The European system of narrower exchange rate margins, commonly referred to as the snake, went into effect in April 1972. While the original plan, embodied in the March 1971 resolution of the Council of the European Communities, was for participating central banks to support snake parities through coordinated dollar intervention, this was abandoned in favor of intervention in snake currencies in the wake of the Smithsonian agreement of December 1971. Because this agreement widened the permissible band of fluctuations around parity for all currencies, the margin of discretion over snake-dollar exchange rates was accordingly enlarged. In this context, it was thought that coordinated dollar intervention in support of intrasnake parities might increase the difficulties associated with the determination of an appropriate exchange rate between the snake and the dollar. Circumventing this potential source of discord, the Basle agreement of April 1972 relegated the valuation of the dollar to the marketplace, by substituting snake currency intervention for dollar intervention as the means for maintaining snake parities.1 Moreover, the significance of this issue increased dramatically in March 1973 with the onset of generalized floating.

Abstract

The European system of narrower exchange rate margins, commonly referred to as the snake, went into effect in April 1972. While the original plan, embodied in the March 1971 resolution of the Council of the European Communities, was for participating central banks to support snake parities through coordinated dollar intervention, this was abandoned in favor of intervention in snake currencies in the wake of the Smithsonian agreement of December 1971. Because this agreement widened the permissible band of fluctuations around parity for all currencies, the margin of discretion over snake-dollar exchange rates was accordingly enlarged. In this context, it was thought that coordinated dollar intervention in support of intrasnake parities might increase the difficulties associated with the determination of an appropriate exchange rate between the snake and the dollar. Circumventing this potential source of discord, the Basle agreement of April 1972 relegated the valuation of the dollar to the marketplace, by substituting snake currency intervention for dollar intervention as the means for maintaining snake parities.1 Moreover, the significance of this issue increased dramatically in March 1973 with the onset of generalized floating.

The European system of narrower exchange rate margins, commonly referred to as the snake, went into effect in April 1972. While the original plan, embodied in the March 1971 resolution of the Council of the European Communities, was for participating central banks to support snake parities through coordinated dollar intervention, this was abandoned in favor of intervention in snake currencies in the wake of the Smithsonian agreement of December 1971. Because this agreement widened the permissible band of fluctuations around parity for all currencies, the margin of discretion over snake-dollar exchange rates was accordingly enlarged. In this context, it was thought that coordinated dollar intervention in support of intrasnake parities might increase the difficulties associated with the determination of an appropriate exchange rate between the snake and the dollar. Circumventing this potential source of discord, the Basle agreement of April 1972 relegated the valuation of the dollar to the marketplace, by substituting snake currency intervention for dollar intervention as the means for maintaining snake parities.1 Moreover, the significance of this issue increased dramatically in March 1973 with the onset of generalized floating.

Recently, however, intervention in U. S. dollars has been used increasingly by European central banks. For our purposes, two points are particularly relevant. First, the dollar intervention is intended primarily to smooth out short-run fluctuations in dollar rates, leaving long-run valuation of the dollar to the market. And, second, the current utilization of dollar intervention is designed mainly to counter shocks experienced directly in the dollar market. Snake currency intervention is still used to maintain intrasnake parities in the face of disturbances affecting only snake currencies. Nevertheless, this loosening of the Community’s commitment to exclusive reliance on the free market mechanism for the valuation of the dollar reopens the question of discretion over the dollar rate and raises the attendant issue of resolving intra-Community differences in the desired dollar rate.

Although several interesting questions of fact and analysis are suggested by this sequence of events, this paper seeks only to clarify the discretionary aspects of alternative modes of intervention in the markets for foreign exchange. To this end, the paper considers two exogenous shocks to European currency markets. The first consists of a shift in demand from one snake currency to another; the second is a shift in demand from dollars to a snake currency. Both shocks are assumed to require intervention to maintain the snake parities, and in response to each shock the effects of the alternative intervention responses—that is, dollar intervention versus pure snake intervention—are analyzed.

The paper concludes that dollar intervention encompasses a broad quantitative spectrum of intervention responses, each with a different impact on snake-dollar rates. One such response, where the intervening European central banks purchase and sell identical quantities of dollars in exchange for their respective currencies, is equivalent to pure snake intervention. Because, in effect, the Community’s scope for altering relative dollar supplies is clearly delineated, this form of response has a systematic impact on the value of the dollar. While other specific modes of dollar intervention are similarly systematic and hence could be relied on to maintain intrasnake parities harmoniously, on the whole, “dollar intervention” connotes the array of these responses and, without quantitative restrictions on its use, does introduce discretion over the Community’s value of the dollar.

