پویایی تورم و جایگزینی ارز در اقتصاد باز کوچک

This paper analyzes the relationship between money and inflation in a small open economy, where domestic and foreign currencies are perfect substitutes as means of payment. It is shown that, if the path of domestic money supply is such that individuals find it optimal to change the currency in which transactions are settled, there will be an adjustment period during which domestic inflation adjusts to equal the foreign inflation rate. The model captures the stylized fact that temporary increases in the inflation rate may have permanent effects in the use of foreign currency, even without the introduction of dollarization costs.

In high inflation countries, the rapid erosion of the value of domestic currency leads agents to substitute it with foreign currency in any or all of the basic functions of money. This phenomenon is usually called currency substitution, or simply dollarization (for a literature review, see Giovannini and Turtelboom, 1994). A question that has attracted increasing attention is that, in some cases, dollarization exhibits hysteresis, with the demand for foreign currency remaining high after stabilization (for empirical evidence, see Guidotti and Rodriguez, 1992, Kamin and Ericsson, 2003, Clements and Schwartz, 1993, Mueller, 1994 and Reding and Morales, 1999). However, this is not always the case. In some economies of Eastern Europe, for example, dollarization has fallen substantially in the aftermath of successful stabilization plans ( Sahay and Végh, 1996). This paper follows Guidotti and Rodriguez, 1992 and Uribe, 1997, and Reding and Morales (1999), addressing the issue of dollarization hysteresis with a small open economy model where domestic and foreign currencies are perfect substitutes as means of payment. Our approach contributes to the literature in two main respects. Firstly, we allow the inflation rate to be endogenous. Secondly, we are able to explain the phenomenon of dollarization hysteresis even in the absence of dollarization costs. Previous explanations for the phenomenon of dollarization hysteresis have emphasized the role of dollarization costs. Dornbusch and Reynoso (1989) and Dornbusch et al. (1990) argued that the process of financial adaptation involves learning costs that, once incurred by economic agents, imply persistence. Guidotti and Rodriguez (1992) present a model with perfect means of payment substitutability, in which agents face costs of adjusting their holdings of foreign currency. These costs result in an inflation band within which agents choose not to switch between currencies. Hence, de-dollarization may only be achieved if domestic inflation decreases enough to offset the switching costs. Perfect means of payment substitutability was also assumed by Uribe (1997) and Reding and Morales (1999), who stressed the role of network externalities as the source of non-linearities in the relationship between money demand and inflation. According to this interpretation, transaction costs faced by individual agents decline with the aggregate level of dollarization, giving rise to multiple equilibria and history dependence in the demand for money. Thus, a temporary high level of inflation can start a dollarization process which will not necessarily be reversed when inflation comes down, as network economies will provide agents with a permanent lowering of transaction costs in the use of foreign currency. Guidotti and Rodriguez, 1992 and Uribe, 1997, and Reding and Morales (1999) have two important shortcomings. First, the inflation rate is assumed to be exogenous. As changes in the money demand impact on the inflation rate, the dynamic analysis made by these authors may be misleading. Second, these approaches involve the specification of dollarization cost functions, which are necessarily ad-hoc. In this paper we argue that the short-term dynamics of money and inflation under the assumption of perfect means of payment substitutability is sufficiently rich to capture the different patterns of dollarization identified in the empirical literature, without the need to introduce dollarization costs. A well known result with perfect substitutability is that, without binding constraints on currency holdings, the exchange rate is undetermined (Kareken and Wallace, 1981). In short, any monetary equilibrium with two currencies coexisting in the same commodity domain requires their user costs (inflation rates) to be equal. Under these circumstances, the demand for each currency is undetermined and so, too, will be the (unchanging) exchange rate. The same mechanism that produces exchange rate indeterminacy in Kareken and Wallace (1981) is at work here, but our model differs in that residents of both countries face binding constraints in the use of currencies. In particular, we assume that foreign residents are not allowed to hold the domestic currency (only asymmetric currency substitution is considered) and a minimum constraint on the amount of domestic money holdings is imposed. 1 In this framework, money demand indeterminacy occurs only when money supplies evolve at the same rate. We show, however, that if money growth rates evolve in such a way that currency substitution becomes optimal, there will be a period of time during which the domestic inflation adjusts to equal the foreign inflation rate. We also show that the demand for foreign currency does not necessarily decline when a disinflation program is implemented in a dollarized economy. If, however, reversibility occurs (partial or total), the demand for foreign currency will decline smoothly, contrasting with the once-and-for-all change in the inflation rate. The remainder of the paper is organized as follows. In the next section, we present the model. In Section 3, we discuss the dynamics of inflation and currency substitution. In Section 4, we discuss the possible outcomes of a disinflation program. Conclusions are presented in Section 5.

In this paper we investigate the dynamics of money demand and inflation in a small open economy, where domestic and foreign currencies are perfect substitutes as means of payment. The main property of this model is that, if the path of money supply is such that individuals find it optimal to change the currency in which transactions are settled, there will be an adjustment period during which the domestic inflation rate adjusts to equal the foreign inflation rate. This implies that, during the currency substitution process, the behavior of money demand is not accounted for by changes in the current inflation rate. In the context of this model, a temporary increase in the rate of money creation gives rise to a demand for foreign currency that does not necessarily decline when the inflation rate comes down. If, however, reversibility occurs (partial or total), the demand for foreign currency will decline smoothly, contrasting with the once-and-for-all change in the inflation rate. Therefore, the model is able to explain the failure of econometric studies in identifying stable money-demand relationships in dollarized economies. The model captures the different patterns of dollarization identified in the literature, namely reversibility in some cases and non-reversibility in others. This does not depend on any ad-hoc dollarization cost function. It is, instead, an obvious consequence of the assumption of perfect substitutability. Reversibility in this context is assured if the domestic currency becomes a better alternative than the foreign currency. Such an event, however, may not be easy to achieve in most developing economies. In these cases, co-existence will be the more likely outcome of a successful stabilization plan, unless administrative actions are taken to banish the foreign currency. It may be argued that the knife-edge equilibria discussed in this paper are not realistic. It is true that in the real world agents do not switch between currencies in response to small inflation differentials. As mentioned above, however, the model can easily be extended so as to better resemble reality, introducing some form of dollarization costs. But there is nothing particularly new in such an extension that makes it more attractive than the version discussed in this paper. On the contrary, by focusing on the simpler model, we stress the main point that there are more fundamental forces driving the behaviour of money demand than dollarization costs, which are not even necessary to explain the stylized facts.