Many economists have started to concede in recent years that contagion and self-fulfilling beliefs of investors have played a crucial role in the emerging market financial crises of the 1990s. Despite the progress on the theoretical side, however, empirical models of currency crises have been shown to perform poorly (Berg and Pattillo 1998) and many economists and policy institutions have been struggling to develop adequate models to predict future financial crises (Kaminsky et al. 1997, Goldstein et al. 2000).
Much of the empirical literature on financial crises, however, still focuses on country-specific macroeconomic factors and has ignored or at least underestimated the importance of contagion--the possibility that the origin of a crisis may lie in the occurrence of a crisis elsewhere in the world rather than with weak domestic fundamentals. As a consequence, economists still lack the answer as to why many crises of the 1990s clustered within regions and affected a broad range of countries almost simultaneously. In other words, the question that remains is how and why crises occurring in different economies are linked and interdependent.
The aim of this paper is to help find an answer to this question. The use of a non-linear Markov- switching model, based on the seminal work by Hamilton (1989, 1990), is suggested in order to enable a systematic comparison of three competing explanations for financial crises: weak economic fundamentals, sunspots--exogenous shifts in agents’ beliefs--and contagion. Contagion in this paper is defined as the transmission of a crisis that is not caused by the affected country's fundamentals (although, of course, the transmission has an impact on the country's fundamentals ex post) but by its ''proximity'' to the country where a crisis occurred.
The paper suggests and develops a new methodology to measure three types of ''proximity'', or channels of contagion. The first one measures the real interdependence among economies through trade competition. A second one analyzes to what extent countries are competing for bank lending in third markets. And the third channel measures the degree of stock market integration across countries.
The paper then conducts three complementary tests on the relative importance of fundamentals, contagion and sunspots. First, the use of Markov-switching models reveals that country-specific fundamentals generally fail to explain the timing as well as the severity of financial crisis in individual countries. Including contagion in the model, however, improves the explanatory power of the model significantly in most cases and even eliminates the need for regime shifts in the Markov-switching framework for some countries. Second, a panel data analysis confirms the robustness of these results for a sample of 24 open emerging markets. The results suggest that the Latin American crisis in 1994-95 and the Asian crisis of 1997 spread across emerging markets not primarily due to the weakness of those countries' fundamentals but rather to a high degree of financial interdependence among affected economies. Third, the model’s ability to predict the Asian crisis is tested. It is shown that taking contagion factors into account would have permitted quite an accurate prediction of countries to which the crisis spread. Overall, these results emphasize that only if we take into account the systemic character of financial crises will we be able to improve our understanding and better predict the occurrence of future crises.