There is by now a large and well-known body of theoretical literature on the relative merits of fixed and flexible exchange rate regimes. This section first describes the salient features of the literature, which has been developed mainly in the context of industrial countries engaged in short-run stabilization of output,4 but which may also be relevant for developing countries not restricted by a severe balance of payments constraint. The discussion focuses on the effectiveness of fixed and flexible exchange rates in insulating the economy from external and domestic shocks. The section concludes with an assessment of the practical implications of this literature for adopting fixed or flexible exchange rates in cases in which the objective is to stabilize output in a low and stable inflation environment.
Optimal Exchange Rate Regime
There is no unambiguous right answer to the question of whether a small country is better off with a fixed or a flexible exchange rate system. In general, the optimal management of the exchange rate depends on the policymakers’ economic objectives, the source of shocks to the economy, and the structural characteristics of the economy in question. Consequently, alternative assumptions about these factors can be expected to alter the optimal degree of exchange-rate flexibility. It is therefore difficult to draw from the theoretical literature guidelines for exchange rate policy which are of general applicability.
The first issue that needs to be addressed is the criterion of optimality. In principle, a standard welfare-related criterion should be specified and applied.5 In practice, however, the modern analysis of this issue has focused on the relatively narrow criterion of macroeconomic stability—defined in terms of minimizing the variance of real output, the price level, or real consumption—in the face of random transitory shocks. The problem with specifying policy objectives in such a fashion is that choosing to stabilize any single macroeconomic variable runs the risk of destabilizing some other variable that may also seem relevant to the general welfare. Such trade-offs could be made only by using an explicit welfare function, but no such comprehensive analysis has as yet been undertaken. The criterion most commonly adopted in the literature has been the stability of real output. The concrete question would in this case be how best to manage the exchange rate so as to minimize the variance of real output around its full capacity level in the face of random shocks arising from diverse external and domestic sources.
An important result of the theoretical literature is that, in general, neither of the extremes of permanently fixed or completely flexible exchange rates is optimal in seeking macroeconomic stability. Instead, some intermediate degree of flexibility is more likely to succeed in stabilizing the economy in response to random shocks. In fact, such “managed” or “flexibly fixed” exchange rates constitute the most common arrangement in developing countries. The question then is what degree of flexibility should be adopted in specific circumstances.
The type of shock to which the economy is likely to be subjected is a key consideration in determining whether the exchange rate should be fixed or adjusted. This view emerges clearly from the longstanding debate on fixed versus flexible exchange rates. For example, early arguments for flexible exchange rates emphasized the insulating properties of exchange rate adjustment in the face of external nominal shocks—that is, faced with movements in the foreign price level, domestic prices can be stabilized by a suitable adjustment in the exchange rate. Thus, when foreign nominal shocks are important, flexibility in exchange rate management has been perceived as desirable.6
The early literature also identified certain criteria for exchange rate management in response to specific domestic shocks. Whether domestic shocks favor exchange rate fixity or flexibility depends on whether the shocks are monetary (i.e., originating in the money market) or real (i.e., originating in the goods market). When domestic shocks are monetary, the conventional view is that maintaining a fixed exchange rate would be more effective in stabilizing output. Given that the money supply is endogenous under a fixed exchange rate, disturbances in the domestic money market would simply be absorbed by changes in international reserves without affecting the supply-demand conditions in the goods market that determine the level of economic activity. In contrast, when the domestic shocks are real, the exchange rate should be adjusted to stabilize output by generating (or withdrawing) external demand. For example, exchange rate appreciation would cushion the effect of a positive shock to domestic demand on output by directing part of the additional demand abroad.7
Generally speaking, to the extent that the policy aims at stabilizing output in the face of transitory shocks, the exchange rate should be adjusted when shocks originate abroad or in the domestic goods market, but should be fixed when shocks originate in the domestic money market. However, the structural characteristics of the economy, such as openness to international trade, the degree of capital mobility, and rigidities in the labor market, affect the insulating properties of the exchange rate regimes.
Prima facie, it might be argued that the more open the economy (i.e., the larger the share of the traded goods sector), the stronger the case for fixing the exchange rate because of potential costs to international transactions of frequent exchange rate adjustments. Furthermore, such openness makes a fixed exchange rate more effective in channeling abroad a domestic monetary shock. On the other hand, openness helps an exchange rate adjustment to stabilize output in the face of external or domestic demand shocks. To the extent that openness makes a country more vulnerable to external shocks, it may necessitate more frequent adjustments of the exchange rate. On the basis of these considerations, it is clear that openness per se does not unambiguously determine the choice of exchange rate regimes.
