Testing a Neoclassical Growth Model with the Institution Effect of Government Size Using Panel Data: Is the Economic Success of Nordic Countries Due to Welfare State?

Abstract

This paper develops and tests a Neoclassical growth model with the institution effect of government size in the framework of a dynamic panel data model, using two Generalized method of moments (GMM) estimators to address endogeneity of explanatory variables in estimation. I find a robust significant negative correlation between government size and economic growth for all OECD countries between 1960 and 2012, with much higher estimates of growth elasticity of government expenditure, tax revenue and total welfare than those obtained in previous empirical studies on this subject. The estimated convergence speed is consistent with those obtained from other growth studies using panel data and the implied capital share is close to one third of real value. Although the negative correlation does not necessarily imply causality, tax revenue as a percentage of GDP is identified as having causal negative impact on growth. This correlation or causality also applies to Nordic countries. The evidences strongly suggest that the economic success of Nordic countries in terms of growth is not due to welfare state characterized by big government, high tax and high welfare, but attributed to other time-invariant and time-varying institutional or policy factors. In the long run, whether Nordic characteristics that are in favor of growth can keep on dominating the negative growth effect of welfare state is still an open question.

Keywords: Government size; Government expenditure; Economic growth; Cross-country regressions; Dynamic panel data model; Fiscal policy; Taxation; Public sector

JEL Classification: E62; H11; H20; H50; O23; O40; O43