A number of countries have adopted dual foreign exchange markets to combat disruptive capital flows. Under such arrangements, commercial transactions typically pass through an officially supported market, and financial transactions through a free market. It has been pointed out in the literature, especially by Fleming,¹ that the separation of the foreign exchange market into two tiers by exchange controls provides an extra policy instrument. It allows net capital flows to be controlled independently of the level of domestic interest rates by discretionary official operations in the free financial market. In particular, a dual market can enable a country to maintain overall payments balance, at least in the short run, if the foreign currency sold (purchased) through support of the commercial rate is purchased (sold) in the free financial market.

In another paper,² I have shown that net capital flows can be controlled under a unitary exchange market if official support is given exclusively to the forward exchange rate and is accompanied by discretionary official operations in a free spot market. An extra policy instrument is already available in a unitary market by the separation of the foreign exchange market according to delivery periods. A supported unitary forward rate achieves the same objective as a supported commercial rate—it provides a stable exchange market accessible to traders. A free unitary spot market achieves the same benefits as a free financial market—it not only prevents speculative short-term capital movements from influencing the reserves and money supply but also provides a market in which official open market operations can be undertaken to control net capital flows by influencing the spread between the supported and free exchange rates rather than by interest rate movements.

The apparent similarity between the potential benefits of a dual market and exclusive forward exchange rate support in a unitary market gives rise to a number of interesting questions. First, if the separation of a unitary exchange market by maturity periods already enables the authorities to control both the exchange rate and reserves, will not the further separation of the exchange market by exchange controls provide additional policy instruments not recognized to date in the literature? Second, since under both systems the possibility exists that individuals may attempt to profit by switching between the supported and free markets, are the resulting official losses under a dual market likely to be less than under a unitary market? If so, will they offset the administrative costs of operating a dual market? If not, will they be offset by any additional benefits achieved by a dual market?

This paper attempts to answer the preceding questions. The paper assumes that the target of official intervention in the free market is the insulation of the money supply by the inducement of capital inflows (outflows) equal to the trade deficit (surplus). Alternative targets of official intervention in the free market are not examined. Section I presents the notational and definitional framework for examining official intervention policy in a multiple-tier, multiple-maturity exchange market. Sections II, III, and IV examine and compare a dual market with official intervention in spot markets only and a unitary market with official intervention in both spot and forward markets. Trade payments flows are assumed to be exogenous until Section V, which compares the impact of leads and lags under the two systems. Section VI examines official intervention policy in additional exchange markets. Finally, Section VII sets forth the problems of evasion in a dual market and purely speculative forward market operations in a unitary market, and Section VIII presents the conclusion.

I. Notation and Definitions

This section presents a notational and definitional framework for examining official intervention policy in a multiple-tier, multiple-maturity exchange market. In later sections, the framework is applied to the simple cases of: (1) a dual market with official intervention confined to spot markets; (2) a unitary market with official intervention in both spot and forward markets; and (3) a dual market with official intervention in both spot and forward markets.

The notation utilizes subscripts to distinguish three characteristics of foreign exchange contracts. They are the maturity period of the contract, the date of the contract, and the tier of the market in which the contract is made. Let D_{i,τ, j} (S_{i,τ, j}) equal the total ex post demand for (supply of) the domestic currency for delivery after i days contracted on day τ in the jth tier exchange market.

Contracts to buy and sell the domestic currency may be classified as being due to current account transactions (V), capital account transactions (C), or official intervention in the foreign exchange market (G). Let the superscript d represent demand, s represent supply, d—s represent net demand, and s − d represent net supply.

Substituting equations (2) and (3) in equation (1):

The combined ex post demand for domestic currency by traders and capital transactors for any maturity at any date and in any tier of the exchange market must equal the ex post supply by the authorities.

Let day T represent the current period. Define the current trade payments balance (TB) as the sum of all maturing current account purchases (+) and sales (−) of domestic currency. It therefore equals the sum over all tiers of the exchange market of all i-day maturity current account transactions contracted i days ago, that is, contracted on day T − i.

