PETER J. MONTIEL AND JONATHAN D. OSTRY
At the time that most industrialized countries abandoned fixed exchange rates, the vast majority of developing countries continued to maintain them, typically linking their currencies to that of a major industrial country, usually the US dollar. In response to increases in international and domestic instability over the past fifteen years, however, policymakers in a number of developing countries believed that the costs associated with maintaining fixed exchange rates (for example, the difficulty of maintaining international competitiveness in the face of high domestic rates of inflation) were becoming large relative to the benefits usually associated with such regimes (relating primarily to the role of the exchange rate as a nominal anchor for the domestic price level).
Thus, among the developing country members of the IMF, the share of countries pegged to a single currency dropped from nearly two thirds in 1976 to less than one third in 1992. The counterpart of this change was a corresponding increase in the share of developing countries adopting more flexible arrangements. In most of these countries, the official exchange rate is changed frequently in accordance with some rule, which often links changes in the official exchange rate to the difference between domestic and foreign rates of inflation. Such rules are justified on the grounds that they help to maintain competitiveness, because they keep the real effective exchange rate close to its purchasing power parity (PPP) level (that is, the level at which a unit of currency can buy the same bundle of goods in all countries). For this reason they are generally referred to as real exchange rate rules.
Most commonly, adoption of a real exchange rate rule in a developing country, frequently in conjunction with an IMF/World Bank-supported adjustment program, entails that policymakers manage the nominal exchange rate in such a way as to keep the real exchange rate from appreciating relative to some initial baseline value. Such a policy can help to maintain the competitiveness of producers of traded goods, and hence is likely to expand a country’s export markets and to increase its growth potential.
For example, active exchange rate management was used in the mid-1980s to prevent real exchange rate appreciation—and indeed to promote depreciation—in Brazil, Chile, Mexico, and Turkey among others—with the beneficial effect of a rapid expansion of exports for these countries In contrast, substantial real appreciations in the late 1970s and early 1980s had a predictably negative impact on export performance in a number of developing countries, including Chile, Mexico, and Nigeria. In addition, reducing the variability of the real exchange rate (as under real exchange rate targeting) appears to have been one of the factors responsible for strong investment and export performance in a number of developing countries.
Despite the apparent advantages of active exchange rate management, certain unfavorable aspects of macroeconomic performance in the context of real exchange rate rules have led governments to adopt stabilization programs where the exchange rate is fixed. These so-called exchange-rate based stabilization programs have usually been justified on the grounds that only under a regime of fixed nominal exchange rates can the latter fulfill its traditional role as nominal anchor for the domestic price level. This may help to explain why, for example, in the high inflation “Southern Cone” countries of Latin America in the late 1970s, PPP-based rules were replaced by a preannounced crawl (the “tablita”); and why in the near-hyperinflation countries of Argentina, Brazil, and Mexico during the mid-1980s, the “heterodox” stabilization programs that were adopted featured a fixed nominal exchange rate as a major component.
In other cases, though, macroeconomic performance appears to have been more favorable, particularly when, as seems to have been the case in a number of developing countries (Colombia), activist exchange rate policies designed to offset inflation rate differentials were accompanied for long periods by fairly stable and moderate inflation. This suggests that the fear that real exchange rate rules always lead to hyperinflation may not be warranted, and that the precise nature of the effects of such rules on inflation merits further scrutiny.
Are such rules hyperinflationary?
The logic of the view that real exchange rate rules lead to a loss of control over the domestic inflationary process is clear enough. Adoption of such rules implies that the nominal exchange rate and, through the balance of payments, the money supply, are both indexed to the domestic price level. Shocks to the domestic price level will therefore tend to be fully accommodated by a faster rate of exchange rate depreciation and a faster rate of monetary growth. Under these conditions, therefore, there is no longer any exogenous nominal anchor to tie down the domestic price level.
While no one would dispute the fact that the adoption of a PPP rule for the nominal exchange rate prevents the latter from serving as a nominal anchor, there is room to disagree about the precise nature of the effect on the domestic inflationary process of the types of disturbances typically experienced by developing countries when a real exchange rate rule is in place. While hyperinflation appears to be a possible response to shocks, other outcomes—including once-off changes in the rate of inflation—are also possible, and perhaps more likely than hyperinflation in the context of many real world cases.
