Recent theories of `endogenous growth', in which economic activity feeds back into technical progress and knowledge, have proposed that redistribution may adversely affect economic activity by distorting labour supply and incentives for capital accumulation. If the politico-economic equilibrium in a democracy is determined by the preferences of the median voter (who is also the median agent in income distribution and generally poorer than the mean), an increase in inequality will make the median poorer relative to the mean, so there will be more redistribution and less growth. This approach implies that a reduction in the political rights of the poor equivalent to making the median voter richer relative to the mean has a positive impact on growth, which contrasts with the fact that the richest countries in the world are all democracies.

In Discussion Paper No. 613, Research Fellows Gilles Saint-Paul and Thierry Verdier develop a model in which the main instrument of redistribution is spending on public education. This investment in human capital enhances growth, and there is also a hereditary component in the intergenerational transmission of knowledge. Public education modifies income distribution in the long run since it is provided in a relatively egalitarian way; so a society can redistribute income in a growth-enhancing way while also investing in income equality in the long run.
Saint-Paul and Verdier contrast the main implications of their model with those of other recent `endogenous growth' theories. They find that initial inequality benefits growth because it increases political support for public education; but this result may not hold if relative poverty is associated with lower participation in elections. Nor does it imply a cross-sectional positive correlation between inequality and growth: since increased spending on public education implies a faster pace of income equalization, any third factor that affects public education will generate a negative correlation between inequality and growth. They also predict that spending on education will eventually decline and so will the growth rate as inequality declines. An extension of political rights to the poor will generate a spurt in growth through increased spending on public education; while an inflow of poor immigrants will first depress educational spending and thus harm growth, and then boost them once the immigrants or their children are granted political rights.

Saint-Paul and Verdier then extend their model to the case where redistribution is very intense and find that its distortionary effects then outweigh its positive effects: to maximize growth, democratization must take effect progressively and in line with income equalization, so it is desirable to wait until society is sufficiently homogeneous before extending political rights to the poor.