The balance sheet of private capital transactors is described by two magnitudes: their spot domestic currency position and their open domestic currency position. Foreign exchange contracts influence open positions immediately they are contracted, but influence spot positions only when they mature. For example, a three-month forward contract to purchase the domestic currency immediately increases the sum of all claims denominated in the domestic currency, that is, it immediately lengthens the open domestic currency position. On maturity, three months later, the forward claim is transformed into a spot claim. The open position remains unchanged, but the spot position is lengthened. Define the current change in the spot domestic currency position of private capital transactors (ΔSDCP) as the sum of all their currently maturing purchases (+) and sales (−) of domestic currency; define the current change in their open domestic currency position (ΔODCP) as the sum of all their currently contracted purchases (+) and sales (−) of domestic currency, both spot and forward

The balance sheet of the government is described by three magnitudes: the spot domestic currency position, that is, the money supply; the spot foreign currency position, that is, the reserves; and the open foreign currency position. Define the current change in the money supply owing to the balance of payments (ΔM) as the sum of all currently maturing official sales (+) and purchases (−) of domestic currency; define the current change in the reserves in units of foreign currency (ΔR) as the sum of all currently maturing official purchases (+) and sales (−) of foreign currency; define the current change in the official open foreign currency position in units of foreign currency (ΔOFCG) as the sum of all currently contracted official purchases (+) and sales (−) of foreign currency, both spot and forward.

where X_{i,τ, j} equals the price of one unit of domestic currency in terms of foreign currency for delivery after i days prevailing on day τ in the jth tier exchange market.

II. Dual Exchange Market with Official Intervention in Spot Markets Only

This section examines the impact of official intervention policy in a dual exchange market in which, although forward markets are not precluded, official intervention is confined to spot markets only. The procedure is, first, to define the nature of the dual market; second, to derive the magnitude of official intervention in the two tiers of the market necessary to maintain both an officially supported exchange rate accessible to traders and immunity of the money supply; third, to derive the ex post impact of the policy on the portfolios of private capital transactors and the government; and, last, to examine the behavioral relationships on which the unexplained variables depend.

Let it be assumed that:

(1) The exchange market is divided into two tiers where the subscript j = 1 denotes the “commercial market” and the subscript j = 2 denotes the “financial market.”

(2) Only current account transactions pass through the commercial market, that is, capital account transactions in the commercial market are zero.

(3) Only capital account transactions pass through the financial market, that is, current account transactions in the financial market are zero.

(4) Official intervention takes place in spot markets only, that is, official intervention in markets for other than immediate delivery is zero.

Such a system is defined by:

The first objective of official intervention in the foreign exchange market is assumed to be the provision of a stable exchange market accessible to traders. Substituting equations (4) and (11) to (14) in equation (5), official sales of domestic currency in the commercial spot market equal

Maintaining an officially supported commercial spot rate requires official sales of domestic currency in the commercial spot market equal to the resulting level of the trade payments balance.

The second objective of official intervention in the foreign exchange market is assumed to be the insulation of the money supply by the inducement of capital inflows (outflows) equal to the trade deficit (surplus). The intervention policy necessary to achieve immunity of the money supply is obtained by making the change in the money supply—equation (8)—equal to zero and substituting therein equations (11), (14), and (15).

Immunity of the money supply requires official purchases of domestic currency in the financial spot market equal to the value of official spot sales of domestic currency necessary to support the commercial spot rate, that is, equal to the trade payments balance.

The ex post impact of such a policy on the balance sheets of private capital transactors and the government is obtained by substituting from equations (4), (11) to (14), and (16) in equations (6) to (10).³

Equation (17) shows that the intervention policy necessarily induces a spot capital inflow equal to the trade payments deficit. Equation (18) shows that capital transactors are necessarily induced to lengthen their open domestic currency position by the value of the trade payments deficit. Together, equations (17) and (18) imply that capital transactors in aggregate undertake uncovered capital inflows equal to the trade payments deficit. Even though the existence of a financial forward market would mean that individual capital transactors could undertake covered capital inflows, nonintervention by the authorities in the financial forward market means that any covered capital inflows must be accommodated by outright speculative forward sales by other private individuals. In aggregate, net forward operations by private capital transactors must be zero, and capital flows may be treated as being uncovered.

Equation (19) shows that a gain or loss to the reserves is incurred equal to the difference between the commercial and financial spot rates multiplied by the trade payments balance, that is, multiplied by the magnitude of official sales (purchases) of domestic currency in the commercial (financial) market. Equation (20) shows that, except for this official gain or loss owing to exchange transactions, the intervention policy immunizes the official open foreign currency position from the trade payments deficit.

