Alexander K. Swoboda
The spate of recent proposals for international monetary reform, ranging from advocacy of a return to the gold standard to calls for adoption of a target zone system, testifies to renewed interest in modifying existing exchange rate arrangements. This renewed interest reflects a degree of dissatisfaction with the present system as well as a perception that current circumstances—including the initiatives for greater cooperation and the convergence of inflation rates at moderate levels—are perhaps more favorable to at least modest steps toward reform than they have been since major currencies began to float in 1973.
The recent discussion of international monetary arrangements has been cast almost exclusively in terms of exchange rate stabilization. Less consideration has been given to the internal logic of alternative exchange rate systems, to the broader implications of that logic for the management of the world’s monetary system, and to the nature of international policy cooperation that each system requires. This article seeks to bring out these implications rather than to provide yet another catalogue of the strengths and weaknesses of the present system or to propose a specific plan for reforming it. Yet the exercise, however simple, does, I believe, have relevance for international monetary reform. Abiding by the logic of exchange rate systems is crucial to the design of an improved monetary system that is credible and viable. Any attempt at reform that violates that logic is bound to fail.
The environment of the 1980s
A credible exchange rate system must recognize the economic realities of the day. The world of the 1980s is not that of the immediate postwar period; nor is it that of the 1960s. Three features of the contemporary economic system are particularly relevant here.
First, the power structure in international financial and economic affairs has changed from what it was until the end of the 1960s. The United States was at the center of the regulating mechanism of what was known as the Bretton Woods system. That system’s breakdown can be partly attributed to the change in the relative economic power of the United States. No single country can, in present circumstances, provide the anchor for the international monetary system: any credible exchange rate arrangement depends on the active support of at least three players, the United States, the Federal Republic of Germany, and Japan.
Second, any viable system must recognize that capital markets, in addition to goods markets, have become closely integrated internationally. Macroeconomic policy, if it is to be stabilizing, must take this fact of life into account. In particular, the assignment of instruments to targets must be consistent with a high degree of capital mobility. For instance, under a managed exchange rate system, monetary policy needs to be devoted mainly to external equilibrium as it directly affects expected international yield differentials. In doing so, policy can stimulate equilibrating capital movements (movements that go in the same direction as the desired changes in the trade account) rather than creating the “need” to fight off capital movements that are the counterpart of a destabilizing assignment of policies. This is all the more desirable since attempting to restrict capital movements that are the reflection of disequilibrating policies is, by and large, a self-defeating task.
Third, the economic interdependence that arises from the integration of markets in goods and assets appears to be as strong under flexible as under fixed exchange rates, even though the nature of the interdependence and the channels of transmission may vary. It is important that this fact be recognized lest the adoption of particular exchange rate regimes lead to misplaced hopes about what they can achieve. Specifically, the autonomy that flexible exchange rates provide is, by and large, limited to more independence in national control of the price level and of nominal interest rates. Whether such national autonomy is worth having, especially in a small open economy, depends on how the rest of the world is behaving.
Two further facts of economic life must be borne in mind in any process of exchange rate reform. They concern, on the one hand, the connection between monetary policy and the exchange rate and, on the other, the connection between the exchange rate, fiscal policy, and the current account of the balance of payments. Under flexible exchange rates, the impact of variations in the money supply falls primarily on the nominal exchange rate in the short run, to some extent on output in the medium run, and largely on prices in the medium to longer run. Such variations also affect trade flows in the short run. For instance, an unexpected expansion of the money stock would tend to lead to a real depreciation of the currency and to an improvement in the trade balance resulting from an expansion of the export volume larger than the increase in import volume. Such an improvement may not be very strong, however, and will be shortlived. It will tend to disappear as domestic prices catch up. The corollary to the primary impact of monetary policy on the exchange rate is that to stabilize nominal exchange rates, monetary policy has to be devoted to that purpose.
To achieve a lasting change in the current account requires a lasting change in the savings-investment balance. This follows from the fact that the current account surplus is identical to the excess of national savings over national investment. This magnitude, in turn, consists of the excess of private savings over investment plus the government budget surplus. It would thus appear that dealing with the current account is a prime task of fiscal policy in an open economy. Consequently, one should not expect a change in the exchange rate to have a strong impact on the current account unless it has a strong effect on the private savings-investment balance or is accompanied by a change in fiscal policy, or both. I would thus argue that one should not expect a substantial improvement in the US current account balance from the recent depreciation of the dollar unless it is accompanied by a reduction in the fiscal deficit, although the depreciation should help expand both exports and imports. This is because, although private savings might increase somewhat as a result of the income expansion that results from export growth, private investment is likely to expand also as a result of improved growth prospects and lower interest rates.
