The relationship between the exchange rate regime and financial discipline is a central theme in discussions of the relative merits of fixed and flexible regimes. A conventional argument in the early literature was that, by adopting a flexible regime, the authorities would be free to formulate their monetary policy in accordance with their domestic objectives, allowing exchange rate adjustment to equilibrate the balance of payments. More recently, however, the two basic premises of this argument have been challenged. First, it has become increasingly clear that a flexible exchange rate does not free the authorities from the external constraint on their domestic policies. In the presence of capital mobility, domestic policies greatly influence movements in the exchange rate and the current account balance, and these movements, in turn, constrain domestic policies.26 The second issue relates to the critical assumption underlying the earlier view that the authorities would effectively use their independence in policymaking to achieve their domestic objectives. This issue is at the heart of the current debate on the relationship between exchange rate regimes and financial stability.
Proponents of fixed exchange rates hold the view that the adoption of such a regime imposes a degree of financial discipline that would be absent under a flexible regime. By discouraging recourse to inflationary finance, a fixed regime would facilitate the attainment of price stability. Advocates of greater exchange rate flexibility, for their part, maintain that financial discipline, if absent, is unlikely to be instilled by the adoption of a fixed exchange rate. Instead, the announcement of a fixed exchange rate would merely result in a succession of financial crises followed by devaluations, introducing a high degree of instability into the behavior of the real exchange rate, thereby increasing uncertainty and disrupting trade and investment flows.
These issues are not easy to settle on empirical grounds. The comparison of the experience of different groups of countries provided in Section II indicates that the inflation performance of the group of countries that have operated under a fixed exchange rate regime has been, on the whole, superior to that of the group operating under more flexible arrangements. On the face of it, this comparison would indicate that a pegged exchange rate arrangement is effective in imposing financial discipline. But there are important reasons for being wary of drawing such an inference from simple comparisons of country experiences under different exchange rate regimes.
First, an examination of the countries that maintained pegged exchange rate arrangements over a given period neglects the experience of the countries that initially adopted a pegged arrangement, but were forced to abandon it because of the loss of competitiveness associated with excessively expansionary domestic policies, that is, because of the failure to maintain financial discipline under a pegged regime. Second, the experiences of the countries in each of the two categories have varied widely. For instance, within the pegged group, many countries in Africa and the Western Hemisphere have exhibited a significant deterioration in their price performance, which has required periodic and large devaluations. By contrast, many of the Asian countries that have adopted flexible arrangements have also achieved very favorable price performance. Thus, the evidence is far from conclusive in supporting the contention that a pegged exchange rate arrangement is either necessary or sufficient for price stability. In any event, such comparisons beg the question of cause and effect: it is by no means obvious whether pegged exchange rates have induced greater price stability, or whether greater price stability has permitted the maintenance of a fixed rate.
Since casual empiricism does not take the analysis very far, this section attempts to evaluate the relationship between the exchange rate regime and financial discipline on analytical grounds, using the perspective of recent developments in macroeconomic theory. The following five questions are addressed: (1) What is the nature of the financial constraint imposed by the maintenance of a fixed exchange rate? (2) Does a fixed exchange rate promote the adoption of financial policies conducive to price stability? (3) Under what conditions can a commitment to a fixed rate be made credible? (4) What are the advantages of using the exchange rate, as opposed to other variables, to anchor the system? (5) What does an exchange rate adjustment cost relative to fiscal contraction in bringing about external adjustment?
Financial Constraints Imposed by a Fixed Exchange Rate
The conditions under which a small open economy can maintain a fixed exchange rate have been studied intensively in the recent literature on “balance of payments crises,” which had its inception with the work of Krugman (1979). To derive these conditions in a general framework, consider a small open economy that maintains a fixed exchange rate for both trade and financial transactions.27 In such circumstances, the world inflation rate determines the domestic inflation rate. Assuming a unitary income elasticity of money demand, the rate of growth of domestic money is determined by the rate of growth of real output plus the rate of world inflation.
