VI Developing Countries: Protection, Liberalization, and Macroeconomic Policy
Author: W. M. Corden1
  • 1 0000000404811396 Monetary Fund


A central issue in the design of policies (and of Fund-supported adjustment programs) is to deal with balance of payments problems. In view of the importance of this subject it will be dealt with at some length here even though there is some repetition of analysis presented earlier.

Protection, the Exchange Rate, and Real Wage Rigidity

A central issue in the design of policies (and of Fund-supported adjustment programs) is to deal with balance of payments problems. In view of the importance of this subject it will be dealt with at some length here even though there is some repetition of analysis presented earlier.

What, if any, is the role of trade restrictions when a current account deficit needs to be reduced? When the exchange rate is fixed but adjustable, are trade restrictions ever to be preferred to real devaluation? The Fund’s position is clear: restrictions should not be increased and, if possible, should be reduced. The exchange rate should be used as the required device that switches the pattern of demand away from foreign on to home-produced goods, supplementing the necessary reduction in real expenditures.

But others often do not agree. Many of the issues discussed earlier enter here. One might accept the proposition that for maximization of national efficiency in the medium and long run the best policy is liberal trade (with perhaps some exceptions on infant industry and other grounds discussed earlier). But is such a liberal policy also best for the short run? The former Cambridge group in Britain, for example, did not agree, and similar views have been propounded in Latin America and elsewhere.11

It is worth analyzing in some detail the case for shortrun protectionism when the current account has to be improved. Such a protectionist argument comes often almost instinctively to policymakers and others. It is explicitly or implicitly an argument against devaluation, or at least against sufficient devaluation. Hence one is really concerned with import restrictions (or, more rarely, tariffs) versus devaluation as switching devices to accompany the necessary and inevitable reduction in real expenditure.

The familiar “orthodox” analysis is that a devaluation raises domestic currency prices of imports and exports. This will switch the pattern of domestic demand away from imports and increase profitability of import-competing and export industries, provided nominal wages do not rise, or do not rise much. This higher profitability will then, in due course, lead to expansion of tradable goods output, which is the desired objective. At the same time, if output was at full capacity initially, demand for nontradables should decline—absorption should fall—to free resources for extra production of tradables.

A short-term argument for using import restrictions in preference to devaluation at a time of balance of payments crisis is that the effects of devaluation work with a lag, so that initially very high devaluation (to an extent that cannot be calculated in advance) may be required to achieve a desired reduction in imports. Temporary quantitative restrictions may then be needed as well.

Here it has to be borne in mind that quantitative restrictions also take time to implement and create administrative problems. If they are associated with price controls on restricted imports, excess demand will be generated and powerful pressures can build up to ease the restrictions. If there is no effective price control, the restrictions will yield the familiar monopoly profits for import license holders (who may be local manufacturers using imported inputs). In both cases vigorous rent seeking may result.

Nevertheless, the validity of such a short-run case cannot be completely denied. The main objection is that restrictions once imposed are not readily removed. There would then be medium- and long-run adverse effects through failure to stimulate exports (hence producing too much import compression), as well as distortions in the pattern of imports as the result of using a nonmarket method of discriminating among imports to be restricted.

In fact, it makes an important difference not only whether the import restrictions are in fact temporary or permanent but also whether they are expected to be temporary or permanent at the time they are imposed. If a country devalues at a time of balance of payments crisis and supplements this with import restrictions that are widely accepted as being temporary, longer-term resource allocation decisions will be influenced by the devaluation and not very much by the restrictions. The main role of the restrictions will be to bring about a rapid reduction in imports (in association with the necessary reduction in real expenditure). This may be desirable if reserves are low or short-term loans cannot be easily obtained. An excessive expansion of import-competing production relative to exporting will then be avoided. But a belief that the restrictions will only be retained for a short period is hard to establish—naturally so on the basis of experience in many countries. Furthermore, if the balance of payments problem is really known to be temporary, one has to ask why it is not possible to draw on reserves, the capital market or, indeed, the Fund.

Leaving aside this very short-run argument for import restrictions at a time of balance of payments crisis a deeper, more sophisticated argument (the Cambridge argument) should now be considered. It hinges on the possibility that there is a tendency to real wage rigidity brought about by formal or informal wage indexation. It is highly likely that for a nominal devaluation to succeed in improving the competitiveness of tradable goods industries and so bringing about real devaluation, real wages must fall. With given nominal wages, the average price level would rise. And this creates the problem that the rise in the domestic price brought about by devaluation would cause nominal wages to rise to restore the initial level of real wages and the benefits of the nominal devaluation might be destroyed completely. The required switching of the pattern of output from nontradables to tradables, and the switching of the pattern of expenditure from foreign to domestic goods, would then not take place.

