Policy, Planningi, and Research WORKING' PAPERS Trade Policy Country Economics Department The World Bank June 1989 WPS 191 Openness, Outward Orientation, Trade Liberalization, and Economic Performance in Developing Countries Sebastian Edwards The paper provides a critical review of the existing empirical literature that deals with the relationship between trade orienta- tion and economic performance. Using a model that avoids the shortcomings of most current measures of trade orientation, the author finds strong support for the hypothesis that, other things being equal, countries with a less distorted external sector grow faster than countries with a more distorted external sector. The Policy. Planning, and Research Complex distributes PPR 'Voding Papers todissemanatethe findings of work in progress and to eneourage the exchange of ideas among Bank staff and aU others interested in development issues. Ihese papers carry the names of the authors, rflect only their views, and should be used and cited accordingly. The rindngs, interpretations, and conclusions are the authors' own. They should not be attributed to the World Bank, ins Board of DLrectors, its management, or any of its member countries. [ilcy, Planning, and Research < ~~Trade Policy A major shortcoming of the current policy middle would be an open economy - one with a debate is its increasingly dichotomized, ideo- neutral trade regime, in which trade represents a logical tone. In the last few ycars, not only have 'arge share of GDP. A liberal regime would fall positions become more rigid, tbut there is also between open economy and export promotion. increasing confusion about the meaning of key concepts such as liberalization and outward * Liberalization being defined as a process orientation. The author detects three main that makes greater use of the price system. This shortcomings in the current debate: makes the trade regime more transparent, and brings domestic prices closer to world prices. * By thinking in dichotomies, we have lost The vagueness of this definition disappears as the notion oi trade regimes as existing along a soon as we provide infornation about the initial continuum. trade regime and the intensity of the liberaliza- tion process. X We no longer understand "liberalization" as a process that can operate at different intensities. Is there a relationship between how liberal- ized a trade regime is (how close it is to laissez * A number of writers confuse "liberaliza- faire) and economic performance? Is there an tion" with "laissez faire," with the result that "optimal" trade regime? If so, is that optimal people sometimes advocate what seem to be regime independent of a country's specific contradictory positions (e.g., favoring openness characteristics? and an outward orientation at the same time as opposing liberalization). The author discusses the new theories of economic growth and assesses their usefulness To rescue the policy debate, trade regimes for analyzing the relationship between trade can be thought of as existing along a continuum, orientation and growth in the developing coun- with the terms tries. Using a growth model that relates trade orientation to the ability to absorb technology * Outward and inward orientation denoting from the rest of the world, he tests the model whether policy emphasis is on the domestic using an index of trade orientation that is (1) market or on intemational trade. objective, (2) allows for a continuum of regime, and (3) comparable across countries. * Trade and closed describing the degrce of an economy's openness. At one (the closed) end The results - using cross-country data- of the spectrum would be import substitution strongly support the hypothesis that, other things (deliberate discrimination against ;mports that being equal, countries with a less distorted compete with domestically produced goods). At extemal sector grow faster than countries with a the other end would be export promotion. In the more distorted external sector. This paper is a productof the Trade Policy Division, Country Economics Department, Copies arc available free from the World Bank, 1818 H Street NW, Washington DC 20433. Please contact Maria Ameal, room N8-073, extension 61466 (60 pages with charts and tables). The PPR Working Paper Series disseminates the findings of work under way in the Bank's Policy, Planring, and Research Complex. An objective of the series is to get these fimdings out quickly, even if presentations are less than fully polished. The findings, interpretations, and conclusions in these papers do not necessarily represent official policy of the Bank. Produced at the PPR Dissemination Center Openness, Outward Orientation, Trade Liberalization and Economic Performance in Ueveloping Countries Sebastian Edwards Table of Contents I. Introduction ......................................4 11. The Policy Literature on Trade Regimes: A Review .................................... 6 Il DefiningaTradeRegime .................................... 6 11.2 The Need for a Continuum .................................... 12 11.3 A Digression on Liberalization, Effective Protection and Domestic Resource Costs .15 iln. The Empirical Evidence on Trade Regimes and Economic Performance in Developing Countries .17 IV. Recent Theoretical Developments on the Relation Between Trade and Growth .34 V. New Evidence on the Relation Between Trade Regine and Economic Performance .39 V.1 Trade Orientation, the Absorption of Technological Progress and Growth ............................ 39 V.2 An Index of Trade Intervention .................................... 45 V.3 Trade Restrictions and Cross Country Differences in Growth: Some New Evidence ................................. 47 VI. Concluding Remarks ................. ................... 5 1 References .................................... 56 This paper is part of a project on "Trade Regimes and Development" being carried out at the Trade Policy Division. I have benefited from comments by Bela Balassa, Vinod Thomas, Jim de Melo, Mohsin Khan, John Pencavel, Peter Montiel, John Nash, Ricardo Faini, Ramon Lopez, Gene Grossman and the partici- pants of the CECTP Trade seminar. I am also grateful to Guiliermo Calvo and Ed Leamer for help! ul discussions, and to Miguel Savastano for very helpful assistance. Sebastian Edwards is of the University of California, Los Angeles, and the National Bureau of Economic Research. 