Panel Discussion on Southern Cone
  • 1 0000000404811396 Monetary Fund

This discussion will concentrate on two major issues. The first concerns the optimal design of economic liberalization programs. The key question to be addressed is the proper sequencing of liberalization measures. Specifically, the order of economic liberalization of international transactions in goods and financial markets will be discussed. The second major issue concerns the optimal design of stabilization programs. The key question to be addressed is the choice between gradualistic or drastic applications of stabilization measures. The discussion will close with a few notes on the pegging of a currency to a basket of foreign exchange.


This discussion will concentrate on two major issues. The first concerns the optimal design of economic liberalization programs. The key question to be addressed is the proper sequencing of liberalization measures. Specifically, the order of economic liberalization of international transactions in goods and financial markets will be discussed. The second major issue concerns the optimal design of stabilization programs. The key question to be addressed is the choice between gradualistic or drastic applications of stabilization measures. The discussion will close with a few notes on the pegging of a currency to a basket of foreign exchange.

This discussion will concentrate on two major issues. The first concerns the optimal design of economic liberalization programs. The key question to be addressed is the proper sequencing of liberalization measures. Specifically, the order of economic liberalization of international transactions in goods and financial markets will be discussed. The second major issue concerns the optimal design of stabilization programs. The key question to be addressed is the choice between gradualistic or drastic applications of stabilization measures. The discussion will close with a few notes on the pegging of a currency to a basket of foreign exchange.

Since this discussant does not claim to possess either specific knowledge of or familiarity with the details of the various economic experiences of the Southern Cone countries, it is relevant to note that the main arguments presented do not apply only to this group of countries. In fact, the arguments are not based on any specific experience of any specific country but rather stem from general basic economic principles. As such, the key issues should also be applicable to other experiences of economic liberalization and stabilization programs.

sequencing of economic liberalization

The question of the optimal process of economic liberalization involves one of the most difficult aspects of the theory of economic policy, namely, the transition toward equilibrium. The prescription of economic theory is very clear when applied to “first-best” situations. Economic theory, however, is much more reserved when it comes to evaluation and comparison of distorted situations. The evaluation of alternative strategies for economic liberalization involves making comparisons among various distorted situations that characterize alternative paths an economy may adopt in its approach to undistorted, first-best equilibrium. Consequently, views about the “proper” sequencing of liberalization measures should be put forward with great care and modesty; economic theory does not provide an unambiguous answer to that important question. With these reservations in mind, this discussion will state what is likely to be the proper order of economic liberalization.

Consider an economy that suffers from numerous distortions in goods and capital markets and suppose that in the past many barriers to international trade in goods and capital were imposed. Suppose that the government wishes to remove the distortions and to liberalize trade in goods and capital. Under these circumstances, the general rule is that the first step in the liberalization process should be removal of domestic distortions in goods and capital markets and attainment of fiscal order, so as to reduce the heavy reliance on inflationary finance. The second step should be the liberalization of the economy’s links with the rest of the world. The external liberalization should also proceed in two phases. The first phase should reduce distortions to the free flow of goods by liberalizing the balance of trade. The capital account of the balance of payments should be opened up and liberalized only in the final phase. In what follows, some arguments are offered in support of the proposed order of liberalization of the trade and the capital accounts of the balance of payments.1

The key distinction between the effects of opening up the trade account and those of opening up the capital account of the balance of payments stems from the fundamental difference between goods and asset markets. As is known, the speed of adjustment in asset markets is much faster than in goods markets. Asset markets are more sensitive to expectations concerning the distant future, and new information that alters expectations is reflected in asset prices much faster than in the prices of goods and services. This intrinsic difference in the basic characteristics of goods and assets has several implications. First, since the economy under consideration has been distorted for a significant period, nobody (including economists) can be certain about the precise paths that its various sectors will follow after liberalization. Under such uncertain circumstances, prudence is called for. A program of liberalization that opens up the trade account first has the virtue of providing policymakers with an opportunity to examine the market’s reaction and to correct any errors found. A program that opens up the capital account first is not likely to provide such an opportunity, since once the capital account is opened up, the initial reaction is likely to take place very quickly and the resulting capital flows are likely to be huge. Thus, although it is eventually desirable to have an open capital account, prudence calls for a gradual transition to openness so as to facilitate the required change in institutions and the economic structure. An analogy can be made between an economy recovering from a long history of distortions and an individual recovering from a heart condition and a by-pass operation. While the latter’s complete recovery may well be anticipated, nobody would recommend that he start his rehabilitation program by running a marathon race.

