The Preannouncement of Exchange Rate Changes as a Stabilization Instrument

International Monetary Fund. Research Dept.
Published Date:
January 1981
  • ShareShare
Show Summary Details

In recent years, some developing countries have adopted a variety of innovative stabilization programs in an effort to deal with high and persistent inflation and with balance of payments difficulties. One component of these programs has been the preannouncement of the future path of the exchange rate. This policy essentially involves the public announcement of a fixed schedule of the exchange rate over a specified period with the explicit understanding that the schedule will not be modified during that period. While a fixed exchange rate policy could potentially fall under this definition, and in fact many of the features of a fixed exchange rate system are similar to those of the preannounced system, this study focuses only on the characteristics and the dynamics of adjustment implied by those programs that have involved continuous (e.g., daily) adjustments of exchange rates, usually during periods of substantial domestic inflation, well above the corresponding foreign rates.

The policy of preannouncing the exchange rate has, as mentioned, always been a part of an overall stabilization package comprising trade, financial, and fiscal reforms designed to slow inflation and improve the balance of payments. The reform programs in general and the policy of preannouncing the exchange rate in particular were expected to generate anti-inflationary effects by opening the domestic economy to increased international trade and capital flows. Increased foreign competition in the product markets was perceived as a means of inhibiting domestic price increases, and greater availability of foreign capital was expected to reduce the real cost of credit to the private sector.

In the next section a short review of the preannouncement programs adopted by a number of countries is presented, together with a general discussion of the expected effects and the channels through which a policy of preannouncing exchange rate movements affects the stabilization process. Special attention is devoted to the relationship between exchange rate policy and the problem of capital inflows.

There is also a discussion of some of the issues involved in selecting the most appropriate preannounced exchange rate path. Section II presents a formal model to analyze the dynamics of the adjustment process when preannouncement is introduced, deriving and discussing the conditions under which the system will be stable. A final section summarizes the conclusions, and the formal derivations are presented in technical appendices.

I. Programs Involving Preannouncement of Exchange Rates and Their Effect on the Rate of Inflation

Three Latin American countries, Argentina, Chile, and Uruguay, have used the policy of preannouncing the exchange rate as an integral part of their stabilization effort.1 The three countries started their programs in 1978, and the announcement schedules were published periodically covering periods ranging from approximately 3 to 12 months.2 With some exceptions, the authorities did not allow for any feedback between domestic and foreign variables and exchange rate movements during the announcement period of the individual exchange rate schedules. In addition, when the announcement of a new exchange rate schedule preceded the termination of an existing schedule, the period of overlap almost always contained the same exchange rate movements. Any feedback between exchange rate policy and domestic or international variables was reflected only in the exchange rate movements that occurred beyond the period of overlap between the old and the new schedules.

To identify the role played by the policy of preannouncing the exchange rate, it is important to emphasize three characteristics of the overall economic programs implemented. First, as already noted, the policy of preannouncing exchange rates has been an integral part of stabilization programs composed of fiscal, trade, and financial reforms designed to open the economies to international influences. This has involved a substantive reduction in tariffs and the elimination of quantitative controls and other trade restrictions. In the financial area, the reforms have included some freeing of domestic interest rates, reductions in required reserve ratios, the relaxation of limitations on competition and entry into the financial system, and opening the economy to international capital movements through a gradual reduction or elimination of exchange controls. Second, the paths for a number of policy instruments have been preannounced, including the price schedules of public sector enterprises, future increases in minimum wage rates, forthcoming reductions in tariff and other trade restrictions, the target rate of expansion of domestic credit, and the future path of the exchange rate. The preannouncement of these measures has been viewed as a means of affecting private sector expectations directly, even prior to actually putting the policies into effect. Third, while the specific paths selected for the exchange rate have been influenced by balance of payments considerations, the exchange rate has been used primarily as an anti-inflation instrument.

To analyze the potential stabilizing effects of preannouncing the exchange rate, and the rationale for using it as an anti-inflation instrument, it is useful to distinguish between the direct impact of such a policy on expectations and the indirect effects on product and financial markets.


To reduce the rate of inflation rapidly without incurring high real costs, the ideal stabilization program would induce a substantial reduction in the public’s expectations concerning future inflation. This is a crucial element because, when inflation has been sustained at a high level for long periods of time, institutional developments, such as excessive inventory holdings, fully indexed wage settlements, and high markups over original cost to reflect higher expected replacement costs, make achieving a reduction in inflation more difficult. In this context, a clear signal from the economic policymakers that definitive measures are being taken to reduce the rate of inflation can, by itself, have a significant impact on inflationary pressures.

The channels through which a change in expectations can affect the inflationary process are diverse. For example, if both managers and workers believe that inflation is going to be reduced, they will be more willing to adopt pricing and compensation schemes that embody lower nominal claims. In addition, expectations of lower inflation imply a lower alternative cost of holding nominal financial assets and therefore will tend to induce an increase in the demand for money.3 For given rates of monetary expansion and of income growth, the increase in the demand for real cash balances will have obvious stabilizing effects on the price level.

To have these stabilizing effects, the signal coming from the stabilization package should be both credible and noticeable. To be credible, it should be consistent with the rest of the program. The preannouncement of a gradual reduction in the rate of depreciation of the exchange rate may meet these conditions if it is taken as indicating a commitment on the part of the monetary authority to reduce inflation to rates consistent with the announced policy (considering also, of course, the rate of inflation in the world economy). Obviously, the success of preannouncement in reducing inflationary expectations depends on the consistency of the complete package and the perceived ability of the authorities to pursue these policies successfully.4


A key component of the stabilization programs has been the opening of the economy through the elimination of trade barriers in order to increase the degree of foreign price competition and thereby to limit the ability of domestic producers to raise prices.