At this point the suggestion is made that import restrictions or tariffs might be used instead on the grounds that they do not require declines in real wages. But these devices will also raise the domestic price level by creating shortages. The element of validity in the argument is that tariffs might raise the price level less than devaluation if there is domestic consumption of exportables, and even less if the revenue raised were offset by equivalent reductions in indirect taxes. In that case the initial real wage declines and hence the subsequent increases in nominal wages would be less.

The essential feature of a devaluation compared with the other devices is that it increases the profitability of exporting, which should in due course—as supply expands and foreign markets are exploited—bear fruit in higher export income (in terms of foreign currency). A measure that makes exporting more profitable might tend to reduce real wages more—and so in due course bring about more compensating rises in nominal wages—than measures that are purely import compressing.

In the medium and long run, export promotion through exchange rate adjustment is clearly what is needed. Tradable goods production should expand both on the import-competing and the export front if the nondiscriminatory signals of the market are to be accepted as a guide to resource allocation, and if excessive import compression is to be avoided. But in the short run export supply is often quite inelastic, especially if exports are primary products or if new manufactured products have to be developed. Hence the rise in export profits resulting from devaluation can be regarded as a windfall which could be dispensed with for the sake of reducing the adverse effect of the switching policy in causing nominal wages to rise.

Compared with the free market solution of devaluation, the use of import restrictions is thus a way of taxing profits of exporters so as to sustain real wage levels when, in the short run, real wages really need to fall. Alternatively, one might argue that tariffs and devaluation have the same or similar effects on the cost of living and real wages, as well as on profits of import-competing industries, but that tariffs bring in revenue to the government (which it may or may not offset with the reduction of other taxes) while devaluation brings in the equivalent revenue to exporters.

Thus sometimes, with rigid real wages, trade restrictions of particular kinds could make the short-run problem easier. This would be so particularly if the restrictions were focused on goods not consumed by wage earners, and if the distributional shift implied by the particular pattern of restrictions (for example, on imports of so-called luxury goods) were thought desirable.

This is a sympathetic summary of the Cambridge argument for import restrictions as part of a policy package to deal with a current account problem. Such a package would involve damaging medium- and long-run prospects for the sake of possible short-run gains, though this trade-off is not usually pointed out by the proponents. The source of the long-term damage is that inadequate incentives are provided for export expansion. Relatively uneconomic import-competing production will be fostered in preference to more economic export production. Furthermore, in the medium and long run protection by developing countries is likely to reduce growth in employment as well as slow up real wage increases. The reason is that developing countries have a comparative advantage in labor-intensive products, so that export expansion resulting from outward-looking policies would tend to be in labor-intensive industries. Growth of labor-intensive industries relative to capital- and resource-intensive industries will tend to raise real wages by increasing demand for labor, and may also increase overall employment. The experiences of the newly industrializing countries of East Asia bear this out.

The distributional effects of the policy choices must also be considered. A rise in the domestic prices of exports brought about by devaluation (but avoided by tariffs or import restrictions) may increase incomes of the poorest sections of the community if exports are produced by peasants. The distributional effects of the choice of “switching” device depend on the particular structure of the economy, and in countries with peasant export sectors proposals for preferring import restrictions over devaluation imply unfavorable income redistribution. With real wage rigidity in the urban sector, import restrictions have then an equivalent effect to a combination of devaluation and taxation of rural incomes to subsidize urban employment. In addition, it is almost inevitable that a system of import quotas will have unfavorable distributional effects through the privileged allocation of licenses.

Short-Run Terms of Trade Effects

So far it has been assumed that if exporting were made more profitable by devaluation, supply might not increase much in the short run, even though it would increase in the long run. Thus in the short run the extra profits in the export industries resulting from devaluation would simply be rents which could be taxed away or forgone by avoiding devaluation and using instead the alternative device of import restrictions.

Even if supply did increase, world demand may be inelastic, possibly because of restrictions (including imposition of “voluntary” export restraints) abroad. Extra exports may be excluded from some industrial country markets by protection, and hence would have to be unloaded at substantially lower prices elsewhere. Hence the country’s terms of trade would deteriorate as a result. Even without protection abroad, short-run demand elasticities are often low, so that increased export supply may well lead to declines rather than increases in the value of exports.

Looking purely at the short run, and even without any wage indexation, there appears then to be a case for preferring tariffs or import quotas to devaluation, this being simply a version of the terms of trade argument for protection. It hinges completely on estimates, implicit or explicit, of elasticities of demand for exports. In the 1950s and 1960s elasticity and export pessimism were one basis for the import substitution bias of development policies in Latin America, India, and elsewhere (but not in East Asia) and this view can still be encountered.

The element of justification is that if voluntary export restraints are imposed upon a country then, in effect, its own optimal policy is to impose export taxes or quotas on the particular products affected when directed to the particular markets concerned. But, apart from that, there are two objections to this line of approach as an argument for conventional protection.

