Extended Summary

A welfare state can be defined as a state that provides minimum standards for its citizens in areas such as income, education, health and housing. The concept, which was first used in Bismarck’s Germany following the social security reforms, became widespread after World War II. Esping-Andersen’s study ‘The Three Worlds of Welfare Capitalism’ categorizes welfare regimes into three main models: market-oriented welfare state in liberal democracies, conservative welfare state in Continental Europe, and social-democratic welfare state. Not content with the basic three classifications, researchers have tried to fill the gaps in Esping-Andersen’s analysis. A fourth welfare regime model was proposed by Leibfried. Leibfried contributed to Esping-Andersen’s classification by defining a different welfare state model in the southern part of the European Union, which he characterized as “Latin Belt Countries”. This study uses Esping-Andersen and Leibfried’s classification of welfare regimes to investigate whether countries in four different welfare regimes differ in terms of inequality using panel data analysis. Other factors that may affect wage and income inequality such as the level of openness, skill composition, income level and labor market institutions are also included in the analysis. Due to the difficulty in accessing data on labor market institutions, the data for 1970-2015 and 1990-2015 are divided into two periods: 1970-2015 and 1990-2015.

While there were fewer explanatory variables in the 1970-2015 period, the analysis was conducted with a richer data set in the 1990-2015 period. In both periods, the dependent variables are the logarithm of the Theil statistic as a measure of wage inequality and the Gini coefficient as a measure of income inequality. Regime dummies are included in the model to capture the effects of welfare regimes on inequality. In the 1970-2015 period, the independent variables are the logarithm of the unionization rate (lunion), the logarithm of per capita income and its square (lpcgdp and sqlpcgdp), the kof globalization index (kof) and the logarithm of the ratio of university graduates to primary school graduates in the population aged 25 and over (lskill). In the 1990-2015 period, in addition to these variables, the scope of collective bargaining agreements (lbarg), the stringency of employment protection regulations for indefinite-term contracts (eplr), the stringency of employment protection regulations for fixed-term contracts (eplt), the share of active labor market program expenditures in GDP (lmp), the share of public social expenditures in GDP (tsh), and the ratio of self-employed and unpaid family workers to the active population (vulemp) are also included in the model.

The results of the regression analysis using a two-way unobservable random effects model indicate that in countries with a social-democratic welfare regime, wage and income inequality are significantly lower in both the 1970-2015 and 1990-2015 periods compared to other welfare regimes. These results are not surprising because the social-democratic regime features the most egalitarian and redistributive institutions. In 1990-2015 period the welfare regime with the highest wage and income inequality compared to other regimes is the Southern European welfare regime. In both the 1970-2015 period and the 1990-2015 period, wage inequality was significantly higher in conservative welfare regimes compared to others. Since the general characteristic of this welfare regime is the implementation of labor programs based on the protection of status differences among employees, high wage inequality is an expected result.

The regression results do not support the Kuznets hypothesis, which predicts that an increase in per capita income will first increase and then decrease inequality. In the wage inequality models, per capita income and its square are found to be insignificant, while in the income inequality models the opposite result of the Kuznets hypothesis is obtained. Income inequality first decreases and then increases with the increase in per capita income.

Concerning the effects of labor market institutions on inequality, the results are consistent with previous studies. Unionization rate, collective bargaining coverage and labor market programs reduce wage and income inequality. Skill composition is significant across multiple models, indicating that an increasing ratio of university graduates to primary school graduates is associated with higher wage and income inequality. The KOF globalization index, included to assess globalization’s impact, is insignificant in the 1970-2015 period but consistently significant and negatively signed in all models for the 1990-2015 period.

As a result, it is observed that wage and income inequality differ in countries with different social welfare regimes. After controlling for other determinants of wage and income inequality such as qualification, globalization and labor market institutions, the welfare regimes of countries are determinant on inequality. While inequality is relatively low in social democratic regimes with redistributive and universalist features, inequality is high in regimes where these mechanisms do not function sufficiently.