Adjusting to real shocks
When governments actively manage the exchange rate to achieve some target for its real value, macroeconomic adjustment to both internal and external disturbances differs sharply from that which occurs under fixed exchange rates. This is because in the latter case, changes in the real exchange rate can help to restore macroeconomic equilibrium when a shock occurs, whereas under real exchange rate targeting, some other variable will have to adjust. Since the types of disturbances experienced by developing countries (for example, terms of trade shocks related to commodity price changes, or various fiscal policy measures) frequently require a shift of resources from nontraded to traded goods, targeting the real exchange rate (which effectively fixes the relative price of these two goods) is likely to fundamentally alter the nature of macroeconomic adjustment in these countries.
To illustrate, consider how a developing country would adjust to an increase in government expenditure that fell primarily on non-traded goods (the nature of the adjustment would be similar under a favorable terms of trade shock, say a coffee or oil price boom). The increase in fiscal spending would tend to create excess demand for nontraded goods. With a fixed nominal exchange rate, and hence a flexible real exchange rate, the relative price of nontraded goods would rise to restore equilibrium. This would occur via an increase in the price of home goods relative to the (fixed—given the fixed exchange rate) price of tradables, that is, an appreciation of the real exchange rate.
Matters are quite different under real exchange rate targeting. In this case, the appreciation of the real exchange rate towards its new higher equilibrium level is no longer possible because the government is committed to adjusting the nominal exchange rate (effectively the domestic price of tradables) in such a way that the price of nontradables relative to tradables (the real exchange rate) is constant. To restore equilibrium in this case requires some substitute for the real exchange rate appreciation that would take place if the country followed a policy of fixed nominal exchange rates. It turns out that this substitute is a reduction in the real value of the private sector’s financial wealth.
For a more technical discussion by the authors, see “External Shocks and Inflation in Developing Countries Under a Real Exchange Rate Rule,” IMF Working Paper WP/92/75, available from the authors.
How does this reduction come about? As the price of nontraded goods rises in response to the increase in government spending, the authorities depreciate the nominal exchange rate to prevent an appreciation of the real exchange rate, thereby causing an adjustment in the overall price level. An increase in the price level indeed lowers the real value of private wealth as long as the latter is not fully indexed to price level changes, thereby reducing demand for all goods, including nontradables, and helping to restore internal balance.
This initial jump in the price level is not sufficient to restore macroeconomic equilibrium, however. The reason is that reductions in wealth are generally associated with increases in private sector savings, which, other things being equal, would generate a surplus in the current account of the balance of payments. The latter implies that the private sector is accumulating financial claims on the rest of the world. Over time, this creates additional excess demand pressure in the market for home goods, the elimination of which requires a continuous increase in the general level of prices. Only such a sustained increase in the price level will be sufficient to offset the impact on the nontraded goods market of the increase in private wealth coming through the balance of payments. Notice crucially, however, that although prices indeed rise continuously, the new inflation is constant in this process. The rise in government spending does not, in this case at least, lead the economy toward hyperinflation.
What, then, are we to make of the view that real exchange rate targeting leads to hyperinflation? The key to understanding how hyperinflation is avoided in the previous example relates to the disposition of the inflation tax on monetary balances, which in the first instance accrues to the central bank. If the central bank keeps the proceeds of the inflation tax (using them to accumulate claims on the rest of the world in its own right) or transfers them to the government (which uses them to increase its expenditures), then we have the outcome presented above, a higher but stable once-off increase in the rate of inflation, but no hyperinflation. If, however, the central bank transfers the proceeds of the inflation tax to the government, which then uses them to reduce its (lump-sum) taxes on the private sector, it is easy to see that hyperinflation can readily occur. This is because with the private sector effectively not paying the inflation tax (given the rebates it receives from the government), the accumulation of wealth that is the counterpart of the current account surplus is no longer offset by an erosion of the monetary component of wealth, that is, the inflation tax. This means that there is no end to the process of wealth accumulation, and hence no end to the excess demand pressure in the market for home goods, and thereby on the price level.