The elimination of trade restrictions and the consequent appearance of a price differential between a particular domestic good and its foreign counterpart does not guarantee, however, that international price arbitrage will take place. Even if firms are fully aware of opportunities for arbitrage, there are various initial costs associated with organization, marketing, and processing that confront an enterprise contemplating entry into external trade, and these costs tend to be higher when international trade has been restricted for a long period. Firms will incur these costs and engage in international arbitrage only if the potential profit opportunities are likely to persist long enough to permit recovery of the cost of entry. When the exchange rate is unstable and/or unpredictable, exchange rate fluctuations increase the risk that profit opportunities emerging from the trade reform will not be maintained, and this reduces the incentive to incur the initial cost entailed in exploiting any opportunity for arbitrage. Since unanticipated exchange rate movements reduce the likelihood that organized arbitrage will take place, the preannouncement of the future path of exchange rates will help to curb inflation via increasing the likelihood that international competition will constrain domestic price-setting decisions.5


In an economy where foreign and domestic assets are close substitutes and where domestic financial markets are fully integrated with those of the rest of the world, domestic interest rates should equal the yields on comparable foreign securities plus the expected rate of exchange rate depreciation.6 The domestic interest rate on loans may, however, deviate from the equivalent foreign rate as a result, mainly, of three elements:7

  • (i) country-specific institutional factors 8 such as different reserve requirements and tax treatment of interest earnings.

  • (ii) country-specific risk such as the risk of the imposition of exchange controls at the time of repatriation and political risk.

  • (iii) exchange risk, which arises because assets are denominated in different currencies; also, there may be uncertainty about future exchange rate movements.

These elements can lead to a premium (or discount) of the yield on domestic compared with foreign assets, which will result in a differential between domestic and foreign interest rates after adjusting for expected exchange rate movements. When the adjusted interest rate differential exactly matches the premium (or discount) arising from the institutional and risk elements, there is no incentive for capital flows to take place.9 Whenever some of those elements are altered, however, the premium necessary to equilibrate the market changes accordingly, and capital flows take place until the appropriate interest rate differential is restored. For example, if the imposition of exchange controls became more likely, the risk premium required to equilibrate the market would increase and, given the existing interest rate differential, an outflow of capital and a rise in the domestic interest rate would be induced. The capital outflow, which constitutes a stock adjustment, would cease when the interest differential is fully adjusted so as to reflect the increased risk premium associated with a greater likelihood of exchange controls.

In this setting, a shift from an exchange rate policy of free floating, managed floating, or an undisclosed crawling peg to a policy of preannounced exchange rate movements would tend to reduce exchange rate uncertainty and thereby the risk premium on domestic assets.10 Assuming that the market was in equilibrium at the moment the path was announced, the prevailing interest rate differential would be too large given the reduction in the risk premium. A capital inflow would be induced and the domestic interest rate would tend to fall until the differential was consistent with the new (lower) risk premium. A credible preannouncement entails, therefore, a once-and-for-all inflow of capital until the adjustment of interest rates is completed.11 Under the assumption of perfectly integrated financial markets with unrestricted capital mobility and flexible interest rates, the capital inflow triggered by the preannouncement of exchange rate movements cannot persist. This capital inflow can nonetheless increase the stock of real credit available in the economy and thereby lower production costs by reducing the real cost of credit.

The dynamics of the adjustment process may, however, be rather different in economies that are in the process of opening their financial markets to international capital flows and are undertaking extensive financial reforms, including the removal of interest rate ceilings and the elimination of exchange controls. It is, in fact, unlikely that after long periods of tight capital controls the domestic capital market will immediately achieve full integration with world financial markets. In particular, it is probable that the domestic market will show some degree of segmentation, since not all financial agents will have the same direct access to international borrowing, especially when new financial institutions specializing in international intermediation are not yet developed. Given the fixed set-up costs and, in some cases, the oligopolistic structure of the domestic banking system, the development of new institutions may require a substantial period of time. In the meantime, borrowers without direct access to international markets must obtain credit from domestic financial intermediaries that can borrow abroad and lend to domestic agents. In such a situation, the domestic interest rate is not determined solely by the foreign interest rate, the expected rate of change in the exchange rate, and any risk premium but is also strongly influenced by domestic market conditions, including the domestic demand and supply for credit, the structure of the domestic financial system, and the state of inflationary expectations.12

It is, therefore, likely that immediately following a financial reform and the opening of the financial market there will be a relatively long adjustment period during which domestic considerations will have substantial weight in the determination of the domestic interest rate. When this is the case, the domestic interest rate may exceed the foreign rate plus the expected devaluation by more than the risk premium, and this apparent profit opportunity will not be immediately arbitraged. Capital inflows will take place but not necessarily at the speed and volume needed to bring the domestic interest rate rapidly down to the level consistent with the foreign cost of capital.

When the economy is in the midst of such an adjustment process, preannouncing exchange rates can have an inflationary effect. By reducing exchange risk, the preannouncement program increases any existing differential between domestic interest rates and the equivalent yields (adjusted for risk) on foreign assets and generates additional capital inflows. These capital inflows will have two effects on domestic interest rates: (i) an availability, or liquidity, effect that tends to reduce domestic interest rates (and the interest rate differential) as the stock of real domestic credit is supplemented by foreign credits 13 and (ii) an expectations effect which tends to increase interest rates and which arises from the inflationary pressures caused by the expansion in the monetary base following the capital inflow.14

When the integration of capital markets is not complete, the higher inflationary expectations can offset the impact of the greater availability of foreign credit and can reverse, temporarily, any initial decline in the domestic nominal yields.15 In this situation, unlike that of a fully integrated economy, the inflow of capital may not be characterized by a once-and-for-all stock adjustment but may persist as long as the speed and magnitude of the decline in domestic interest rates are limited by the institutional structure of the market. Thus, the capital inflow may sustain domestic inflation even in the face of other anti-inflation measures. While an increasing degree of market integration will ensure that the greater availability of credit ultimately leads to a decline in domestic interest rates and to a reduction in the capital inflow, the monetary effects of the capital inflow may significantly lengthen the required duration of the anti-inflation program and erode the credibility of the preannouncement program.

