Several economic models have described the theoretical causes and onsequences of 'credit rationing' and 'under-investment'; one string of this literature shows the long-run effects of initial wealth distribution and entrepreneurial ability on the process of occupational choice and performance, and its consequences on inequality. Surprisingly, there is very little micro-level evidence on the existence and effects of 'credit rationing' in the context of developing countries. This is the contribution of the current paper. Using a survey of 4,553 entrepreneurs in 51 slums in Rio de Janeiro, this paper uses mean and quantile regression estimates to shows the effects of the type of initial capital, credit constraints, and human capital factors on entrepreneurs' performance. The main findings of the the paper are that entrepreneurs that were able to self-finance their business start-up presented earnings 16% greater than entrepreneurs that had to borrow their initial capital. In addition, entrepreneurs that explicitly claimed to be credit constrained performed substantially worse than their observationally identical counterparts, even if they were credit worthy. Both, initial source of funding and liquidity constraint presented greater effects on the highest quantiles. In terms of human capital, the current study shows positive and statistically significant returns for both years of schooling and experience, with higher returns on the lowest quantiles, indicating the potential role of these factors on inequality reduction.