Capital mobility, which directly links domestic and foreign interest rates, also influences the effectiveness of exchange rate adjustment in stabilizing the economy.8 Consider, for example, a rise in external demand associated with an expansionary monetary policy and a corresponding decline in foreign interest rates. Under a fixed regime and in the presence of capital mobility, domestic interest rates would also fall, reinforcing the destabilizing effects of the higher external demand operating through the current account. In this case, the exchange rate needs to be appreciated to stabilize domestic output. By contrast, if the rise in external demand is associated with expansionary fiscal policy, the concomitant rise in interest rates would tend to offset the destabilizing effect of higher external demand on domestic output. In fact, depending on the relative size of the various interest elasticities of aggregate demand and supply functions, the net impact of such an external shock on domestic output may be negative. In such a case, the appropriate policy response would be to depreciate the exchange rate. Thus, the nature of the expansionary external shock would affect the specific exchange rate action. A similar line of reasoning demonstrates that, under a fixed regime, a high degree of capital mobility amplifies the destabilizing effects of a given domestic demand shock (by preventing a change in domestic interest rates, which would otherwise tend to stabilize demand). In the case of a domestic monetary shock, a high degree of capital mobility makes a fixed rate more effective in stabilizing output by limiting the destabilizing movements of domestic interest rates.
The degree of wage rigidity also has a bearing on the effectiveness of exchange rate policy. The effects of a nominal devaluation on macroeconomic variables depend critically on how nominal wages and prices respond to a devaluation. The larger the increase in nominal wages in response to a devaluation, the smaller the change in real wages. Thus, when the degree of wage indexation to the general price level is high, and when the economy is very open (so that traded goods represent an important component of the general price index), the effect of a change in the nominal exchange rate on the real wage—and thus on output—will be small.9 The effects of wage indexation on the desirability of exchange rate flexibility can therefore be expected to be significant.10
Practical Policy Implications
The above discussion has suggested a number of criteria for moving the exchange rate depending on the nature of the shock to the economy, as well as the economic structural characteristics. In practice, however, it is difficult to apply these criteria to specific cases. These difficulties are associated not only with the practical problems involved in discerning the source of shocks and identifying the relevant structural characteristics, but also with potential conflicts among various policy objectives.
Identifying the nature of shocks responsible for macroeconomic instability is arduous. The economy is often affected simultaneously by various shocks from different sources, which cannot be accurately identified, let alone quantified. For example, in the absence of a complete model of the economy, it is generally impossible to determine the extent to which a slowdown in economic activity can be attributed to sluggish domestic demand, tight monetary conditions, or weak external demand. Determining whether a particular shock, say a deterioration in the terms of trade, is a transitory or a permanent phenomenon further complicates the analysis. Thus, while in some particular cases—such as a severe drought—the shock to the economy may be readily identified, in many cases it would not be possible to identify the source of the shock.
The criteria for choosing the appropriate exchange rate regime also depend on the policy objectives adopted by the authorities. The literature on the merits of fixed and flexible exchange rates has, by and large, been relevant to the case of a small country that is not experiencing severe inflation and is broadly in external equilibrium, but that is subjected to transitory shocks. In this framework, it may indeed be appropriate to focus on the stability of output, rather than the balance of payments, as the main policy objective; in fact, output is stabilized precisely because the balance of payments acts as a shock absorber to dampen the impact of various disturbances on domestic output. However, a policy conflict would arise when the balance of payments is itself a policy objective.
The nature of the above policy conflict can be highlighted by considering some of the specific cases discussed earlier. For example, fixing the exchange rate to channel abroad a positive monetary shock would not be feasible if international reserves are already below the desired level. Similarly, an appreciation of the exchange rate aimed at neutralizing a rise in domestic demand would inevitably add to the deterioration in the balance of payments, which might not be acceptable. There are, of course, cases in which the attainment of domestic and of external objectives would be compatible. For example, a negative supply shock at home would require an exchange rate depreciation to stabilize output as well as to improve the external position. In many cases, however, output stability, inflation, and the balance of payments are likely to present conflicting policy objectives. This conflict is likely to be particularly pronounced in the case of developing countries, which are frequently short of foreign exchange. To the extent that the external position imposes a binding constraint, the exchange rate has to be used as an instrument for attaining the balance of payments objective rather than stabilizing output. On the other hand, if inflation constitutes the most pressing problem, exchange rate policy may have to be assigned the task of assisting in the stabilization of the price level. The subsequent sections of the paper focus on the role of exchange rate policy in the process of external adjustment and in bringing about financial stability.
A very useful recent contribution by Genberg (1989) covers many of the relevant issues. Earlier work on the subject includes Frenkel and Aizenman (1982). See also Argy and De Grauwe (1990).
See Friedman (1976) and Johnson (1969). A more formal analysis is undertaken by Turnovsky (1976).
An early analysis leading to these conclusions is provided by Boyer (1978).
See, for example, Flood (1979).
Under the extreme assumption of full indexation, an exchange rate adjustment would be completely ineffective in stabilizing output in the face of domestic shocks. In such an extreme case, the choice of exchange rate regime would be immaterial from the perspective of output stabilization. See Marston (1982) and Tumovsky (1985).
For a more complete discussion, see Aizenman and Frenkel (1985).