Since capital inflows can threaten to undermine the program, it is important to analyze in detail the conditions under which a stabilization program that includes the preannouncement of the exchange rate as one of its components will indeed result in a reduction of inflationary pressures. In the next section a model is developed that will enable us to consider this question. It should be clarified, however, that in addition to the problem of identifying the stability conditions, which is the main concern of this paper, there is the additional issue of selecting the “best” path to be announced, in terms of achieving the authorities’ objectives regarding inflation, the balance of payments, and growth. The specification of the most appropriate preannounced exchange rate path would involve expanding the short-run model (given in the next section) to allow for variable output and employment (including a specification of wage behavior) and the authorities’ objective function. The “optimal” exchange rate path would then be derived, along with such paths for the other policy instruments, as part of a solution to an optimal control problem. While the solution of this optimal control problem could be quite complex (especially if there were a large number of instruments and objectives), there are certain characteristics and determinants of the preannnounced exchange rate path that are likely to hold regardless of the exact nature of the authorities’ objective function.

First, any preannounced exchange rate path would be closely related to the paths selected by the fiscal, trade, and financial policy instruments. This coordination should encompass the selection of both the initial level of the exchange rate and the rate at which it would be depreciated. As will be shown in the next section, this policy coordination is one of the key requirements for ensuring a stable adjustment process, especially when the preannouncement of the path for the exchange rate is directed toward achieving a reduction in inflation. While the preannouncement of the exchange rate path can strengthen the linkages between domestic and international prices, this type of policy is most likely to be successful if the trade reform has already significantly reduced the barriers to price arbitrage between domestic and international prices and fiscal reforms have been implemented to reduce the issuance of base money needed to finance budgetary deficits.

In setting the path for the exchange rate, one particular risk is that a seriously overvalued exchange rate may develop. The stability analysis in Section II is particularly concerned with this issue. This danger can possibly be minimized by combining an initial discrete exchange rate depreciation with an ensuing period during which the exchange rate depreciates at a rate slower than the differential between domestic and foreign inflation. Such a policy may make it possible both to eliminate any initial overvaluation of the exchange rate (an effect on the price level) and to slow inflation (an effect on the rate of change of the price level).

Equally important considerations in setting the path for the exchange rate are the effects this program will have on the level of employment and the rate of growth of the economy. Special difficulties may be created for the anti-inflation program by the presence of comprehensive wage indexing schemes. If these indexing arrangements keep real wages constant during the adjustment period, this may induce more severe employment and output declines and also make it more difficult to eliminate any existing overvaluation of the exchange rate through an initial discrete depreciation. This suggests that, to minimize any adverse employment effects, the use of the preannouncement of the exchange rate may require some changes in the existing wage indexing arrangement (e.g., allowing for only a partial cost of living adjustment for some fixed period).

II. The Model

In this section, the effects of a stabilization program based on trade and financial reforms are modeled and the impact of preannouncing the future exchange rate path on the stability of the adjustment process is analyzed. The trade reforms included in the program encompass the removal of import quotas and the lowering of tariffs; the financial reforms are based on the freeing of interest rates and the elimination of exchange controls. To analyze the effect of these measures and, in particular, the role of the preannounced exchange rate policy, first one has to consider the determinants of portfolio and spending behavior in the financial and goods markets.


In the financial markets, a distinction is made between non-bank and bank behavior. The nonbank sector consists of firms and households that hold bank deposits and capital as assets and bank loans as liabilities. Since the focus is on the short-run adjustment process, holdings of capital are taken as fixed. Money (M) consists solely of bank deposits that earn a rate of interest rM.16 The rate of interest on bank loans is rL.

In the short run, the overall size of the nonbank sector’s portfolio is fixed, but it is assumed that the mix of liquid assets (money) and liabilities (loans) can be varied continuously. Over time, the size of the portfolio can be altered, and portfolio behavior would be determined by the attempts to achieve both a desired portfolio size and an optimal mix of assets and liabilities.

Given a fixed portfolio size, the desired mix of money relative to loan holdings is a positive function of the expected real return on money and the expected real cost of loans:17


M=bank deposits
L=bank loans
YM=nominal interest rate on bank deposits
rL=nominal interest rate on bank loans
πe=expected rate of inflation
hi=partial derivative of h with respect to the ith argument

Even though it is assumed that wealth holders can rearrange their existing portfolio quite rapidly, this does not imply that they will continuously achieve the desired overall size of their portfolio. Since portfolio size can increase only through net savings or capital gains owing to the revaluation of existing assets, the attainment of the optimal size of the overall portfolio may require adjustments of expenditure plans and take considerable time. To reflect this possibility, the specification of portfolio behavior can be completed by an equation for the nonbank sector’s desired holdings of money or loans and an assumption about the process of adjustment of actual to desired stocks.18 Let the nominal demand for money be a function of the level of income, the level of prices, and the expected real rate of return on money holdings:19


Md=desired money holdings
Y=real income
P=domestic price level

This demand for money, however, will not necessarily be satisfied at each moment in time, since changes in the stock require adjustments in spending and borrowing plans that may occupy an extended period of time. The exact nature of these spending and financing adjustments will be specified in the discussions of goods market and balance of payments behavior.