The preceding discussion implies that (1) capital market integration and the exchange rate/money link make it imperative to adhere to the proper assignment of policies to targets, whatever the exchange rate regime, and (2) one cannot expect the exchange rate regime to do everything.
Logic of alternative regimes
The logic of an exchange rate regime defines what it can do and how it does it. At its simplest, and taking a longer-run perspective, the essential logic of an exchange rate regime concerns the joint determination of the price level, money stock, nominal exchange rate, and the real exchange rate. The logical characteristics of the purely flexible and the strictly fixed exchange rate systems are easy to define; those of “in-between” systems are not as clear-cut.
Flexible exchange rates. Under such a system, national monetary policy determines the national price level, and the nominal exchange rate adjusts (given foreign prices) to bring the real exchange rate to a level that is consistent with underlying tastes, technology, and factor endowments. That is, under flexible rates a country’s money stock is an exogenous, nationally determined, variable, while the domestic price level, and the nominal and real exchange rates, are endogenous variables determined by demand and supply, given the money stock and the independently set foreign prices. Given tastes, technology, and factor endowments, money is neutral in the long run; a doubling of the national money stock eventually leads to a doubling of the national price level and to a halving of the exchange value of the currency, leaving the real exchange rate unchanged. In the short run, of course, monetary changes, especially if unexpected, are likely to have real effects; a monetary expansion would in all likelihood lead to a depreciation of the currency before it leads to an increase in prices, depreciating the real value of the currency in the process. Moreover, taking interest rate effects and capital mobility into account, the currency may initially depreciate by more than the percentage increase in the money stock, an instance of the so-called overshooting phenomenon. If the country in question is large, these changes will produce real effects abroad as well as at home; flexible exchange rates insulate countries from foreign monetary shocks in the long run only.
Fixed exchange rates. The role of the money stock and the exchange rate is reversed under such a regime: money is endogenous, the exchange rate exogenous. Given foreign prices, once the nominal exchange rate is fixed, the national price level adjusts to yield a real exchange rate that is consistent with existing tastes, technology, and endowments; the domestic stock of money, in turn, adjusts to the level required to achieve payments equilibrium at the chosen parity. The monetary instrument that the authorities can control is the supply of domestic credit, not the money stock (at least in a small open economy). An increase in domestic credit will eventually result, again for a small economy, in an equivalent loss of international reserves leaving real variables unchanged. Again, there are likely to be real effects in the short run, but these will be particularly short-lived if capital mobility is high.
One important difference from flexible exchange rates is that the foreign price level cannot be assumed to be given independently of the actions of large countries. Under fixed exchange rates it makes sense to think of a world price level (to which national price levels are closely linked) whose long-run behavior is governed by movements in the world money stock (the sum of national money stocks converted into a common currency unit at the going fixed exchange rates). The evolution of the world money stock in turn depends on the specific exchange rate pegging arrangements that are adopted and on the assets that are held as international reserves by national monetary authorities. This is closely related to the so-called n-1 problem. In a world of n countries there are only n-1 exchange rates. Suppose that the monetary policies of n-1 countries are used to fix the n-1 exchange rates; then the monetary policy of the nth country must be used for some other purpose. Under a gold standard, the nth policy implicitly determines the price of gold; in a system where a national currency is used as an international reserve asset that country is the nth country and its monetary policy determines its price level, the world money stock, and the world price level.
Lest this appear too abstract a proposition, consider the de facto dollar standard that characterized the period extending from the early 1960s to early 1973. The fact that major industrial countries pegged to the dollar and that the major part of changes in their international reserves took the form of changes in their holdings of dollar assets implied that US monetary policy by and large determined the course of the world money stock and of the world’s rate of inflation in the medium run during that period. The main point here is that the way in which the n-1 problem is solved is an integral part of any fixed exchange rate regime; it is a problem that cannot be avoided.
In-between systems. A large variety of exchange rate regimes lie somewhere between the two prototype pure systems. They range from wide-band systems, which behave like floating rate systems within the band and like fixed rate systems once the limits of the band are reached, to gliding parities that crawl at a predetermined rate and behave like fixed rate systems but allow (or force) national inflation rates to diverge by the rate of the crawl. They include “fixed but frequently adjustable exchange rates” which attempt to strike a compromise between discipline and autonomy that, if not sufficiently credible, is likely to be severely tested by markets. They include, finally, target zones which have attracted considerable attention recently and therefore deserve some discussion.
