Official Intervention on the Forward Exchange Market: A Simplified Analysis

INTEREST IN THE PROBLEMS of official intervention in forward exchange markets has received impetus from recent changes of practice on the part of the U.S. monetary authorities. In March 1961, the U.S. Treasury intervened for the first time on the exchange markets in support of the forward dollar; and since February 1962, the Federal Reserve System has concluded a number of swap arrangements with other central banks.1 In view of these developments, it appears timely to consider what, in theory, are the short-run and long-run effects of official transactions in forward exchange, and under what circumstances such transactions are likely to serve a useful purpose. Tentative answers are given in this paper to such questions as the following: Should intervention be “limited” in extent and/or duration, and, if so, in what sense, and why? In what sort of payments situation is official support of the forward exchange appropriate? Need it be confined to meeting “speculative” attacks?

Abstract

INTEREST IN THE PROBLEMS of official intervention in forward exchange markets has received impetus from recent changes of practice on the part of the U.S. monetary authorities. In March 1961, the U.S. Treasury intervened for the first time on the exchange markets in support of the forward dollar; and since February 1962, the Federal Reserve System has concluded a number of swap arrangements with other central banks.1 In view of these developments, it appears timely to consider what, in theory, are the short-run and long-run effects of official transactions in forward exchange, and under what circumstances such transactions are likely to serve a useful purpose. Tentative answers are given in this paper to such questions as the following: Should intervention be “limited” in extent and/or duration, and, if so, in what sense, and why? In what sort of payments situation is official support of the forward exchange appropriate? Need it be confined to meeting “speculative” attacks?

INTEREST IN THE PROBLEMS of official intervention in forward exchange markets has received impetus from recent changes of practice on the part of the U.S. monetary authorities. In March 1961, the U.S. Treasury intervened for the first time on the exchange markets in support of the forward dollar; and since February 1962, the Federal Reserve System has concluded a number of swap arrangements with other central banks.1 In view of these developments, it appears timely to consider what, in theory, are the short-run and long-run effects of official transactions in forward exchange, and under what circumstances such transactions are likely to serve a useful purpose. Tentative answers are given in this paper to such questions as the following: Should intervention be “limited” in extent and/or duration, and, if so, in what sense, and why? In what sort of payments situation is official support of the forward exchange appropriate? Need it be confined to meeting “speculative” attacks?

The approach in this paper is a simplified one.2 The purpose is to provide a straightforward account of the theory of intervention and to use it to discuss the problems just raised. To do this, a new classification of the forces determining the forward rate is developed: the rate must be such as to equalize the Net Speculative Position (defined as the sum of the net assets—spot or forward—of residents of country A in other currencies less the sum of the net assets—spot or forward—of residents of other countries in A’s currency) and the Net Loan Position (defined as the sum of the net lending of residents of A to residents of other countries). This presentation is believed to have certain expositional merits.

Determination of Equilibrium

To simplify the analysis, complications arising from the multiplicity of foreign currencies are ignored; official intervention is thought of as taking place on the forward market for the domestic currency versus “foreign currencies in general,” which, for convenience, will generally be referred to as though they constituted a single currency.

It is further assumed, initially, that spot rates are rigidly fixed. Finally, the complications arising from the variety of maturity dates of forward exchange contracts are left out of account; all forward contracts are treated as if they had the same maturity, and all forward rates as if there were only a single rate.

The forward exchange market may be conveniently treated in terms of stocks rather than of flows; that is, the forward exchange rate is taken as reconciling the desires of market participants with respect to the holding—rather than the changing—of forward exchange positions. In every forward exchange contract, each of the contracting parties holds a forward asset in one currency and a forward liability in another currency. Forward exchange positions are forward positions (assets or liabilities) in currencies other than that of the country of which the holder is a resident. Spot exchange positions are assets or liabilities, other than forward positions, in currencies other than that of the country of which the holder is a resident.

If a resident of country A lends to a resident of country non-A, either the A resident acquires an asset in non-A currency, or the non-A resident acquires a liability in A currency; in either case, there is an increase in the net exchange position of the A resident less the net exchange position of the non-A resident. On the other hand, if an A resident concludes a forward exchange contract with a non-A resident, there is no change in net lending between the residents of the two areas; the net forward exchange position (positive or negative) acquired by the A resident in non-A currency will be balanced by an equal net forward exchange position acquired by the non-A resident in A currency. It follows that the sum of the net assets (spot or forward) of A residents in non-A currency less the sum of the net assets (spot or forward) of non-A residents in A currency, which may be termed the Net Speculative Position (NSP) of country A, is necessarily equal to the sum of the net lending of A residents to non-A residents, which may be termed the Net Loan Position (NLP) of country A.