1 I. Introduction Tnere is by now agreement among a large proportion of the economics profession that countries that have relied on "outward oriented" development strategies have done better over the medium and longer run than those countries that have adopted "inward looking" strategies. Recently, and to the surprise of many, even ECLA/CEPAL -- not exactly known for its endorse- ment of outward policies -- has recognized that the excesses of import substitution have been very costly for Latin America; some of its senior staff members have recommended that in the future export promotion should play a more central role in Latin American development policies (see Bianchi et al. 1987). There seems to be relatively less agreement, however, on whother "trade liberalization" packages have played an important role in the performance of the outward orienred economies. For example, in a recent paper presl3nted at the World Bank-IMF Conference on adjustment with growth, Sachs (1987:' ques- tioned the idea that trade liberalizations are indeed a required component of successful outward oriented strategies. Making reference to the experiences of the East-Asian countries -- Japan, Korea, Singapore, l'aiwan, and Hong Kong -- Sachs argues that these countries' success was to a large extent due to an active role of government in promoting exports in ar. environment where imports had not yet been fully liberalized, and where macroeconomic (and especially fiscal) equilibrium was fostered (see also Sachs, 1988). Whether one agrees with Sachs depends on how outward orientation, export promotion and trade liberalization are defined. In fact, in crder to have a meaningful discussion about the role of trade regimes and commercial policies it is crucial to know exactly what we mean by each of these terms. 2 Unfortunately this doesn't seem to happen anymore. Somehow in the midst of the policy debate of the last few years we seem to have lost our control over language and what for some is a "liberalized economy" for others is a "nonliberalized" economy; for some "outward oriented" means a certain thing, while for others it means a different one. There is little doubt that the policy discussion on trade regimes has become highly 4deologized. Perhaps the best example of the current confus- ing state of affairs is that the case of South Korea is now considered an example of the validity of different (almost opposing) views regarding commercial policy. For some (i.e., World Bank 1987) Korea is the best example of an outward oriented liberalized economy, while for others (Collins and Park 1988, Sachs 1987) Korea is a prime example that in order for a small developing economy to grow (very) fast it should avoid an abrupt liberalization. Certainly, either one of these camps is not right in its assessment, or the participants of the debate are giving a different meaning to the same terms. Recently, we have witnessed a shift in the meaning of "liberalization". In the more traditional policy literature of the 1960s and 1970s trade liberalization was defined in a very general way; what economists usually meant was some relaxation of trade and exchange controls. In fact, in the by now classical NBER study on trade regimes directed by Bhagwati (1978) and Krueger (1978) a liberalization episode was defined as a more extensive use of the price mechanism that would reduce the anti-export bias of the trade regime. In her review article on the problems of liberalization presented at the World Bank conference on the dynamics of liberalization, Krueger (1986) reaffirmed this general definition and argued that even a (real) devaluation in the presence of QRs constituted a liberalization episode. 3 These are indeed very mild definitions of liberalization. In fact today very few people will raise an eyebrow about them. However, more recently, and in some quarters "trade liberalization" has acquired a more drastic connotation, meaning an elimination of QRs coupled with a severe reduction of import tariffs to a uniform level of around 10 percent. Moreover, the term "economic liberalization" has in many ways, become synonymous of free market oriented policies with minimum or no government intervention at any level. In a way it has become synonymous of "laissez faire". The difference between the "old" and "new" definitions of trade liberalization is, to a large extent, one of degree or intensity. A devalu- ation in the presence of QRs, or the replacement of QRs by (quasi) equivalent tariffs is a mild form of liberalization. However, the reduction of tariffs (with no QRs) to a uniform 10% or, for that matter, the complete elimination of tariffs is a very drastic liberalization. In order to clear- ly understand the different issues involved in policy discussions it is, then, crucial to specify the intensity of liberalization we are referring to. Unfortunately this is not always done these days. As a result of the ideologization of the policy debate, more and more people are discussing these issues as if they were either "black" or "white", with no room for different shades of grey in between. However, an increasing number of authors have recently begun to recognize these problems. For instance in an interesting paper on the experience of the East Asian nations, Bradford and Branson (1987) say: Part of the controversy undoubtedly derives from the use of loosely fashioned phrases which sound like dichotomous typologies when in fact more rigorous specifications of meaning would reveal that they define different points along a spectrum of policy regimes rather than stark alternatives. (page 17) 4 Bhagwati (1986) has also recognized the need to sharpen the debate by clearly defining the key terms. He says: What exactly is meant by an export promoting trade strategy? Unless we are clear on that critical question, we cannot properly debate the merits of the strategy and its alternatives. Clarification of the question is therefore important, especially as everyday usage of this phrase echoes many different notions that are wnolly unrelated. (page 11) Richard Cooper (1987) has referred to the dangers of the debate on trade regimes becoming excessively ideological. He argues that the imprecision of language is at the very heart of this danger. Since his view captures in a very clear way the shortcomings of the