Regarding the banking sector, its assets include reserves and loans and its liabilities are deposits of the nonbank sector and foreign borrowing. Assuming that banks are subject to a required reserve ratio of K per cent and hold no excess reserves, the banks’ balance sheet constraint is given by


K = required reserve ratio

X = exchange rate

F = foreign borrowing by banks

It is assumed that the banking system is not perfectly competitive and that international and domestic financial markets are not perfectly integrated in the sense that all domestic economic units do not have the same access to international borrowing. The model reflects these characteristics as follows. First, it is assumed that the competitive structure of the financial system allows banks to follow a markup pricing scheme such that20


rF=cost of foreign borrowing (inclusive of expected exchange rate movements and country risk premiums)
ɸ=markup factor

This implies that the loan rate exceeds the cost of bank funds by a constant factor φ. The cost of bank funds equals the deposit rate weighted by the proportion of loans financed by deposits plus the corresponding weighted cost of foreign funds. The cost of foreign funds is


rF=foreign interest rate
e=expected rate of change in exchange rate
θ=risk premium associated with foreign exchange transactions

To reflect the lack of full integration of domestic and international capital markets, it is assumed that only banks can borrow abroad. In addition, because of imperfect competition in the banking sector, it is assumed that banks arbitrage only gradually any differential between the domestic and the adjusted foreign interest rate. Let f = XF/L, then the banks’ process for arbitrage is given by 21

Thus, banks move gradually toward the elimination of any differential between the domestic loan rate and the covered foreign interest rate. The value of μ reflects the degree of capital mobility, and it is affected by the existing limitations to the free movement of capital. Therefore, under a very extreme regime of exchange controls, μ may approach zero and capital movements will be low or nonexistent even if fd ≥ f. The value of μ will be increased by financial reforms that reduce or eliminate exchange controls.

The financial market behavior represented by equations (1) through (6) implies that, although domestic interest rates adjust so that the nonbank sector achieves the desired distribution of its existing stock of wealth between money and loans at each instant in time (“instantaneous financial market equilibrium”), the nonbank sector will not necessarily have achieved its desired stock of wealth as implicit in the demand for money equation (“full portfolio equilibrium”). In combination with the specification of bank behavior, the market-clearing loan and deposit interest rates can be expressed as functions of the required reserve ratio, the expected rate of inflation, the “covered” foreign interest rate, and the ratio of bank foreign borrowing to bank loans. As shown in Appendix II,



Let there be both foreign and domestic goods. The prices of foreign goods are determined by world prices adjusted for the exchange rate (and, in the case of import goods, tariffs). The rate of increase in the prices of domestic goods is assumed to be influenced by internal excess demand as well as by the rate of change of the domestic price of foreign goods. Postulating a monotonic relation between domestic excess supply and domestic monetary disequilibrium,22


π=rate of change in price of domestic goods
πF=foreign inflation
=rate of change in exchange rate

The coefficient τ measures the effect of domestic monetary conditions on relative commodity prices and is determined by the proportion of foreign to domestic commodities in the economy as well as by the expenditure elasticity of domestic commodities. If all goods are internationally traded, τ will be zero, reflecting complete integration of the domestic with the foreign market and therefore full commodity arbitrage and convergence of the domestic to the foreign inflation rate. Clearly, τ will not be constant over time but will tend to decline as the proportion of internationally traded goods increases and arbitrage becomes more extensive. Therefore, a trade reform including the reduction of tariffs and the elimination of trade restrictions implies a lower value of τ which will further decline as the institutional framework of importers and wholesalers is developed and as trade and exchange risk are reduced. Specifically, it is assumed that τ declines over time following the initial trade reforms and is also a positive function of the degree of exchange risk associated with financial or trade credit activities (θ). Thus,


tt0 time since the initial trade reform

This formulation implies that as the degree of exchange risk declines there will be greater integration between the domestic and the foreign goods market, and this will be reflected by a decline in the impact of domestic monetary disequilibrium on domestic prices, with a correspondingly greater spillover onto international markets. This assumption about the distribution of monetary disequilibrium between prices and the balance of payments also bears on the relationship between the domestic and foreign sources of monetary growth.

Using H for the monetary base,

Thus, all the base money is used to satisfy the demand for bank reserves. The two sources of base money growth are the central bank accumulation of foreign exchange generated through the balance of payments () and its provision of domestic credit ().23 In terms of growth, this implies

The balance of payments is assumed to be determined by both international price and interest rate arbitrage and monetary disequilibrium. Thus,24

The first two terms in equation (13) indicate that foreign reserves will accumulate whenever the foreign rate of inflation (adjusted for exchange rate changes) rises relative to the domestic rate of inflation and when the domestic loan rate rises relative to the adjusted foreign interest rate.25 In addition, if the increase in the money supply generated by the central bank’s issuance of domestic credit exceeds the increase in the nominal demand for money (brought about by increases in the price level or in the expected real rate of return on deposits), there will be a reduction in the level of reserves in order to restore monetary balance. Note that, as the economy becomes more integrated with the rest of the world, τ falls and the effects of excess money supply spill more onto the balance of payments and less onto domestic prices. As τ approaches 0, complete arbitrage takes place, and monetary disequilibrium is fully resolved through the balance of payments.

It is important to notice that the rate of domestic credit expansion is closely related to the size of the fiscal deficit. Therefore, the magnitude of monetary disequilibrium in the economy, which bears on the nature of the adjustment process and on the stability of the system under preannouncement, is largely determined by the government’s ability to control the fiscal accounts.

To close the basic structure of the model, assumptions about the formation of expectations are required. It will be assumed that the private sector formulates its expectations about inflation and exchange rate changes on the basis of its best estimates of the effects on these variables of future reform and stabilization policies. In our deterministic model, this “rational” expectations assumption implies perfect foresight, which requires



One can now examine the stability of the adjustment process implied by the stabilization program and consider the effects on the nature of the process of including or excluding a pre-announced exchange rate policy. In addition to trade and financial reforms, the government is taken as generating a lower and fixed rate of growth for domestic credit and as maintaining a given rate of depreciation of the exchange rate. Initially, it will be assumed that the future course of the exchange rate is not preannounced. Immediately following the implementation of the trade and financial reforms, the authorities are likely to be confronted with rapid inflation (reflecting past monetary expansion), high nominal and real interest rates (reflecting a high expected inflation rate and a shortage of real credit), and balance of payments disequilibrium. The issue first considered is under what conditions the mix of policy measures postulated yields a stable adjustment process leading toward a lower rate of inflation.