The implications of the different speeds of adjustment of the trade and capital accounts can be illustrated picturesquely by a carriage pulled by two horses. Suppose that one of the horses is a fast runner while the other is a slow runner. As is obvious, if the two horses were to run at different speeds, the carriage would turn over. To avoid a disaster, the speeds of the two horses must be equalized. This can be achieved by speeding up the slow horse or by slowing down the fast one. It stands to reason that the former solution is not sustainable while the latter can be achieved with little effort. Analogously, the overall balance of payments constraint requires that the trade and the capital accounts be brought into line with each other, and thus, even though the two accounts tend to respond at different speeds, the overall constraint implies that the two speeds of adjustment must be harmonized. It seems that such consistency could be achieved with less effort by slowing down capital flows than it could by speeding up trade flows.

Second, it is easier (and less costly to society) to reverse wrong portfolio decisions than to reverse wrong real investment decisions. This potential difference in social costs also supports the thesis that the proposed sequence of liberalization measures is the correct one. Once the authorities remove distortions in the commodity markets and open up the trade account, real investment in the economy will take place in a less distorted environment—and thus will be more consistent with long-run patterns of real investment. Portfolio investments will continue to be based on a distorted capital market as long as the capital account is not opened up. Once the capital account is liberalized, some portfolio decisions will have to be reversed, but such a reversal is not likely to be very costly. On the other hand, if the capital account is opened up first, portfolio decisions are likely to reflect more accurately the undistorted long-run conditions but real investment decisions will still reflect a distorted environment as long as the trade account is not opened up. Owing to the distortions, the social costs of investments are likely to exceed the private costs. Such real investment decisions will have to be reversed once the trade account is liberalized. Therefore, the trade account should be liberalized before the capital account. In this context, it should be noted that distortions in the flow of goods affect both the trade account and the capital account and, by the same token, distortions in the flow of capital also affect both accounts. However, it is likely that the relative effects of each distortion on the two accounts differ; distortions in the flow of goods are likely to have a stronger effect on the trade account than on the capital account, while distortions in the flow of capital are likely to have a stronger effect on the capital account than on the trade account. This difference in the relative sensitivities of the two balance of payments accounts to the two distortions underlies the recommended order of liberalization.

Third, the cost of a distortion depends on the distortion itself and on the volume of transactions that take place in the presence of the distortion. Thus, when the trade account is opened up first, the cost of the remaining distortion (i.e., the closed capital account) is proportional to the volume of trade, which, owing to the slow adjustment of the market for goods, is likely to be relatively small. One the other hand, when the capital account is opened up first, the cost of the remaining distortion (i.e., the closed trade account) is proportional to the volume of capital flows, which, owing to the high speed of adjustment in asset markets, is likely to be very large. Thus, a comparison of the costs of distortions also supports the proposition that the trade account should be opened up first.

The various arguments that were outlined above rationalized the different treatment of the trade and capital accounts by reference to the different speeds of adjustment in goods and asset markets. In view of the different speeds of adjustment, the synchronization of the two balance of payments accounts was brought about by staggering the liberalization measures. An alternative strategy for coping with the different characteristics of the two balance of payments accounts could be adoption of a dual exchange rate regime. Such an exchange rate system allows for one exchange rate for commercial transactions in goods and services and another (typically a floating) exchange rate for capital account transactions. With this system, the different speeds of adjustment in goods and asset markets are reconciled through an adjustment of the relation between the two exchange rates. It is relevant to note, however, that while a dual exchange rate system may solve the synchronization problem, it starts the economy moving along the potentially dangerous path to multiple exchange rate regimes. The danger of such multiple exchange rate regimes is that they typically involve a complex set of distortions a complicated bureaucratic system, an elaborate structure of rent-seeking activities, and a costly allocation of rents. And, as is typically the case, it is much easier to create distortions than to remove them. Furthermore, experience suggests that a dual exchange rate regime cannot be sustained if the difference between the exchange rates for commercial transactions and for capital transactions becomes too large. In other words, it seems that the dual exchange rate system may be an additional viable effective instrument as long as reliance on that instrument is not too heavy. Since, however, the synchronization problem might be severe during a liberalization program, it is likely that the sustainability, and therefore the usefulness, of a dual exchange rate system might be limited. Under these circumstances, the proposed sequencing of liberalization measures might be more viable.

The removal of various distortions in the market for goods—including various taxes, subsidies, and tariffs—will yield a new equilibrium price level, a new nominal exchange rate, and a new real exchange rate. It is pertinent to note that prior to setting out on such a path, the initial conditions have to be set correctly, so as to reflect the new equilibrium exchange rate and the prices that will prevail following the removal of the distortions. In this context, it is useful to recall that several attempted liberalizations have resulted in uncontrolled inflation. These outcomes can be explained, in part, by the nature of indexation clauses. In order to avoid such failures, it is critical that the initial (once-and-for-all) changes in prices that result from the removal of subsidies and other distortions be excluded from the indices that are used for wage indexation.

