Private Bag 92019, Auckland, New Zealand.

Barro (2000) reports how the growth-inequality relationship varies between the rich and poor countries. We provide an economic explanation for that finding in a purely neo-classical growth model with discrete occupational choice and redistributive taxes. In our model a fiscal redistributive tax program directly impacts the steady state distribution of human capital, which in turn determine income inequality and the growth rate by influencing the occupational choice of the agents. The proportion of innovators in the economy and the redistributive tax rate determine the steady state growth rate and income inequality. There are multiple steady states. Across those steady states, the correlation between growth rates and income inequality depends crucially on the degree of redistribution, output elasticity of unskilled labor and the degree of institutional barrier to knowledge diffusion. (Prepared for presentation at the World Congress of the Econometric Society, Seattle an