negative significance in terms of egalitarianism or not. As mentioned before, Protestantism was taking a special part in the reasoning of Fish. If he is right, Protestants should show some negative significance in terms of their choice for equality in society. However, as the Figure 10 shows, these expectations of Fish are not met.

Figure 10: Regression of the inequality perception measure on the Protestant dummy variable with the same controls: no statistical significance observed

Similarly, the Catholic variable is also not statistically significant, even though its direction is different than Muslim and Protestant dummy variables. (See Figure 11). Compared to the Muslim and Protestant cases, Catholics seem to be more supportive of egalitarian policies (See Figure 10), as the theoretical literature was expecting.

Figure 11 Regression analysis of the relationship between being an individual from a Catholic predominant country and egalitarianism

Not Religion but Religiosity

Now I want to test my theoretical argument that we should get some significant results when we add an interactive religiosity variable into this picture. To start with, after going through the literature, I formed the following correlation table (See Figure 12) to have an idea about the potentially relevant variables in explaining egalitarianism. This analysis from Norris` data set provided me the opportunity to see the significant correlations between various variables and the dependent variable.

Figure 12: Some Important Correlations of Various Variables with Egalitarianism

Figure 12, where I shared correlation statistics of egalitarianism variable with some other critical variables, gives us some important clues about the potentially important determinants of inequality perceptions including religiosity and religion in general. First of all, non-significant correlation results with GDP and Gini scores show that economy might not be the key determining factor. On the other hand, maybe in an unexpected way, being a business owner makes people much more likely to be egalitarian: the correlation is among the highest. Also democracy rate and gender empowerment variables show significant correlations with egalitarianism. Religiosity shows importance as I was expecting, while predominant religion of the country shows no importance. More importantly, the interactive variables that I added by multiplying the effect of religiosity with Muslims, Catholics and Protestants, I find that the multiplicative effect of religiosity and Muslim dummy is statistically significant in the direction that Fish expected them to be: supportive of egalitarian policies. Another important and related result comes from the group of people who disapprove abortion: they show strong egalitarian tendencies. This result can also indicate the relevance of religiosity, even more than the religion itself.
Looking at the correlation statistics is not an end in itself in understanding what the determinants of egalitarianism are. This is why I want to conduct another regression analysis of egalitarianism, with the same controls but this time with an interactive variable of religiosity and Muslim country dummy. I expect to see statistical significance once religiosity is integrated to the picture. The results are in the expected direction. The interactive Muslim dummy variable is significant now in the existence of religiosity. The interactive variable coefficient is significant indicating us that Muslim religion variable has indeed have an interactive relationship with religiosity which should be taken into account. Also, the increase in the R2 from 28% to 59% can be taken as a good sign in terms of better model fit, even though not sufficient to conclude. The result shows us that, being an individual from a predominantly Muslim country, on average, make people less egalitarian except if they are religious. In that case, as Fish was expecting them to be, they become more pro equality. This result makes intuitive sense in reference to Fish`s causal mechanism argument also. His textual references can only be applied to people who take religion seriously. Those are the people who should be expected to have communitarian views inspired by religion. Just being from a predominantly Muslim country is not an explanatory factor in people`s preferences, just like it does not explain much for other religions. Measuring the effect of religion can only be meaningful when our population is religious. This is why interactive model worked better than the additive model.

Figure 13 Religious Muslims are significantly more egalitarian than religious non-Muslims

In interactive models like this one looking at the standard regression tables outputs is never enough to see the marginal effect of the independent variable on dependent variable only when another independent variable is present. That is why I am also sharing a marginal effect graph for being Muslim on being egalitarian in the presence of religiosity (See Figure 13).

Figure 13 Muslims gets more egalitarian as they get more religious

The graphical interpretation of the interactive relationship between being Muslim and religious can be clearly observed in Figure 13, which summarizes the conditional relationship in a clear way. It shows us that being Muslim have a positive relationship with support for equality in the presence of religiosity. This result shows that Fish was actually right in his intuition that there is something different about Muslims in terms of egalitarianism. However, he made a mistake by assuming that religion has a homogenous effect on all Muslims on average. This study shows us that the reality is far from this expectation: religious Muslims are actually more similar to religious Catholics than less religious Muslims in terms of economy related ideals. This finding, in a way, confirms Davis and Robinson`s theory(2007) that religious conservative groups from Abrahamic religious have some common characteristics shaping their behavior and expectations.

Limitations One of the potentially important limitations of my research is that the dummy variable of post- Communist country is not included in my regression analysis as a control variable. However, when we look at the regression analysis of Fish (See Figure 2 Model 4) this control variable does not make much of a difference in the significance level of the Muslim element. This is why; I do not think that including that control to my analysis would make a significant difference. Second relevant issue is that the measurement of Muslim is percentage in Fish`s analysis while it is a categorical dummy variable in my dataset. I do not have the percentage version of the same measurement to test and see whether this is a critical difference or not but, again, considering the fact that I achieved to reach the significance level that Fish reached in my regression analysis, I assume that this would be a good sign of non-biased estimates that I received when additional controls were implemented. Another problem in this proposal is that, as mentioned before, the increase in the R2 when going from the UNDP measure of inequality to World Bank measure is too big to be true. Even though theoretical approaches point overfitting problem, the fact that I did not add any new control variables to the regression analysis, make the situation puzzling. Solving this problem is beyond the scope of this paper. However, I assume that a deeper analysis of these two different measures of inequality could provide some clue in understanding the source of the problem.
Last but not least importantly, there is a problem related to the data collection structure of the WVS. There are five waves of this survey, and not every survey has all the questions. Also the numbers of countries change from one wave to the other. To reach the best result from the dataset I used the 5-wave aggregated mean scores in measuring the egalitarianism dependent variable. Even though this choice maximized the sample size of my research, it might have biased some country results because the independent variables are not necessarily from the same years. I used the latest year possible in controls to have the most updated information, however since not all the variables are from the same exact year, there might be some mismatch in the relationship between the changing context and the perceptions of inequality. I am assuming that the perceptions of individuals are embedded in their belief systems and they do not change in the short run. However, this assumption does not always hold. For example, after catastrophic events or big financial crises people`s redistributive expectations might change in a short term period in significant amounts. Controlling for a “historically significant event” variable might potentially solve this problem.