First, insofar as it has some validity, it might justify export taxes or restrictions, the rates of tax being higher where the demand elasticities are believed to be lower. In other words, the first-best short-term policy (ignoring the medium and long run) is to impose export taxes or restrictions differentially between different exports. Some exports, especially of manufactures, may face very high elasticities of demand and no foreign restrictions, so that even significant expansion of exports by many developing countries at the same time would require only taking up a small share of a very large industrial country market. In these cases there would be no need for prices to be reduced much, if at all, in order to expand sales in world markets, and hence there would be no significant terms of trade effect. The optimal export tax would then be zero. In other cases, where a country is a significant world supplier of a product, there might be some shortrun benefit from export restriction.

It might be noted here that the current weakening of commodity prices, explained to some extent by expansion of supply in earlier years, suggests, at least with hindsight, that world market prospects have not always been adequately foreseen, whether by governments or by private decision makers. In particular cases less supply expansion than actually took place would clearly have been desirable. But this was more a question of the difficulty or inadequacy of forecasting than of a divergence of private and social interests justifying export taxes or restrictions.

The second objection concerns the application of the argument to actual and prospective exports of manufactures in general. Whatever the short-run case, the evidence has shown that elasticities in world markets tend to be high in the medium and long run. Thus there has been a very significant expansion of exports of manufactures from developing to industrial countries during the 1960s and continuing after the first oil shock. This has not just been a volume but also a real value expansion. Hence medium-run elasticity pessimism for exports of manufactures is not justified on the basis of recent experience.12

It is then a matter of balancing possible short-run gains from import restrictions against the medium- and long-term losses that result from failure to seize export opportunities. If fundamental or structural improvements are sought it is clearly necessary to bear in mind adverse medium-run effects of policies.

Trade Liberalization: Exchange Rates and Timing

The most important point about large-scale unilateral trade liberalization is that it must be associated with real devaluation if the current account is not to deteriorate and if the employment losses in protected import-substituting industries are to be compensated by employment gains elsewhere, especially in export industries. Normally nominal devaluation will be needed to bring about the required real devaluation. A liberalization program must, therefore, be part of a policy package which includes exchange rate adjustment. The appropriate exchange rate adjustment will be hard to judge in advance, but it is important to bear in mind that the longer-run equilibrium real exchange rate does depend on the degree of trade liberalization.

The question is often asked where the extra jobs would come from when liberalization brings about employment losses in highly protected import-competing industries. In a general context of growth, liberalization may lead not to absolute losses in employment in these industries but rather to a reduced rate of growth. But the question is relevant even in that case.

In the main, the extra jobs will be generated by the real depreciation which is part of the policy package. It will make exporting more profitable and also improve the competitiveness of import-competing industries. Some import substitution may become economic that was previously uneconomic—and hence generate extra employment. Those import-competing industries which had received low protection, or none at all, will benefit.

Industries that use imported inputs that were not readily available before liberalization or that used inputs produced locally at high cost owing to protection will find that their costs have fallen and may expand employment. Against this must be set the higher costs of imports resulting from the devaluation, which will affect some industries adversely. In addition, the improvement in resource allocation in the tradable sector as a whole is likely to increase real national income, leading to more consumption spending and hence employment generation in domestic industries producing for the home market. The higher real incomes will yield more tax revenue and so improve the fiscal balance. If tax rates are reduced to restore the initial fiscal balance, increased private consumption or investment will generate extra jobs, or alternatively, increased government expenditure financed by the higher tax revenue would do so.

The sequencing of liberalization and the associated exchange rate adjustment is also a matter of some complexity. A choice, essentially political, has to be made between gradualism and sudden measures, and how much advance announcement there should be. The longer the period between the announcement of a credible program of liberalization and the actual changes in tariffs and quotas, the easier it is for protected industries to make the necessary adjustments and for potential gainers—notably export industries—to gear up for expansion. But the announcement has to be credible. On the other hand, it has been argued that the longer the lag between the formal commitment to a program and its actual implementation, the greater the opportunity for interest groups to slow up or even halt liberalization and the more likely it is that expectations about the credibility of the authorities’ intentions will be undermined.13

The exchange rate should be adjusted early even at the cost of generating temporarily excess profits in export industries and in import-competing industries that are not protected. The beneficial effects of depreciation on exports are likely to develop with a lag, while an increase in imports resulting from liberalization could be quite quick. A firm, credible assurance that a program of liberalization will be followed should discourage the flow of resources out of nontradables into highly-protected industries during the transitional period when the protected industries are excessively profitable because the exchange rate has already been devalued while the liberalization process is not complete.

Is a time of balance of payments difficulties the right time to liberalize trade? This important issue arises currently and needs to be considered in relation to Fund programs. It is, of course, not possible to resolve this issue here or arrive at conclusions appropriate for all countries but some considerations can be set out.

