Within Convergence : Estimating Unique Steady States Using a Dynamic Panel

A problematic outcome of the cross sectional study of convergence is that developing economies are often estimated to be above their steady states, implying that they have too much capital per worker. This paper introduces a method to estimate steady states in a way that can explain this counterintuitive result. The method uses the coefficient estimates of an augmented Solow growth model and, through the use of panel data, can include total factor productivity as a country-specific effect. In sum, it is found that countries beginning a period of development above their steady states do so because of low total factor productivity. In the panel setting this can explain what it means to have too much capital per worker: these economies simply do not have enough productivity to maintain their initial levels of per worker capital. This implies that development through investment can only succeed when there is convergence from below the steady state. Above the steady state, total factor productivity is too low to sustain the relatively high levels of capital. Empirical results confirm that these growth patterns can offer a satisfactory explanation of the observed cross-country income divergence of developed and developing economies