debate, it is worth- while quoting at some length from him: [L]abels may become slogans. Words such as "liberalization" and "privatization" and "adjustment and growth" may become slogans used by both proponents and opponents of particular courses of action. If this happens, the debate can occur without sufficient refinements and therefore without being really joined. (page 516) Regarding the intensity of liberalization Cooper says: [E]conomics ... has continuous gradations and therefore gives rise to both subtle distinctions and to the possibilities for compromise. That is especially true of such terzs as "liberalization" ... (page 516) And more importantly: [I]t is necessary to distinguish between different types of liberalization to make clear that liberalization can be viewed as a proqess rather than a state and to disassociate liberalization from laissez-faire. (page 518) The purpose of this paper is to analyze at a broad level the question of the relation between trade orientation, liberalization and economic performance avoiding the ideological overtones that have plagued so much bf the recent literature. In particular, this paper tries to synthesize the discussion by establishing some general organizing principles. In doing this, the literature on the subject is analyzed. While emphasis is placed 5 mainly on the empirical evidence, the new -- and quite exciting -- theoretical literature on growth is briefly reviewed. Also, I present new evidence on the relation between trade regimes and economic performance. The novelty of this impirical analysis is that, contrary to the existing literature, It uses at- objective index of trade intervention. The paper is organized in the following way. Section II deals with policy literature on trade regimes. Here the literature is reviewed and it is pointed out that, at some time, the notion of a continuum of regimes was replaced by a dichotomized view. It is strongly argued that in order to make progtess in the policy discussion it is very important that we return to the original notion of a continuum of regimes. Also in this section the increas- ingly controversial concept of "liberalization" is discussed. Section III reviews (some of) the empirical literature on trade orientation and economic performance in developing countries. Here the question is what e-o we know about this relationship, if anything. The approach taken focuses on scrutinizing the methodology used by different studies. Section IV reviews the new theoretical literature on growth, and discusses what these papers have to offer to the debate on the relation between trade policy and growth. Section V presents some preliminary new evidence on the relationship between trade orientation and economic performance. The empirical analysis presented here tries to overcome two important limitations of previous studies: (1) the lack of an objective (as opposed to subjective) and contitnuous indicator of the extent of distortions of a trade regime; and (2) the difficulty in comparing any indicator of this kind across countries. Finally, the paper closes with a summary and with a brief discussion on directions for future research. 6 II. The Policy Litgrature on Trade RegLaes: A Review The abundant policy literature on trade regimes and economic performance has mainly focused on two issues: (a) defining alternative trade regimes; and (b) determining the relationshlp between trade regimes and growth. In this section I provide a general and selective review of this literature, focusing on these two questions. I also discuss briefly the different measures of "protection" that have been used in studies on tradb reforms. I1.1 Defining a Trade Regime Tho purpose of this subsection is to discuss alternative definitions of trade regimes that have been used ir. the literature. A particularly important objective of the section is to illustrE_e the way in which the meaning of "liberalization" has changed in the last few years. Little, Scitovsky and Scott (1970) and Balassa (1971) are the pioneering modern studies on trade orientation and economic performance. These influential works discussed the characteristics of two broadly defined trade regimes -- import substitution and exp.rt promotion -- and dealt with some important issues related to the transitional period from one regime to another. No effort was made in these studies, however, to provide a detailed taxonomy of trade regimes. In the classic NBER study directed by Krueger (1978) and Bhagwati (1978) we find for the first time a systematic devel6pment of concepts that allow a more or less formal classification of trade regimes. In this study two key concepts wero introduced into the discussion: (1) the degree of bias of the trade regime; and (2) the premi that is created by the existence of quantitative restrictions (QRs) in the importables market. 7 The "bias" of the trade regime is defined as the ratio of the exchange rate effectively received by exporters (EERx) to the exchange rate effect- ively paid by importers (EERm). If this ratio is lower than one it is said that the trade regime is biased against exports. More specifically the effective exchange rate for exports is calculated as the nominal exchange rate applied to exports (Ex) corrected by export subsidies s and other incentives to exports r:1 EER - Ex (l+s+r). On the other hand, the effective exchange rate for imports is defined as the nominal exchange rate applicable to imports EM, corrected by the average (effective) import tariff (t), other import surcharges (n) and the prem- ium associated to the existence of quantitative restrictions (P):2 EERM - EM(l+t+n+P). Naturally, if there are unified nominal exchange rates for commercial transactions then Ex - EM - E. The degree of bias of the trade regime is given by: EER E (l+s+r) B E^- x- (1) EZRM EM,(l+t+n+P) Two points are worth making at this point. First, this definition of bias is based on averg&2 incentives. We can thus have a country that protects some sectors but that still, on average, does not exhibit an anti- export bias (B < 1). Second, this definition of bias implicitly allows for 1Naturally, if exports are taxed, s and/or r will be negative. 2Krueger's (1978) definition is slightly different. We have decided to use the definition corresponding to Bhagwati's (1986) reinterpretation of Krueger because it facilitates the discussion. 