Financial instability in emerging markets

In this dissertation, I analyze the causes of financial instability and policy initiatives designed to reduce that instability. Chapter 1 studies the correlation between twin crises and exchange rate regimes. I apply a two-equation Probit model to separate the effects of those regimes into three different channels: those directly affecting banking or currency crises respectively, and those affecting the correlation between banking and currency crises. I find that intermediate regimes do not affect the probabilities of either currency crises or banking crises. But intermediate regimes introduce a positive correlation between banking and currency crises, and thus increase the probabilities of twin crisis and no crisis. These findings thus support augments both for and against the corners hypothesis.;Following the Mexican and Asian crises, there has been a proliferation of international initiatives to encourage banks, firms and governments to disclose more information about their financial affairs. Chapter 2 studies whether these initiatives affect macroeconomic forecasts. I find that disclosure standards improve the forecast accuracy of both private sectors (the EIU) and international organizations (the IMF). By applying forecast encompassing tests, I further find that IMF forecasts become more informative to the EIU after standards' implementation. I then propose an information model to illustrate circumstances under which public disclosure could make IMF's forecasts more informative to others.;Chapter 3 shows that the impact of transparency initiatives may be more limited than often thought in that public disclosure crowds out private investments in information. I first develop a theoretical model of the incentive to invest in information and the impact of public disclosure. I then analyze a panel data set of stock market analysts' forecasts for sixty countries for the period 1990--2002. I find that disclosure standards enhance forecast accuracy directly but at the same time reduce the number of analysts per stock (proxy for private investments in information). The net effect of disclosure standards on forecast accuracy and dispersion thus ranges from weak to nonexistent.