What variables should be included in a dynamic EWS model in order to have the best fitting model for financial crises (currency, banking and debt crises) ?

Details

Supervisors

Candelon, Bertrand

Faculty

Louvain School of Management

Degree label

Master [120] : ingénieur de gestion, à finalité spécialisée

Abstract

Being in an era where unpredictable financial crises happen, it is necessary to understand which financial indicators should be included win an Early Warning System (EWS). The purpose of our master thesis is to identify the key variables that enhance the accuracy of an EWS for currency, banking and debt crises. In first stage, we reviewed the current literature in order to have a full understanding of the difference between types of crises and to acknowledge the different EWS methods that exists. Secondly, we proceeded in an empirical research where we started by testing the relationship between the variables using a bivariate regression, a correlation and multicollinearity analysis. Subsequently, we proceeded in the implementation of a logit regression model to establish our EWS, followed by a specification model. Our results showed that for the currency crises the most significant variables are the CPI energy, short-term interest rate, exchange rate, industrial production and the 3-Months treasury bill. For banking crises, we identified the exchange rate, CPI, short-term interest rate, 3-Months treasury bill and the 10-Year constant maturity. For debt crises, the most significant variable are the CPI, exchange rate 10-Year constant maturity and the 3-Month Treasury Bill. Once we have identified the most significant variables, we have performed forecasts for each type of crises for the next decade, and conducted model evaluations.