The results obtained so far have been ex post. No behavioral relationships have been examined. Assuming that the trade payments balance resulting from the level of the supported commercial spot rate is determined exogenously, the only unexplained variable present in the preceding results is the financial spot rate; the factor that must explain its level is the behavior of uncovered capital flows.

Uncovered capital flows are subject to exchange risk. Therefore, assuming that private capital transactors are averse to risk, the inducement of an increase in the quantity of uncovered funds held in the domestic currency will require an increase in the expected rate of return on such funds. Official sales of domestic currency in the free financial market will depreciate the financial rate sufficiently to secure the increase in expected return necessary to induce the required uncovered capital inflows. The actual depreciation of the financial rate will depend on the impact of a depreciation on the expected return and on the confidence with which expectations are held.

A depreciation in the financial rate has two effects on the expected return from converting funds into the domestic currency on an uncovered basis.⁴ First, assuming that interest earnings pass through the commercial market, a depreciation of 1 per cent in the financial rate will increase the expected interest earnings by approximately 1 per cent multiplied by the domestic interest rate, provided that expectations concerning the future commercial rate remain unchanged. Being the product of two percentages, this rate of interest effect may be assumed to be insignificant except when both the rate of interest and the depreciation in the financial rate are large percentages. Second, a depreciation of 1 per cent in the financial rate affects the expected capital gain by a magnitude that depends on the elasticity of the expected future financial rate with respect to the current financial rate. If the elasticity is zero, that is, expectations are unaffected by the movement in the current financial rate, a depreciation of 1 per cent in the financial rate will increase the expected capital gain by approximately 1 percentage point; if the elasticity is one, that is, a depreciation of 1 per cent is accompanied by a reduction of 1 per cent in the expected future financial rate, the expected capital gain will remain approximately unchanged; if the elasticity is greater than one, that is, expectations are extrapolative with a depreciation leading to the expectation of a further depreciation, a depreciation of 1 per cent will reduce the expected capital gain and give rise to local instability.

Official sales of domestic currency in the financial market will therefore depreciate the financial rate by a magnitude that depends on the elasticity of expectations. The lower is the elasticity, the lower will be the depreciation in the financial rate necessary to secure a given increase in the expected total return on uncovered funds; the lower is the certainty with which expectations are actually held, the greater will be the increase in the expected total return necessary to induce the required uncovered capital flows. Except when expectations are held with certainty, a rising expected return will be necessary to induce a continuing inflow of funds to finance a trade payments deficit. Given that the official loss owing to exchange transactions depends on the difference between the financial and commercial rates, the inducement of a continuing inflow of funds will result in a growing (declining) official loss (gain) on exchange transactions.

III. Unitary Exchange Market with Official Intervention in Spot and Forward Markets

This section examines the impact of official intervention policy in a unitary exchange market in which official intervention is confined to the spot market and a single forward market. The procedure is similar to that in Section II: first, the nature of the unitary market is defined; second, the required magnitudes of official intervention in the forward and spot markets are derived; third, the ex post impacts of the policy on the portfolios of private capital transactors and the government are derived; and fourth, the behavioral relationships on which the unexplained variables depend are examined.

Let it be assumed that:

(1) The exchange market contains only a single tier.

(2) Official intervention is confined to the spot market and the t-day forward market, that is, official intervention is zero in markets for maturities other than i = 0 and i = t.

(3) To abstract from the problems of leads and lags that are considered in Section V, the trade payments balance is fully reflected in traders’ demand in the t-day forward market, that is, the net demand for domestic currency by traders is zero for maturities of other than t days.

(4) To abstract from problems arising from historical operations by private capital transactors, the net demand for domestic currency in the past, that is, on days τ > T, by private capital transactors is zero.

Such a system is defined by:

The official objective of providing a stable exchange market accessible to traders is achieved by official support of the t-day forward market. Substituting equations (4) and (21) to (24) in equation (5), official sales of domestic currency in the t-day forward market on day T − t can be obtained. As there is no need to distinguish between separate tiers of the exchange market, the j subscript is omitted.

Maintaining an officially supported forward rate requires official forward sales of domestic currency on day T − t equal to the current trade payments balance.

The intervention policy to achieve immunity of the money supply is obtained by making the change in the money supply—equation (8)—equal to zero and substituting therein equations (21), (22), and (25).