While these Net Positions are necessarily equal ex post, they are not necessarily equal ex ante. The desire to lend abroad is not necessarily matched individually or collectively by the desire to hold a foreign currency. A decision to borrow or lend abroad is, in principle at least, different from a decision to undertake an exchange risk. Let us assume that, where neither the lender nor the borrower wishes to undertake the foreign exchange risk that is necessarily involved in the loan, the party that first takes the risk seeks to cover it by concluding a forward exchange contract. In this event, there will be a net demand in the forward market for A currency in terms of non-A currency, or vice versa, and the forward exchange rate will move until the equilibrium of the market is restored and the desired NSP equals the desired NLP.3

The manner in which, given the spot exchange rate, the forward exchange rate serves to bring about equality between the desired NSP and the desired NLP is illustrated in Diagram 1. In that diagram, the balance of claims (which can be regarded, ex post, either as the NLP or the NSP) is measured on the horizontal axis, while the forward exchange rate is measured on the vertical axis. Both NLP and NSP are considered from the standpoint of country A. The forward rate represents the forward price of A currency in terms of non-A currency. The curves are drawn on the assumption that the spot rate of exchange, the anticipated future spot rates of exchange, and the rates of interest in A and non-A are not affected by changes in the forward exchange rate.

As the diagram shows, country A’s NLP slopes negatively, and its NSP positively, with respect to the forward price of its currency. (Strictly speaking, while the NSP varies directly with the outright forward rate, the NLP varies inversely with the forward premium, i.e., the excess of the forward value of A’s currency over its spot value. But with the spot exchange rate fixed, this is equivalent to saying that it varies, inversely, with the forward rate.) It is easy to see why the NLP schedule, as a function of the forward rate, slopes negatively. The higher the forward rate (and hence the greater the forward premium on A currency), the cheaper it is for non-A lenders or A borrowers (whoever is bearing the exchange risk) to cover that exchange risk by a forward contract. The higher the forward rate, the greater will be the incentive for non-A residents to lend to A residents on a covered basis, and the amount of such lending will tend to increase. For analogous reasons, the cost of covered lending from A residents to non-A residents will increase, and the amount of such lending will tend to diminish. Moreover, the changes in loan positions need not take the form of changes in covered exchange positions; they may take the form of switches between spot positions (positive or negative) and forward positions of the same sign. For example, a rise in the forward exchange rate will tend to induce A residents with positive spot positions in non-A currency to switch to forward positions, and non-A residents with positive forward positions in A currency to switch to spot positions.

There is another way in which a rise in the forward rate will reduce the NLP. Such a rise will lead to an increase in the cost of covering, and hedging, the foreign exchange risk involved in A’s exports and a decline in the cost of covering, and hedging, the risk involved in A’s imports. (We use the term “covering” to mean offsetting by a forward contract the exchange risk involved in any trade credit that extends over the period from the date when ownership passes to the date of payment; we use “hedging” to mean a similar offsetting of the exchange risk involved in the export contract over the period from the date of contract to the date when ownership passes.) Insofar as this results in a decline in A’s exports and a rise in A’s imports, which in turn involves a decline in A’s covered trade credit to non-A and a rise in non-A’s covered trade credit to A, there will be a decline in the NLP other than that already taken into account.

The NSP schedule normally has a positive slope. On the assumption that changes in the forward rate have no effect on expectations regarding future spot rates, a rise in the forward rate (i.e., in the forward price of A currency in terms of non-A currency) will induce an increase in the net positions of A residents in non-A currency and a decline in the net positions of non-A residents in A currency. Residents of A will have an incentive to increase their forward positions in non-A currency whether or not they have negative spot positions in non-A currency which they can cover by so doing; analogously, non-A residents will reduce their forward holdings of A currency even though this may involve leaving spot positions uncovered or other risks unhedged. To some extent, as we have seen, these shifts in forward positions will be matched by corresponding shifts in the opposite direction in spot positions. Spot positions, however, involving as they do the opportunity cost of tying up funds in a particular use, are not perfect substitutes for forward positions, and a change in the forward rate, i.e., in the price of forward positions, will involve a net change in the sum of spot and forward positions taken together, i.e., a change in the NSP.

Insofar as the rise in the forward rate worsens the trade balance, it is likely not merely, as we have seen, to reduce the NLP by reducing the balance of covered trade credits but also to raise the NSP by reducing the hedging of the currency risks associated with export contracts (prior to the time of actual export) and increasing the hedging of currency risks associated with import contracts.

Equilibrium between the NSP and the NLP is determined at the point Q, common to both schedules, where the market is willing to maintain the Net Speculative Position implied by the desired Net Loan Position. The equilibrium point determines both the equilibrium forward rate, P0, and the equilibrium balance of claims, C0.