From the narrow but popular partial view it certainly appears to be the wrong time. Traditionally, a balance of payments crisis has led to the imposition or tightening of import restrictions since it is noneconomists’ common sense that when imports are too high in relation to exports the proper policy is to restrict imports.

Two immediate answers can be given. Firstly, liberalization will allow, possibly for the first time, the ready availability of cheap imported inputs required for exports. This aspect of liberalization would improve the balance of payments by raising exports even if the exchange rate stayed constant. There may be a lag before all the benefits come through since it takes time to expand exports, find new markets, and so on, but at least there is a favorable and direct balance of payments effect.

The more important answer is that the alternative to import restrictions is not to do nothing but to depreciate the exchange rate. Hence one is back to the choice already highlighted several times in this paper between two “switching” devices, one of which discriminates between imports and in favor of import substitution relative to export expansion, while the other—exchange rate adjustment—is nondiscriminatory. The improvement in the current account requires both a reduction in aggregate spending and depreciation of the exchange rate to switch demand from foreign toward home-produced goods and, within the latter, toward tradables. If there is simultaneously some liberalization of imports, even more depreciation is required.

Restrictions may be so widespread, complex, and dislocating that their removal can have fairly immediate beneficial effects on incentives and output. This would particularly be so when imported inputs for manufacturing production with export potential are subject to licensing. It would then also become easier to solve the balance of payments problem, because an increase in supply would modify the extent of the reduction in aggregate demand required, and hence reduce the painfulness of the required adjustment.

In other cases the beneficial effects of removing restrictions combined with adequate devaluation would only show up in the medium run, as new investment is directed into more productive channels, rent seeking declines, and so on. In the short run, liberalization associated with exchange rate adjustment may give rise to dislocations and to localized or industry-specific unemployment as profitability of some industries declines while that of others improves. The question then arises whether the short-run problems of liberalization should be added to the problems involved in bringing about the reduction in real expenditures needed for restoring macroeconomic stability.

In considering whether to liberalize at a time of balance of payments crisis it is then a matter of trading off additional short-term problems against medium- or long-run benefits; there may be short-run adjustment costs, and the short-run problems these engender may be harder to bear when at the same time the needs of the balance of payments call for a major reduction in aggregate spending. Liberalization may also risk being unjustly blamed for the many problems—including often unemployment—caused by the need to reduce spending in order to improve the balance of payments. On the other hand, some countries seem to have continuous balance of payments difficulties which have underlying structural causes and call for longer-term measures to increase output, as well as appropriate adjustment of demand to available resources, domestic and foreign. In these cases it may be best to focus on supply-side measures—including liberalization designed to improve resource allocation, reduce rent seeking and so on.

Finally, trade liberalization might be attempted when a country suffers from high inflation. Countries may certainly have balance of payments problems without at the same time suffering from high inflation, and may also, though less commonly, suffer from high inflation without having a balance of payments problem. For a country to sustain inflation higher than that of its trading partners without a balance of payments problem emerging, the nominal exchange rate would need to be continuously or frequently depreciated. Normally high inflation is associated with many distortions, notably an exchange rate that is not adjusted sufficiently so that it becomes overvalued, nominal interest rates that are too low owing to controls of various kinds, and import controls designed to compensate for the failure to adjust the exchange rate sufficiently.

The question arises whether trade liberalization is possible if the fundamental factors causing high inflation (usually a high budget deficit financed by money creation) are not eliminated. The answer has to be that it is technically possible provided the nominal exchange rate is depreciated even more than it needs to be to compensate for the inflation. If there is continued high inflation, as well as a current account deficit that has to be reduced, the addition of trade liberalization to the policy program will require continuous nominal depreciation to compensate for continuous inflation, and in addition nominal depreciation to bring about sufficient real devaluation. The real devaluation must be large enough both to improve the current account as required (to switch expenditure from foreign towards home-produced goods and to switch the pattern of output from nontradables toward tradables) and to compensate for the employment and current account effects of trade liberalization. In addition, real expenditure must, of course, be reduced.

In practice high-inflation countries have often failed to depreciate their currencies sufficiently, and high inflation gives rise to the distortions mentioned above. These are caused essentially by attempts to deal with symptoms rather than causes—to control interest rates, prices, imports and so on. It is then important to tackle the fundamental causes, and either to associate liberalization with a credible and adequate inflation stabilization program or even to make the attainment of reasonable stabilization a precondition for substantial trade liberalization.

Liberalization with Fixed Exchange Rates

A special problem arises with countries that are part of a currency zone and where, therefore, the nominal exchange rate cannot be unilaterally devalued. Such countries are short of a policy instrument. It is assumed here that even a once-for-all exchange rate adjustment is ruled out. Trade liberalization will still require real devaluation, but this cannot be brought about by nominal devaluation.

