Social security expenditure and GDP in OECD countries: A cointegrated panel analysis

Abstract

Using panel data unit root tests and panel cointegration tests, as well as estimation techniques appropriate for heterogeneous panels such as the full modified OLS, this paper re-examines the long-run co-movement and the causal relationship between GDP and social security expenditure in a bivariate model, employing data on 25 OECD countries from 1980 to 2001. Our cointegration test results show strong evidence in favour of the existence of a long-run equilibrium cointegrating relationship between GDP and social security expenditure after allowing for a heterogeneous country effect. Regarding the panel-based error correction model, we find that GDP and social security expenditure lack short-run causality, but reveal the existence of long-run bidirectional causality. This shows that, in the long run, economic growth must be based on a social welfare policy that should be carried out, and economic growth can facilitate contiguous development in a social welfare policy. Lastly, we also provide evidence to support that social security expenditure can affect growth through the savings and human capital accumulation in OECD countries.