Suppose now that there is a rise in the expected exchange value of foreign currencies in terms of the domestic currency. This will lead to a tendency toward an uncovered capital movement, expressed in a shift of both the NSP and the NLP schedules to the right by an equal amount (assumed to be QR). But, in addition, there will be a tendency toward increased speculation in the forward market, expressed in a further shift in the NSP schedule which has no counterpart in the NLP schedule. At the original forward rate, P0, the increase in spot speculation is QR and the increase in forward speculation RS. The NLP line, therefore, shifts to the right, by QR, to NLP′, while the NSP line shifts to the right, by QR + RS = QS, to NSP′. At the original forward rate, therefore, there is an excess of the NSP over the NLP, which is equivalent to an excess supply of the domestic currency for forward delivery. The forward exchange rate then falls until the NSP is reduced and the NLP is increased to the point where the two are again equal. The new equilibrium is at Q′, where the new schedules intersect. At Q′, the forward rate (P1) is lower, and the balance of claims of residents on nonresidents (C1) has increased.

The results of other changes are easily discovered. An increase of interest rates at home relative to those abroad will tend to stimulate an inward capital movement because of the higher return both on a covered and on an uncovered basis. The tendency toward an uncovered capital movement will be reflected in a shift of both the NLP and the NSP curves by equal amounts, but the tendency to a shift in covered arbitrage will involve a further shift of the NLP schedule to the left. The result will be a decline both in the forward exchange rate and in the balance of claims.

Counterpart of Intervention

Official intervention in the forward exchange market can be analyzed by regarding the authorities either as part of the market or as distinct from it. If the NSP is defined exclusive of official intervention, then the new equilibrium requires not equality of the NSP and NLP but a difference between the two equal to the amount of the intervention itself. On the other hand, if the government position is included as part of the market, government purchases or sales of forward currency represent a change in speculation. We shall adopt the first alternative.

Let us continue to assume, initially, that neither the official intervention itself nor the associated change in the forward rate has any effect on expectations with regard to the future course of the exchange rate. Now consider the effect of forward intervention by A’s Government in support of A’s currency, i.e., a purchase of A’s currency forward. The NSP schedule as a whole will move to the left by the amount of the official intervention; the slope of that schedule will not alter, nor will the NLP schedule be affected in any way. The forward purchases of domestic currency by the authorities will create an excess demand for A’s currency, i.e., a tendency for the NSP to fall short of the NLP, which can be relieved only by an increase in price. The forward rate will therefore rise until the NLP has fallen and the NSP has risen, along their respective schedules, sufficiently to restore equality between the two. To put it another way, the rate must rise until the private market is willing to supply (against foreign currency) the extra amount of forward domestic currency demanded by the authorities.

The effect of the intervention is illustrated by Diagram 2. Starting from an initial equilibrium at Q, government intervention to the amount G = SQ (in the form of forward sales of foreign currency in exchange for domestic currency) gives rise to a new NSP curve, NSP’, which yields a new equilibrium at Q′. The counterpart of the official intervention can then be seen to be ST and TQ, the former representing an increase in unofficial net speculation against (or a reduction in unofficial net speculation in favor of) the domestic currency, and the latter a reduction in net claims (or an increase in net liabilities) of domestic residents vis-à-vis residents abroad.

It will be noticed that the inflow of funds arising from the intervention is necessarily less than the amount of the intervention. Moreover, the more inelastic is the NLP line (i.e., the less the international mobility of funds) and the more elastic the NSP line (i.e., the more assured are people’s anticipations about future exchange rates and the greater is their willingness to speculate on them), the less will be the proportion of funds attracted to the amount of the intervention.4

“Swap” Transactions and Extramarket Transactions

We have dealt with official intervention in the forward exchange market as if it took the form exclusively of outright forward transactions unaccompanied by any spot transaction. In fact, forward intervention may take the form of “swap” transactions which involve simultaneous operation in the spot and forward markets such as to leave the Net Speculative Position of the authorities unchanged. Thus, a forward sale of foreign (and a forward purchase of domestic) currency would be accompanied by a spot purchase of foreign (and a spot sale of domestic) currency.

As long as we adhere to the assumption that spot exchange rates are rigidly fixed, the “spot” element of the swap will have no effect on the forward rate, which will be affected, in the manner described above, exclusively by the forward element in the transaction. The effect of the “spot” element will depend on the mechanism through which the spot rate is held constant. For example, assume that the spot rate for dollars in terms of foreign currencies is held constant by the action of the U.S. authorities. If these authorities were to undertake a swap transaction, selling foreign currencies forward and buying them spot, they would immediately be obliged, in order to stabilize the rate, to sell these currencies spot for dollars, thus rendering the spot element in the swap null and void. But a more realistic assumption is that the spot rate is held constant by the foreign authorities who, however, could call upon the U.S. authorities to convert dollars into gold at a fixed rate. The effect of the U.S. spot purchase of foreign currencies then would be partly to reduce U.S. official holdings of gold and partly to increase foreign official holdings of dollars. The latter effect could be represented by a shift to the left of both the NLP and NSP curves of the United States by the amount of the increased foreign official holdings, leaving the forward rate unchanged.

