Is external debt sustainable? A probabilistic approach

We develop a probabilistic approach to measure a country's external debt sustainability. Using data on international investment position and balance of payments from the International Monetary Fund, we estimate a vector autoregressive model for 38 countries (11 developed and 27 developing). Using the estimated parameters, we perform a Monte Carlo simulation to compute the distribution of the capacity to repay for each country. A large portion of the projected distribution to the right of current debt is a warning indicator, signalling the need for devaluation. We provide simulations for each country. One scenario is where the discount factor is lower than 1. According to the literature, this situation should prevail in the medium and long run. A quite different situation is where external sustainability is achieved because of a simulated discount factor of more than 1. Here, interest on debt is lower than GDP growth. This situation is associated with dynamic inefficiency. The results suggest that flight to safety is penalising some developing countries, whilst benefiting some developed countries.