There are, then, two possible approaches to the problem. The first is to rely on gradual liberalization, keeping the domestic rate of inflation below the inflation rate in trading partner countries. Liberalization would, in the first instance, reduce demand for domestically produced goods, and the moderation of increases in domestic wages and prices that might result would then restore competitiveness, bringing about the required real devaluation. To avoid significant output losses it would be necessary for the liberalization to be gradual.

This approach is likely to work if there is significant inflation abroad. Without this, some downward flexibility of domestic nominal wages and prices would be needed.

The second approach is to reduce or eliminate distortions not by removing trade restrictions but rather by establishing a uniform ad valorem tariff combined with a uniform ad valorem export subsidy at the same rate. This package of policies would have effects similar to that of a devaluation, at least on trade and output of goods (given that usually it cannot be applied to services). But it may also present administrative problems, and it might be difficult to attain complete uniformity and to resist pressures from sectoral interests to provide lower or higher tariffs in particular cases.

Quantitative restrictions might at first be replaced by tariffs and then the tariffs might be adjusted either quickly or slowly in the direction of uniformity. At the same time, export subsidies would be necessary to avoid an import substitution bias. The revenue from tariffs would finance the export subsidies and, if there is a trade deficit, there would still be a net revenue yield from the tax-subsidy system. If the system could not be applied to services, some distortion would remain. Possibly the whole level of tariffs and subsidies could be gradually reduced in time, and even eliminated eventually, if there is reasonable flexibility of domestic wages and prices.

Capital Market and Trade Liberalization

For the more usual cases, where exchange rates can be altered, a matter that has been much discussed has been the relationship between trade liberalization and capital market liberalization. This discussion has been stimulated by the experiences of Argentina, Chile, and Uruguay where some liberalization of both kinds took place in various orders. In Argentina capital market liberalization (for a limited period) came first and in Chile trade liberalization.14

It is clear that one kind of liberalization is possible without the other. Some countries have very open capital markets but restrictive trade regimes while others have extensive international capital controls but relatively free trade. Among industrial countries during the Bretton Woods era controls on international capital movements were the norm while trade was progressively liberalized, and this has also been true until very recently within the European Community. It is striking that in recent years the tendency to increased protection or protectionist pressures in some major industrial countries has coincided with the rapid growth of the international capital market and a general tendency to capital market liberalization.

There are three important links between the two kinds of liberalization.

First, capital market liberalization involving the freeing of domestic interest rates and the removal of controls on inward and outward capital flows may lead to greater capital inflows than before. Not only would removal of controls on inflows, including direct investment, encourage this, but removal of controls on outflows (provided the liberalization is expected to last and economic conditions support the policy thrust) might also, since it would reduce the risk that capital cannot be repatriated. With more foreign capital available domestically it is then particularly desirable that the relative profitability of domestic industries gives a true indication of social profitability, so that investment is directed in optimal directions. Hence some trade liberalization should ideally precede capital market liberalization if the existing protection system is very distorting.

The need to get the signals right also applies when new investment is wholly domestically financed, but the argument is strengthened when major capital inflows are in prospect. It is unfortunate if foreign capital flows primarily into heavily protected industries so that low benefits to the country result, and possibly there could be a social loss, the local consumers of the protected products in effect subsidizing foreign capital. In addition, foreign companies become yet another interest group in support of maintaining protection.

Second, the process of capital market liberalization is likely to affect the real exchange rate, possibly quite sharply for a limited period, as a portfolio adjustment takes place. If domestic interest rates had been held down by controls and are now raised, capital will flow in, or at least there will be pressures in that direction. This effect will be strengthened if investors’ perceptions of the security (and opportunity to repatriate) of investment in the country improve. The nominal exchange rate and, with it, the real rate may then appreciate. The exchange rate may, of course, depreciate if the portfolio adjustment involves net capital outflow, which might occur if controls on outflows were initially severe or if decontrol were expected to be temporary. But the more common experience has been for the exchange rate to appreciate consequent upon financial liberalization.

If the real exchange rate appreciates, this will make trade liberalization inconsistent with current account balance. The adverse effects of appreciation on import-competing industries will intensify the effect of the trade liberalization. Of course the real appreciation caused by capital market liberalization will be temporary, but it does create problems. Furthermore, the appreciation will render exporting less instead of more profitable.

If capital tends to flow out after financial liberalization, the exchange rate will move in the right direction for the current account (by depreciating) but it will overshoot, since the extent of depreciation required for the current account to be maintained with trade liberalization is less than that required for a temporary current account surplus to accommodate capital outflow.

Third, if the capital market has already been liberalized, trade liberalization, or even just the expectation of it, may give rise to capital movements which then affect the exchange rate.

