The wealth effect on new business sartups in a developing economy

<p><p>Various theoretical models have postulated nonlinearities in the wealth effect on self-employment stemming from start-up costs and / or liquidity constraints. Nonlinearity implies that the extent of entrepreneurial activity in an economy depends on the distribution of wealth, though in potentially complex ways. To test for nonlinearities, we estimate both a non-parametric linear probability model and a parametric nonlinear model of the choice to be self employed amongst return migrants in Tunisia. Controls for heterogeneity are included, and tests are made for selection bias and separability between wealth and the controls. The relationship between the probability of starting up a business and savings repatriated at return is concave for almost all the range of our data, though we find weak evidence of a convex relationship at very low wealth levels. Our results indicate that the aggregate self-employment rate is an increasing function of aggregate wealth, but a decreasing function of wealth inequality.</p></p>