One motivation for target zones is the perception that the current system has given rise to serious and lasting misalignments of exchange rates. The degree of misalignment is determined by comparing the current exchange rate with what has been termed the fundamental equilibrium exchange rate. The latter is the real exchange rate which would bring about a current account balance compatible with “underlying” capital flows at a level of national income that corresponds to internal balance at home and abroad. Putting a band around that “equilibrium” rate defines the target zone within which the authorities should strive to keep the actual real exchange rate. One can distinguish several versions of the target zone proposal depending, notably, on the precise method used to calculate the fundamental rate, the width of the band, whether the band is announced or not, and the degree of commitment to keeping the actual exchange rate within the zone.
In terms of the theme of this essay, the main difficulties with the target zone proposals are that their internal consistency is not obvious and they could exacerbate policy conflicts and tax the ability and willingness of governments to coordinate macroeconomic policies. First, there are many technical difficulties in defining the fundamental equilibrium exchange rate; the choice of a base period and the price indices to be used in the calculation are but two examples. Even if these difficulties could be solved, there is no compelling reason why the real exchange rate should be chosen as either an objective or an intermediate target of policy. If the objective is to balance the current account, then it should be targeted directly using actual instruments of macroeconomic policy—fiscal policy for instance.
Second, there are no clear means for achieving the target real exchange rate once it has been selected. Advocates of target zones suggest that monetary policy (possibly supplemented with foreign exchange market intervention) can be used for the purpose since it affects the nominal exchange rate and, with sticky prices, the real exchange rate. But the relationship between nominal and real exchange rates can, at best, be exploited by monetary or intervention policy in the short run only, a horizon that is shorter than the medium- to long-term perspective that underlies the calculation of the fundamental equilibrium real exchange rate.
Third, no market mechanism is provided by which the real exchange rate can adjust to the changes in underlying technology, tastes, endowments, and government policies that affect real variables; such changes must be constantly monitored by the authorities and their implications for the equilibrium real exchange rate must be continuously calculated so that the target zones can be adjusted. This has two consequences. On the one hand, the procedure maximizes the informational requirements for correct policy making, hardly a desirable property in a world of uncertainty and limited information. On the other hand, if the authorities fail to calculate the fundamental real rate correctly, the pursuit of target zones will be fundamentally destabilizing, leading to continuous inflation in those countries for which the target rate has been set too low, and to continuous deflation where it has been set too high.
Finally, the degree of international agreement and cooperation required to operate a target zone system properly may be greater than can realistically be expected and than would be required to run a system of strictly fixed exchange rates. One reason is that the “fundamental equilibrium exchange rate” is not an objective variable, just as what constitutes a “fundamental disequilibrium in the balance of payments” is a matter of judgment over which analysts equipped with the same basic information can honestly disagree. The calculation of the fundamental rate requires not only detailed knowledge of the workings of the home and foreign economies but also agreement as to what all the policies that impinge on the determination of the world economies’ main relative prices (notably the terms of trade and the relative price of traded and nontraded goods) should be. For the fundamental rate is not invariant with respect to these policies. Nor is the “normal underlying capital flow” a concept on whose value countries are likely to agree. Furthermore, target zone plans typically do not address the n-1 problem explicitly. That problem is likely to be exceedingly difficult to solve in a system that takes real exchange rates as targets.
Conflict and cooperation
Alternative exchange rate regimes require different degrees and kinds of policy coordination to function properly and they resolve potential policy conflicts in different ways.
On the surface, a system of flexible exchange rates requires no policy coordination to operate: conflicts could, in principle, be resolved automatically by market forces. This may well be a virtue when international monetary cooperation is, for one reason or another, impossible.
Policy conflicts do exist under flexible rates notably because, as argued above, money is not neutral in that regime in the short run. One country might be tempted, for instance, to stimulate output through an expansion of exports gained via a competitive depreciation of its currency. It might thus gain employment at the expense of the rest of the world; hence, monetary policy may under certain circumstances be a “beggar-thy-neighbor” policy. Furthermore, with closely integrated capital markets, the impact of one large country’s policies on the real rate of interest becomes a matter of international concern. In general, instability in any one country becomes a matter of international concern as it will destabilize exchange rates even if the rest of the world behaves in perfectly stable fashion. The implication is that the right type of policy coordination (and policy stability in general) is desirable in a world of floating rates as it can lead to better outcomes for all. Such coordination, however, is difficult to achieve in practice as detailed knowledge of the dynamics of transmission is lacking and as the incentives to adopt nationalistic policies that maximize short-run gains can be high. It may be, then, that the best way to run a floating exchange rate regime is to refrain from overly refined attempts at policy coordination, to try to achieve stability of individual national macroeconomic policies, and to observe a few simple rules of conduct, discussed further below.