Official swap transactions are frequently undertaken not on the open market but by direct arrangement with foreign monetary authorities or with commercial banks. Such operations are likely to differ from swaps on the open market in that they exercise a greater influence on the Net Loan Position, and little, if any, influence on the forward rate and on the Net Speculative Position.

In the case of direct intercentral bank or intergovernmental transactions where the foreign authorities refrain from passing on the swap to commercial banks or other private parties, there is no reason why the forward rate or the private Net Loan and Net Speculative Positions should be influenced at all. Suppose that the authorities of country A sell foreign currencies forward and buy them spot through spot transactions with the authorities of other countries. The loan position of the foreign authorities will be affected directly by the transaction, and the NLP curve will shift to the left by an amount equal to the intervention by the authorities in country A. The speculative position of the foreign authorities will not be directly affected by the transaction itself, but the NSP curve will shift to the left by the same amount as the NLP curve because of the change in the speculative position of the authorities in A. Indirectly, however, a part of the inflow of funds brought about by the transaction may be offset if the foreign authorities are induced by the swap transaction to reduce their uncovered reserve holdings in A currency below the level at which they would otherwise have maintained them.

Even if the foreign authorities pass on the swap to their commercial banks by direct negotiation at a forward rate (and premium over spot) more favorable than the rate prevailing in the market, they may be able to ensure that the balances in A currency thus acquired by the commercial banks on a covered basis are largely additional to those they would otherwise have held and covered at the market rate. If they were completely “additional,” there would be a leftward shift in the NLP curve precisely equal to the amount of the swaps. However, unless the foreign banks would otherwise have held no A currency on a covered basis, it is unlikely that this condition would be fulfilled and it would be difficult to avoid some decline in the amount of these holdings covered through the market.

Balance of Payments Effects

In considering the effects on the balance of payments of official intervention of the sort under discussion, we must distinguish nonrecurrent (“stock”) effects from continuous (“flow”) effects. Both types take time—say, a number of months—to manifest themselves fully, a fact which tends to blur the distinction between them. For sharpness of analysis, however, we shall assume that the “stock” effects act almost instantaneously.

The principal kinds of “stock” effect arise from once-for-all readjustments of asset and liability positions:

(a) To the extent that the official sales of forward exchange evoke a fall in the NLP, i.e., an inflow of foreign funds or a repatriation of domestic funds on a covered basis, there will be, at fixed exchange rates, a corresponding rise in gross reserves.

(b) On the other hand, the forward exchange liabilities acquired by the authorities may themselves be regarded as weakening the external liquidity of the country, as would any other short-term liability to foreign holders.

Now, as we have seen, the increase in official forward exchange liabilities must always exceed the induced net inward capital movement by an amount equal to the induced increase in private (forward) exchange speculation. The increase in gross reserves will therefore fall short of the increase in official forward exchange liabilities. Under our present assumption of fixed speculative anticipation, however, this fact has no importance from the standpoint of external liquidity. There would be an enhanced danger of a future drain on reserves only to the extent that there was an enhanced danger of a future outflow of funds on a covered basis. This danger, in turn, would arise only with respect to that part of the counterpart of the official intervention which consisted in an influx of funds on a covered basis, i.e., a decline in the NLP; and that part of the counterpart has been fully matched by an increase in reserves.

Moreover, on the assumption that the authorities are willing to renew their forward liabilities on maturity, such a withdrawal of funds on a covered basis would be likely to occur only if there were a loss of confidence in the willingness or ability of the authorities to fulfill these renewed contracts—a situation that seems remote.

For the reasons mentioned above, the portion of the official forward exchange liabilities that serves as counterpart to the adverse forward speculation induced by the rise in the forward exchange value of the domestic currency gives rise to little or no objective liquidity risk. If the proportions in which the official intervention evoked an addition to the unofficial Net Speculative Position and a diminution in the Net Loan Position, respectively, were exactly known, the proportion of official forward liabilities taken up by adverse speculation could simply be ignored. Since, however, the proportion is not exactly known, the authorities are likely to regard their reserve needs as enhanced by some fraction of their forward liabilities, irrespective of whether these in fact correspond to additional unofficial speculation or to inward capital movements. It therefore becomes important that the proportion of induced capital movement to induced speculation should, in fact, be as large as possible.