I. Theory of Exchange Rate Determination

Prior to considering the comparative effects of alternate intervention responses, it is useful to review the mechanism by which changes in exchange rates equilibrate the markets for currencies. Economic theory indicates that the demand for money is a function of income, wealth, and the expected rate of return on alternative assets, the latter being the opportunity cost of holding money. While a change in the exchange rate does affect the former variables through its effect on aggregate demand, these effects are, by and large, realized only over time, and cannot be relied upon to equilibrate money markets in the short run. Rather, recent research2 suggests that exchange rate changes clear markets for currencies by providing the expectation of capital gains or losses on holdings of foreign exchange.

In its simplest form, the mechanism is as follows. Consider a shift in demand for money away from Netherlands guilders to deutsche mark.3 In order to induce the absorption of the resulting excess supply of guilders into investors’ portfolios, the opportunity cost of holding guilders must decline. Guilder-denominated securities being particularly relevant alternatives to guilder cash balances, we would expect a reduction in Dutch interest rates. This, of course, would actuate an incipient flight of capital from the Netherlands, as well as an excess supply of Dutch securities. These pressures are relieved, however, by a depreciation of the guilder to a level below that which the “market” expects to prevail in the future.4 In this way, non-Dutch purchasers of guilder-denominated securities are compensated for the reduced interest rate by the anticipated capital gain on the guilder. Similarly, Dutch investors find other European and U. S. securities to be less attractive because they anticipate an appreciation of the guilder prior to the repatriation of their funds.5 At the same time, the converse occurs in the Federal Republic of Germany; the increased demand for deutsche mark induces a rise in that country’s interest rates. This precipitates an incipient capital inflow and excess demand for its securities, which are choked off by an appreciation of the deutsche mark above its expected value.

The guilder depreciates and the deutsche mark appreciates vis-à-vis all currencies in response to the shift in demand away from guilders toward deutsche mark. These cross exchange rate changes occur in response to the changes in the interest rates of the Federal Republic of Germany and the Netherlands and are required to extinguish incipient capital flows between those countries and the rest of the world. Moreover, because the interest differential between those countries is widened the most, the exchange rate between their two currencies is most affected.

These changes in direct and cross exchange rates are consistent with the cross currency arbitrage conditions. In equilibrium it must be that one cannot make profits by buying one currency, selling it for another, and then selling that for the original currency. In terms of guilders (f.), dollars ($), and deutsche mark (DM), this implies that the cross rates are related in the following way:

f.DM=f.$.$DM(1)

Thus, as the guilder floats down (f.DMrises), equation (1) is satisfied by the depreciation of the guilder vis-à-vis the dollar (f.$rises) and all other currencies, and the appreciation of the deutsche mark vis-à-vis the dollar ($DMrises) and all other currencies.

Over all, this analysis suggests that interest rates and exchange rates adjust simultaneously to equilibrate the markets for currencies and securities. In general, both rates will change in response to an autonomous change in the supply of or the demand for a currency whether this be caused by central bank intervention or by shifts in the asset preferences of private investors.6 Moreover, because of the close-knit fabric of the world financial community, changes originating in any one major financial center are quickly felt in others. This means that there is, in effect, a single foreign exchange market and a single securities market, and—from an analytic perspective—the geographic origin of the disturbance and the nationality of the buyer or seller of a currency or security are of little moment.

In the longer run, as traders react to the altered configuration of exchange rates, changes are induced that narrow the gap between the actual and expected exchange rates, in a process that allows the expected capital gains (losses) on currency transactions to materialize. In particular, in the depreciating country, aggregate demand is increased.7 This raises prices and reduces the real money supply. The interest rate rises in response. In the initially appreciating country, the converse occurs. The international interest rate differential thereby narrows, and, as a result, a smaller expected capital gain on currency transactions is required to induce the holding of securities denominated in the initially depreciating currency. Hence, the exchange rate will show a tendency toward a countervailing appreciation. This process continues until prices have risen sufficiently in the initially depreciating country to choke off excess aggregate demand, and have fallen sufficiently in the initially appreciating country to restore aggregate demand with the exchange rate equal to its long-run expected value.8