Conclusions The goal of this paper was to investigate the relationship between religion egalitarianism and religiosity. I was inspired by Fish (2011) in conducting this analysis because some potential flaws in his ambitious findings provided me to opportunity to investigate this issue deeper. Specifically, this paper was about the relationship between being a member of a Muslim country and holding income equality friendly ideas. Relying on his OLS regression results which were showing that predominantly Muslim countries are significantly more equal than Christian as well as all non- Muslim countries, Fish was claiming that this is because Muslim culture is inherently more egalitarian than Christian culture. I found Fish`s arguments theoretically weak and empirically suspicious because they were overly relying on personal anecdotes. I assumed that his anecdotal references do not necessarily represent a statistically important part of a culture, which is widely spread around the world and far from being homogenous. As an alternative way of approaching this issue, I claimed that religion alone is not explanatory but its multiplicative effect with religiosity on egalitarianism can be supportive of Fish`s expectations. I found supporting evidence to my claims. Before testing Fish`s causal mechanism arguments, I replicated his analysis on income inequality. When I included additional control for ideology, I got insignificant results for the predominantly Muslim country dummy variable. More importantly, when I changed the measure of inequality from one source to another, the results on which he formed his arguments on became obsolete. This situation raises the question of reliability of different datasets in measuring inequality. When I checked correlation of inequality with these two different datasets, I saw a significantly different picture. I even observed different effects of these two measures of inequality on the main dependent variable in terms of significance and direction. When I tested Fish`s causal mechanism argument, I found the results I was estimating to find: coming from a Muslim country does not make you more egalitarian in comparison to the rest of the world: actually the opposite (See Figure 14 for the mean of egalitarianism variable as a continuous variable). Following regression analyses confirmed these descriptive findings. Neither religion in general nor predominantly Muslim country variable showed statistical significance in explaining individuals` preferences of equality over incentives.

Figure 14 Muslim vs. non-Muslim Egalitarianism Statistics

Finally I asked the question: “if not being Muslim, then what?” and shared correlation statistics of egalitarianism variable with some other potentially critical variables. While economic variables did not show any statistical significance except the business owners group, social and political determinants showed some significantly interesting correlation with egalitarianism. Among these, religiosity variable is an important one. It ended up being correlated significantly with egalitarianism, while religion in general and Muslim, Catholic and Protestant variables separately did not show any significance. Religiosity changed the picture dramatically and showed that religious Muslims are the ones who are significantly more egalitarian than the rest, including non-religious Muslims.

References

Alesina, A. and G.M. Angeletos (2005a), Fairness and Redistribution: US vs. Europe, American Economic Review, 95: 913-35.

Alesina, A. and G.M. Angeletos (2005b), Redistribution, Corruption and Fairness, Journal of Monetary Economics, 1227-44.

Alesina, Alberto F. and Giuliano, Paola. (2009). Preferences for Redistribution. NBER Working Paper Series, 14825: 1-36.

Alesina, A. and N. Fuchs-Schuendeln (2007), Good Bye Lenin (or not?) The Effect of Communism on People’s Preferences, American Economic Review, 97:1507-1528.

Alesina, A. and E. Glaeser (2004), Fighting Poverty in the US and Europe: A World of Difference, Oxford University Press, Oxford UK.

Andreoni, J. and L. Vesterlund (2001), Which is the Fair Sex? Gender Differences in Altruism, Quarterly Journal of Economics.

Benabou, R. and E. Ok (2001), Social Mobility and the Demand for Redistribution: the POUM Hypothesis, Quarterly Journal of Economics, 116 (2001), 447-487.

Benabou, R. and J. Tirole (2006), Beliefs in a Just World and Redistributive Politics, Quarterly Journal of Economics, 121(2), 699-746.

Corneo, G. and P.H. Gruner (2002), Individual Preferences for Political Redistribution, Journal of Public Economics 83, 83-107

Nancy, J., & Robert, V. (2006). The Egalitarian Face of Islamic Orthodoxy : Support for Islamic Law and Economic Justice in Seven Muslim-Majority Nations. American Sociological Review 71(2).

Esping- Andersen. (1999). Social Foundations of Post Industrial Economies. Oxford University Press, Oxford, UK.

Fish, Steven M. (2011). Are Muslims Distinctive? A Look at the Evidence. Oxford University Press NY.

Guiso, L., Sapienza, P., & Zingales, L. (2003). People’s opium? Religion and Economic Attitudes. Journal of Monetary Economics Vol. 50, pp. 225-282.

Piketty, Thomas. (1995). Social Mobility and Redistributive Politics. The Quarterly Journal of Economics 110 (3): 551-584

Todd E. (1985). The Explanation of Ideology: Family Structures and Social Systems, Oxford: Basil Blackwell.

Appendix

Regression Analysis with the additional control for region (UNDP) No significant change is observed

--------------------------------------------
[ 1 ]. See Koran, T. (1994). Zakat and Inequality: Some Evidence from Pakistan, Review of Income and Wealth (2), 205-217 for an interesting research on the issue.
[ 2 ]. See http://www.hks.harvard.edu/fs/pnorris/Data/Data.htm for access