The case of hyperinflation, though, is likely to be a theoretical curiosum, since it assumes that the inflation tax will be rebated to the private sector in the form of reductions in other lump-sum taxes, something that is not very common in developing countries, where the inflation tax is typically used to finance the government’s budget. Thus, from an empirical standpoint, it would seem that once-off changes in the inflation rate are more likely to be observed than hyperinflation.
Role of monetary policy
When the exchange rate cannot be a nominal anchor for the domestic price level, as under real exchange rate targeting, it is natural to think of monetary policy as substituting for the exchange rate in this role. Surely, since “inflation is everywhere and always a monetary phenomenon,” it should be possible to design some monetary policy that would be able to offset the destabilizing effects on the price level that arise from the pursuit of a real exchange rate target.
It turns out, however, that there is relatively little that monetary policy can do here. There are really two cases to consider, depending on whether capital is mobile or immobile internationally. In the case of high capital mobility, the only available monetary policy instrument is the stock of credit. Suppose that the government decides to fix this stock in nominal terms, that is to impose a credit freeze on the economy, in order to bring about price level stability. This is admittedly an extreme example but it illustrates the issues. With a credit freeze in place and with the private sector having access to the international capital market, borrowers who find themselves unable to borrow domestically can borrow abroad instead. Because the private sector’s overall demand for financial assets is unaffected by the credit freeze, the latter’s only effect is to increase foreign borrowing by the private sector (i.e., to increase the surplus in the capital account of the balance of payments). The inflow that comes through this channel—like the inflows coming through the external current account—add to the money supply. Since inflation is indeed a monetary phenomenon, and since inflows through the balance of payments (which are the sum of the inflows in the current and capital accounts) will be the same with or without the credit freeze (because the demand for them is the same in either case and the perfect capital mobility assumption guarantees an infinitely elastic supply); monetary policy, therefore, will be powerless to affect the inflation rate under real exchange rate targeting.
Now it might be countered that the assumption of perfect capital mobility is to blame here, and, furthermore, that such an assumption is not really applicable to many if not most developing countries. This is not the case, however. Even with a closed capital account, monetary policy is unlikely to have anything but a short-run effect on the inflationary process under real exchange rate targeting. Although the argument is a little more complicated in this case, an example is nevertheless illustrative. Suppose policymakers impose complete capital controls, that is, they prohibit all international borrowing and lending. The likely outcome will be the emergence of an unofficial market (or parallel market) in which foreign exchange is traded at a market-determined price. With capital controls, policymakers can control the supply of money in the domestic economy because they can sterilize inflows that come through the current account of the balance of payments; they do not need to sterilize capital flows (which were potentially infinite under perfect capital mobility) because the capital account is assumed to be closed.
Suppose, then, in analogy to the previous case, that the government targets a zero growth in the nominal money supply in order to achieve price stability. In the short run, such a policy may indeed succeed in lowering inflation. But ultimately the private sector will attempt to satisfy its demand for money by trading in the parallel market. This will cause the market-determined price of foreign exchange to differ from the official price. Moreover, the difference between these two prices will tend to grow ever larger because, assuming that the underlying change in fiscal policy is permanent, the resulting inflow through the balance of payments (which is sterilized by the authorities) will be permanent. The need for perpetual credit contraction to offset reserve inflows means that the pressure on the market-determined exchange rate is continuous. Eventually, the ever-widening gap between official and market exchange rates must compromise the government’s ability to maintain capital controls. At such a point, the capital controls have to be abandoned and we return to a world of capital mobility in which, as previously argued, the inflation rate is unaffected by the government’s chosen monetary policy.
To conclude …
Real exchange rate targeting is likely to have a direct impact on the inflation process in developing countries. Monetary policy, how ever, is unlikely to be successful in mitigating these effects, except in the short run. Nonetheless, following a real exchange rate rule does not of itself imply that the country will experience hyperinflation in the presence of real shocks.