In one important sense, and contrary to popular opinion, a fixed exchange rate system requires little deliberate coordination of macroeconomic policies to function properly. “All” that is required is that national monetary policies be devoted to the maintenance of existing parities or, equivalently, to the maintenance of overall balance of payments equilibrium at the given parities. (In fact, setting targets cooperatively for national money or domestic credit growth rates may be positively harmful, as these rates are unlikely to be those required to maintain payments equilibrium in a world subjected to a wide variety of shocks.) That is, fixed exchange rates require the abandonment of monetary autonomy in the simple sense that the choice of a parity is, by and large, the choice of a domestic price level and rate of inflation in the long run, given foreign prices and the foreign rate of inflation. The loss of autonomy may well be acceptable if the foreign rate of inflation is acceptable to the countries that are part of the fixed exchange rate area. That rate of inflation will depend on the way in which the n-1, or anchor, problem is solved. There are many solutions to that problem, ranging from the gold standard to internationalist solutions a la Keynes or Triffin, based on the collective creation of an outside asset such as the “Bancor.” Each requires a specific kind and degree of international monetary cooperation. Indeed, it is the necessity of solving the n-1 problem rather than the fixing of exchange rates per se which requires agreement, or international cooperation, for a fixed rate system to be viable.
In fact, one may attribute a large part of the success of the Bretton Woods system until about 1966-68 to basic agreement with the broad stance of US macroeconomic, particularly monetary, policy and its implications for the growth of the world money stock and price level. Similarly, one can attribute a large part of the breakdown of that system to increasing disagreement with the inflationary bias imparted to the world economy by the rise in US monetary expansion that took place in the late 1960s and spread to the rest of the world through the mechanics of the dollar standard.
Whichever way the n-1 problem is solved, whatever exchange rate regime is adopted, credibility is the key to viability. As already indicated, to be credible, an exchange rate regime must recognize economic realities, be run in a fashion that is aligned with its internal logic, and be consistent with the existing degree of international monetary cooperation. There should also be enough commitment to the rules of the game if the regime is not to break down as markets test the resolve of the authorities. Rules have an important role to play in insuring the stability of any regime that is adopted; too much discretion leads to instability.
To be workable in practice, however, rules must possess certain desirable properties. They should be simple so as to be monitorable and enforceable. They should pursue objectives that are attainable with available instruments. And they should be consistent with the political will of governments, lest they be abandoned whenever they are perceived to entail short-run costs. This last point provides one argument (there are others) for contingent rather than rigid rules; it is unlikely that governments would adhere to rigid rules, whatever the contingency. For instance, a fixed money supply growth rule is unlikely to be followed in the midst of a severe depression; neither is a balanced budget rule particularly appealing at every point of the cycle. It is much better to design rules that achieve monetary or budgetary stability on average over the cycle, while specifying clearly and simply the deviations that are permissible over the cycle.
Two types of rules are relevant here. The first concerns the principles that should guide the evolution of the instruments of macroeconomic policy over time. One example of a useful contingent rule under fixed exchange rates is Mundell’s famed assignment of monetary policy to external balance and fiscal policy to internal balance. This type of rule has three advantages: it is consistent with the logic of the exchange rate system in which it is embedded; its application is contingent on the state of the economy (on whether there is inflation or recession, a payments surplus or deficit); and its informational requirements are low in that a detailed knowledge of the economy is not required for the rule to lead, eventually, to the desired policy outcome. Such a rule also has the advantage of leaving it to national discretion to decide the precise rate at which the values of policy instruments should be changed to reach ultimate targets. (That rule is, however, insufficient in the medium run; it should be supplemented by achieving, on average, that budget position which is compatible with a sustainable current account over the longer run.)
The second type of rule concerns appropriate broad principles of behavior beyond the setting of the values of macroeconomic policy instruments. As an example of such a broad code of conduct, one could argue that any agreement on reforming the exchange rate system could include the following four principles: (1) be limited to three (or perhaps five) major countries; (2) give countries outside the agreement the choice either to peg to a currency of their choice or to float (while obeying some rules); (3) include a number of simple rules of behavior in the form of “do’s and don’ts”: for example, do not engineer competitive depreciation under flexible rates; do not attempt to use monetary policy for domestic goals except over the business cycle under fixed exchange rates; and (4) reaffirm and honor strictly the commitment to open trade in goods and assets that the system is designed to foster.
A final remark. It may have become apparent that I would favor a pure exchange rate system over an in-between system, especially one that attempts to target real exchange rates, and also that I have some preference for a fixed rate regime over a flexible one because it is simple to run, provides for discipline, can help provide an anchor for expectations, and may be easier to stabilize. This is not the place to develop those arguments. The important point is that for any regime to work requires a sufficient degree of consensus and commitment, and rules of conduct that are consistent with its internal logic.