8 a conti of regimes; B can be high, low, somewhat high or rel'tively low. Even though for analytical purposes we can (and should) define threshold values for B -- B - one being the more natural threshold -- this approach is not based on a dichotomized view of trade regimes. In order to evaluate the effect of trade policies the Krueger-Bhagwati study combined the concept of bin with the definition of five phases in the evolution of trade regimes. Phase I is characterized by across-the-board imposition of quantitative controls, usually associated to a balance of pay- ments crisis. During Phase II the control system bec more complex and discriminatory, increasing the anti-export bias of the regime. Phase 3 is the beginning of the liberalization process and is characterized by the implementation of a (nominal) devaluation and relaxation of some QRs. During Phase 4 further steps towards replacing quotas by tariffs are implemented. In Phase 5 the economy has become fully liberalized: current account tr'&nsactions are fully convertible and QRs are not used any longer. In a much needed synthetic paper Jagdish Bhagwati (1986) makes an effort to further clarify the definition of trade regimes, using the framework developed by him and Krueger in the NBER study. According to Bhagwati it is useful to distinguish between three basic trade strategies: 1. Import Substitution (IS) Strateg: Bias < 1 2. Export Promotion (EP) Strategv: Bias 8 1 3. Ultra Export Promotion Strategy: Bias > 1. Three points are worth noting. First, within the tradition of the NBER study there still is room for a continuum of regimes. We can face a mild IS strategy -- where B is slightly below 1 -- or we can have a regime heavily oriented towards IS -- with a bias index well below one. Second, Bhagwati has chosen to label the neutral regime (B - 1) "export promotion". 9 Naturally if B 3 1 there is no bias against exports, but there isn't either a bias in favor of exports as the label EP usually implies. Third, and perhaps more importantly, neither of these regimes is necessarily related to a "hands off" or "laissez-faire" strategy. In fact an EP strategy can be achieved with a highly interventionist government that, on average does not distort the ratio (1+s+r)/(l+r+n+P). Moreover, an ultra export promoting strategy, where B > 1, may require significant government intervention. As discussed below, this fact has recently prompted some experts to make an additional distinction between a negtral trade regime and a liberal trade regime. It was argued in Section I that the notion of "trade liberalization" has changed significantly in the last decade. In the original NBER study liberalization was defined in a very general way as a more extensive use of the price mechanism. Anne Krueger has, in fact, maintained that position throughout the years. At the specific operational level, in the synthesis volume of the NBER project Krueger defined trade liberalization as a reduc- tion in the premium resulting from the use of QRs. There is no mention of zero or even very low tariffs. In fact, it must be noticad that according to this definition we can have a "liberaliced" economy with very high tariffs. This, indeed, is a possibility that Krueger recognizes: Inspection of the definitions of bias and liberalization shows that there is no necessary reason, at least in theory, for a connection between the two. A regime could be fully liberalized and yet employ exceedingly high tariffs in order to encourage import substitution. The regime would then be liberalized and highly biased. (1978, p. 89, emphasis added) This definition of liberalization is indeed very different to the way in which the concept has been used (as synonymous with free trade) in the 3Krueger (1978, 1981, 1984). 10 last few years. An important implication of Krueger's definition is that a devaluation in the presence of QRs constitutes a step towards liberalization. The reason, of course, is that if there are QRs a devaluation will reduce the premium obtained by those that have an access to the quota. On the other hand, as noted before, the definition of an unbiased or neutral trade regime (B = 1) does not preclude in any way an active role for the government. We can attain a B - 1 either with high levels of s, t, r, n, and P or with no government intervention in international trade. This has recently prompted a number of authors to make a distinc- tion between a neutral trade regime (B - 1) and a liberal regime where there is almost no government intervention. According to this view governments should not only aim at reducing the anti-export bias -- and definitely should not try to promote exports via intervention -- but should try to implement a liberal strategy.5 Lal and Rajapatirana (1987) are, perhaps the clearest representatives of the new liberal view regarding trade regimes. They -eject government intervention both to protect imports and to promote exports: Of course, exDort promotion can be as inefficient and chaotic as protection .... The liberal position on trade and growth (which we support) is different. As a first step it entails a neutral trade regime. (p. 208, emphasis added) And they go on to say: The case for a liberal trade regime ... becomes part of the general case for mar1gets against mandarins .... [I]f one accepts the need for restraints on the natural and often irrational 4Notice that until now we haven't said anything about other sectors such as the labor market, the capital market or the capital account. Some of the issues that arise when we look at many markets are addressed below. 5See, for example, World Bank (1987), Lal and Rajapatirana (1987). 