Immunity of the money supply requires official purchases of domestic currency in the spot market equal to maturing official forward sales of domestic currency, that is, equal to the trade payments balance.

The impact of such a policy on the balance sheets of private capital transactors and the government is obtained by substituting from equations (4) and (21) to (26) in equations (6) to (10).

Equation (27) shows that a capital inflow equal to the trade payments deficit is necessarily induced. Equation (28) indicates that the capital flow may be uncovered or covered, depending on the magnitude of current forward purchases of domestic currency by private capital transactors. Equation (29) shows that a gain or loss to the reserves is incurred equal to the difference between the past forward rate and the current spot rate multiplied by the trade payments balance, that is, multiplied by the magnitude of official sales (purchases) of domestic currency in the past forward (current spot) market. Equation (30) shows that the change in the official open foreign currency position is approximately equal to the change in the reserves plus the magnitude of current forward purchases of domestic currency by private capital transactors.⁵

Two unexplained variables exist in the foregoing results. They are the current spot rate and the magnitude of current forward purchases of domestic currency by private capital transactors. Since purely speculative forward operations are examined in Section VII, it is assumed here that they are zero. Current forward operations by private capital transactors are therefore assumed to represent only exchange risk covering operations.

In the absence of default risk, borrowing foreign currency for t days and relending it in the domestic currency on a covered basis also for t days is riskless. Whenever the covered interest rate differential is nonzero, such covered interest arbitrage operations yield a riskless profit, and uncovered capital flows will therefore be inefficient. With an unchanged officially supported forward rate, official operations in the spot market produce a tendency for the covered interest rate differential to change. Official purchases of domestic currency in the spot market tend to appreciate the spot rate, decrease the implicit interest rate of the forward premium, and therefore worsen the covered interest rate differential. Similarly, official sales of domestic currency will tend to improve the covered interest rate differential. In the absence of default risk, it must be assumed that such operations induce covered capital flows, that is, the change in the open domestic currency position of private capital transactors—equation (28)—is zero, and that the relationship between the spot and forward rates is maintained in the vicinity of interest parity. Such assumptions enable the unexplained variables to be determined:

When all capital flows are covered, the change in the official open foreign currency position approximately equals the gain or loss to the reserves plus the foreign currency value of the trade payments balance. Official reserve losses incurred by the operation of the unitary market system and deficits in the trade payments balance both shorten the official open foreign currency position.

Given interest parity, the current spot rate is determined by the current level of the supported forward rate and the current interest rate differential. If the allowed rate of depreciation of the supported forward rate is equal to the interest rate differential in favor of the domestic currency, the current spot rate will equal the past forward rate, and official losses to the reserves will be eliminated.

IV. Comparisons Between the Dual and Unitary Market Systems

The results of the previous analysis indicate important differences between the dual and unitary market systems.

(1) Under both systems, a capital inflow equal to the trade payments deficit is induced. However, under a dual market with official intervention confined to spot markets, the capital inflow is necessarily uncovered; under the unitary market system, the capital inflow is likely to be covered. Under normal assumptions, uncovered capital flows are more risky than covered capital flows.⁶ Under not unreasonable assumptions—namely, the absence of default risk—covered capital flows are riskless. Other things being equal, a smaller change in expected return is necessary to increase the size of covered balances than is necessary to increase the size of uncovered balances. Furthermore, a larger depreciation of the unsupported rate is likely to be necessary to provide a given increase in expected return under the dual market than under the unitary market. Unless the elasticity of expectations of the future financial rate with respect to the current financial rate is zero, a depreciation of 1 per cent in the financial rate will increase expected return by less than 1 percentage point. An elasticity greater than one produces instability. Under the unitary market system, on the other hand, a depreciation of 1 per cent in the spot rate leads unequivocally to an increase of 1 percentage point in the forward premium and, therefore, to an increase of 1 percentage point in the profitability of covered interest arbitrage. The lower risk of covered rather than uncovered capital flows and the greater impact of a depreciation of the free rate on expected return under the unitary market system rather than the dual market system combine to imply that smaller changes in the free rate are necessary to induce capital flows under the unitary market system than under the dual market system.

