This thesis includes two essays that analyze some features of the past financial crises. In the first chapter I study the possible reasons why investors reduced their holdings of foreign equities, and, at the same time, they increased their holdings of short-term government bonds, during the 2007 financial crisis that first hit the U.S. economy and soon became a world crisis. More precisely I analyze how the increases in uncertainty during the crisis affected capital flows. I use a two country DSGE model and I assume that there is trade in both goods and financial assets. I assume that each country is allowed to issue equities and government bonds, and I assume that each economy is hit by three types of shocks: Preference, productivity and government spending shocks. I proxy the increase in uncertainty with the introduction of uncertainty shocks, i.e. I allow the variances of the shocks to be time-varying. My findings show that uncertainty is a source of portfolio-dynamics that can contribute to explain, together with the other sources already identified in the literature, deviations of the portfolio from its steady-state. Investors choose their portfolio with the goal to smooth consumption. Therefore they want to hold assets with returns that display a negative covariance with consumption. When uncertainty shocks hit, the way in which the real variables of the model covary with asset returns changes. As a consequence, agents need to re-adjust their portfolios until when the shock has disappeared.

In the second chapter I study the currency denomination of the debt in emerging countries. Empirical studies have shown that emerging countries are often characterized by the presence of a high share of foreign currency denominated debt. As the debt crises of the 1990s show, the presence of foreign currency debt can be risky because, beyond creating a mismatch in the domestic firms balance sheets, it also constraints the traditional domestic policy instruments in dealing with home and foreign economic shocks. The reasons why such risky forms of international finance arise in the first place remain an open question. If foreign debt is so dangerous-as it is-it may be worth trying to give a micro-foundation to its emergence. Such a high share of foreign currency debt should be at least in part justified by the presence of some private benefits for the agents that choose this form of finance. The goal of this chapter is to rationalize the choice to borrow in dollars rather than in domestic currency on the international markets. In order to do so, I study how informational asymmetries and heterogeneous expectations can affect the choice of a borrower to expose herself to a currency risk. Furthermore I look at the policy implications of my findings to understand which policies could reduce the incentive of agents to dollarize. My model is a portfolio choice model with asymmetric information that analyzes how agents choose the currency denomination of their debt. The main findings of my model show that when domestic agents have a high informational advantage and/or there is a low level of transparency on international markets, an increase in the degree of dollarization might be observed, if the fundamentals are relatively strong. Alternatively, if there is endogeneity between the exchange rate policy implemented by the monetary authority and domestic agents decisions, a certain degree of complementarity in borrowers choices may arise, thus creating a phenomenon of moral hazard. If domestic agents know that a high share of dollar debt in the economy makes the exchange rate more rigid, they may want to coordinate on the equilibrium where all the corporate debt in the economy is denominated in the same currency, even when the fundamentals of the economy are relatively weak. These results have interesting policy implications. A benevolent central bank that strongly bases her policy on the degree of dollarization in the economy, can generate a coordination mechanism among the domestic borrowers that results in a risky degree of dollarization. The solution would be to ex-ante choose a central banker with a strong preference for a flexible exchange rate. (Abstract shortened by UMI.)