11 dirigisme of mandarins in most developing countries, then the adoption of a liberal trade regime ... becomes an important means to this end. (p. 209) Naturally, according to this view liberalization is different from the milder and more specific notion employed by Krueger. For the advocates of liberal regimes, liberalization is (almost) synonymous with free trade and "laissez faire". Although Lal and Rajapatirana recognize that "it is important to distinguish between the degre and pattern of protection" (p. 197, emphasis added), their analysis still has a distinctive dichotomized flavor. Criticizing some of the implications of this "liberal' view, Jeffrey Sachs, Rudiger Dornbusch and other prominent researchers have recently argued against "liberalization" in the developing countries. For example, commenting on the links between liberalization and the resolution of the debt crisis Sachs (1987) notes: The current focus on liberalization is distracting attention from the more urgent needs of the debtor countries. ... (T]he attempts to stimulate exports at all costs through trade liberalizations or aggressive depreciations of the exshange rate can often undermine a stabilization program and thus postpone a resolution of the debt crisis" (p. 294) And later he says: [Elxport orientation can be pursued without an across-the-board import liberalization and can be fostered by an activist government. (1987, pp. 322-3) It is clear that Sachs is not using the term liberalization in the traditional Krueger sense. He is not even allowing for the possibility of different degrees of liberalization. It would seem fair to say that for him liberalization has become an all-or-nothing event, where if you opt for it 12 the government has to ompletely relinquish its role.6 The Need for a Continuum Recently, Bradford and Branson (1987) and Edwards (1988) have argued that in order to rescue the policy debate from its current overly ideologized overtones it is necessary to thin.k, once again, of trade regimes as a continuum. Table 1 contains the typology suggested by Bradford and Branson (1987). They argue that outward and inward orientation are very general categories that serve to denote whether the policy emphasis is put on the domestic markets or on international trade. Thus they are placed in the 1u.ddle of the continuum. They then use the terms trade and closed economies to cap- ture the degree of openness of a country. In their analysis these concepts are not necessarily the result of policies but rather lendogenous economic outcomes". An economy can be closed -- that is, total trade (imports plus exports) does not exceed 5% -- either because its resource endowment is similar to the world's endowment or because of trade impediments. An 2Pe economy is one with a neutral trade regime and a large share of trade in GDP. Notice that there are various categories of open economy, depending on the role of policies in other sectors. Towards the closed economy end of the spectrum the authors place import substitution (IS). This is a regime that deliberately discriminates against imports that compete with-domestic- ally produced goods. Again the authors distinguish between different shades of IS. On the other extreme of the spectrum Bradford and Branson have placed different types of export promotion. This regime is characterized 6However, in a recent paper Sachs (1988) criticizes "significant" liberalizations. Still, he does not make an explicit distinction between the effects of different types of liberalization programs. 13 TABLE 1 Definition of Trade Regimes: The Bradford-Branson Continuum Autarky No trade Dirigisme "Delinking" Self-reliance Closed Exports and imports less than 5% as a share of GDP Economy Import (a) Discriminates against all imports through Substitution controls: EERm > EERx (b) Selective discrimination (c) Mild and limited applications ("left wing deviations') Inward Priority given to the domestic economy Markets Orientation Outward Priority given to exports Markets Orientation Trade Exports 15% or more as a share of GDP Economy Open Economy Internal liberalization EERx - EERm (a) tradable goods (b) (a) + nontradable goods (c) (a) + (b) - macroeconomic variables Export (a) Uniform subsidies for all exports: promotion EERx > EERm (b) Selective subsidies: Industrial policy Import substitution ("right wing deviations") Dirigisme NOTE: EERx and EERm are the real effective exchange rates for exports and imports, respectively. Source: Bradford and Branson (1987, p. 16). 14 by active government intervention aimed at boosting export performance. Although not complete, this taxonomy is very useful. One of the missing elements, however, refers to placing the "liberal regime" a la Lal and Rajapatirana within this spectrum. This type of regime should be placed in between the open economy and export romotion categories of the table. In general, there is little doubt that considering a trade regime as a continuum rather than as a dichotomy will greatly improve the quality of the debate. However, while it is not difficult to define such a continuum at an analytical level, it is quite hard to find empirical counterparts to these concepts. As will be argued in detail in Section III, one of the main limitations of the existing empirical evidence on the relation between trade regimes and economic performance has been precisely its inability to design and construct adequate indexes of trade orientation. Ideally, such indexes should satisfy the following three criteria: (a) Objectivity. It is necessary that the index used to classify a country's trade regime (or to locate it within the spectrum) is based on relatively objective criteria and not on subjective preconceptions brought in by the researchers.7 (b) Cont- inuity. The ideal index should allow for different shades of grey in between the extreme regimes. We would like to be able to determine whether or not small movements in either direction in the trade orientation spectrum will affect economic performance. (c) Comparability. The index should be useful for cross-country comparisons. Unfortunately, gigantic data limitations have precluded researchers to construct indexes of regimes that satisfy these three characteristics. In Section IV of this paper, however, an attempt is made to capture the effects 7The main problem with the subjective approach rests on how to classify the countries in the first place. Where should Korea or Brazil be placed? 