As was emphasized earlier, it is essential that the liberalization program be preceded by restoration of a fiscal order that reduces the government’s dependence on inflationary finance. This restoration is an especially important prerequisite to opening up the capital account. The liberalization of the capital account is likely to encourage the process of “currency substitution.” This phenomenon results in an effective reduction of the “tax base” for inflationary finance. Unless the need for inflationary finance is reduced, the opening up of the capital account and the resultant shrinkage of the inflationary tax base may result in an accelerated rate of inflation as the government attempts to collect the needed revenue from a smaller tax base.

It should be recognized that during a liberalization program, some sectors in the economy are harmed more than others. These differential effects are likely to create strong pressures for accommodative monetary and exchange rate policies. It is important, therefore, to emphasize that monetary and exchange rate policies should not be designed to deal with such sectoral difficulties. Monetary and exchange rate policies are aggregate policies that should not be guided by intersectoral considerations. These inter-sectoral considerations are extremely important, however, and some of them should be dealt with. The proper policy instruments for dealing with intersectoral considerations are fiscal policies rather than monetary or exchange rate policies.

Finally, since the success of the liberalization program depends crucially on expectations concerning the feasibility and credibility of the program, it is important that the various measures that are undertaken hang together in a consistent way. Trust and confidence that has been built up with great difficulty can be destroyed very easily. Therefore, it is important not to start the program until all loose ends are tied up. The economic system, through the mechanism of memory that builds itself into expectations, shows little tolerance of errors. Therefore, while a slight delay in the introduction of a program will not be fatal, the premature introduction of an inconsistent plan might be.

Gradualism versus drastic measures during stabilization programs

In many cases, the decision to embark upon a liberalization program coincides with the decision to embark upon a stabilization program. The reason for this coincidence is easy to imagine. Both liberalization and stabilization programs are associated with short-run costs, and both require political courage as well as broad political support. The circumstances likely to generate these requisites are typically serious deep economic difficulties, wide-spread distortions, and galloping inflation. While liberalization and stabilization programs have tended to be introduced together, it is pertinent to note that the two programs are distinct and need not accompany each other.

One of the central questions concerning the initiation of a stabilization program involves the choice between gradual or drastic applications of the stabilization measures. The case for gradualism is well known; it rests on prudence, and on the fine tuning and mid-course corrections that it makes possible. In what follows, some arguments in favor of drastic treatment are outlined. It should be emphasized, however, that as is the case with liberalization, economic theory does not provide unambiguous guidance to those who must choose between gradualism and drastic treatments. Therefore, it is especially inappropriate to approach this issue dogmatically.2

The first argument for drastic treatment is that since there are numerous problems of unsynchronized changes of money and wages in the economy, a major policy shock will minimize the total output costs of seriously reducing the inflation rate. A sharp reversal of policy is likely to ensure that many preexisting contracts will be reopened and that economic agents will adapt their behavior to the likely course of policy; if the policy is well planned, the initial conditions it will produce should be close to the new steady-state equilibrium for the economy.

The second argument for drastic treatment is that gradualism may not, in practice, reduce the inflation rate. Underlying gradualism is the fear that attempts at restrictive policy will produce unemployment. It is precisely these fears that make it likely that any future inflationary shock will not be resisted, but rather accommodated. Under such circumstances, gradualism may well end up gradually increasing the inflation rate rather than reducing it, since inflationary shocks are accommodated and deflationary shocks are allowed to feed into output at the same inflation rate.

The third argument for drastic treatment rests on the recognition that reducing the budget deficit is a vital part of any program to reduce inflation. There is little doubt that the single most important factor in reducing a galloping rate of inflation is a significant cut in the budget deficit. Such cuts are, by their nature, very difficult to implement. They are likely to be even more difficult to implement gradually. The short-run pressures are usually to expand the budget deficit for one worthy cause or another. Unless the problem is faced up to directly and as a whole, the budget deficit is not likely to be cut. But once the problem is faced up to, the opportunity is at hand for making a concentrated attack on the inflation rate.

One of the major arguments against drastic measures is that there are uncertainties about the effects of the anti-inflationary package that is put in place. But the basic fact is that any policies that are followed in an economy in a need for a stabilization program have a highly uncertain outcome. It should not be thought that the consequences of pursuing a gradualist policy are certain either. In fact, it is, as argued above, quite possible that gradualism will increase the inflation rate and, consequently, merely postpone the application of drastic treatment.

The central factor determining the success of the stabilization program is its impact on the public’s expectations. It is crucial that the introduction of a new policy should be consistent, credible, and broadly based. The main components of any stabilization plans are the fiscal and monetary policy measures that are adopted. It is essential that these measures be viewed as viable, consistent, and sustainable.