(2) Under the dual market system, the relationship between the financial and commercial rates is determined by the behavior of uncovered capital flows, with expectations playing the crucial role. When the trade balance is weak, expectations are likely to be such that the financial rate will stand at a discount to the supported commercial rate. Under the unitary market system, on the other hand, the relationship between the spot and forward exchange rates is determined by the behavior of covered capital flows, the forward premium reflecting the interest rate differential against the domestic currency. When a weak trade balance is caused by a relatively high rate of monetary expansion that is reflected in relatively high domestic interest rates, a forward discount is likely to exist, that is, the spot rate is likely to stand at a premium over the forward rate. The contrast emerges that, when the trade balance is weak, the free exchange rate is likely to stand at a discount to the supported exchange rate under the dual market but at a premium under the unitary market. The implications of this are that official sales (purchases) of domestic currency (reserves) in the free market to finance a weak trade balance are likely (a) to result in official losses under the dual market but in official gains under the unitary market and (b) to pry the exchange rates apart under the dual market but to tend to bring them closer together under the unitary market.

(3) Total official gains or losses include not only the exchange transactions gains or losses to the reserves caused by differences between the free and supported exchange rates but also revaluation profits and losses. The magnitude of the profit owing to a revaluation of the reserves (devaluation of the domestic currency) depends on the official open foreign currency position. Equations (20) and (33) show that, under the dual market, the official open foreign currency position is immunized except for the gains or losses to the reserves; the authorities prevent a trade deficit from shortening the official open foreign currency position by inducing private capital transactors to lengthen their open domestic currency positions. Under the unitary market, the official open foreign currency position reflects the trade balance as well as gains or losses to the reserves. A depreciation of the domestic currency following a period of deficit in the trade balance will therefore result in a bigger revaluation profit under the dual market.

V. Leads and Lags in Trade Payments

The previous analysis has assumed that trade payments flows are exogenous. This section examines the impact of leads and lags in trade payments on the dual and unitary market systems. It is found that: (1) changes by traders in their exposure to exchange risk are fully reflected in the trade payments balance under the dual market system but are less than fully reflected in the trade payments balance under the unitary market system; (2) changes in the trade payments balance owing to leads and lags have a bigger impact on the spread between the supported and free rates under the dual market system; (3) changes in the spread between the supported and free rates have a destabilizing influence on the trade payments balance under the dual market system, but a stabilizing influence under the unitary market system.

(1) A trader can alter his exposure to exchange risk either by changing his covering operations in the forward market or by changing the timing of his spot foreign exchange contracts.⁷ In a dual market with official intervention in spot markets only, traders are the only operators in the commercial forward market. They are, therefore, unable in aggregate to alter their exposure to exchange risk through operations in the forward market. Their only means of altering this exposure is by changing the timing of their spot foreign exchange contracts.⁸ Such changes are immediately and fully reflected in the trade payments balance.

In contrast, under the unitary market system traders in aggregate are able to alter their exposure to exchange risk by operations in the forward market. Changes in the value of trade contracts covered by operations in the forward market rather than met by operations in the future spot market have no effect, however, on the trade payments balance, that is, they have no effect on the maturity date of traders’ foreign exchange contracts. Furthermore, such changes have no effect on the spread between the forward and spot rates. Increased forward sales of domestic currency to the authorities by traders increase the required official sales of domestic currency in the future spot market necessary to immunize the money supply, but they also correspondingly reduce the need for traders to meet their trade contracts by future spot sales of domestic currency. The total net demand for spot domestic currency by traders and the authorities is unaffected under the unitary market system by the proportion of trade contracts covered in the forward market rather than met by future spot contracts.

The trade payments balance is affected under the unitary market system by changes in the timing of traders’ operations in the spot market. Given that the spot rate is free, however, such changes automatically induce offsetting capital flows. To the extent that traders change their forward covering operations rather than change the timing of their spot foreign exchange contracts, a given change in the exposure of traders to exchange risk will be less than fully reflected in the trade payments balance.