One possibility is that depreciation comes to be expected—since, as noted earlier, it would eventually be required if the current account is not to change once trade liberalization takes place. Hence the exchange rate depreciates in advance of actual trade liberalization. As also noted above, this might be desirable, bearing in mind that export expansion will be required and is likely to take time. But the opposite possibility must also be allowed for.

The acceptance of proposals for trade liberalization, possibly combined with other structural reforms, may make local and foreign investors more “bullish” about the economy, expecting higher and more secure profits. Hence the tendency will be for capital to flow in and for the exchange rate to appreciate. This is similar to the effect discussed above where the capital inflow and the appreciation were caused by capital market liberalization. In the present case the capital market is already liberalized and the same effect results from expectations stimulated by trade liberalization and other structural reforms. In both cases the appreciation can create problems by making industries producing tradable goods in general less profitable at a time when industries that are losing their protection will suffer a loss of profitability in any case while export industries need to become more profitable.

To sum up, opening the domestic capital market to the world market is likely to make it more difficult to manage the exchange rate. The rate will be put under capital-market-determined pressures, and this presents problems if it is desired to fine tune the exchange rate as part of a major trade liberalization exercise. On the other hand, there seems little reason to slow down capital market liberalization if trade liberalization is piecemeal and gradual. Furthermore, sometimes capital market liberalization may be inevitable because of the breakdown or high administrative costs of controls.

One might also note the case where the nominal exchange rate is kept fixed or, at least, the more common case where there is some degree of intervention designed to moderate exchange rate changes. If capital market liberalization or trade liberalization stimulates capital inflow, as seems quite possible, there will then be a buildup of reserves and, if the effects are not sterilized, a domestic monetary expansion, and hence domestic inflationary effects. In the latter case there would be a real appreciation with the same kinds of adverse effects already discussed. If the capital movements are short term it will clearly be desirable for the monetary effects to be sterilized.

The problem would be greater if the net effect of capital market liberalization and of the expectation of trade liberalization (and hence of eventual devaluation) were for capital to flow out rather than in. The reserves will then decline, and—if the exchange rate regime is not changed—the balance of payments situation may then inhibit the trade liberalization process.

A similar difficulty can arise when an import liberalization is not expected to last—when a government has not succeeded in making the program credible to private traders. Imports may then flood in, in expectation of the reimposition of quotas. An immediate, though temporary, balance of payments problem will then be created. If depreciation of the exchange rate is ruled out, a tightening-up of monetary policy may be needed in all these cases.

Protection in Industrial and Developing Countries

The argument is often heard in developing countries that the recent revival of protectionism in the industrial countries justifies a reluctance to liberalize by the developing countries. Does protection in the industrial countries, combined with the need to improve current accounts because of the debt situation, call for inward-looking policies by the developing countries? This involves the general question whether protection in one group of countries can justify or even necessitate the protection policies of another group.

In very broad terms, protection overall is much higher in almost all developing countries than in industrial countries. On the other hand, again in broad terms, protection in industrial countries has been increasing, at least since 1980, and the threat is of further increases, while protection in developing countries has on the whole not changed much (with some exceptions, where there has been liberalization) and all the proposals, if not prospects, are for further liberalization.

The question has then been raised whether there are or will be “inequities in global liberalization.” Why should one part of the world move in one direction—a direction that is favorable for the world system—when another part (the source of most of the preaching) is moving in the opposite direction?

One approach to this question focuses on prospective current accounts. It is said that industrial countries as a group are not willing to live with current account deficits (excluding interest payments) especially if the United States eliminates its deficit; hence the developing countries cannot have or sustain the non interest surpluses required to meet their interest obligations and eventually even repay some of their debts. So there is no point in developing countries pushing exports, and growth will have to be associated with import substitution.

The argument is fallacious on the basis of the discussion earlier: once the possibility of exchange rate adjustment is allowed for, current accounts do not depend on protection; rather, protection determines whether a given current account outcome is obtained with more or less import substitution relative to export expansion.

If industrial countries do not allow developing countries to improve their current accounts, a choice still has to be made by developing countries between expanding exports and hence also being able to import more, or (taking a particular case) keeping exports and then also imports constant, hence lowering the ratio of imports to GNP in the process of growth. And this choice raises the familiar protection issues that have nothing to do with current accounts. For any given current account balance, choices can still be made between inward-looking and outward-looking growth, and, as noted earlier, the empirical evidence as well as economic analysis suggest that outward-looking growth is generally better for developing countries.

It should also be borne in mind that industrial country markets are still quite open for most products. So far, their protectionism has not in fact stopped a steady rise in the share of developing countries’ exports of manufactures in total consumption of manufactures in industrial countries (even in clothing and textiles). Furthermore, except in clothing, the share is generally quite small, so that the scope for expansion is considerable. In 1983 in the United States the share was only 3 percent for all manufactures and 15 percent for clothing. In 1973 the percentages were respectively 1.1 and 5.6 (Balassa and Michalopoulos, 1985).

