Social Insurance, Payroll Tax Structure, and Saving Rates: An International Comparison for OECD Countries

The motivation for this research is to examine the debate about the effect of social insurance on private savings, Feldstein [1974] and Leimmer and Lesony [1982]. I use time series data and alternative models to examine the hypothesis. With this objective I use an estimation model, as I describe below. Since programs and demographic among the OECD counties are significantly different, I first analyze and compare them, with special attention to similarities among the members of the European Union, which share common markets and have reciprocity programs for social security. In this context, I compare the basic statistics for wage income thresholds subject to payroll tax, their lower and upper thresholds, the marginal tax rates, and the respective shares of the employee-employer contributions for the countries in the sample. Since some OECD countries allow a maximum contribution and possible additional contributions, and have different provisions for deductibility of the payroll tax, these features of the programs are also compared. Since there are variations in social insurance tax rates for self-employed, we also compare them. Furthermore, we compare age eligibility, work history requirements, and the social security income and health care benefits across those countries. Since some countries have reciprocity for social security program in the European Union, their policies will also be considered. With such fundamental investigations, we then develop the model for empirical investigation, which is a multivariate. Both linear and log-linear specifications are considered. The dependent variable is household saving rates [per National Income and Product Account], during 1992-2011 [will be expand the time as data become available). The independent variables are: social security contributions as percentage of GDPs, (for comparison, the percentage of total tax revenues, in comparison to the OECD average are also considered), and relevant population demographics, macroeconomic factors such as the real rate of interest and the terms of trade, the Gini Coefficient, and expected social security benefits upon eligibility. We pay attention to the possibility of Collinearity and heteroscedasticity in our variables. Our preliminary results reject the hypothesis that social insurance replaces private savings. We have reviewed the relevant literature for theories of saving behaviors and social security. Our research is a new empirical contribution to the literature.