Suppose now that the monetary authorities require that some minimum fraction of the domestic money supply be backed by foreign exchange reserves in order to defend the fixed parity. If foreign exchange reserves fall below this critical level, the fixed parity would be abandoned. It then follows that the rate of growth of domestic credit cannot permanently exceed the rate of growth of the nominal demand for money. This is so because, if domestic credit grows faster than the demand for money for an extended period of time, domestically created money would gradually replace foreign exchange reserves in the central bank’s balance sheet, and the authorities’ reserve threshold would soon be breached.28
According to this literature, viability of a fixed exchange rate requires that the underlying rate of growth of domestic credit not exceed a specified upper bound. The upper bound in question is determined by the rate of growth of the demand for money, which depends on the world rate of inflation, the rate of growth of real income, and the income elasticity of demand for money. The permissible rate of growth of domestic credit would be larger (i.e., the degree of imposed financial discipline would be less) for (1) a higher rate of world inflation, (2) a higher rate of growth of real income, and (3) a larger income elasticity of demand for money.
A fixed exchange rate regime imposes a limit only on the long-run rate of growth of credit. Transitory episodes of rapid credit expansion are not incompatible with the maintenance of a fixed parity, as long as such episodes are perceived by the private sector to be reversible. The duration of periods of rapid credit expansion—and thus the effective degree of discipline—depends both on the initial stock of reserves and on the authorities’ reserve threshold. The larger the initial stock of reserves relative to the target, the longer the period that excessive credit growth can be sustained, and thus the weaker the degree of discipline imposed by the fixed regime.
The financial discipline imposed by the reserve constraint may be weakened to the extent that the central bank can resort to external borrowing to replenish reserves and sustain the exchange rate parity. In such a circumstance, the reserve threshold would not be binding, and credit expansion could proceed at an excessively rapid pace. This process, however, cannot continue unabated because foreign creditors will not be willing to finance the authorities’ intervention operations to defend the rate indefinitely.29 In general, creditors will be willing to lend to the government only as long as it is perceived to be financially solvent. The solvency requirement ultimately sets the financial conditions under which a fixed exchange rate will be viable, and captures the financial discipline imposed on the economy.
The public sector will be perceived to be solvent if the present value of its anticipated future debt-service payments (amortization plus interest) is at least equal to the face value of its existing debt. The resources available to service debt, in turn, consist of the stream of anticipated future primary surpluses of the public sector plus anticipated future seignorage revenue. Thus, the external financing in question will be available as long as the present value of anticipated primary surpluses plus seignorage is at least as great as the face value of the public sector’s net debt.
In practice, an assessment of the solvency of the public sector is not an easy task because it involves a judgment on the government’s power to control its net worth through tax and expenditure legislation. The purpose here is not to delve into this complicated issue, but rather to point out the fundamental link between the government’s fiscal operations and the viability of a fixed exchange rate. By setting the economy’s steady-state inflation rate equal to that of its trading partners, a fixed exchange rate sets a limit on the future revenue available to the public sector from the inflation tax. Given the initial stock of debt and the solvency requirement, this limit on seignorage puts a lower bound on the present value of the stream of future primary surpluses of the public sector. Were the exchange rate to depreciate, smaller surpluses could be accommodated by printing money and accepting a higher domestic rate of inflation. If the solvency condition is violated under a fixed exchange rate, the authorities would not be able to borrow, and the exchange rate could be sustained only as long as the authorities’ own reserves are not depleted. In the absence of a policy correction, a speculative attack would soon ensue—much as predicted in the “balance of payments crisis” literature—and the fixed rate could no longer be defended.
In summary, a fixed exchange rate requires that a country be able to maintain fiscal discipline, in the sense that its primary surpluses satisfy an intertemporal budget constraint. This constraint will be looser, the greater the initial level of international reserves, the smaller the public sector’s initial net debt, and the greater the revenue from seignorage.
