Investigation into currency and banking crises: a novel approach to the identification and prediction of twin crises.

This thesis focuses on providing novel insights into the relationship between currency crises and banking crises and building a tool to identify and predict the crises. Even though currency and banking crises have occurred periodically, the nature of twin crises is still unclear. There are still debates on whether currency crises trigger banking crises. The dates of twin crises are still difficult to identify due to the limitation of the existing technique. In addition, economists have difficulty in examining the risk of the crises as there is no consensus on how to define them. To address the issue, we examine the twin crises literature using the systematic literature review methodology. We then identify the pressure dynamics of the twin crises in Latin American and East Asian countries during the period 1980-2007. Finally, we examine the crisis risk of the currency and banking crises in 80 countries during 1970-2016. The literature suggests that banking crises often precede currency crises. However, on the contrary, we show that currency crises often precede banking crises by minimising the bias in the identification techniques. While the literature argues that foreign liabilities are responsible for twin crises, we explain that liquidity shortages and the insolvencies of banks may also trigger twin crises. In addition, we argue that currency crises may also trigger bank crises. Thus, twin crises should be examined as a two-way relationship. Furthermore, we combine the Exchange Market Pressure Index and the Money Market Pressure Index into a c-index to evaluate the twin crises episodes in the existing literature. We find that the model is able to pinpoint the dates of the twin crises episodes in our sample countries. Finally, we divide the crises into four levels as there is no consensus on how to define the crises. We demonstrate that the c-index can predict the probability of any given condition to shift to the ‘next crisis level’ in the next two years. The findings also suggest that regulators and investors are risk takers in low-pressure periods and become risk-averse when conditions worsen.