15 of the inclusion of two indexes that satisfy (some of) these requirements on a regression that explains the economic performance of a group of developing countries. II.3 A Digression on Liberalization. Effective Protection and Do1estic Resource Costs Mos: of the indicators commonly employed to characterize the: trade regime prevailing in a particular country -- such as the bias index B -- reflect, at best, the avejage level of protection given to two broadly defined sectors of the economy. However, almost every country 1?rotects different sectors at different rates. In particular, final goods usually have different tariffs and QRs than capital goods and intermediate inputs. A number of authors have indeed recognized that having information on both the average level and the complete structure of protection are important. As a result, most studies on trade liberalization -- including the Bhagwati- Krueger (1978), the Balassa (1982) and the Krneger (1983) projects -- have tried to compute Effective Rates of Protection (ERPs) and Domestic Resource Costs indexes (DRCs) to measure more accurately the way in which input- output relations affect the extent of "true" protection granted to a particular sector of the economy. The concept of effective protection was pioneered by Corden (1966), Balassa (1965) and Johnson (1965) and tries to capture in a single indicator the rate of protection granted to value added in a given industry.8 During the 1960s and early 1970s the literature on measurement and implications of 8The rate of effective protection to industry j is defined as r - (VAj-VA*)/VA*, where VA is domestic value added, and VA* is "world value added taken to be a proxy of the most efficient way of producing J. Assuming a linear relation between inputs and outputs -- where a denotes an input-output coefficient -- the ERP for industry J can be rewitten as: J - (tJ-Eaijti)/(l-Eaij), where ti is the tariff on input i. 16 ERPs developed extensively and a number of studies computed ERPs for large groups of countries.9 In spite of the efforts made by a number of experts to g.neralize the applicability of ERPs and DRCs, both concepts have important limitations within a general equilibrium framework. Perhaps the most serious problem is that in a world with more than two final goods ERPs give no information on the way in which resources will be reallocated in the case of changes in the tariff structure. This limitation of ERPs has led a number of experts to argue that the concept of effective rate of protection should be abandoned altogether (see for example Dixit and Norman 1980 and Dixit 1986; Bhagwati and Srinivasan 1983 provide a fairly detailed discussion on the properties and limitations of ERPs and DCRs; see Balassa and Schydlowsky on the relation between ERPs and DRCs). Although both ERPs and DRCs are much more limited than what it once was thought, they do provide some valuable information; abandoning them would be clearly an overreaction. More specifically, ERPs and DCRs give us informa- tion on the extent of inefficiency "society", or perhaps more accurately the "government", is willing to grant a particular sector.'° Even though this is a much more modest role than that assigned to these indicators in the past, it is a particularly important one in political economy studies of trade policy. What is critically important, however, is to be aware of the limitations of these concepts. 9See, for example the studies directed by Balassa (1971, 1982). 10The reason for this interpretation follows directly from the definition of ERP as the percentage deviation of domestic unit value added from world (Ind efficiently produced) unit value added in industry J. 17 An important implication of the shortcomings of ERPs and DRCs is that traditional policy recommendations based on these concepts should be reanalyzed to verify if they still hold under fully specified general equilibrium frameworks. For instance, one recommendation that does not survive in a general equilibrium setting is the 'uniform tariff" proposal. Generally speaking, in equilibrium models with intermediate inputs, government budget constraints and other realistic features a uniform tariff structure will be optimal only by mere coincidence. 1 I11. The Emnirical Evidence Trade Resimes and Economic Performnce in Developing Countrieg3L The purpose of this section is to critically review the empirical evidence provided by the literature on policy orientation and economic performance. Most studies in this area have asked a variant of the following question: "What does the empirical evidence say regarding the relationship between trade orientation, exports and economic growth?"13 In reviewing this literature it is convenient to classify the different contributions in four broad and related categories: "1There may be, however, some practical reason to advocate a uniform tariff. One such reason considered by Harberger (1984) is that only if there are uniform tariffs (and subsidies on exportable inputs) we can know for sure the ERPs structure; they will all be equal. What is not clear, however, is why we should worry about the structure of ERPsl A more convincing reason for advocating a uniform tariff stems from the political economy of protection. aI m grateful to Miguel Savastano for his assistance in preparing this section. 13 Most of these studies have been characterized by a lack of rigorous theoretical underpinnings. Naturally, the most important contributors to this literature are perfectly aware of its lack of theoretical base. See, for example, Krueger (1983, p. 42). 18 (a) kulticountry studies (primarily syntheses volumes of large-scale projects); (b) Studies that investigate specifically the relation between exports and output growth; (c) Studies based on the Kravis (1970) decomposition method that try to determine the relative contribution of external demand, competitiveness and diversification factors in the evolution of a country's exports; and (d) Studies that compare the Heckscher-Ohlin predictions regarding the factors composition of trade with the actual trade patterns. Table 2 lists 17 recent empirical studies classified in these four groups. As pointed out, a common problem faced by all these studies consists of the lack of an appropriate indicator of rtrade orientation". Any attempt to relate trade regimes and growth requires data on either how a trade regime evolves through time in a particular country, or data that can be compared across countries.i This, of course, is not an easy task. For example, the recent Choksi, Michaely and Papagergiou (1986) project at the World Bank constructed time series of a subiective index of trade liberalization for 19 developing countries. These indexes, however, were not comparable across countries and, thus, could not be used to investigate empirically whether different degrees of trade liberalization can explain cross country differences in economic performance. In the Krueger-Bhagwati project, on the other hand, the researchers faced difficulties in constructing a series of the Bias (B) index discussed in the previous section. The reason, of 14Balassa (1985, 1988) derived a trade orientation index as the percentage difference between the ac6tual and predicted per capita exports. *. *r 19 TABLE 2 Classification of Selected Empirical Studies on Trade Regimes Group A: Summary Results of Multicountry Projects on Trade Regimes No. of Countries Period 1) Balassa, B. (1971) 2) Krueger, A. Liberalization Attempts and Conseau- 10 1950-72 (1978) ences, 1st Synthesis volume of NBER project on "Trade Regimes and Economic Development". 