When the exchange rate is not preannounced, the adjustment path for the economy can be described by combining equations (1) through (15) to yield two differential equations, with the rate of change of domestic prices (π) and the ratio of banks’ foreign borrowing to domestic loans (f) as state variables.26 These equations will have the form 27


This implies that the adjustment process will converge to a lower inflation rate only if


Given the signs of the partial derivatives in equations (16) and (17), a sufficient condition for stability is that π˙/∂π0 or, equivalently,

To understand the implications of the sufficient condition, it is useful to examine the effects that each of the four terms represents. The first term is the effect of higher domestic inflation on the current account of the balance of payments. A rise in π reduces the rate of growth of foreign reserves, thereby slowing the rate of monetary expansion. This will help to generate a path of declining inflation and hence will be stabilizing. The second term (C2τg2) is the impact of higher inflation on the expected inflation rate and thereby on domestic interest rates and the interest arbitrage term in the balance of payments. As πe rises with any increase in π, this will lead to an increase in the domestic loan rate (rL), which will stimulate a capital inflow. The resulting accumulation of reserves will raise the rate of monetary expansion and drive up or sustain the current rate of inflation. This effect is obviously destabilizing. The third term (−τ) reflects the fact that a higher π implies that the private sector must hold more nominal money to support the same level of transactions. This works to reduce the excess supply of money and hence any balance of payments surplus. The resulting decline in the rate of monetary expansion will help to slow inflation, and is thus stabilizing. The final term represents the effect of a higher π on the demand for money via its impact on the real return on money holdings. As π increases, and for a given e the loan rate rises, stimulating a capital inflow and raising f, the proportion of bank loans financed by foreign borrowing. This has the effect of lowering the deposit rate and reducing the demand for money.29 A reduction in money demand stimulates domestic inflation (via equation (9)). This is partly offset, however, by the fact that the decline in the demand for money leads to an outflow of reserves and thereby reduces the rate of monetary expansion. This slower growth of monetary aggregates would help to reduce inflation. This fourth term represents the net impact of these two effects and indicates that the inflationary effect dominates. Thus, the impact of a higher f is to raise the rate of inflation, which tends to be destabilizing.

The analysis of the stability condition indicates that an initial decline in π may not be sustained if capital inflows result in a rate of monetary expansion that overwhelms the stabilizing effects associated with international price arbitrage. This implies that an important relationship may exist between the stability of the adjustment process and the sequencing of the trade and financial reforms. A financial liberalization encompassing both the removal of ceilings on interest rates and the absence of controls on capital inflows can create potentially destabilizing capital inflows unless there is a significant opening of the domestic economy to international trade and price competition.

A stable adjustment process is illustrated by Figure 1. The π˙=0 curve represents the combinations of π and f that will result in a given rate of inflation, and the f˙=0 curve describes the combinations of π and f that imply a given ratio of foreign borrowing to domestic loans. At the moment of the implementation of the reforms, the economy may be in a position such as A, with π above its longer-term value π and f below its equilibrium value f. The elimination of trade and financial restrictions will move the system from the off-equilibrium position at A toward equilibrium at B. The adjustment toward the stable equilibrium position, with a lower rate of inflation and a higher ratio of foreign borrowing to domestic loans, may be either direct or cyclical. If the adjustment is cyclical, then there possibly will be a period of a rising rate of inflation and increasing f followed by a falling rate of inflation and a declining value of f.30


Note that the π˙=0 curve is drawn for given values of τ. The initial trade liberalization will rotate the curve counterclockwise. As import and wholesale institutions develop following the trade reform, τ declines further and eventually approaches zero. This shifts π˙=0 down and ultimately makes the slope equal zero (see Figure 2). Thus, the economy will eventually tend toward an equilibrium with π equal to the sum of the foreign rate of inflation plus the rate of depreciation of the exchange rate and f taking on a value consistent with the absence of any differential between the domestic and foreign interest rates (apart from a risk premium).


If the economy’s path to this long-run equilibrium is stable and noncyclical, then the adjustment process can be portrayed using the solid lines in Figure 3. The trade and financial reforms, which occur at time t0, will have the effect of producing a discrete reduction in inflation and a jump in bank borrowing from abroad. This reflects the impact of international price arbitrage on domestic price formation once import quota restrictions are removed and the attractiveness of foreign borrowing in the face of a substantial differential between domestic and foreign interest rates. As the reform proceeds, τ declines and ultimately the economy will achieve a rate of inflation equal to πF + ẋ with f = fL.



The authorities’ objective in using the policy of preannouncing the future exchange rate is to bring about a more rapid adjustment to a lower rate of inflation. By lowering exchange risk, it is argued that this policy will encourage greater, and faster, international arbitrage between domestic and foreign prices and interest rates. The arbitrage in the goods market will lower inflation more rapidly to the sum of the world rate plus the rate of depreciation of the exchange rate, and the interest rate arbitrage will increase the supply of real credit in the economy and lower domestic interest rates. In terms of Figure 3, this argument implies that combining the trade and financial reforms with the preannouncement of the future exchange rate movements will lead to the paths for π and f denoted by the dashed lines. The decline in π would thus be more rapid than without the preannouncement, whereas f would rise more rapidly and achieve a higher long-term level owing to the reduced exchange risks associated with foreign borrowing.

To analyze whether preannouncing the exchange rate will have this favorable effect, it is first necessary to introduce variable exchange risk into the model. As described above, the interest rate that foreign financial institutions charge on loans to domestic banks includes a risk premium (θ) associated with such factors as exchange and political risk. Similarly, the proportion τ of excess supply of money spilling over onto the goods market as opposed to the balance of payments is also influenced by the level of exchange risk.