The morality argument (as it might be called) should also be dismissed. Liberalization by developing countries would benefit industrial countries and the world system, just as liberalization by industrial countries would benefit developing countries and the world system. It is then asked why developing countries should generate these benefits when the industrial countries are failing to do so, and, in fact, are moving in the opposite direction. The answer is that liberalization by developing countries would also benefit the developing countries themselves, and most debates, like the discussion in this paper, are concerned with defining these benefits.

The broad point can be put as follows. Protection by industrial (developed) countries reduces the gains from trade in both parts of the world. It damages both the residents of the industrial countries in the aggregate (though particular sectors may benefit) and it damages the developing countries, especially when the protection discriminates against their exports. Adding protection by developing countries further reduces the gains from trade in both parts. It is this broad point that is the key one: even if protection in industrial countries does increase, a bad example being set and the interests of the developing countries being damaged as a result, it would not be in the developing countries’ interests to forgo their own liberalization for that reason.

The adverse effects on developing countries of a significant increase in protection by the industrial countries need hardly be restated. The developing countries’ terms of trade would deteriorate as a result. With export revenue lower than otherwise and real incomes reduced, the tax base would decline and hence the task of attaining fiscal balance would become more burdensome. A current account improvement would require more import compression than otherwise. If there is some rigidity of real wages, unemployment would probably increase.

For any given level of protection or liberalization in developing countries, an increase in protection by industrial countries, if directed against the exports of the developing countries, will shift the relative profitability of exporting and import substitution in developing countries in a trade-compressing direction. It will make exporting relatively less profitable and import-substitution relatively more profitable. Thus development will tend to become more inward looking. This would be an appropriate response for the developing countries which would result even if all resource allocation decisions were based on the market signals facing them—including the signals distorted by the protection of the industrial countries. Protection by the developing countries themselves does not have to increase for this result.

If simultaneously there is actually some liberalization by developing countries, the reduction in trade and the shift to inward-looking development would be modified and possibly offset. Their own policies would have become more outward-looking but because of the shift in the opposite direction in the policies of the industrial countries, their development pattern—governed by the relative prices facing domestic producers—would not necessarily, on balance, shift in an outward-looking direction.

The question remains as to how the benefits to the developing countries of liberalization (or possibly of protection) by the developing countries themselves are affected by the protection policies of the industrial countries.

First, protection by developing countries can conceivably be used as a bargaining device to reduce industrial country protection. Sometimes it may then be justified to postpone unilateral liberalization if there is a chance that a good reciprocal bargain can be struck. The developing country would gain from unilateral liberalization, but it would gain even more if, as a result of its own willingness to liberalize, developed countries engaged in some reciprocal liberalization of their own restrictions against the developing country’s exports. Nevertheless, the possibility of using a system of protection as a bargaining chip is a doubtful argument for providing protection in the first place. Protection generates domestic interest groups that will oppose liberalization and that will not be happy to see the basis of their profits and employment transformed into a bargaining chip, even if scope for such bargaining exists.

Second, the threat of import restrictions by industrial countries can justify the governments of developing countries imposing voluntary export restraints. These are themselves a form of trade-restricting intervention. Such restraints are acceptable to developing countries either because they provide a means of improving their terms of trade or—more commonly—because the alternative is the imposition of import restrictions by the importing countries themselves. The objective of the industrial countries in seeking these restraints from foreign suppliers is to protect their own producers, and this could alternatively be achieved by imposing tariffs or import quotas. The revenue from tariffs goes to the treasury of the industrial country and that from quotas goes to the license-holder, who is usually a local trader or producer (unless quota rights are sold). By contrast, excess profits from voluntary export restraints at least go to the developing countries’ exporters or, alternatively, to their governments in the form of revenue from export taxes or the sale of quota rights.

As noted earlier, there is no argument here for the imposition of general restrictions on imports by developing countries. Nor is there an argument for general restriction of exports to all destinations. The argument is for export restraints only in particular cases where there is a threat of import restrictions by developed countries.

Returning to the main issue and leaving aside the voluntary export restraint cases which apply only to a limited group of products, if tariffs and import restrictions in industrial countries are given and unaffected by how much protection there is in developing countries, they do not alter the case for liberalization by developing countries.

This conclusion ignores terms of trade effects, which may produce some gains for one group at the expense of the other group. But such gains for developing countries would only be short term, since developing countries are relatively small in supplying total world consumption, other than in the case of a limited number of primary products. The major exception, of course, is oil. Hence one should not expect much medium-term gain in the terms of trade to result from their intervention policies.


The Cambridge Economic Policy Group produced their argument for protection when a country has a current account problem in Cambridge Economic Policy Group (1976) and elsewhere, and the argument is analyzed in detail in Corden (1985), on which this discussion is based.