Implications of an Announced Fixed Exchange Rate for Financial Stability
A fixed exchange rate imposes financial discipline on the authorities only if the exchange rate is fixed permanently and not adjusted periodically. Recurrent devaluations would enable the government to wipe out, from time to time, that part of its debt denominated in domestic currency, thereby easing its solvency constraint. The issue, then, is whether the incentive structure under a pegged arrangement is conducive to the adoption of an immutably fixed exchange rate (with attendant financial discipline) or to periodic devaluations.30
Consider the case in which a government is cognizant of the cost of inflation arising both from the adverse effects of the associated rise in uncertainty on domestic investment and output over the long run as well as from unfavorable distributional effects. Notwithstanding these longer-run costs of inflation, the government may still be tempted to generate an inflationary “surprise” to raise domestic output and employment in the short run. To the extent that wage contracts are based on the expected rate of inflation, an inflationary surprise (which would result in an actual inflation rate in excess of the anticipated rate) would lead to lower real wages, a greater demand for labor, and a higher supply of output. In an open economy, such inflationary surprises would lead to recurrent devaluations of the currency.
The public would, of course, be inevitably aware of the authorities’ incentive to generate inflationary surprises which would be validated by devaluations. In such an environment, the expected rate of inflation would depend on whether the authorities, in fact, have the discretion to generate inflationary surprises. The more discretion the authorities are perceived to have, the higher the expected rate of inflation, and the higher the actual rate of inflation that the authorities actually have to generate to engineer an “inflation surprise.” If the marginal welfare cost of an increase in inflation rises with higher rates of inflation, then the economy will ultimately settle at an inflation rate sufficiently high so that the cost of a marginal increase in “surprise” inflation is just equal to its benefits. Only in such a high-inflation equilibrium will the authorities fail to have an incentive to deviate from the expected inflation path, and will therefore be able to convince the public that they will not allow the inflation rate to rise further. Ironically, the authorities ultimately forgo any benefits associated with inflationary surprises but only after the inflation rate has risen to (a high) equilibrium level. Of course, since inflation is assumed to affect social welfare adversely, the authorities would have attained a higher level of welfare if, to begin with, they had eschewed their discretion to generate inflationary surprises in a credible manner and thus maintained price stability.
When policymakers are trying to maximize economic welfare, the nature of the incentives facing them may be such that a stable price equilibrium would not in fact emerge. As long as a marginal “surprise” deviation from stable prices appears to be beneficial, the private sector will expect the authorities to try to produce such inflationary surprises. In this setting, the authorities’ efforts to stabilize prices would introduce negative inflationary surprises (i.e., actual inflation would fall below expected inflation), leading to higher-than-expected real wages and a correspondingly lower supply of output. Thus, the authorities may find it preferable to accept a high-inflation outcome with recurrent devaluations, rather than attempt to “squeeze out” inflation that would result in a contraction of output.
The solution to the problem would be for policymakers to convince the private sector that it will forgo the option to inflate—in other words, to precommit to a policy of price stability. The problem with this solution, however, is that the authorities may find it difficult to make their commitment credible. Unless they can find a way to do so, disbelief on the part of the private sector may result in worse economic outcomes than a high-inflation equilibrium. Thus, policymakers may find it desirable to rule out inflationary surprises by undertaking to fix the exchange rate indefinitely if they could do so convincingly.
Achievement of Credibility
An obvious way to make the fixed exchange rate regime credible would be for the authorities to surrender the power to alter the exchange rate. This could be achieved, for example, by forming a currency union under which a group of countries adopts a common currency and fixes its parity against a major currency—the CFA franc zone provides an example of such an arrangement. A less rigid arrangement would be one similar to the European Monetary System (EMS) under which the exchange rate can be adjusted periodically, although such adjustments would require international agreement.31 The important point is that, to be credible, such arrangements will have to be based on certain institutional arrangements that make it costly to alter the exchange rate.