The “recurrent” or “flow” effects on the balance of payments of official support of the domestic currency in the forward market fall into three main categories:

(1) To the extent that domestic and foreign merchants carrying on the foreign trade of the country concerned are accustomed to cover on the forward market the exchange risks incidental to such trade, the appreciation of the forward value of the currency resulting from official support will tend to raise the price of exports to the foreign importer and lower the price of imports to the domestic importer. This will bring about a fall in exports and a rise in imports which, with normal foreign trade elasticities, will involve a deterioration in the balance of payments on current account.

(2) To the extent that the support of the forward exchange rate involves a decline in the NLP—and on the assumption that interest rates at home and abroad are kept constant—there will be a decline in receipt of interest from abroad and/or an increase in payments of interest to foreigners.

(3) The fact that forward exchange rates diverge from the spot rates as they ultimately turn out to be at the date of maturity of the forward contracts gives rise to profits and losses, some of which may enter into the balance of payments. To the extent that the sale of forward exchange by the authorities evokes additional purchases— whether by way of speculation or covered interest arbitrage—on the part of foreigners, rather than residents, the balance of payments will show a gain or a loss, depending on whether the forward value of the domestic currency at the supported level is below or above the spot value. Again if, apart from the official purchases, residents have a positive (negative) net position in foreign currencies on forward account,5 the rise in the forward value of the domestic currency will bring about an improvement (deterioration) in the balance of payments on current account equal to the extent of the appreciation times the net forward position.

Of the adverse effects on the current balance resulting from official support of the forward exchange value of the domestic currency, the effect on the balance of trade (category 1, above) is likely to be more important than the interest cost (category 2). Take the United States as an example: Suppose that an appreciation by 1 per cent a year of the forward dollar and forward dollar margin vis-à-vis all foreign currencies evokes a once-for-all inward flow of foreign and U.S. funds (decline in the NLP) of the order of $500 million, and that an appreciation of spot and forward dollars by 1 per cent reduces exports by 2½ per cent of $20 billion, i.e., $500 million a year, and increases imports by 1 per cent of $15 billion, i.e., $150 million a year, or $650 million a year in all. However, since an appreciation of 1 per cent a year on a 3-month forward contract represents an absolute appreciation of only ¼ per cent, this figure is cut to $160 million. Moreover, if we assume that only 60 per cent of U.S. trade is covered, the adverse effect on the U.S. balance of trade would be reduced to something like $100 million.

Even $100 million a year, however, is a considerable payments loss to incur in addition to the interest cost of, say, $20 million on an inward movement of arbitrage funds of the order of $500 million.6

Effect on Speculative Anticipations

We must now reconsider the assumption, earlier adopted, that official support of the forward exchange value of the currency has no effect on expectations as to the future course of the spot rate. The precise effect on such expectations is uncertain; there are possible effects in both directions. Much depends on the situation, and particularly on what is generally believed to be the extent of official intervention. Insofar as those participating in foreign exchange markets are ignorant of the fact or extent of official support operations, the impression given by the appreciation of the forward rate itself that other people are feeling less bearish about the currency is likely to impose a more optimistic view about the future spot value of the supported currency. Knowledge by some of the potential operators that the authorities are supporting the forward market will likewise tend to encourage bullishness insofar as it is interpreted as evidence of official determination to defend the currency and of official willingness to continue to have resort to this particular technique of increasing reserves. On the other hand, knowledge of the past accumulation of official forward liabilities as such will have a bearish or pessimistic influence, and the intervention as such may be interpreted as a sign of weakness. In considering the relative strength of these influences, it should be borne in mind that many of those whose expectations concerning rates are important (i.e., many potential “speculators”) are not market operators at all, but traders, who are likely to be more influenced by visible criteria, such as reserve movements, than by changes in official forward exchange liabilities, data on which are not published and the magnitude of which is unknown. Danger would arise only if the market were to assign an undeserved importance to that part of the official forward liabilities that corresponds to the induced adverse speculation, or if the total magnitude of forward liabilities were to be exaggerated by rumor.

On the whole, it appears that official support of the forward market is more likely to enhance than to reduce confidence in the future spot rate of the currency in question, particularly in the short run, and provided that the operations are not believed to be too large. The restoration of confidence will, of course, be merely temporary if the underlying deficit persists.

Insofar as official support of the forward market leads to more bullish expectations regarding the future spot rate, causing a shift in spot speculation, both the NLP and the NSP curves will lie further to the left than in Diagram 2; and insofar as official support affects forward speculation, it will result in a further shift to the left in the NSP’ curve. The reduction in the balance of claims (the favorable capital movement) will thus be greater, and the forward rate higher, than in Diagram 2.