The issue of the process of trade liberalization and how it relates to macroeconomic and other policies is currently being researched in a World Bank project involving the study of 37 liberalization episodes in 19 countries. For a preliminary report, see Papageorgiou, Michaely and Choksi (1986). Many of the liberalization issues are discussed in contributions to Choksi and Papageorgiou (1986).

A Review of Analytical Issues
Author: W. M. Corden
  • Anjaria, S. J., N. Kirmani, and A. B. Petersen, Trade Policy Issues and Developments, International Monetary Fund Occasional Paper 38 (Washington: IMF, 1985).

    • Search Google Scholar
    • Export Citation
  • Balassa, Bela, Outward Orientation,” World Bank Discussion Paper (Washington: World Bank, 1985).

  • Balassa, Bela, The Importance of Trade for Developing Countries,” World Bank Discussion Paper (Washington: World Bank, 1986).

  • Balassa, Bela, and Constantine Michalopoulos,Liberalizing World Trade,” Development Policy Issues Series Report (Washington: World Bank, 1985).

    • Search Google Scholar
    • Export Citation
  • Bergsten, C. Fred and William R. Cline, The United States-Japan Economic Problem, Policy Analyses in International Economics 13 (Washington: Institute for International Economics, 1985).

    • Search Google Scholar
    • Export Citation
  • Bergsten, C. Fred and William R. Cline, and John Williamson,Exchange Rates and Trade Policy,” in Trade Policy in the 1980s, ed. by C. Fred Bergsten and William R. Cline (Washington: Institute for International Economics, 1983).

    • Search Google Scholar
    • Export Citation
  • Cambridge Economic Policy Group, Economic Policy Review (Cambridge: Cambridge University Press, 1976).

  • Choksi, Armeane M. and Demetris Papageorgiou (eds.), Economic Liberalization in Developing Countries (Oxford: Basil Blackwell, 1986).

  • Cline, William R. (ed.) Trade Policy in the 1980s (Washington: Institute for International Economics, 1983).

  • Cline, William R. The Future of World Trade in Textiles and Apparel (Washington: Institute for International Economics, 1987).

  • Corden, W. Max, Trade Policy and Economic Welfare (London: Oxford University Press, 1974).

  • Corden, W. Max, The Normative Theory of International Trade,” in Handbook of International Economics, ed. by P. B. Kenen and R. W. Jones (Amsterdam: North-Holland, 1984).

    • Search Google Scholar
    • Export Citation
  • Corden, W. Max, Protection, Growth and Trade: Essays in International Economics (Oxford: Basil Blackwell, 1985).

  • Edwards, Sebastian, The Order of Liberalization of the External Sector in Developing Countries, Essay in International Finance 156 (Princeton: Princeton University, 1984).

    • Search Google Scholar
    • Export Citation
  • Eichengreen, Barry, The Political Economy of the Smoot-Hawley Tariff, National Bureau of Economic Research Working Paper 2001 (Cambridge, Massachusetts: NBER, 1986).

    • Search Google Scholar
    • Export Citation
  • Finger, J.M. and Andrzej Olechowski,Trade Barriers: Who Does What To Whom,” in Free Trade in the World Economy, ed. by Herbert Giersch (Tübingen: J.C.B. Mohr, 1987).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, Annual Report on Exchange Arrangements and Exchange Restrictions (Washington: IMF, various issues).

  • Johnson, Harry G., Optimal Trade Intervention in the Presence of Domestic Distortions,” in Trade, Growth and the Balance of Payments, ed. by Robert E. Baldwin et al. (Amsterdam: North-Holland, 1965).

    • Search Google Scholar
    • Export Citation
  • Lawrence, Robert Z. and Robert E. Litan, Saving Free Trade: A Pragmatic Approach (Washington: The Brookings Institution, 1986).

  • Nogues, Julio J., Andrzej Olechowski, and L. Alan Winters,The Extent of Nontariff Barriers to Imports of Industrial Countries,” World Bank Economic Review (Washington), Vol. 1 (September 1986).

    • Search Google Scholar
    • Export Citation
  • Organization of Economic Cooperation and Development, Costs and Benefits of Protection (Paris: OECD, 1985).

  • Papageorgiou, Demetris, Michael Michaely, Armeane Choksi,The Phasing of a Trade Liberalization Policy: Preliminary Evidence,” World Bank Discussion Paper (Washington: World Bank, 1987).

    • Search Google Scholar
    • Export Citation
  • Salvatore, Dominick (ed.), The New Protectionist Threat to World Welfare (Amsterdam: North-Holland, 1987).

  • Stern, Robert M. (ed.), U.S. Trade Policies in a Changing World Economy (Cambridge: MIT Press, 1987).

  • World Bank, World Development Report (Washington: World Bank, various issues).