(2) Under both the dual and unitary market systems, official intervention policy ensures that changes in the trade payments balance are accompanied automatically by offsetting changes in capital flows. Under the dual market system, the capital flow is uncovered; under the unitary market system, it is likely to be covered. When a change in the trade payments balance is caused by leads and lags, the distinction as to whether the capital flow is uncovered or covered is especially important. A worsening in the trade payments balance owing to leads and lags reflects (a) the hedging operations of traders who lead their sales of domestic currency in order to avoid the risk of a depreciation of the domestic currency; and/or (b) the speculative operations of traders who lag their purchases of domestic currency in the hope of a depreciation. At a time when traders are hedging against the risk of a depreciation, it is likely that capital transactors will also wish to cover the risk of a depreciation; at a time when traders are speculating against the domestic currency, it is unlikely that capital transactors will wish to speculate in favor of the domestic currency. Under the dual market system, capital transactors are unable in aggregate to cover the exchange risk involved in holding the domestic currency; under the unitary market system, they are able in aggregate to obtain cover. Under the dual market system, the spread between the supported and free rates must widen sufficiently to induce capital transactors to move into the currency on an uncovered basis, that is, to induce speculation in favor of the currency by capital transactors when traders are speculating against the currency; under the unitary market system, the spread must widen sufficiently to induce offsetting covered interest arbitrage flows. A danger of instability lies in the dual market system. Owing to leads and lags, the need for the authorities to sell domestic currency in the financial market will increase at a time when the expectations of capital transactors are deteriorating and the financial rate is depreciating. No danger of instability is present under the unitary market system. The return on covered interest arbitrage depends in no way on expectations concerning exchange rate changes.

(3) In addition to the fact that autonomous changes in the trade payments balance have an effect on the spread between the supported and free rates under the dual and unitary market systems, changes in the spread are themselves likely to have an effect on the trade payments balance.

In a dual market, the return to traders—including the profitability of leads and lags—depends only on the current and future commercial rates. The financial rate will affect the operations of traders only to the extent that it influences their expectations concerning the future commercial rate. In fact, the financial rate is likely to be taken by traders as a prime indicator of the future movement of the commercial rate. Given that a depreciation of the financial rate is likely to lead to a worsening in expectations concerning the future commercial rate, it will encourage traders to lead their spot sales, and to lag their spot purchases, of domestic currency. Physical export flows may be postponed and import flows brought forward by the prospect of a depreciation of the commercial rate. The worsening in the trade payments balance that is likely to accompany a depreciation of the financial rate will mean that any depreciation of the financial rate that accompanies official sales of domestic currency in the financial market will increase the scale of official purchases of domestic currency necessary in the commercial market and, as a result, will require further official sales of domestic currency in the financial market. If capital inflows are less responsive to a depreciation of the financial rate than is the trade payments deficit, instability will result.

In a unitary market, the return to traders—including the profitability of leads and lags—depends on both the spot and forward exchange rates. For trade flows that are normally met by spot foreign exchange contracts, a depreciation of the spot rate will provide a stimulus to exports and a deterrent to imports. Furthermore, a depreciation of the spot rate is likely to improve the trade payments balance owing to leads and lags. Forward sales of domestic currency (foreign currency) are encouraged (discouraged) as an alternative to advanced spot operations as the means of hedging exchange risk; stabilizing trader arbitrage ⁹ is encouraged by the improvement in the covered interest rate differential resulting from the depreciation of the spot rate. The improvement in the trade payments balance that is likely to accompany a depreciation of the spot rate is a stabilizing factor in the operation of the unitary market system.

VI. Official Intervention in Additional Exchange Markets

The paper has so far confined its attention to official intervention in two exchange markets separated by either exchange controls or maturity period. In both a dual and a unitary exchange market, however, official intervention may take place in more than two separate exchange markets. Under a dual market, official intervention may take place in forward markets as well as spot markets; under a unitary market, official intervention may take place in additional forward markets. This section investigates whether additional targets are attainable by official intervention in more than two exchange markets. It examines: (1) forward intervention in the commercial market under the dual market system; (2) forward intervention in the financial market under the dual market system; (3) forward intervention in additional forward markets under the unitary market system.

It is found that, for as long as spot and forward markets are linked by perfectly elastic covered interest arbitrage flows, no additional targets are attainable under the unitary market. Under the dual market, however, the authorities are permitted the dubious additional benefit of being able to speculate against their own currency while simultaneously supporting the commercial rate and immunizing the money supply.