Insofar as the official support of the forward market leads to less bullish, or more bearish, anticipations about the spot rate, the NLP curve will lie to the right and the NSP’ curve still further to the right, of the corresponding curves in Diagram 2. Even if the effect on forward speculation is insufficient to outweigh the initial official intervention, and the forward rate therefore rises above the initial position, the effect on spot speculation may be such that the balance of claims actually increases (i.e., the net effect on capital is unfavorable).

Official Forward Support with Floating Spot Rates

We have thus far been assuming that spot exchange rates are fixed by official action. In the context of the sort of arrangements that prevail in the world for keeping exchange rates in the vicinity of par values, this is equivalent to assuming that the balance of spot market transactions is such as to hold the spot exchange value of the domestic currency at rates which evoke the stabilizing intervention of domestic or foreign monetary authorities. At any one time, this is likely to be true of the spot exchange rates relative to some foreign currencies but not relative to others.

If the spot value of the domestic currency is not at an official buying or selling point, the effect of official support of the currency on the forward market will be different from what has been described. The consequences of forward intervention under a floating spot rate system would take too long to describe fully here. The following summary treatment is impressionistic rather than precise. The tendency, which exists when the spot rate is fixed, for the rise in the forward value of the currency to evoke a decline in the NLP will operate, under a floating spot rate, to promote a rise in the spot value of the currency. Thus, the effect of supporting the forward rate will be to raise both forward and spot rates of exchange.

If it is assumed that any short-run effect of the rise in the spot rate on the balance of current payments is of negligible proportions, the rise in the spot rate must, in fact, be sufficient to choke off entirely the tendency toward a decline in the NLP that results from the rise in the forward rate. This is accomplished in two ways: the rise in the spot value of the currency (given that anticipations regarding future spot rates remain the same) evokes adverse spot speculation; and the inward shift in covered interest arbitrage is reduced because the forward discount on the currency declines less with a floating than with a fixed spot rate. This effect on covered interest arbitrage means that the forward rate will have to rise more with a floating than with a fixed spot rate in order to evoke the addition to forward speculation required to equilibrate the forward market. In the end, the entire negative official speculative position in forward exchange will be balanced by an increase in private forward speculation.

Even if the spot value of the currency is free to appreciate only over a narrow range, the effect of this freedom will be to reduce the benefit which the country’s gross reserves will derive from the favorable effect on capital movements. Any inward arbitrage will be at least partially offset by outward uncovered capital movements, unless, indeed, the rise in the spot rate has the effect of raising the anticipated future level of the spot rate.

Conclusions

Official support of the foreign exchange value of a currency is in some respects analogous to official borrowing from abroad on a short-term but renewable basis, in foreign currency or accompanied by a foreign exchange guarantee. From the standpoint of the country that seeks to attract funds from abroad, such intervention in the forward exchange market differs from direct official borrowing chiefly in that it is less conspicuous, and engages less fully and less openly the credit of the authorities; on the other hand, it tends, in a way which may be unwelcome, to worsen the trade balance. From the standpoint of the lending countries, it has the characteristic, in contrast to intergovernmental or intercentral bank loans, of attracting private funds and thus reducing the liquidity of the commercial banking system. Quite apart from these comparative advantages or disadvantages of the technique, a country’s readiness to have recourse to such operations makes available a source of external liquidity which, though not large, is for the most part additional to those otherwise available. From what has been said it is clear that the most appropriate use of the technique will be to meet a temporary deficit which, if successfully financed, would be likely to be followed, within a period of time measured in months rather than years, by a corresponding surplus.

A good example would be an outflow of funds caused by a temporary loss of confidence in a currency that was basically sound, in the sense that in the long run it could be satisfactorily maintained without devaluation, though there might be some temporary weakness in the basic balance of payments. Such an outflow of funds—provoked, say, by some political event or by a misunderstanding of the policy intentions of the government—would be not only temporary but inherently reversible. While it lasted, it would probably be associated with a discount in the forward exchange value of the currency, which would be giving a temporary stimulus to the balance of trade. In these circumstances, official support of the forward market would offset the effect on reserves of the outflow of funds; the eventual reflux of these funds would enable the authorities to shed their forward liabilities; in the meantime, a large part of the speculative outflow could be prevented without raising the forward rate above the spot rate, i.e., without exchange loss to the authorities or even, probably, to the country. The stimulus to the balance of trade afforded by the forward discount would be removed; in certain circumstances, however, this may not be undesirable.

The next most promising occasion for applying the policy of forward support would be to deal with a short-term capital outflow stimulated by a relative decline in a country’s interest rates, likely to be followed at some not too distant date by a reflux attributable to a relative increase in the interest rates. Such a situation might arise if a country whose balance of payments was in fundamental equilibrium was suffering from a recession when other countries were prosperous.