One of the difficulties in implementing a drastic remedy is that the existing economic system has grown up over many years and has become deeply embedded. Consequently, any drastic change is necessarily disruptive because it renders many existing investment projects uneconomical. It seems, therefore, that the desirable solution to the problem of the choice between gradualism and drastic measures should be a plan that incorporates the best elements of each option. An example of such a solution could be a preannouncement of a feasible path for the policy instruments. Thus, if stabilization involves a cut in government spending, a reduction in the rate of monetary growth, and the like, gradual reduction coupled with preannouncement of the entire path will avoid the undesirable consequences of immediate drastic cuts, while the feasibility and the consistency of the plan will indicate the government’s long-term commitment, will enhance the plan’s credibility, and promote stabilizing expectations.

Finally, any stabilization program must have two chief aims and areas of operation. First, the program must adopt policy measures that rapidly reduce the high rate of inflation. Second, the program must begin the process of institutional reform to reduce the economy’s susceptibility to future inflationary shocks and to make it possible to keep future inflation under better control. A key factor in keeping future inflation under better control is the presence of a nominal anchor guiding the course of monetary policy. Such an anchor could be the money supply, the nominal exchange rate, or another nominal quantity.

basket pegging

The large variability of bilateral nominal and real exchange rates resulted in proposals that some of the Southern Cone countries peg their currencies to a basket of foreign currencies rather than to a specific leading currency like the U.S. dollar. These proposals have many positive features, but they also raise some important practical and conceptual issues that have not been fully addressed. In what follows, some of these issues are raised.

First, as should be obvious, all weighted averages (with positive weights that sum to unity) vary to a lesser degree than some of their individual components. Therefore, it is clear that pegging the currency to a basket of foreign currencies will yield a smaller variance of the exchange rate expressed in terms of the basket. But, as is also clear, averages tend to mask great diversity, and the relevant question is whether introducing stability into the weighted average should indeed be the objective regardless of the degree of instability in the components of the average. Specifically, shifting from a bilateral pegging to a basket pegging will stabilize the value of the currency in terms of the basket but will destabilize its value in terms of the individual currency to which it was previously pegged. Is it obvious that the stability of a weighted exchange rate is to be preferred to the stability of a bilateral exchange rate vis-á-vis a major currency? In analyzing this question, it is worth remembering that discussions of the welfare cost of inflation deal with related matters. In those discussions, the distinction is drawn between the cost associated with the variability of the price level and the cost associated with the variability of relative prices. Drawing on this analogy, one wonders whether those who emphasize the need for stability of the weighted exchange rate instead of the stability of bilateral rates would also place less emphasis on the cost of fluctuations of individual relative prices than on the cost of fluctuations of the aggregate price level. Prior to adoption of a new pegging strategy, the costs and benefits of the relevant alternatives should be specified.

Second, pegging to a basket per se will not eliminate many of the difficulties with, and objections to, flexible exchange rates unless there is an actual asset that represents the basket and unless that asset can be traded freely in futures markets.

Third, there is the practical question of choosing weights for the currency basket. The typical procedure has been the adoption of a trade-weighted basket. The choice of the proper weights, however, is not just a technical question, since it begs the conceptual problem of the role of the capital account. Specifically, since a large fraction of international transactions involve international capital markets, what role should these transactions play in determining the proper weights of the currencies that comprise the basket?

The success of a new exchange rate regime, like the success of liberalization and stabilization programs, depends on the adoption of a consistent set of policy tools and on a reasonable understanding of the implications of each course of action. It would be irresponsible to experiment with new systems just to learn how they work out. As is true of liberalization and stabilization programs, the cost of delaying the adoption of a new pegging arrangements until their full implications are understood is likely to be small relative to the cost of a premature adoption.

Rudiger Dornbusch

There are really three ways of stabilizing an economy. One is with financial orthodoxy, another with interventionism, and a third way is that adopted by Latin America, a consequence of which has been that in the last year economic performance in the region has been the worst since the 1930s. Whatever went wrong must have gone wrong in the large countries—Argentina, Brazil, Chile, and Venzuela. What went wrong was that the mix of stabilization and liberalization policies that these countries followed had nothing in common with either the interventionist methods that have often gone wrong or the financially orthodox methods that have often worked. The latter were fashionable in the inter-war period, in the 1950s, and even in the 1960s but fell into disrepute when everybody started criticizing the Fund. In fact, the Fund’s financially orthodox stabilization programs have been right in terms of comparative statics, though often not in terms of dynamics.