VII. Evasion and Forward Market Speculation

The paper has so far abstracted from a problem that many may consider a serious drawback to both the dual and unitary market systems. The problem is that of individuals attempting to profit by switching between the supported and free markets. Under the dual market system, there is the danger that official intervention in the financial market will encourage individuals to disguise commercial transactions for financial transactions, and vice versa. Under the unitary market system, there is the danger that official support of the forward rate will encourage individuals to buy currency in the forward market in the hope of being able to sell it at a profit in the future spot market. This section investigates the problems of evasion under the dual market system and forward market speculation under the unitary market system. The section does not consider, however, the administrative controls that are used to separate the foreign exchange market into two tiers or the various methods by which evasion can take place in a dual market.¹⁰ Rather, it simply considers one method of evasion—the underinvoicing or overinvoicing of trade flows—as representive of the means whereby commercial transactions can be channeled through the financial market, and vice versa. This section examines, first, the impact of evasion and forward market speculation on the spread between the free and supported rates, and, second, the profitability and magnitude of evasion and forward market speculation.

VIII. Conclusion

This paper has compared two mechanisms of official intervention in the foreign exchange market that provide a stable exchange market accessible to traders while achieving immunity of the money supply by the inducement of capital inflows equal to the trade deficit. The first policy is as follows: the establishment of a dual market; official support given exclusively to the commercial market; and official sales (purchases) of domestic currency resulting from support operations in the commercial market offset by official purchases (sales) of domestic currency in the free financial market. The second policy is as follows: exclusive forward exchange rate support in a unitary market; a free spot exchange rate; and official forward sales (purchases) of domestic currency resulting from support operations in the forward market offset, on maturity, by official purchases (sales) of domestic currency in the free spot market.

The major differences found between the dual market system with official intervention confined to spot markets and the unitary market system are as follows:

(1) Uncovered capital flows are induced under the dual market system; covered capital flows are induced under the unitary market system.

(2) Larger changes in the spread between the free and supported exchange rates are necessary to induce capital flows under the dual market system than under the unitary market system.

(3) Under the dual market system, the official operations in the financial market will normally push the exchange rates apart, resulting in official losses on exchange transactions; under the unitary market system, the official operations in the spot market will normally pull the exchange rates together, resulting in official profits on exchange transactions.

(4) Larger foreign currency revaluation profits will be reaped under the dual market system than under the unitary market system if the domestic currency is allowed to depreciate.

(5) Changes in the spread are likely to have a destabilizing effect on leads and lags under the dual market system, but a stabilizing effect under the unitary market system. Instability will result under the dual market system if capital inflows are less responsive to a depreciation in the financial rate than is the worsening in the trade payments balance owing to leads and lags.

(6) Under the dual market system, the profitability of evasion depends on the current spread between the free and supported rates. Evasion is likely to become an increasingly serious problem as the spread increases. Under the unitary market system, the profitability of forward market speculation does not depend on the current spread but rather on the future unsupported spot rate.

(7) To deter forward market speculation under the unitary market system would require administrative controls to monitor only forward contracts; to deter evasion under the dual market system would require more extensive administrative controls that monitored both spot and forward contracts.

Emerging from the findings of the paper is a fundamental difference between the dual and unitary market systems. The dual market system relies on the inducement of speculative, that is, uncovered, capital flows but is threatened by arbitrage between the supported and free markets. In contrast, the unitary market system relies on the inducement of arbitrage capital flows but is threatened by speculation between the supported and free markets. Given that individuals—especially banks—are averse to risk, a system that relies on riskless operations and is threatened by risky operations will work more effectively than a system that relies on risky operations and is threatened by riskless operations.

Although a number of countries have adopted dual markets, they have been reluctant to undertake large-scale operations in the free financial market. The large changes in the financial rate that may be necessary to induce capital inflows, together with the resulting detrimental effects on both leads and lags and evasion, are the probable reasons. It has been shown in this paper, however, that these problems can be lessened by forward intervention policy in the dual market. Destabilizing leads and lags can be discouraged by forward intervention in the commercial market; changes in the financial rate can be minimized by undertaking swaps in the financial market rather than by outright spot operations. When such problems are overcome by forward intervention policy, however, the dual market system becomes similar to the unitary market system. The only advantage to such a dual market system is that speculators would still be operating in a free exchange market rather than in a supported exchange market. The cost of this advantage is the administrative cost of separating the markets. Given that the administrative costs of preventing purely speculative forward operations in a unitary market would be less than the administrative costs of a dual market, the unitary market system can achieve the same results as the dual market system but at lower cost.