3) Bhagwati, J. Anatomy and Conseguencas of Exchange 10 1950-72 (1978) Control Regimes, 2nd Synthesis volume of NBER project on "Trade Regimes and Economic Development". 4) Balassa, B. Development Strategies and Economic 11 1960-73 (1982) Performance, Synthesis of World Bank Project on "Development Strategies in Semi-Industrial Economies". 4a) Balassa, B. "Exports and Economic Growth: Further 11 1966-73 (1978) Evidence,fn Journal of Development Economics. 5) World Bank World Development Report, summarizes 41 1963-73 (1987) preliminary results of World Bank project 1973-85 on "Timing and Sequencing of Trade Liber- alization Reforms". Group B: Studies Measuring Contribution of ExRorts to Growth in LDCs in a Neoclassical Production Function-Tyge Framework 6) Tyler, W. "Growth and Export Expansion in 55 1966-77 (1981) Developing Countries," Journal of Development Economics. 7) Feder, G. "On Exports and Economic Growth," Journal 19 1964-73 (1983) of Development Economics. 31 (2 groups) 8) Kavoussi, R. "Export Expansion and Economic Growth: 73 1960-73 (1984) Further Empirical Evidence," JouXnal of DeveloRment Economics. 9) Ram, R. "Exports and Economic Growth: Some 73 1960-70 (1985) Additional Evidence," Economic Develop- 1970-77 ment and Cultural Change. 20 Table 2 (continued) No. of Cnotrles Period 10) Balassa, B. "Exports, Policy Choices and Economic 43 1973-79 (1985) Growth in Developing Countries After the 1973 Oil Shock," JouXnal gf Develjpment Economics. 11) Balassa, B. "The Newly Industrialized Developing 12 1973-78 (1981) Economies After the Oil Crisis," Weltwirtschaftliches Archiv. 12) Jung, W. & "Exports, Growth and Causality in 37 1950-81 Marshall, P. Developing Countries," Journal of (1985) Development Economics. 13) Hutchinson, M. "Exports and Growth in Developing 34 1950-81 and Singh, N. Economies: Identifying Externality (1987) Effects," UCSB Working Papers. Group C: Studies Using Kravis (1970) "Decomposition Method" to Analvze Exgort Performance in IDCs 14) Love, J. "External Market Conditions, Compe- 27 1950-78 (1984) titiveness and Diversification in LDCs Exports," Journal of DeveloRment Econ. 15) Kavcussi, R. "International Trade and Economic 52 1967-73 (1985) Development: Recent Experience of 1973-77 Developing Countries," JournaL of Developing Areas. 16) Singer, H. and "Trade Policy and Growth of Develop- 52 1967-73 Gray, P. (1988) ing Countries: Some New Data," 61 1977-83 World Developmnent. Grouo D: Studies Using Heckscher-Ohlin Framework to Analyze Pattern of Trade and Factor Productivity 17) Krueger, A. Trade and Emnlgvment in Developing 12 1960-75 (1983) Countriea, Synthesis volume of NBER project on "Alternative Trade Strateg- ies and Employment. 18) Havrylyshyn, 0. "The Direction of Developing Country 45 1963-78 (1985) Trade: Empirical Evidence," Journal of Development Economics. 21 course, had to do with the problems involved in computing the premium associated to QRs, as well as with the absence of reliable data on import tariffs. Researchers have developed two types of strategies to deal with this problem. (a) Some, as the Krueger-Bhagwati study, the World Bank (1987) and Choski et al. (1988), have resorted to the construction of subjective indexes of trade orientation. (b) Another group of researchers have chosen to decompose the question of the effects of trade orientation on economic performance into two stages. The first stage basically amounts to assuming -- without testing rigorously -- that a more liberalized regime will encourage exports via a reduction of the antiexport bias.16 At the second stage, then, the researcher usually tests whether higher exports (or a more rapid growth in exports) have indeed been associated with a higher rate of output growth (Michaely 1977, Balassa 1978, 1982). Multicountry Studies The Krueger-Bhagwati NBER project dealt with 10 countries -- Brazil, Chile, Colombia, Egypt, Ghana, India, Israel, Philippines, South Korea and Turkey -- during the period 1950-1972. Detailed data on a number of variables, including ERPs and DRCs rere used to classify each country at a particular moment in time in one of -Aie Phases previously defined in the 15The subjectivity of the indexes employed in the Krueger-Bhagwati study is, however, rooted on some objective indicators. In particular, as mentioned below, in the empirical analysis Krueger tries to control for the different degrees of distortions prevailing in each country. 160f course the difficulty in computing B is at the heart of them assuming that this relation between B and exports holds. 22 project.17 Three groups were then formed: (i) countries that seemed to have prograssed steadily from phases I to IV (Brazil, Israel, South Korea); (ii) countries that cycled back and forward between phases II and IV (Chile, Colombia, Ghana, and the Philippines); and (iii) countries that remained in Phase II for long periods (India, Turkey, and Egypt). Using the wealth of information generated in the ten country studies Krueger (1978) and Bhagwati (1978) attempted to determine whether, on average, the relaxation of the anti-export bias had affected exports and the overall economic performance. We will review first Krueger's synthesis of this project. An important finding is that devaluations -- that is, the movement from Phase II to Phase III -- generally resulted in important reduc- tions in the premium on import licenses and, thus, in a lower anti-export bias. She also investigated the extent to which changes in the trade orientation of a country contributed to an expansion of exports and whether this phase movement implied some costs in terms of output growth. From a regression analysis Krueger then found that while the real effective exchange rate played a significant role in determining exports, the dummy variables included to control for the different liberalization phases were not significant. With respect to real output she found that on average, higher exports were associated with higher GNP, with no significant costs being related to the transition towards a more liberalized trade regime. Perhaps Krueger's most important conclusion from this monumental cross-country study was that in order to attain a sustainable growth in exports (and, thus, GNP) 17Recall from Section II that the phases were defined as follows: (I) systematic imposition of across-the-board QRs; (II) control system becomes more discriminatory with strong antiexport bias; (III) devaluation and relaxation of some QRs; (IV) continued liberalizativ , with greater emphasis on the price mechanism than on QRs; (V) fully liberalized and convertible regime, no QRs remain. 