In this case, if there were no concurrent loss of confidence in the currency, the outflow of funds, though draining the country’s reserves, would lead to an appreciation in the forward value of the currency, given the fact that at least part of the outflow was on a covered basis. Official support of the forward value of the currency, by reducing or even reversing the outflow of covered interest arbitrage, would reduce or check the over-all capital outflow. If, however, as might be the case, such support should create or accentuate a forward premium on the currency, the authorities would incur losses from their forward operations, partly to the benefit of foreign participants on the forward market. The substantial rise to be expected in the forward premium would, of course, have an adverse effect on the foreign balance, which might be unwelcome from a cyclical standpoint though it would probably merely involve a diminution in the improvement that would otherwise have occurred as a result of the recession.

Such use of forward support to counteract interest-motivated capital movements is unlikely to be as successful as in the case previously considered. In the first place, there will always be uncertainty concerning when the tide of capital movements will turn and whether it will turn to the full extent. The period over which intervention would have to be maintained might have to be long. Secondly, since no one would credit the possibility of revaluation in the circumstances envisaged, the rise in the forward premium would evoke a large amount of adverse forward speculation. This means that considerable intervention would have to be undertaken, and large losses incurred, in order to achieve a given effect on the movement of capital.

A temporary deficit in the current balance of payments that was due to some such event as a harvest failure would not constitute a good occasion for the application of the technique of forward support. The current account deficit resulting from this cause, while it would be temporary, would not be succeeded by any subsequent surplus in the balance of payments. The deficit in question could be offset, as far as the effect on reserves is concerned, by an influx of capital brought about by forward exchange support; but since there would be no ensuing surplus, the official debit position on forward exchange would have to be renewed indefinitely. This, as we have seen, involves some deterioration in the current balance as a consequence of the appreciation of the forward exchange rate. In any event, a permanent debit position on forward exchange would constitute a permanent weakness in the reserve position, so that it would be preferable that such a contingency should be met by long-term borrowing.

The only situation in which a case might be made for forward exchange support operations to meet a current deficit would be one in which it appeared probable that the current deficit would, within a reasonable time, be replaced by a surplus without the necessity for resorting to devaluation. If the turnaround in the current balance could be achieved only by devaluation, it would be preferable that the forward support not be undertaken until after devaluation has taken place—otherwise the authorities would incur a loss as a result of their operations, part of which would accrue in the form of windfall gains to nonresidents.

Analyse simplifiée de l’intervention officielle sur le marché de change à terme

Résumé

L’intérêt porté aux problèmes que pose l’intervention officielle sur les marchés de change à terme, intérêt qui allait croissant depuis un certain nombre d’années, a été encore renforcé par la nouvelle ligne de conduite récemment adoptée par les autorités monétaires des Etats-Unis. En mars 1961, le Trésor américain est intervenu pour la première fois sur les marchés de change pour soutenir le dollar à terme. En outre, le Federal Reserve System a conclu, depuis février 1962, plusieurs accords concernant les opérations sur devises avec d’autres banques centrales. Il semble donc opportun de considérer quels sont, en théorie, les effets à court et à long terme des transactions officielles sur le change à terme, et dans quelles conditions ces transactions peuvent être efficaces et utiles. Le présent mémoire s’efforce de répondre, au moins à titre provisoire, à diverses questions telles que les suivantes. L’intervention doit-elle être “limitée” quant à sa portée et à sa durée, et, dans l’affirmative, dans quel sens et pour quelles raisons doit-elle l’être? Dans quelles situations de paiements le soutien officiel du change à terme est-il opportun? Ce soutien doit-il se limiter à la defense de la monnaie contre les attaques “spéculatives”?

Ce mémoire envisage le sujet d’une façon simplifiée à dessein. Il a pour but d’exposer clairement et succinctement la théorie de I’intervention, et d’examiner à la lumière de cette théorie les problèmes qui viennent d’être évoqués. Dans ce but, une nouvelle classification des éléments qui déterminent le taux de change à terme est présentée; ce taux doit être tel qu’il assure l’équilibre entre d’une part la Position spéculative nette (qui par définition est égale à la somme des avoirs nets, au comptant ou à terme, des résidents d’un pays A dans d’autres monnaies, moins la somme des avoirs nets, au comptant ou à terme, des résidents d’autres pays en monnaie du pays A) et d’autre part la position nette en matière de prêts (qui par définition est égale à la somme des prêts nets des résidents du pays A aux résidents d’autres pays). Il semble que cette manière d’envisager le sujet offre certains avantages du point de vue de la clarté de l’exposition.