Looking at the experiences of Argentina, Brazil, and Chile, one observes two problems—a high real interest rate and an over-valued currency. In Argentina and Chile, these problems have been partially solved; in Brazil, similar developments may be expected within the next six months. Each country failed to recognize that an effective stabilization program must (1) correct the fiscal problem, and (2) simultaneously achieve a real depreciation of the currency. In Chile, the fiscal reform was very successful, but policymakers allowed the currency to become more and more overvalued each year, until they were forced to raise the real interest rate extremely high just to finance the current account imbalance. Such a situation was not sustainable, of course, and investors began speculating about how long it would last. While the real interest rate remained high, people praised the new liberalization program, which entailed an across-the-board import tariff of only 10 percent. All things considered, the real exchange rate today (some would even say the exchange rate that prevailed late in 1981) is the lowest for decades. Consequently, now is a good time to acquire importables before the real exchange rate returns to its former level. Not surprisingly, there has been considerable borrowing to finance the acquisition of consumer durables or equipment. The likely near-term consequences are an exchange crisis, a real depreciation of the currency, and a return to the economic conditions of 1975. In the Chilean case, there is little doubt that confusion of stabilization and liberalization, and failure to recognize that fiscal reform must be accompanied by a real depreciation of the currency, have led to the current exchange rate crisis.

In Argentina, commercial liberalization was less extensive, fiscal reform never took place, and an increasing real wage was financed by external borrowing. Of course, the situation was untenable and, in fact, was more severe than in Chile because the real appreciation of the currency had gone much further and the political conditions made it much more difficult to lower real wages enough to offset not only the initial real appreciation but also the debt accumulated during the war in 1982 and the deterioration in external competitiveness brought about the efforts of other developing countries to achieve real depreciation. Looking at Argentina today, one can say that although real wages may have fallen 20 percent, the fiscal deficit has not stabilized and there will certainly have to be another real depreciation to strengthen the current account sufficiently to generate confidence that the economy can be effectively stabilized. Whatever else a country does, an economy cannot be stabilized unless the current account position convinces people that the exchange rate is realistic. Financing, regardless of the amount obtained, is just a gimmick; in the end, the current account will have to service the debt, and the exchange rate will stabilize only when the current account deficit is reduced to manageable proportions.

The preceding remarks provide background for a discussion of Brazil, which really is now in a situation analogous to those in Mexico, Chile, or Argentina before their recent crisis occurred. The Brazilian Government has allowed the real exchange rate to appreciate over the last four years. There has been no real depreciation to make up for the oil shock, the debt accumulation, the higher real interest rates, Southeast Asia’s gain in competitiveness, or reduced economic growth rates in the Organization for Economic Cooperation and Development (OECD) countries. The real exchange rate is clearly overvalued, if one allows for the current cyclical situation, though it should be kept in mind that the balance of payments situation is manageable only because the country is experiencing its first recession in forty years. Attainment of a sustainable real exchange rate that was consistent will full employment would probably require a substantial real depreciation.

What happens in these circumstances? It is generally recognized that, with central bank reserves at their current low levels, the real exchange rate will have to fall eventually. The only question is whether the currency can be devalued fast enough in real terms to discourage people from getting into dollars by taking advantage of leads and lags or other means. Again, the problems are caused by failure to achieve fiscal stabilization and a futile attempt to maintain the real wage by exchange manipulation, which leads over time to an incredible overvaluation—incredible because everybody knows that although the deficit can be financed for one year, it certainly cannot be financed for another year.

Central banks have in some way become complacent about such excessive financing and have consequently been able to delay real exchange rate adjustment. Eventually, however, this has led to exchange crises, as have occurred in Mexico, Chile, Argentina, and—most recently—Brazil.

Returning to the question of liberalization and stabilization, it is fair to say there has been insufficient financial orthodoxy in the countries that have been discussed. The traditional approach to stabilization is to correct the budget, to bring about a big real depreciation of the currency, and to establish a controlled capital account program. Of these three goals, Latin American countries have had the most success with fiscal stabilization. Since cutting government subsidies and lowering the real wage are extremely difficult politically, these countries have always tended to avoid lowering the real exchange rate sufficiently, partly because immediate reductions in inflation can be achieved if depreciation is not too fast. This behavior leads first to overvaluation, then to higher real interest rates, and finally—when even these can no longer maintain the exchange rate—to an exchange crisis. The pattern has become very predictable, highlighting the fact that any successful stabilization program must start by undervaluing the currency. When the currency is undervalued, there is credibility in employment; once the latter has been established, the country can afford fiscal cuts. Countries can maintain an overvalued currency only by financing the resulting deficit and by forcing up real interest rates; however, in the end, they cannot avoid an exchange crisis.

Not much has been said about liberalization. In all these countries, the problem is one of stabilizing the budget and living with the current account. The worst thing to do is to liberalize the capital account to allow the public to get into dollars before the necessary real depreciation has been achieved. And it is certainly a grave error to try to promote exports and to destroy import-competing industries in world recession. Liberalization attempts have been counterproductive in Chile and would be so in Brazil. They have fostered expectations that real exchange rates were too high and, therefore, have promoted financial instability.

William H. Branson

This discussion will focus on a diagram called Arriazu’s box, which looks like an Edgeworth box. It has four sides, along which one measures the degrees of openness in trade and the capital account. One corner of the box represents a completely closed economy, and at the opposite corner a completely open economy. One could even imagine something like a contract curve through the box that would give the optimum degree of openness or the optimum degree of change in openness if he wanted to talk about liberalization in terms of changes in states.