II. Intervention Responses to a Snake-Snake Shock

Diagramatically, the effects of the shock and the subsequent intervention are readily depicted. Figure 1 portrays the demand for guilders as a function of Dutch interest rates in the right-hand quadrant and the demand for deutsche mark as a function of interest rates in the Federal Republic of Germany in the left-hand quadrant. The initial supplies of guilders and deutsche mark are given by f.1 and DM1, respectively. The posited exogenous shift in demand away from guilders toward deutsche mark is represented by the dotted curves. At the prevailing interest rates, rDM1 = rf.1 there is an excess supply of guilders and an excess demand for deutsche mark. These money markets could be cleared by an increase in the interest rate in the Federal Republic of Germany to rDM2 and a decrease in the Dutch interest rate to rf.2. However, as outlined earlier, these adjustments would actuate changes in the market for bonds and other financial instruments, which would lead to a depreciation of the guilder vis-à-vis all currencies and an appreciation of the deutsche mark vis-à-vis all currencies. These exchange rate repercussions could be prevented, or reversed, by a change in the supplies of guilders and deutsche mark. In particular, if f.1-f.2 guilders were purchased by the authorities in exchange for DM2-DM1 deutsche mark, the national interest rates would return to rf.1 and rDM1, respectively, restoring equilibrium to the bond market at the initial configuration of exchange rates.

The same effects could be established by using the dollar, or another third currency, as an intermediary. For example, f.1-f.2 guilders could be purchased with dollars by De Nederlandsche Bank (the Dutch central bank) at the same time that DM2-DM1 deutsche mark were sold for dollars by the Deutsche Bundesbank (the central bank of the Federal Republic of Germany). The key point is that as long as the shift in demand is accommodated by an appropriate alteration in relevant currency supplies, the preshock configuration of interest rates and exchange rates can prevail, and longer-run adjustments are neither required nor actuated.

Nevertheless, if the intervention response were something other than a simultaneous reduction in the money supply of the Netherlands by f.1-f.2 and an increase in the money supply of the Federal Republic of Germany by DM2-DM1, the configuration of interest rates and exchange rates would be affected, and a dynamic adjustment process begun. In the present context, several alternative intervention responses seem particularly germane: (1) Simultaneous snake intervention by De Nederlandsche Bank and the Deutsche Bundesbank to keep the declining guilder-deutsche mark rate from crossing the snake band. Both buy guilders and sell deutsche mark. (2) Unilateral dollar intervention by De Nederlandsche Bank to prevent the declining guilder-deutsche mark rate from reaching the limits of the snake. It alone buys guilders and sells dollars. (3) Simultaneous dollar intervention by De Nederlandsche Bank and the Deutsche Bundesbank to prevent the guilder-deutsche mark rate from reaching the snake bands; the former buys guilders and sells dollars, and the latter sells deutsche mark and buys dollars. (4) The U. S. Federal Reserve System sells dollar-denominated securities and buys dollars in response to (2) or (3) to keep the supply of dollars in private banks unchanged.

pure snake intervention

If the preshock exchange rate between the guilder and the deutsche mark is within the required band, then the exchange rate must be allowed to float to this limit before intervention is warranted. In terms of Figure 1, this implies that fewer than f.1-f.2 guilders are bought and fewer than DM2-DM1 deutsche mark are sold. The Dutch interest rate remains below rf.1 but will be above rf.2. Similarly, the interest rate in the Federal Republic of Germany will lie above rDM1 but below rDM2. These interest rate changes indicate that the guilder depreciates, while the deutsche mark appreciates vis-à-vis all third currencies including the dollar. Other snake currencies may rise somewhat with the deutsche mark or fall with the guilder, but, over all, there is no compelling reason to expect a systematic change in snake-dollar rates.

unilateral dollar intervention

To illustrate this case, it is assumed that De Nederlandsche Bank responds to the shift in demand away from guilders toward deutsche mark by a unilateral purchase of guilders in exchange for dollars. Because the supply of dollars is increased with no change in demand, we can say unequivocally that U. S. interest rates fall, and the dollar depreciates vis-à-vis all other currencies. Furthermore, since the demand for deutsche mark has increased with no change in supply, interest rates rise in the Federal Republic of Germany and the deutsche mark appreciates vis-à-vis all currencies.

