We investigate whether a higher financial integration with the rest of the world can help the African countries reduce their production inefficiency and/or push up their efficient production frontier. Since the traditional literature focuses on growth, not much is known about the financial integration/production efficiency link, especially in a macro-perspective for the African countries. Thus, we propose estimates that allow for heterogeneous financial openness-production frontier nexus across countries, working with a panel of 45 Sub-Saharan African countries over the period 1996-2014.

We first explore which subsets of explanatory variables have significant effects on GDP per-capita (excluding the financial integration variables) by using a Bayesian Model Averaging technique. We find that GDP per-capita is significantly influenced by the sectoral composition of the GDP, demographic variables, policy variables and a few governance variables.

Second, we investigate the effect of financial integration on GDP per-capita by using Stochastic Frontier (SF) models. Our results show little evidence of a systematic positive effect of international financial integration on GDP per-capita efficiency scores. We find instead that the inflows of FDI and equity portfolio, debt liabilities, official development assistance or flows of capital between domestic and foreign banks can frustrate higher economies of scale in production, thereby implying that efficiency scores do not necessarily improve. A salient feature of our result is that financial integration increases the distribution of technical efficiency when the African countries are compared to each other. These findings play against the dominant view that one size fits all. In other words, the international financial integration can increase or decrease African countries' standard of living. The policymakers might be more cautious in moving further to achieve regional financial integration within Africa.

Third, we attempt to understand the observed heterogeneity across countries in terms of production efficiency behaviors. Maybe the countries with the highest efficiency scores make a better combination of financial integration with the other determinants of GDP than those with lower efficiency rates. Thus, we allow parameter heterogeneity in the GDP per-capita equation (which is not possible with the SF approach by using the Quantile estimators to understand a complete characterization of the conditional distribution of the frontier because they are robust to outliers in a panel data context. We find that disbursements from official donors is the financial integration that helps the countries catching-up with the most efficient ones - in terms of reaching the highest GDP per-capita.