There are, however, costs associated with forgoing the use of devaluation. The reason is that, as outlined in previous sections, it may sometimes be optimal to alter the exchange rate in response to permanent exogenous shocks. By forgoing the use of the exchange rate instrument, the authorities are forced to rely entirely on financial policies to realign relative prices of traded and nontraded goods in the event of a permanent change in the equilibrium real exchange rate resulting from, say, a terms of trade shock. Such an approach may entail a prolonged period of contraction and price deflation, which may prove more costly than a once-and-for-all adjustment in the exchange rate. (This point is considered more fully below.) Thus, while it would be desirable for policymakers to attach a greater weight to price stability than that perceived by society as a whole, this weight should not be so large that the exchange rate is never moved.
In principle, an exchange rate rule could be designed according to which a devaluation would be undertaken only in response to large (permanent) exogenous shocks, but not in instances of misalignments in the real exchange rate arising from expansionary financial policies. Thus, a period of unsustainable fiscal deficits would be countered by fiscal overcorrection, rather than by exchange rate adjustment. In practice, however, it would be difficult to implement such exchange rate rules, which must be contingent on the specific state of the economy. It is not always possible to separate the impact on competitiveness of expansionary financial policies from that of exogenous shocks. Furthermore, even if earlier fiscal expansions are at the root of the exchange rate overvaluation, it may be optimal to use the exchange rate instrument to realign relative prices if the costs of fiscal overcorrection exceed the cost of undermining credibility by adjusting the exchange rate.
Paradoxically, it would be advantageous to rely on the exchange rate instrument in the two extreme cases in which the authorities either lack credibility completely, or have established it firmly. In the case that the authorities lack credibility, they may find it preferable to accept frequent devaluations and a high level of inflation rather than to encounter the output cost of deflationary surprises. Alternatively, if credibility is firmly established— for example, through a change in economic administration—the authorities may be able to devalue at the beginning of a program and still convince the public that the devaluation will not be repeated. More generally, a practical solution for countries falling in between these two extreme cases may be to grant considerable autonomy to a central bank that has a reputation for financial conservatism.32 This approach would demonstrate that the authorities attach a greater weight to price stability than that warranted by the social cost of inflation, but still retain the flexibility to respond to external shocks and to exogenous changes in the policymaking environment.
The degree to which the authorities may wish to “tie their hands” in the use of the exchange rate instrument depends partly on the expected frequency and scale of external shocks. Assuming that the economy is not subjected to large and frequent adverse shocks, a case can be made for the authorities to precommit to a fixed exchange rate as a means of demonstrating their resolve to maintain financial discipline. It may therefore be desirable for the authorities to adopt an institutional arrangement that imposes a large political cost to reneging on such precommitment.
The Choice of Anchor for Financial Stability
The preceding discussion has focused on the role of a fixed exchange rate in establishing the authorities’ credibility in pursuing noninflationary policies. But it can be argued that such credibility could also be gained through other means, such as announcing an inflation or a monetary target. Clearly, the authorities need to target all three nominal variables—inflation, money supply, and the exchange rate—in a consistent manner, but the question is whether an announced commitment to target a particular variable (rather than the others) would be more binding and therefore more credible. Although it is difficult to be unequivocal on this issue, an announced fixing of the nominal exchange rate may be preferable to a target either for inflation or for the money supply.
The most direct means for the authorities to publicize their intention to refrain from inflationary policies is, of course, to announce an inflation target. However, the rate of inflation is not under the direct control of the authorities. Consequently, an inflation target that is not linked to specific policy commitments which can be readily monitored is not likely to be credible. By contrast, a monetary target can be defined in terms of specific behavior of monetary aggregates. The relationship between various monetary aggregates and inflation, however, is quite complex. Insofar as short-term variations in velocity are likely to exceed those in the equilibrium real exchange rate, a nominal exchange rate rule may be more desirable than a monetary rule. The main advantage of an exchange rate rule is its transparency. The exchange rate is readily observable at any instant, as opposed to inflation and money supply data which are provided by the authorities with a lag (which is sometimes quite long).33 By announcing a fixed nominal exchange rate, the authorities undertake to put into place all the necessary policies to establish and maintain price stability.