The effect of this intervention on Dutch interest rates and the configuration of exchange rates is somewhat more complicated. If the Dutch intervene to peg the guilder-deutsche mark rate at its preshock level, then the Dutch interest rate will have to rise to equality with the rate in the Federal Republic of Germany. In terms of Figure 1, f.1-f.3 francs must be bought to raise the postshock Dutch interest rate to rDM2. This, of course, means that the guilder appreciates with the deutsche mark, vis-à-vis all other currencies, including those in the snake, which also may be required to intervene to maintain the required alignment.

Variations on this intervention response can be similarly analyzed with the aid of Figure 1. In particular, if the Dutch purchase fewer than f.1-f.3 guilders, the guilder will depreciate vis-à-vis the deutsche mark. Moreover, if it were the Federal Republic of Germany that intervened unilaterally in dollars, it would have to sell DM3-DM1 deutsche mark to maintain the guilder-deutsche mark rate. In this case, of course, the dollar would appreciate vis-à-vis all currencies in response to the constriction of its supply.

simultaneous dollar intervention

This case is intermediate between the two just considered. De Nederlandsche Bank buys guilders and sells dollars, while the Deutsche Bundesbank sells deutsche mark and buys dollars. In the event that the two central banks buy and sell equal quantities of dollars, there is clearly no net change in the supply of dollars in private hands. The effect on interest rates and exchange rates is then the same as if each central bank had simply bought guilders and sold deutsche mark, and the pure snake intervention results, outlined earlier, obtain.

Nevertheless, the central banks are not constrained to buy and sell equal quantities of dollars. If, for example, De Nederlandsche Bank sells more dollars than the Deutsche Bundesbank buys, then the quantity of dollars in private hands increases, and the same qualitative results derived above for unilateral dollar intervention by the Dutch prevail. Although the snake appreciates vis-à-vis the dollar, the magnitude of the appreciation is diminished. In fact, the size of the appreciation is related directly to the net injection of dollars into the market. Conversely, if the supply of dollars decreases because the Deutsche Bundesbank buys more than De Nederlandsche Bank sells, the snake depreciates vis-à-vis the dollar. Similarly, the magnitude of the depreciation depends on the magnitude of the net dollar shrinkage.

U. S. Federal Reserve System intervenes in securities

If dollar intervention were to result in an increase in the quantity of dollars in private hands, the U. S. Federal Reserve System could remove them by an open market sale of dollar-denominated securities. This would restore U. S. interest rates to their previous level, consistent with the initial supply of dollars. However, by altering the supply of bonds, other repercussions ensue.

To be specific, consider the case of the unilateral sale of dollars by the Dutch, analyzed earlier. With interest rates in the Netherlands and the Federal Republic of Germany above their preshock levels, the demand for bonds in those countries exceeds the preshock levels. These demands may be satisfied by the newly marketed U. S. securities, but only at a comparable rate of return. Thus, the dollar depreciates below its expected level in order to generate the required expected capital gains. Third currencies depreciate as well, because otherwise there would be a decrease in demand for their securities. Moreover, even if the Dutch unilateral dollar intervention is incomplete, that is, the guilder depreciates vis-à-vis the deutsche mark, a similar process occurs.

Thus, if dollar intervention increases the supply of dollars in private hands, the dollar depreciates vis-à-vis all currencies. An open market sale by the U. S. Federal Reserve System does not reverse this depreciation; it simply causes other currencies to decline with the dollar. This may actuate intervention by third-party snake currencies, which decline with the dollar vis-à-vis the deutsche mark.

III. Intervention Responses to a Snake-Dollar Shock

In this section we consider the effects of the different intervention responses to a shift in demand from guilders to dollars. Figure 2 illustrates this shift, with the demand for guilders as a function of Dutch interest rates in the right-hand quadrant and the demand for dollars as a function of U. S. interest rates in the left-hand quadrant. As before, the dotted curves represent the postshock demands. The initial supplies of guilders and dollars are f.1 and $1, respectively. The shift in demand causes U. S. interest rates to rise to r$2, and Dutch interest rates to fall to rf.2. As outlined earlier, this divergence in interest rates results in an appreciation of the dollar vis-à-vis all other currencies, and a depreciation of the guilder vis-à-vis all other currencies, including other snake participants. It is the latter that evokes the intervention response, if the guilder-deutsche mark rate was at the snake limits prior to the shock.

pure snake intervention

The simultaneous sale of deutsche mark and purchase of guilders lowers interest rates in the Federal Republic of Germany and raises Dutch interest rates, causing the guilder to appreciate vis-à-vis the deutsche mark and the dollar, and the deutsche mark to depreciate vis-à-vis the dollar.