There seems to be a problem in the ordering of liberalization. One can imagine, for instance, a scenerio beginning with liberalization of the capital account, which would generate increased confidence in the country doing so. This, in turn, would generate capital inflows and a real currency appreciation and would be bad for trade, especially in manufactures. On the other hand, if current account were liberalized first, one could imagine a situation where imports would surge. (It takes time to develop export markets.) That would lead to a current account deficit, reserve losses, depreciation, inflation, and the creation of a vicious circle that would destroy the whole experiment. It seems here is a problem that needs to be worked on, and that could be thought about in terms of Arriazu’s box.

A funny analytical problem appears when one thinks about the stability of a system that has adopted an accelerating peg. Essentially, when one adopts an accelerating or decelerating peg, he states a schedule of depreciation according to which the rate of depreciation itself is gradually reduced, on the presumption that a stable situation lies at the end of the road. So, basically, one is stating a schedule of the rate of change of the rate of change of the exchange rate. This involves a dynamic analysis that will inherently be very difficult. In technical terms, one will solve second-order differential equations. Making exchange rate stability under such a regime consistent with a financial market equilibrium as the economy is being opened would seem to be an awfully complicated problem.

Ricardo H. Arriazu

I will concentrate my comments on just a few of the questions most frequently raised in relation to these liberalization problems. I will, obviously, base my comments on my knowledge of Argentina.

Basically there are four questions that I would like to deal with. One is this question of what should be open first, the trade or the capital account? Second is the question of preannounced exchange rates. Third is the question of the relationship between fiscal deficits and exchange rates. The fourth is related to the definition of the current account, to which Dornbusch assigns so much emphasis.

Starting with trade or capital liberalization, conventional wisdom says that the trade account should be opened first and only then the capital account, if we want to avoid problems. I share that wisdom, but I would like to emphasize that first we should define clearly what we mean by trade account liberalization. We have at least two possibilities or definitions of trade liberalization: one definition relates to making the economy more open to trade, in the sense of increasing the proportion of goods subject to international competition (proportion of tradables in the economy), and the other relates to liberalization directed at improving the allocation of resources by reducing the effective protection of some sectors while increasing that of others.

The first of these definitions has little to do with the level of effective protection in a country and is much more related to the level of redundant protection (”water”) in the tariff structure. As a matter of fact, it is possible to have a product with an effective protection of 500 percent, fully used, which is tradable in the sense of being subject to international competition, while other products with a nominal effective protection of 20 percent behave as nontradable as this protection is not fully used, at least in the short run.

Therefore, when we talk about the dangers of opening the capital account first, rather than the trade account, what we are talking is about the danger of opening the capital account when still a lot of products can be considered as nontradables as a consequence of widespread redundant protection. In these circumstances, if you allow me to use the same box that Branson has just described, we run the risk of moving into the upper left corner, which extreme case defines economies totally closed to the movements of goods and totally open to movements of capital. In this extreme case, the price level may become totally undetermined. Why? Because, under its assumptions, the price level is exclusively determined by the forces of domestic demand and supply, while the money supply is endogenous to the system, which implies that under certain conditions, capital flows may finance any level of domestic prices, even if domestic credit is rigidly controlled. In this sense, it is totally correct to say that to avoid this possibility it is better to open first the trade account (elimination of “water” in the tariffs), then the capital account. It is important to emphasize that this reform has nothing to do with a reform to improve the allocation of resources and which may, or may not, increase the proportion of tradables in the economy. The liberalization process in Argentina, for example, went much further in eliminating redundant protection than in reducing dispersion or effective protection. As a matter of fact, dispersion was one of the main problems of the tariff structure, which provided an average effective protection of 39 percent for the industrial sector, but with wide variations between products from minus 40 percent to plus 400 percent; in addition, a large proportion of the industrial sector confronted negative effective protection; it just happened that the extreme cases of negative and positive protection occurred in the same productive sector. As a matter of fact, existing effective protection structures are the result of accumulated political decisions over time rather than rational decisions on optimal—or preferable—industrial and production structures. The recent tariff reform in Argentina was basically directed in the initial stage to reducing the dispersion of protection between sectors and the level of redundant protection, and only marginally to reducing average protection. One problem in attempting a reform as described is the fact that the authorities normally do not know which sectors have “water” in their tariffs so that efficient reforms take time. The conventional wisdom of opening first the trade account is therefore not easy to implement.

We come now to the second point. Should we control capital movements? And if the answer is affirmative can we really control capital movements? Dornbusch says yes to both questions but do the authorities control capital movements or does the public? My own experience in Argentina makes me very doubtful on the possibilities of controlling capital flows. In 1975, for example, Argentina introduced one of the more severe and extensive sets of capital controls in force in the international economy; during that year the current account registered a deficit exceeding $1 billion, at the same time that Argentines accumulated foreign assets abroad also exceeding $1 billion. How could this accumulation take place with so many controls? Basically, through overinvoicing and underinvoicing.