It is postulated now that the preannouncement of the future path for the exchange rate will affect θ in two ways. First, the initial announcement of the next exchange rate program would bring about some reduction in uncertainty regarding future exchange rate movements and thereby reduce θ directly. The extent of the initial reduction in θ would depend on market participants’ perceptions of how likely it is that the authorities will follow the proposed exchange rate path. Second, in the period following the initial preannouncement, θ will be influenced by how well the authorities maintain the announced exchange rate path and by the private sector’s perception of whether economic or political developments will force the authorities to depart from the preannounced path. To simplify, it is assumed that the private sector’s perception of the sustain-ability of the preannouncement program is related to the length of time that the preannouncement schedule has been maintained (given by α0(tt0))31 and the deviation from relative purchasing power parity induced by the exchange rate policy adopted. Thus,32

Hence, the private sector’s evaluation of the sustainability of the preannouncement program at any moment in time is based on how long the program has been in effect and whether or not this policy is leading to the creation of an overvalued currency.33

Consider first the initial impact of the establishment of a preannouncement program and then analyze whether this policy affects the stability of the adjustment process. Let the announcement of the new exchange rate program bring about an initial discrete reduction in perceived exchange risk equal to θ˙0=θ0θ¯0, where θ is the level of exchange risk at the instant prior to the announcement. While such a reduction in exchange risk will induce a discrete increase in commercial bank foreign borrowing, it will have an ambiguous effect on domestic inflation. Lower exchange risk enlarges any existing differential between the domestic loan rate and the risk-adjusted cost of foreign borrowing. This larger interest rate differential will lead to a once-and-for-all rise in f, which causes an inflow of reserves and thereby speeds up the rate of money growth with the consequent inflationary effect. Offsetting this effect on inflation, however, is the fact that the initial decline in θ induces a fall in τ implying that a greater proportion of any excess money supply will spill over onto the balance of payments rather than onto the domestic goods market and the rate of inflation. Since it is not clear which of these effects will dominate, the preannounced exchange rate policy could initially either raise or lower the domestic rate of inflation.34

Given the ambiguity surrounding the initial impact of the preannouncement policy, there is naturally the question of whether or not it will enhance or weaken the stability of the adjustment process. This issue can be analyzed by examining the adjustment process that is generated when the preannouncement is part of the program and comparing it with the case where it is not included. The adjustment process that incorporates the effects of variable exchange risk (θ) can be described by combining equations in π, f, and θ of the form35

The adjustment process will be stable if


Since θ˙/θf=θ˙/θ=0, these stability conditions will be satisfied if π˙/∂π0 and

Clearly, this last condition will be satisfied only if π˙/∂θ is negative.

How likely is it that both π˙/∂θ and π˙/∂π will be negative? π˙/∂θ reflects the influence of two separate factors. A rise in exchange risk (θ) tends to lower inflation by reducing capital inflows and the rate of monetary expansion. At the same time, a higher θ will raise τ, which means that any given rate of monetary expansion will be more inflationary. π˙/∂θ will be negative only if the interest rate differential effect dominates.

Even if π˙/∂θ is negative, π˙/∂π is less likely to be negative than in the case of no preannouncement, because in the current case π˙/∂π0 implies

As in the previous case, this condition is more likely to be fulfilled the more open is the economy, that is, the lower is τ. However, a comparison with the condition for π˙/∂π0 when there is no preannouncement (see equation (20)) indicates that equation (28) includes the addition of the terms −τβ1q3α + ατ2 In(M/Md), both of which reduce the likelihood of π˙/∂π0. The first term reflects the fact that a higher π raises exchange risk, and thus the cost of foreign borrowing, leading to a lower deposit rate and a lower demand for money. A lower demand for money works to generate a higher rate of inflation. The second term represents the fact that a higher rate of inflation increases exchange risk and therefore pushes up the value of τ, increasing the impact of any excess supply of money on inflation. Taken together, these effects imply that the interaction between inflation, exchange risk, and the preannouncement policy will be less likely to result in a stable adjustment process compared with the case in which the preannouncement is absent. Furthermore, the possibility of an unstable result increases the larger is the level of monetary disequilibrium and consequently the size of the fiscal deficit and the more rapid is the increase in the level of perceived exchange risk as the domestic currency becomes overvalued (α). A clear implication is that a stable program involving pre-announcement requires the implementation of consistent and coordinated fiscal, trade, and monetary policies.

There is, thus, an implicit trade-off involved in using a preannounced exchange rate policy: a higher probability of a faster reduction in inflation can be obtained at the cost of a higher probability that the adjustment process may be unstable. This also suggests that the authorities can improve this trade-off if they utilize this exchange rate policy only after substantive measures have been taken to reduce or eliminate domestically created monetary disequilibrium and to open the economy to international trade.

III. Summary

This paper has discussed the rationale for using a policy of preannouncing the exchange rate as an anti-inflationary instrument and has analyzed, using a stylized theoretical model, the conditions under which a stabilization program that includes the preannouncement of the exchange rate will indeed result in a reduction of inflationary pressures.

It is clear that the preannouncement of the exchange rate can be used only as an integral part of a comprehensive stabilization program. It may be a useful instrument, particularly within the framework of anti-inflationary programs that rely heavily on the opening of goods and financial markets to international influences. In this context, a reduction of exchange risk may speed up the integration of domestic and foreign markets, bringing about a more rapid reduction in the pace of inflation. However, the introduction of preannouncement increases the potential instability of the system due mainly to the possibility that capital flows could create monetary pressures that reduce the credibility of the system and make it unsustainable. That is particularly possible when the trade reforms, including import liberalization, have not been extensively implemented and when the government’s financing requirements imply a fiscal deficit and therefore a rate of expansion of domestic credit inconsistent with the preannounced rate of devaluation.

While preannouncement of the exchange rate path can generate these potential benefits, it also reduces the authorities’ flexibility in meeting unanticipated transitory shocks to the economy, especially those originating abroad but also those associated with unforeseen domestic events. The “freezing” of the path for the exchange rate means that external shocks to the balance of payments must be absorbed into the money supply, which may well require an offsetting monetary stabilization action. The reduction of uncertainty in the foreign exchange market may, therefore, be obtained at the cost of increased instability in the domestic money market.36

Even when a policy of preannouncing exchange rates leads to a decline in the rate of inflation, the specification of an appropriate exchange rate path is indeed difficult. When the exchange rate is used as part of an anti-inflation program, a relatively low rate of depreciation is clearly an attractive policy option. For a preannounced exchange rate path to be credible and sustainable, however, the rate of depreciation should be consistent with the rate of domestic credit creation implied by the fiscal requirements and should not result in persistent and growing departures from purchasing power parity and in the emergence of a large disequilibrium in the balance of payments. A growing overvaluation of the domestic currency can have serious effects on the balance on current account and on employment. Such an overvaluation will also militate against the redistribution of resources from the import competing sector to the export sector that the authorities work to encourage via reductions in tariffs and in other restrictions on imports.