Assuming that a country chooses to adopt a fixed exchange rate, there still remains the choice between pegging to a single major currency or to a basket of currencies. In a world of generalized floating among the major currencies, pegging to a single currency (or for that matter to a basket of currencies) would still result in exchange rate fluctuations vis-à-vis the other currencies. Such fluctuations will impose certain costs by increasing the exchange rate risk for trade and capital transactions and by altering relative prices affecting production and investment decisions.34
Fixing to a basket of currencies has the advantage of reducing the average fluctuation of the domestic currency vis-à-vis other currencies, thus reducing the risk for those who have to take an open position in various currencies. However, a basket peg implies that all traders would bear the exchange rate risk. By contrast, if the domestic currency is pegged to, say, the U.S. dollar, the risk for market participants dealing in dollars would be eliminated, while all those dealing in other major currencies could cover themselves through the use of financial instruments available for forward transactions between the dollar and other major currencies.35 Consequently, a country without well-developed financial markets may find it advantageous to peg to a single major currency and thereby effectively expand the domain of its currency by allowing market participants to take advantage of services available for that major currency.36
Another consideration relevant to a single currency versus a basket peg is that a country may experience less inflation if it pegs to the currency of a major country that has a very low inflation performance. However, domestic inflation goals may not be met if the major currency to which the domestic currency is pegged experiences large real exchange rate movements against other major-currency countries that are also important trading partners of the home country. In this case, a basket peg may be preferable, with the inflation goal sought by varying the peg of the domestic currency relative to the basket.
On the whole, the choice of pegging to a single currency or a basket of currencies is likely to be of secondary importance, as long as the underlying financial policies are consistent with maintaining the announced peg. If a country’s predominant share of trade is in one major currency, a good case can be made for pegging to that major currency, particularly if domestic financial markets are not well developed. However, a case for pegging to a trade-weighted basket of currencies can be made if trade is highly diversified and real exchange rates among the major trading partners tend to undergo large movements. To the extent that a basket peg is adopted, it would be preferable to choose a basket composed of only a few major currencies, such as the SDR.37 Such an arrangement would be simpler to operate and would provide a more transparent signal of the authorities’ determination to use the exchange rate as an anchor for financial stability.
Exchange Rate Adjustment Versus Fiscal Contraction
The trade-off between exchange rate adjustment and fiscal contraction frequently presents a dilemma for the developing countries formulating adjustment programs. The issue is to what extent fiscal contraction could substitute for devaluation in achieving external adjustment, if the exchange rate is to assume the role of an anchor and remain fixed. While no definitive answer can be given to this question in the abstract, the following arguments can shed some light on this important policy issue.
Consider a country that has undertaken expansionary fiscal policies for a period of time, while maintaining a fixed exchange rate. Further assume that the associated rise in domestic inflation has led to an overvaluation of the real exchange rate, resulting in a deterioration in the external position. Successful adjustment in this stylized example would require, in the first place, a reduction in the fiscal deficit to a level that is sustainable over the longer term. Such a fiscal contraction would inevitably entail some output cost in the short run. In addition, the correction of the external imbalance would require an increase in the relative price of tradables in terms of nontradables—that is, a depreciation of the real exchange rate. Such a realignment of relative prices can be achieved either through a devaluation, which would directly raise the price of tradables, or through a period of fiscal “overcorrection,” which would lower the price of nontradables.
The decision on the appropriate course of adjustment would clearly depend on the relative costs of devaluation and fiscal overcorrection. The conventional wisdom on this issue holds that this decision would be critically influenced by the degree of flexibility in wages and prices. The more rigid wages and prices, the longer (and thus, the costlier) the required period of fiscal contraction. In such a situation, a devaluation, which corrects relative prices immediately, would avoid much of the output cost associated with a prolonged period of fiscal overcorrection, although it would still entail some contractionary costs in the short term.
In contrast, when domestic wages and prices are completely flexible, the contraction of aggregate demand brought about through a restrictive fiscal policy would lead to a downward adjustment of the domestic price of nontradables within a relatively short time. Thus, the real exchange rate would depreciate without an intervening nominal devaluation. The lower relative price of nontradables would increase demand for such goods, thus mitigating the associated output cost. Moreover, because improved competitiveness would contribute to external adjustment, the initial degree of fiscal contraction would not need to be as severe as in the sticky-price case to achieve the desired improvement in the external position. While arguments can be made for an initial exchange rate adjustment even in such a setting, the case for devaluation is clearly less compelling when domestic wages and prices are flexible.