Análisis simplificado de la intervención oficial en el mercado de cambio a término

Resumen

El interés que han suscitado los problemas relativos a la intervención oficial en los mercados de cambio a término, el cual viene aumentando desde hace varios años, ha recibido un nuevo impulso con motivo de las recientes alteraciones efectuadas en las prácticas seguidas por las autoridades monetarias estadounidenses. En marzo de 1961, por vez primera, la Tesorería de Estados Unidos intervino en los mercados cambiarios con el objeto de prestar apoyo al dólar a término. Además, desde febrero de 1962 el Sistema de Reserva Federal ha concertado con otros bancos centrales varios acuerdos “swap.” Por lo tanto, parece oportuno reflexionar sobre cuales son, en teoría, los efectos a largo y a corto plazo de las transacciones del gobierno en divisas a término, y bajo qué circunstancias resulta probable que esas transacciones logren una finalidad útil. En este artículo se procura contestar tentativamente a preguntas como las siguientes: ¿Debe limitarse el alcance y/o la duración de la acción interventora y, de ser así, en qué sentido y por qué razón? ¿En qué género de situación de pagos resulta apropiado proporcionar apoyo oficial al cambio a término? ¿Debe la intervención limitarse a combatir los ataques de índole “especulativa”?

El enfoque empleado en este artículo es llano. Lo que se procura es presentar una exposición sin ambages de la teoría respecto a la intervención y aplicarla para examinar los problemas que acabamos de señalar. A ese efecto se elabora una nueva clasificación de aquellas fuerzas que determinan el tipo de cambio a término: es preciso que dicho tipo de cambio sea capaz de equilibrar la Posición Neta Especulativa (la cual se define como el total de los activos netos—ya sea al contado o a término—que posean los residentes del país A en otras monedas menos el total de los activos netos—al contado o a término— de los residentes de otros países en la moneda del país A) así como la Posición Neta de Préstamos Otorgados (que se define como el total de los préstamos netos concedidos por los residentes del país A a los residentes de otros países). Parece que dicho método de presentación reune ciertas ventajas de carácter ilustrativo.

*

Mr. Fleming, Assistant Director in the Research and Statistics Department, is a graduate of Edinburgh University. He was formerly a member of the League of Nations Secretariat, Deputy-Director of the Economic Section of the U.K. Cabinet Offices, U.K. representative on the Economic and Employment Commission of the United Nations, and Visiting Professor of Economics at Columbia University. He is the author of numerous articles in economic journals.

Mr. Mundell, Professor of Economics, McGill University, was formerly economist in the Special Studies Division of the Fund. He is the author of numerous articles on international trade and economic theory.

1

Charles A. Coombs, “Treasury and Federal Reserve Foreign Exchange Operations,” Federal Reserve Bulletin, September 1963, pp. 1216-23, and Board of Governors of the Federal Reserve System, Press Release, October 31, 1963.

2

For a mathematical exposition of the interrelationships determining forward exchange rates under more general assumptions, see S. C. Tsiang, “The Theory of Forward Exchange and Effects of Government Intervention on the Forward Exchange Market,” Staff Papers, Vol. VII (1959-60), pp. 75-106. See also, William H. White, “Interest Rate Differences, Forward Exchange Mechanism, and Scope for Short-Term Capital Movements,” Staff Papers, Vol. X (1963), pp. 485-503.

3

A change in the forward exchange rate is not, in fact, the only way in which an ex ante disequilibrium between NSP and NLP manifests itself. In the absence of a forward exchange market, such a disequilibrium would manifest itself in a change in the interest rate at which A residents lend to non-A residents (or vice versa) in A currency, compared with that at which they lend in non-A currency.

4
The ratio of the funds attracted to the amount of the intervention is uniquely determined by the elasticities of the NLP and NSP schedules. Specifically, the change in claims (C) as a fraction of a (small) government intervention (G) in support of the home currency is as follows:
CG=ηLηLηS
where ηL and ηS are, respectively, the elasticities of the NLP and NSP schedules. Since ηS > 0 and ηL < 0, an increase in G (intervention in support of the home currency) reduces C. But it is easily seen that the reduction in the balance of claims is always less than the amount of the intervention; CG is always a (negative) fraction.
5

This implies that foreign residents would have had a corresponding negative (positive) net position in domestic currency on forward account.

6

From the standpoint of real national income, the two figures are on a very different footing. The loss of $20 million in interest is a net loss to the country even if full employment is preserved—though it may be “worthwhile” if the $500 million inflow of funds is added to the real capital of the country. The $100 million loss on the trade balance may involve no loss at all in real income if employment is preserved by a corresponding addition to domestic expenditure. Indeed, there may be a net gain from the improved terms of trade. However, both the $20 million and the $100 million are apt to generate declines in income and employment which may intensify the real loss involved. In most cases in which official intervention on the forward market in support of the domestic currency arises, a decline in competitiveness on world markets would be unwelcome.