Debt, corruption, R&D and growth in developing countries

This paper analyses optimal paths in a one-sector growth model when the technology is not convex. In such a case, we prove that optimal paths converge to the upper steady state iff the initial wealth is above a critical level. Then we first show that thanks to debt and/or R&D the poverty trap may be avoided. Second, we introduce a distortion: corruption which mostly has dramatic consequences on growth. These results may explain why empirical works lead to the conclusion of non convergence in large cross-country samples.