The experience of 1978 also tends to confirm this fact. At the beginning of 1978, the authorities decided to tighten domestic credit as part of their anti-inflationary program. As reserves had been increasing and covered domestic interest rates had been higher than abroad, the public started to bring capital from abroad at unprecedented rates, in this way avoiding the restrictions on credit. In order to stop these flows, the government decided to introduce an interest equalization tax in the form of a deposit equivalent to 20 percent of the flows—with no remuneration whatsoever; in a situation of 6 percent monthly inflation, this deposit implies a substantial tax. Of course, with this measure, capital flows disappeared as such, but capital continued flowing in, now in the form of exports prepayments. As a deposit was also required for these flows, capital then began to flow in the form of export prefinancing, and when these flows were also restricted, the movements took the form of overinvoicing and underinvoicing. Exchange sales by tourists also increased during the period, even though these “tourists” were obviously Argentines.

After these examples the question remains: can we really control capital movements? It is possible that for some countries the answer would be positive, but in my opinion not for Argentina. This leads us to a final point on this subject. When a country like Argentina has repeatedly introduced strong economic shocks, causing many of its people either to completely lose their incomes and their wealth in very short periods of time, or to accumulate extraordinary profits as a consequence of these shocks, then the association of what is moral and what is legal tends to disappear. The people themselves start to decide what is moral and legal and on many occasions their definition does not coincide with what the government considers to be moral and legal. In these circumstances, it is not rare for the public to challenge what the government decides as illegal if they consider that the decision is not moral, and therefore they do what they consider more convenient. For example, at this moment Argentina has re-established controls on capital outflows, including a dual exchange market, and the country is experiencing very large capital outflows. Unfortunately, persistent economic shocks and frequent income transfers also tend to encourage economic anarchy, which is something very dangerous for the whole social structure of a country. The question is then, if you have capital flows and you are not able to control them, what do you do? Perhaps if you are in the United States or in other countries less used to frequent upheavals, controls could be improved, but if capital flows take place and you cannot stop them, what is the optimal strategy? I think this is a very important problem to discuss, and economists generally tend to forget this structural constraint.

The second subject I want to deal with is related to the meaning of preannounced exchange rates. I have read recently many papers on the subject and most tend to argue that the basic objective of the system is to change expectations in such a way as to help reduce inflation. Since I have been an advisor on this policy in Argentina, let me say that I feel that this aspect of the policy has been overemphasized. Actually, the basic idea was much simpler and comes from the theory of fixed exchange rates. If a country wants to have a fixed rate system and it confronts a large fiscal deficit, it is clear that the end result would be reserve losses and periodic balance of payments crises—with devaluations. In these circumstances, if you cannot have a fixed rate but you want the advantages of a fixed rate system, in the sense of having more transparent markets, less uncertainty, and items like that, then one possibility is to proceed as follows: start by estimating how large the fiscal deficit is going to be and how much monetary base will be created as a consequence of this and other factors. This allows an estimate of how large a flow demand for monetary base is required to be able to maintain balance of payments equilibrium, given the fiscal deficit. Once this disequilibrium is estimated, the reasoning continues as follows: when a country devalues, the prices of tradables also increase as a result of cost increases, indexation practices, and other adjustment forces. This increase in prices will normally be accompanied by an increase in the nominal demand for money, which, in turn, determines an increase in the flow demand for base money. The problem is, therefore, how to estimate the size of a devaluation needed to generate the required flow demand for base money to compensate for the effects of the fiscal deficit. To this estimate, governments always add a security margin just in case the calculations prove wrong.

Once the required devaluation is estimated, the question is whether to keep it secret in order to guide the exchange rate policy without restricting the freedom of the authorities, or to preannounce a schedule of periodic devaluations—making the periods between adjustments as short as possible to reduce speculation—in order to give more certainties to the market. In this sense, preannouncement of exchange rates is equivalent to preannouncement of monetary targets, in countries with floating exchange rates.

Now if it happens that the resulting schedule of adjustments is a declining one, because the calculations show that in monetary terms the country does not need devaluations as large as the ones being implemented, then the declining schedule might contribute to dampen inflationary expectations. However, to try introducing a declining schedule of devaluations, with the express objective of influencing expectations totally inconsistent with the adjustment needs brought about by the fiscal deficit, would have disastrous economic consequences and would be equivalent to maintaining a fixed exchange rate in similar circumstances.

