Explaining and forecasting currency crises in developed and emerging markets' economies
The series of banking and currency crises occurring in the 1990s have stimulated the study of financial crises. The research for this thesis has been conducted in the midst of this cluster of events: the Asian flu in 1997-98 and the disturbances in the Russian and Brazilian markets in 1998 and 1999, respectively. The aim of this work is to provide some empirical evidence on the general and systematic factors driving currency crises. After a summary of the literature on currency crises conducted in Chapter 1, the second chapter analyses the determinants of currency crises for 20 OECD countries for the period from 1970 to 1997. We use duration models in order to investigate the causes behind the duration of non-crises periods. Fundamentals are revealed as important determinants in assessing the likelihood of currency crises. Variables concerning the state of the external sector (exports, imports, degree of openness), the REER, foreign portfolio investment and net claims on central government help explain the onset of currency crises. Following historical events, the subsequent chapters in this thesis study currency crises in developing and emerging markets' economies. Chapter 3 develops an indicator called the Emerging Markets Risk Indicator, whose monthly scores reflect the currency risk for 36 emerging markets. We evaluate the contribution of the explanatory variables of this model to the probability of the main crisis events. In order to judge the forecasting power of this model, we estimate two reduced samples: the first one until December 1996 and the second one until December 1997. With these two models we can study the predictive power of the model on the onset of the Asian crisis in 1997 and the Russian and Brazilian crises in 1998-99, respectively. The results shows that had this model been used at those times, it would have predicted those crises. Chapters 4 and 5 analyse the joint occurrence of banking and currency crises, i.e. contribute to the debate on twin crises. In Chapter 4, we develop a currency crisis model with explicit reference to banking crisis indicators as possible determinants of currency crises. Deepening the subject of twin crises. Chapter 5 endogenises the banking crisis variable and jointly estimates two equations, one for banking crises and another for currency crises. This procedure allows us to fully test the interdependence of banking and currency crises, i.e. the direction of a causal link (if any) between both types of crisis. The results indicates the existence of bidirectional causal link between banking and currency crises. The phenomenon of twin crises is rather a case of banking and currency crises being closely intertwined. Which crisis type occurs first is a matter of circumstance.