The important point, however, is that the snake intervention in support of the guilder-deutsche mark rate affords no discretion over the dollar rate. This can be verified by referring to equation (1). In order to hold the guilder-deutsche mark rate fixed (f.DMconstant), either both currencies must depreciate by the same percentage (f.$rises,$DMfalls), appreciate by the same percentage (f.$falls,$DMrises), or not change at all. The sale of deutsche mark and purchase of guilders, however, move their respective dollar rates in opposite directions. Nevertheless, this inverse movement compensates for the greater depreciation of the guilder vis-à-vis the dollar in response to the initial shock, such that a unique pair of guilder-dollar and deutsche mark-dollar rates emerges. Other snake currencies, not at the intervention points, simply depreciate against the dollar in response to the shock.

unilateral dollar intervention

This case is easily handled using Figure 2. De Nederlandsche Bank simply buys f.1-f.2 guilders and sells $2-$1 dollars. U. S. and Dutch interest rates return to their preshock levels, that is, r$1-rf.1. The shock is completely accommodated by a change in the supplies of the affected currencies, and the preshock configuration of exchange rates is restored.

However, instead of the Netherlands selling dollars and buying guilders in support of the guilder-deutsche mark rate, the Federal Republic of Germany could sell deutsche mark and buy dollars. This would cause the deutsche mark to depreciate fully in tandem with the guilder vis-à-vis the dollar. Other snake participants may be forced to buy dollars in exchange for their currencies to maintain the snake alignment.

simultaneous dollar intervention

If De Nederlandsche Bank and the Deutsche Bundesbank, respectively, sell and buy equal quantities of dollars, then the same results obtain for pure snake intervention. On the other hand, if more dollars are bought (deutsche mark sold) than sold (guilders bought), then the snake’s depreciation vis-à-vis the dollar is greater than under pure snake intervention. Conversely, if more dollars are sold than bought, the snake’s depreciation is intermediate between the pure snake intervention case and the case of unilateral dollar intervention by the Netherlands.

U. S. Federal Reserve System intervenes in securities

If, in response to unilateral dollar intervention by the Netherlands, the U. S. Federal Reserve System were to sell securities and buy dollars, U. S. interest rates would return to r$2, but the Dutch interest rate would remain at rf.1. This would entail an appreciation of the dollar relative to all currencies, with intrasnake rates remaining intact. Because this response increases the supply of securities and decreases the nominal supply of currency in private hands, the long-run equilibrium is characterized by interest rates above rf.1 = r$1.

IV. Summary and Conclusions

Table 1 summarizes the principal findings of this paper. The entries indicate the effects on the snake-dollar exchange rates of the posited shocks and intervention responses. For example, consider a shift in demand away from guilders to dollars, which would cause the guilder to depreciate vis-à-vis the deutsche mark and all other currencies, if no intervention response were taken. (1) If the Netherlands unilaterally sells dollars and buys guilders such that the initial guilder-deutsche mark rate is maintained, then the guilder-dollar rate and all snake-dollar rates are likewise maintained. (2)-(4) If the guilder-deutsche mark rate is maintained by joint dollar intervention of the Netherlands and the Federal Republic of Germany, then the dollar appreciates vis-à-vis the guilder and the deutsche mark. Moreover, the size of the appreciation is related inversely to the net sales of dollars. (5) If the Federal Republic of Germany alone intervenes to maintain the guilder-deutsche mark rate by buying dollars and selling deutsche mark, the dollar appreciates vis-à-vis the guilder and the deutsche mark even more than in the previous cases. (Other snake participants may be forced to buy dollars and sell their own currencies to maintain their place within the snake. This withdrawal of dollars would cause further appreciation of the dollar vis-à-vis the guilder and the deutsche mark, and is more likely the greater is the appreciation of the dollar induced by the shock plus intervention.) Hence, in each case except the first, the dollar appreciates vis-à-vis snake currencies (a > 0, b > 0, c > 0, d > 0), and the appreciation is larger the smaller the net increase in the supply of dollars (d > c > b > a > 0).

Table 1.