In assessing individual countries’ experiences, therefore, the question to answer is whether the authorities were right in their assessment of how much inflationary tax was needed, given the fiscal deficit and the public’s hoarding desires, to equilibrate the balance of payments. In the case of Argentina, if you take Harberger’s assessment, which coincides with that of the majority of Argentine economists, the answer appears to be that it was wrong, as the fiscal deficit required much larger devaluations than those implemented; the deterioration of the current account and a declining exchange rate are shown as proofs of these points. However, these points are not enough proof to conclude that nominal exchange rate adjustments are insufficient, since other problems of economic policies can generate similar results. Here I would like to make use of some of the points raised in my theoretical paper.

Current account deficits are the counterpart of excesses of expenditures in relation to national income. Dividing total domestic expenditures into public and private expenditures, and adding and subtracting the level of taxes, we can divide the relationship between the current account, expenditures, and national income into two parts: a public sector budget (public expenditures minus taxes) and a private sector budget (private expenditures plus taxes minus national income). Since these are real budgets in the sense of measuring real flows, it is clear that the concept of taxes includes both nominal and inflationary taxes; inflationary taxes—as stated in my paper—produce the same macroeconomic effects as normal taxes.

The only way of verifying whether the current account deficit originates in undervaluations of the nominal exchange rate, or originates in other factors is to see whether the deficit originates in the public or the private sector budget and, in order to incorporate the effects of the inflationary tax, we have to go below the line and look at assets and liabilities—duly adjusted for inflation—and observe their evolution through time. Disequilibrium must be reflected in variations in each sector’s net financial wealth. A deterioration in the net financial wealth—adjusted for inflation—of the public sector is evidence that taxes—including the inflationary tax—were not sufficient, and as such is a priori evidence that nominal exchange rate movements are insufficient to collect the required amount of inflationary tax.

One important question is how to treat the accounts of the Central Bank. If the Central Bank holds reserves and, as a consequence, most of the devaluation taxes accrue to the Central Bank, how are these profits treated? From the accounting system developed in the paper, it follows that the central bank figures should be consolidated with those of the public sector, especially if subsidies are hidden in the central bank accounts. Surprisingly, when these calculations are performed for Argentina, for the period under discussion, the figures show an improvement in the net financial wealth of the consolidated public sector, which implies that, adjusted for inflation, the Government was in surplus. Obviously, there was a nominal deficit, but the amount of inflationary taxes collected was larger than was required to compensate for the deficit. How then did the current account turn into deficit and the real exchange rate appreciate so sharply?

Let me try to give you a possible explanation of what happened in Argentina and what, in my opinion, went wrong. During the last decade, public sector expenditures increased drastically (more than 10 percent of GDP if subsidies hidden in the Central Bank are included), at the same time that taxes—including inflationary taxes—increased even more. This increase in taxes generated the surplus mentioned a moment ago, but increased the tax burden of the private sector, at the same time that—from 1979 on—the Government was pushing real wages upward. This policy coincided with a sharp decline in the terms of trade and an unprecedented high international real rate of interest. This was typical of the Government wanting more guns and more butter, just at a moment when the international crisis required a reduction in expenditures.

The result for the productive sectors caught in the middle was disastrous, as their profits turned to losses, forcing them to borrow to survive. Real interest rates shot up, both as a result of the international environment and as a result of this borrowing, and capital flowed in, financing the emerging disequilibrium. As a consequence, a current account deficit emerged and, since expenditures exceeded income, real exchange rates tended to fall. The system proved inconsistent.

In a situation like this, characterized at the same time by frequent wage adjustments based on past inflation, faster movement of the nominal exchange rate would only have provided the Government with more inflationary taxes, leaving the real exchange rate exactly as before. Obviously, abandoning wage indexation would have allowed the real exchange rate to recuperate, at the expense of real wages. Here, Dornbusch is right in the sense that, if the Government is going to increase public expenditures, real wages have to decline in order to allow the real exchange rate to remain constant.

The final question is related to the definition of the current account. Very briefly, if we are talking about real flows and not nominal flows, we should also talk of the current account in real terms. The balance of payments, as all nominal accounting, is subject to distortions introduced by inflation. Nominal interest payments are composed of two elements: a maintenance-of-value element and a real-interest-rate payment. In measuring real flows, only the latter should be included in the definition of the current account, treating the rest as advance payment of real capital.


This order of economic liberalization is advocated in Ronald I. McKinnon, “The Order of Economic Liberalization: Lessons from Chile and Argentina,” in Economic Policy in a World of Change, ed. by Karl Brunner and Allan H. Meltzer, Carnegie-Rochester Conference Series on Public Policy, Vol. 17 (Amsterdam, 1982), pp. 159–86. The subsequent arguments are adapted from Jacob A. Frenkel, “The Order of Economic Liberalization: Lessons from Chile and Argentina: A Comment,” pp. 199–201 in the same book.


Some of the subsequent arguments in this section draw on Stanley Fischer and Jacob A. Frenkel, “Stabilization Policy in Israel” (unpublished, Bank of Israel, July 1981).