Dollarization in post-stabilization economies: Is there hysteresis

Recent experience from Latin America, Eastern Europe, and countries of the Former Soviet Union suggests that the ratio of foreign-denominated to total deposits in the domestic banking system—called the dollarization ratio—grows sharply during periods of high inflation. When inflation subsides, however, the dollarization ratio often remains high—a phenomenon that some describe as hysteresis. Interestingly, hysteresis seems to be present in most of Latin America and absent in Eastern Europe and Former Soviet economies.; There are two main competing explanations for these observed patterns in dollarization. Uribe (1997) and Peiers (1994) develop micro-theoretic models of hysteresis due to network externalities in transactions. These models imply that agents may continue using foreign currency even when relatively higher rates of return on domestic-denominated deposits prevail. On the other hand, dollarization may be caused by movements in rates of return.{09}Sahay and Vegh (1996) reveal that positive real rates of return in domestic-denominated assets are correlated with the recent drop in dollarization in Eastern Europe.{09}Battaile (1996), also finds dollarization rates for post-inflationary Bolivia and Peru to be well explained by standard ‘portfolio balance’ variables.; The primary aim of this dissertation is to incorporate both network externalities and portfolio motives into an empirical model to determine the importance of each effect in explaining recent patterns in dollarization.; This goal is achieved in several steps.{09}First, data on dollarization for fourteen Latin American, Eastern European and former Soviet economies are presented graphically so that possible patterns can be tentatively identified. Next, an ad-hoc foreign-currency demand equation is estimated for Bolivia using a Vector-Error-Correction Model, after incorporating ratchet terms. The estimates provide some support for the portfolio balance theory, but no support for hysteresis theory in post-stabilization dollarization. Then, a monies-in-the-utility function model that includes domestic and foreign-denominated interest-bearing assets and network externality is developed. The model, which reflects portfolio balance and hysteresis considerations, is empirically estimated using Hansen's generalized method-of-moments procedure. The results show that network externalities are insignificant in explaining post-stabilization dollarization patterns in both Bolivia and Argentina.