I. Notation
M=nominal stock of money (bank deposits)
rM=rate of interest on bank deposits
rL=rate of interest on bank loans
L=stock of bank loans
P=domestic price level
π=rate of change in domestic price level
πe=expected rate of change in domestic price level
Y=real income
K=required reserve ratio
X=exchange rate (domestic currency/foreign currency)
=rate of change in exchange rate
e=expected rate of change in exchange rate
F=foreign borrowing by banks
rF=cost of foreign borrowing inclusive of expected exchange rate movements and country risk premiums
πF=foreign rate of inflation
H=stock of base money
R=stock of foreign exchange reserves (in units of foreign currency)
N=stock of domestic credit
θ=degree of exchange risk
ϕ=markup factor
II. Derivation of Reduced Form Equations for rL and rM

To derive the reduced form equations for rL and rM, equations (1), (3), and (4) are used to yield

with f = XF/L

If these expressions are linearized about their steady-state values (rL, rM), then (where a bar denotes the steady-state value of a variable),

The determinant of this matrix will be given by W = −h2(1 − f) − (1 − K)h1 ≤ 0.

Solving for rL and rM,



which is taken as ≥ 0 but, since K is held constant in the analysis, could be accommodated ≤ 0.

which is taken as ≤ 0 as the most likely case.

III. Adjustment Process Without Preannounced Exchange Rate Policy

Combining equations (1) through (15):

A = 1 + τ2(q2 − 1)β1, is the determinant of the above matrix. A ≥ 0 is taken as the most likely case.

Solving for π˙ and f:

ḟ = Ψ2(π, f)

IV. Adjustment Process with Preannounced Exchange Rate Policy

Combining equations (1) through (15) and equation (20):

Assuming that the determinant of the coefficient matrix is positive as in Appendix III (A = 1 + τ2(q2 − 1)β1 ≥ 0). This implies

ḟ = Ψ4(π, f, θ)


    BuiterWillem H.1980“Implications for the Adjustment Process of International Asset Risks: Exchange Controls, Intervention and Policy Risk, and Sovereign Risk,”Working Paper No. 516 National Bureau of Economic Research (Cambridge, Mass. July).

    GerakisAndreas S. and DeborahDanker“Forward Markets: A Review of Theory, Practice, and Recent Developments” (unpublishedInternational Monetary FundJanuary271977).

    McKinnonRonald I.“Monetary Control and the Crawling Peg,” in The Crawling Peg: Past Performance and Future Prospectsed. byJohnWilliamson (it is scheduled to be published byCambridge University Press).

    McKinnonRonald I. and Donald J.Mathieson“Foreign Exchange and Financial Policies for Repressed Economies” (unpublishedInternational Monetary FundSeptember1981).

    TobinJames“A General Equilibrium Approach to Monetary Theory,”Journal of Money Credit and BankingVol. 1 (February1969) pp. 1529.

Mr. Blejer, an economist in the Financial Studies Division of the Research Department, is a graduate of the Hebrew University of Jerusalem and of the University of Chicago. He has been on the faculty of the Hebrew University and of Boston University.

Mr. Mathieson, Assistant Chief of the Financial Studies Division in the Research Department, holds degrees from the University of Illinois and Stanford University. He has taught at Columbia University.

The authors would like to thank Eliana Cardoso, Ronald I. McKinnon, and Mohsin S. Khan for their comments on an earlier draft of this paper. An earlier version was presented at the Econometric Society Meetings in Rio de Janeiro on July 17, 1981.

A number of other countries, including Brazil, Israel, Peru, and Portugal, have at various times implemented crawling peg systems with preannounced long-run targets (such as the maintenance of approximate purchasing power parity). These cases, however, do not fall within the definition of a policy of preannounced exchange rates, because the rules governing the rate of crawl were stated only in general terms and allowance was made for an explicit feedback between the exchange rate and domestic and foreign variables.

The Argentine program started in December 1978 and was abandoned in June 1981. It encompassed a number of announcements that gradually moved away from a fixed to a more indefinite time horizon for the announcement schedule. This eliminated one particular difficulty with a fixed time horizon schedule, namely, the uncertainty that could be created as the end of the time schedule approached. Some slight changes from the rates implied by earlier schedules were announced in late 1980 and the discrete devaluations that took place in February and March of 1981 were clear departures from the previously announced policy.

Chile began to use a policy of preannouncing exchange rates in February 1978. Prior to that time, Chile had a crawling peg system coupled with occasional discrete adjustments of the exchange rate. The schedules were followed until June 1979, when the peso was devalued by 5.7 per cent and fixed at the new level. This depreciation essentially involved moving the exchange rate to a value close to that scheduled to be attained at the end of 1979.

Since December 1978, the Central Bank of Uruguay has periodically published schedules of daily exchange rates for the Uruguayan peso in terms of the U. S. dollar. To increase the credibility of the policy, Uruguay has provided exchange insurance and guarantees based on the announced schedule.

If the nominal interest rates on all financial assets are market determined and reflect the expected rate of inflation, then the impact of a lower rate of inflation would be to increase currency holdings. If interest rates on other financial assets do not fully reflect expected inflation or a lower rate of inflation is also accompanied by a lower variance of inflation, then all financial assets will be stimulated.

