THE DETERMINANTS OF PREFERENCES FOR REDISTRIBUTION
Savaş Çevik – Erol Turan
One of the most important functions of modern states is to redistribute income through tax and spending systems to ensure relative justice in the distribution of income among members of society. Even if there is no conscious policy to improve the income distribution, even the usual tax collection and spending activities of the state have potential redistributive consequences. This function of states is as much a matter of debate as is common. To what extent the redistribution would be good or the desirability that the state undertakes such a function is a matter of controversy among politicians as well as economists. Further to that debate, the differences between countries in the redistributive policies implemented by them and the redistributive results of state activities in general are also needed to explain. Considering that the existing practices are formed - at least partly- by the citizens’ preferences in democratic administrations, it has been an important research topic among economics and political scientists in recent years that how preferences of individuals about the level of redistribution are determined and what determines these preferences. So, the aim of the study is to examine the factors shaping the redistribution preferences at the individual level.
The data used in the study were obtained from the combination of data sets collected under the World Values Survey (WVS) and the European Values Survey (EVS). More and more countries are targeted through combination to reach a large sampling. Both surveys have been conducted in a number of countries from 1981 to today (From the first wave, both surveys were applied cumulatively in 113 countries / regions representing 90% of the world population). They are applied to examine the values, beliefs, political, social and economic attitudes of individuals and the change in them through samples representing the country. Most recently, the sixth wave of WVS (2010-2014) and the 4th wave of EVS (2008-2010) have been implemented. When the two data sets were combined, a total of 7 waves were obtained.
In the analysis of the data, estimates were made according to two econometric methods. For simplicity, least squares regressions (OLS) were used. The results are largely consistent with each other. Tables report results from both estimates. In all predicted equations, country and wave fixed effects are included to control the impacts of the country and the EVS / WVS wave. In addition, robust standard errors were used in OLS estimations to more accurately measure significance.
First, the basic demographic variables and the effect of income are analyzed as the forecast strategy. The effect of individual values and norms were then included in the basic model predicted in this way. Finally, the effect of variables on social identities and ideology has been tested. In the design of the models, high-correlation variables were avoided as much as possible in the same model.
Analysis of the effects of demographic variables and income are presented in tables. Age is statistically significant in most of the equations. As the age increases, the redistribution rate increases, albeit with small coefficients. In other equations, the birth year was used to evaluate the effect of the birth cohort and was found to have a negative effect. Accordingly, newer generations favor less redistribution compared to the old ones.
In parallel with other studies, being a woman is a positive and important variable in all equations. Women clearly have a higher preference for redistribution.
Again, being married according to expectations shows a lower redistribution bias compared to the other two categories. Divorced / widows / separated men and women, and singles are more in favor of redistribution than married ones.
In the first four equations, the income was evaluated by the absolute level and it was found to have a negative coefficient. In the other four equations, in order to evaluate the effect of the relative income position of the individual, the difference of his/her income from the median income was evaluated, and it was found to be similarly effective and significant.
While unemployed people favored higher income distribution, those who are self-employed showed less redistribution preference than expected, in connection with the possible risk preferences. The coefficients for both variables are significant.
Conclusions and Discussion
The study aimed to provide a detailed review of the determinants of the redistribution preferences in light of the theoretical and empirical literature and the data from the WVS / EVS. The findings of the study were largely in line with the hypotheses and findings of other studies. The results show that demographic variables such as age, marital status, gender are effective on re-distribution preferences. The birth cohort, which represents similar shock and change experiences outside of age, is also a significant variable.
Income is associated with less redistribution, both in terms of its absolute level and as a difference from the median income. In order to test the POUM hypothesis, the difference between the level of education and the level of education of the father have a negative effect on the preference for redistribution in line with the predictions of the hypothesis. “Being self-employed, which can be seen as a reflection of the risk preferences of individuals, is again tied with less redistribution preference.
One of the most important variables is the opinions about the cause of poverty. Indirectly, the coefficients of this variable, which can be seen as an indicator of attitudes about whether the beneficiaries of the redistribution will deserve, are important and significant. Accordingly, those who see the cause of poverty as a natural consequence of laziness and modern developments are less in favor of redistribution than those who see it as unluckiness and injustice in society. Besides, life control perception (fatalism), reciprocity motivation, altruism and trust were found as individual values and norms in terms of determining the preference for redistribution.
Those who express themselves as rightist in the political spectrum are less in favor of redistribution than those who are leftists. An interesting finding of the study is that the religious categories to which the individuals belong are effective both in influencing social identities and individual values, and in the re-distribution preferences in order to see insurance function. According to this, all religious categories are more favorable to income inequalities than those who do not belong to any religion. The interaction between religion and religiousness also yielded significant results.
The study found that social identities are important, in this sense, subjective social class belonging and national pride affect the preferences of redistribution. A negative coefficient was found regarding the national pride variable. This nationalism can be explained by more individual sacrifice and activity bias and reluctance to share with others “in ethnically heterogeneous or high-migrant countries. The issue can be further investigated with a study that takes into account the ethnicity of countries.