Three essays on financial and currency crises in emerging markets

The thesis consists of three essays which investigate the causes of financial and currency crises.; In the first chapter I show that there exist self-fulfilling currency crises when the domestic banking competition is at an intermediate level. Intuitively, when the real interest rates are higher due to a lack of banking competition, this distortion makes the central bank more willing to devalue the currency. The model and its implications for promoting the domestic banking competition fit the facts related to the Finnish devaluation crisis of 1992. Moreover, the model explains why the banking competition declined and the credit growth slowed down before the crisis, observations that have been unexplained by previous models.; In the second chapter I show that when there exists a possibility for a corporate bankruptcy, the occurrence of self-fulfilling currency crises depends on the financial intermediary's ability to collect the debt back from a bankrupt firm. The corporate bankruptcy also decreases the effectiveness of a devaluation in boosting the real output level, which implies that the devaluation crisis equilibrium coincides with a recession, an empirical fact which has not been explained by models so far. Finally, I show that a conditional tax on the financial intermediary's foreign exchange transactions will prevent self-fulfilling crises if and only if the tax is implemented only when nominal interest rates exceed a previously specified threshold level and the tax revenues are also distributed to the firm.; In the third chapter I show that the Estonian stock market crash of 1997 can have resulted from adaptive learning behaviour by private investors. I estimate the stock return distribution as a Pareto and show that an unexpectedly low single observation can lead to a drastic revision of the mean and variance for stock return estimators, because investors conclude that the sample so far has been biased and there will thus be many more low returns forthcoming. This results in a stock market crash. The result implies that transition economies with limited data available are more vulnerable to sudden financial crises and that the assumption of normally distributed shocks may have hampered the empirical performance of financial crisis models.