The above criterion for choosing between an exchange rate adjustment or a fiscal contraction depends critically on the assumption that the degree of wage and price flexibility is a structural feature of the economy and is independent of the adjustment strategy pursued by the authorities. The analysis presented earlier in this section, however, has suggested that this assumption may not be valid. Nominal wages and prices tend to be set in a forward-looking fashion via explicit or implicit contracts. An important factor considered by economic agents in setting future wages and prices is the anticipated behavior of aggregate demand, that is, the likely evolution of the aggregate price level, which is directly related to the expected stance of macroeconomic policies. A demonstrated willingness by the authorities to adjust the exchange rate, rather than curtail demand, so as to correct an overvaluation of the exchange rate would inevitably induce private agents to resist a reduction in their wages. In other words, there is an element of self-fulfilling prophecy in the strategy which relies on exchange rate adjustment to avoid the recessionary effects of fiscal contraction in the presence of wage and price rigidities. That is, a predilection by the authorities to adjust the exchange rate frequently may reduce the credibility of an announced policy of demand restraint, thus creating the downward rigidity of wages and prices that would make a nominal devaluation preferable to fiscal adjustment.
An operational consequence of the above discussion is that while a commitment to manage the exchange rate flexibly may provide helpful assurances to producers of tradables (thereby supporting external adjustment), it may undermine the credibility of the government in adopting restrictive demand policies and thus make it more difficult to lower inflation without imposing an output cost. A corollary of this argument is that the effectiveness of pursuing a noninflationary adjustment strategy would be critically influenced by the policy record of the authorities. A country with a longstanding record of prudent financial policies could resort to a once-and-for-all devaluation, in response to, say, a substantial deterioration in its terms of trade, while maintaining financial stability. By contrast, a country with a long history of inflationary policies would find it difficult to fix the exchange rate after an initial devaluation unless it is prepared to accept a period of recession until it establishes its credibility in maintaining price stability. In this context, a visible structural change in the policy environment, such as a change in the economic administration or the adoption of an IMF-supported program that emphasizes conservative financial policies, could significantly enhance the credibility of the government in establishing financial stability, thus promoting wage-price flexibility and reducing the output cost of adjustment.
See Poole (1970), Turnovsky (1976), and Boyer (1978).
A crawling peg arrangement under which the rate of exchange rate depreciation is announced and fixed can, in principle, be regarded as a fixed regime. The analysis of this section can readily be extended to such crawling arrangements.
One insight of the “crisis” literature is that, with well-informed agents, the reserve threshold would be reached suddenly via a speculative attack; see Krugman (1979) and Obstfeld (1984).
This issue has been addressed by Buiter (1986). Notice that the currency can be defended by borrowing either abroad or at home. In the former case, the excess supply of domestic currency is exchanged for foreign exchange, while in the latter case it is exchanged for central bank liabilities (domestic public debt). Domestic borrowing by the central bank, however, is equivalent to reduced domestic credit expansion.
For a general theoretical treatment of this issue see Kydland and Prescott (1977), Barro and Gordon (1983), Canzoneri (1985), and Calvo and Guidotti (1989).
See Giavazzi and Pagano (1986) for a discussion of the EMS.
See Rogoff (1985).
It may also be argued that the international community is more aware of, and thus likely to scrutinize more closely, the behavior of the exchange rate than that of the monetary aggregates of a small country.
For a survey of the literature on the choice of an optimal peg, see Williamson (1982). Also, for a discussion of these issues in the context of the Asian countries, see Aghevli (1981).
See Mussa (1978).
In principle, under a basket peg, traders can buy a basket of currencies forward, but this arrangement is quite impractical.
Discussions of the SDR as an optimal peg are contained in Crockett and Nsouli (1976) and Williamson (1982).