Understanding the Timing of Disaster Recovery

This thesis proposes an econometric analysis of the probability of recovery after economic disasters. The study is conducted taking into consideration international data since late 19th Century, for 42 countries (25 of them being OECD countries). The purpose is to build a predictive measure of the probability of recovery and to understand which variables have a greater explanatory power over a time horizon of ten years from the occurrence of the disaster. In particular, the predictors that we consider are grouped in a macroeconomic category – including the distance-weighted average of foreign crisis, the country consumption growth, and worldwide consumption growth – and a financial category – including the US dividend-price ratio, the US realized equity volatility, and the US term spread. The estimates are obtained by implementing a Probit regression for each category of regressors, from either the year of each disaster or from the previous year. The analysis also distinguishes between disasters that occurred in OECD countries and non-OECD countries as well as those that occurred before and after World War II.