Effects on Snake-Dollar Rate of Shocks in Conjunction with Various Intervention Responses

article image

The other shock considered is a shift in demand away from guilders to deutsche mark, which causes the guilder to depreciate and the deutsche mark to appreciate vis-à-vis all currencies. (1) If the Netherlands sells dollars and buys guilders to restore the initial guilder-deutsche mark rate, the dollar depreciates vis-à-vis all currencies. (2) If the Netherlands and the Federal Republic of Germany intervene simultaneously but unequally such that there is a net increase in the supply of dollars, the dollar still depreciates but less than in the previous case. (3) Equal dollar sales and purchases by the Netherlands and the Federal Republic of Germany have no effect on the supply of dollars and no effect on the value of the dollar. (4) Simultaneous intervention in which the Federal Republic of Germany buys more dollars than the Netherlands sells induces an appreciation of the dollar. (5) With the Federal Republic of Germany alone intervening, the supply of dollars is reduced and the dollar appreciates a fortiori. In these cases, since the initial shock affects the dollar only peripherally, the ultimate valuation of the dollar depends on whether the intervention in support of the guilder-deutsche mark rate increases or decreases the supply of dollars. Thus, the dollar depreciates when the Netherlands alone intervenes by selling dollars (e < 0), and when the Netherlands intervenes more than the Federal Republic of Germany (f < 0). The dollar appreciates when the Federal Republic of Germany intervenes more than the Netherlands (g > 0), and when the Federal Republic of Germany alone intervenes (h > 0).

Clearly, the different intervention responses yield different effects on the Community’s valuation of the dollar. This, of course, is the point of the question raised at the outset of this paper, tentatively resolved by the Basle agreement to rely on snake intervention. Electing this convention for maintaining the snake alignment provides for a uniform impact on the snake-dollar rate. In contrast, dollar intervention in support of snake parities encompasses five distinct levels of intervention, each of which has a different impact on the valuation of the dollar. Community agreement on the level for the dollar would, of course, eliminate this discretionary source of contention and, in effect, limit the range of intervention responses. Alternatively, specific rules governing the use of dollar intervention would decrease the operational latitude associated with it, and prevent it from being a potential source of ongoing contention among snake participants who happened to disagree on the desired intra-Community dollar rate.

*

Ms. Salop, economist in the Financial Studies Division of the Research Department, holds degrees from the University of Pennsylvania and Columbia University.

1

The pre-1973 mechanism for maintaining snake parities is described and analyzed in Patrick de Fontenay and Giorgio Ragazzi, “The EEC Narrow Margin Agreement” (unpublished, International Monetary Fund, January 19, 1973).

2

Stanley W. Black, International Money Markets and Flexible Exchange Rates, Princeton Studies in International Finance, No. 32, International Finance Section, Princeton University (1973); Lance Girton and Dale W. Henderson, “Central Bank Operations in Foreign and Domestic Assets Under Fixed and Flexible Exchange Rates,” International Finance Discussion Paper No. 83, Board of Governors of the Federal Reserve System (May 1976).

3

This hypothetical shift could be the result of a shift in portfolio demand for money or a reflection of altered trade conditions that affect the transactions demand for currency. In particular, if commodity demand shifts from products of the Netherlands to products of the Federal Republic of Germany, there ensues an increase in demand for deutsche mark transactions balances and a decrease in demand for guilder transactions balances.

4

The market’s expectations of the exchange rate’s future value may be affected by the shock itself. Nevertheless, the guiding principle remains that of interest parity. Accordingly, interest rate differentials are reflected in expected exchange rate changes.

5

In other words, those who value their wealth in guilders now face a lower expected rate of return, regardless of whether they hold assets denominated in guilders or in other currencies.

6

The model implicit in this paper is one in which securities are perfect substitutes for each other in portfolio; therefore, a given investor is indifferent as to the currency of denomination of securities as long as they offer identical expected rates of return. Alternatively, one could assume that investors viewed securities of different countries as “gross substitutes.” In such a model, we could speak of a shift in demand from securities of the Netherlands to securities of the Federal Republic of Germany. The effect would be a depreciation of the guilder vis-à-vis the deutsche mark; currency intervention would affect interest rates and exchange rates in the same qualitative manner outlined in the text.

7

Initially, the depreciation causes the real money supply to fall, if imported goods are included in the price deflator. This immediately raises the interest rate somewhat, and reduces the magnitude of the required capital gain. Subsequently, as the lowering of relative prices affects trade, aggregate demand necessarily rises if the elasticity of demand for imports exceeds unity.

8

For a formal analysis of this process in the small country case, see Rudiger Dornbusch, “Expectations and Exchange Rate Dynamics,” Journal of Political Economy, Vol. 84 (December 1976), pp. 1161–76.