Another option available to policymakers to reduce inflationary expectations is the announcement of rates of growth of key monetary aggregates together with a policy of letting the exchange rate be determined by market forces. Such a policy is more typical of the type pursued in the major industrial countries. It has been claimed that this option was not followed in the developing countries, apparently because there was concern about how smoothly the foreign exchange market would function after a long period of control and whether problems would be created by seasonality and lumpiness of transactions, which would increase the variability of the exchange rate and thereby raise exchange rate uncertainty in the absence of well-developed forward foreign exchange markets.

A similar effect could be obtained by moving to a fixed exchange rate system. However, pegging the exchange rate is not a credible option when the stabilization program is confronted with a very high and persistent internal rate of inflation. If the private sector does not believe that such an exchange rate policy could be sustained in the face of high domestic inflation, there would be little reduction in perceived exchange rate risk. A similar problem would arise if a policy of preannouncing the exchange rate creates a degree of overvaluation of the domestic currency that is viewed as unsustainable. This last point will be taken up in detail later.

In the presence of a forward foreign exchange market, the domestic interest rate should equal (assuming no risk premium) the covered foreign rate, that is, the foreign nominal rate plus the forward premium.

For a discussion of the effects of international asset risks, see Buiter (1980).

These country-specific factors will affect only the domestic level of interest rates and not the international rates.

This abstracts from the effects of economic growth. In a world of economies growing at different rates, there could be continuing capital flows, even with an equilibrium interest rate differential, as a result of the expansion of private portfolios containing both domestic and foreign assets.

If an efficient forward market has been operating prior to the announcement, the reduction in risk would be much smaller. In most developing countries, however, forward foreign exchange markets are not well developed. See Gerakis and Danker (1977).

This assumes that the preannounced path implies a rate of devaluation equal to the one previously expected, on average, by the market. It does not reduce, therefore, the expected rate of devaluation but reduces the risk premium, because under the new policy, expectations are held with more certainty. If the policy also implies a lower rate of devaluation than previously expected, there will be a further once-and-for-all capital inflow and a larger reduction in the domestic interest rate and in the differential, provided again that the subsequent announced path is credible.

The relative importance of domestic and foreign variables in the determination of the interest rate depends on the degree of segmentation of the domestic market and the extent of the restrictions on capital movements.

This is similar to the effect in the case of full market integration, and the smaller interest differential reflects the lower risk premium.

This inflow of capital reflects primarily capital market considerations and may not be consistent with the private sector’s desired accumulation of monetary assets. However, if the magnitude of the capital inflow is determined solely by the increase in the money demand caused by a reduction in inflationary expectations following the preannouncement, no inflationary pressures need arise.

It should be mentioned, however, that, in those cases where the preannouncement itself has the immediate effect of reducing inflationary expectations, this effect will tend to reinforce rather than offset the liquidity effect. As the monetary base increases, however, the expectations effect will tend to affect interest rates as described above. It is also possible that government intervention in the domestic credit market (direct or through government enterprises) will tend to sustain the interest differential for a period longer than that expected. This has been observed in a number of countries following their opening of the capital market.

Currency holdings could be added with little difficulty as long as the currency-to-deposit ratio is taken as fixed.

See Appendix I for a summary of notation.

Since the stock of capital is taken as fixed, there cannot be independent adjustment equations for both loans and deposits. The specification of the adjustment equation for either portfolio component, along with equation (1), necessarily describes the adjustment equation for the other component.

Although the analysis would be more complex if the demand for money was also a negative function of the expected real loan rate, the results would not be greatly altered as long as money and loans were viewed as gross substitutes in the portfolio and the effect of both rates is not fully symmetrical (see Tobin (1969)).

It is assumed that the central bank does not pay interest on bank reserves.

fd reflects banking system portfolio behavior in the sense that the fd(0) reflects the equilibrium ratio of foreign liabilities to loans. This ratio would also be affected by a variety of risk and institutional factors. The markup factor (φ) may also vary as interest rate ceilings are removed, entry of banks takes place, and the reserve ratio is lowered. These factors are taken as exogenous.

Since the analysis is focused on the medium-term stability of the adjustment process, especially as it relates to the rate of inflation, it is assumed that the level of real domestic output is given.

It is assumed that the central bank issuance of net domestic credit is associated with financing the government deficit. There is no direct lending by the central bank to the private sector.

ε1 is a constant which allows for the possibility of a secular divergence between domestic and international inflation. The analysis would be unaffected by setting ε1 = 0.

Equation (13) implies that capital inflows are determined by the differential between the domestic loan rate and the foreign interest rate. Ronald McKinnon has pointed out to the authors that in a financially repressed economy with a high required reserve ratio there may be simultaneously an inflow associated with a domestic loan rate that exceeds the equivalent foreign rate and an outflow associated with a domestic deposit rate that is below the comparable foreign rate. The focus here is on the capital inflow, since this has been the most common net outcome.

See Appendix III for the derivation of these equations. Since the exchange rate path is not being preannounced, the level of exchange risk (θ) is being taken as a constant. This assumption will be relaxed in the next subsection.

The values of the partial derivatives are given in Appendix III.


Clearly, even if π˙/∂π0, the system will be stable π˙/∂π+f˙/f0, which is equivalent to

This reflects the fact that an oligopolistic banking system will attempt to maintain its level of profits in the face of a higher cost of funds by reducing the time deposit rate.

If π˙/∂π+f˙/f0, then an unstable adjustment process would be generated. In this situation, capital inflows would generate a rate of monetary expansion inconsistent with a declining rate of inflation.

t0 represents time at the beginning of the reform.

This implies that the level of exchange risk (θ) is a cumulative function of the deviation from relative purchasing power parity.

At time T, when full portfolio equilibrium is achieved, it will be assumed that α0(Tt0) = αε2.

In terms of Figure 1, this means that the economy will initially move from point A to a point to the right, but that new point could be either above or below the old point A.

See Appendix IV for the derivation of equations (22) through (24) and for the definition of the partial derivatives.

This problem is, of course, not confined to programs including the preannouncement of exchange rates but arises whenever exchange rates are not perfectly flexible. This also ignores the possibility that some fiscal policy measures could also be used to counter external shocks.

Other Resources Citing This Publication