복합금융위기에 대한 조기경보

초록

The nature of a financial crisis tends to vary over time depending upon its causes as the world has witnessed over the last decades. The conventional signals approach that attempts to predict the latest global crisis based upon models tuned for signaling either the currency crisis or banking crisis appears to have experienced the difficulty in predicting the 2007/2008 global financial crisis with the relevant lead time. As such this paper employs the idea of Balakrishnan et al. (2009) who proposed the new way of defining a complex financial crisis and examines whether the conventional model can still be useful as a main device for signaling a future crisis. The empirical results suggest that expanding the set of leading indicators to include the theoretically and potentially important variables in the model would have been helpful in predicting a future complex financial crisis instead of keeping only the compact set of significant variables identified from models aiming at predicting a simple definition of a crisis. This paper also demonstrates that the standardization of the leading indicators using a moving window substantially improves the ability of predicting both the beginning and ending of a complex financial crisis. Finally, this paper shows that the improvement is not confined to a particular country.

키워드