23 it is very important to make a credible commitment to maittain a high (i.e., depreciated) real effective exchange rate for exports. 8 On the other hand, in his synthesis volume, Bhagwati (1978) analyzed in detail the static and dynamic gains from trade liberalization. He argued that the pre-reform data on ERPs and DRCs clearly suggest that liberaliza- tion in these countries generated important static gains ln the form of efficiency improvements. He also confirms Krueger's finding that biased regimes negatively affect exports growth and that these are generally responsive to price incentives. However, regarding the dynamic gains of a change in trade orientation, however, he found no evidence linking a smaller bias to either technological superiority or to higher savings ratios. Balassa (1978, 1982) summarizes a large World Bank project that focused on 11 countries -- Argentina, Brazil, Chile, Colombia, Mexico, Israel, Yugo- slavia, India, South Korea, Singapore and Taiwan -- for the period 1960-73. He establishes a fourfold classification according to the timing and extent of the export promotion policies adopted by each country. He then evaluates the effectiveness of these policies using the two-step methodology described above. First Balassa analyzed whether export incentives fostered exports growth and, in the second stage, he investigated the effects of an expansion of exports on output growth. In the first stage of the analysis Balassa makes use of two proxies for quantifying export incentives: the rate of growth of manufacturing exports and the change in the export output ratio in manufacturing. He found that these variables were consistently higher in those countries that followed sustained export promotion policies. In the 18She then argues that the best way to make sure that high RER will be maintained is by supplementing liberalization with increased foreign borrowing and a crawling peg. 24 second stage of the analysis Balassa employod simple rank coefficients and cross country regressions. His main finding was that outward orientation, represented by a high growth rate of exports, had positive effects on output growth. Although it was backed by individual country studies Balassa's synthetic piece had some limitations that, in fact, have been present in most of the subsequent literature. Some of these limitations refer to the arbitrary definition of export incentives, the lack of a role for real exchange rates in the explanation of export performance and the inability to deal convincingly with causality issues. The recent World Bank project directed by Choksi, Michaely and Papageorgiou is another ambitious multicountry study that has analyzed in great detail the liberalization experiences of 19 countries since World War II. Since this project has not been completed, it is not possible yet to discuss all its findings. However, some preliminary results have provided important information regarding the dynamics of trade liberalization. For example, Michaely (1988) reports that in only one of 36 liberalization episodes the trade reform resulted in an increase in the aggregate rate of unemployment. This indeed is a striking result that suggests that trade reforms do not have, even in the short run, a negative effect on output and employment. Also, this study confirms previous findings that indicate that consistently higher real exchange rates that support a liberalization process result in faster output growth. In addition to the generalized measurement problem, this study seems to have some limitations related to the lack of firm theoretical basis, as well as to uneven empirical investi- gations across countries.19 The World Development Reyort (1987) presents 19I say that it "seems" to have these limitations, because the final report has not yet been released. 25 some preliminary results obtained in this project, suggesting that countries with a more liberal trade regime have outperformed the more inward oriented economies in the last 25 years. Exports and Output Growth A number of studies focused their attention on the relationship between exports expansion (see, in particular, Balassa 1978, 1982) project and related work by Michaely (1977). This literature has focused on testing the robustness of the findings that indicate a positive effect of exports on growth. This has been mainly done by controlling for the effects ef addi- tional variables, extending the sample of countries, and improving on the measurement of the variables employed. It also provided a (weak) theoret- ical base to the empirical work on this area by postulating that exports should be considered as an additional factor of production in the aggregate neoclassical-type production function of an open economy. The rationale for adding exports as another factor is based on possible positive externalities and technologicat diffusion effects they generated by more rapid exports. In this sense, most of these studies derived their estimated reduced forms from an augmented neoclassical production function such as: Y - f(K,L,X) (2) where Y is the level of output,20 K is the capital stock (the growth of which was usually proxied by the level of investment, I), L is the labor force and X are exports. Table 3 provides a detailed summary of four representative articles of this growing literature. The table shows the sample and period covered by 20To account for the fact that exports are a component of GNP these measures of output usually referred to the non-export GNP, i.e., GNP - total exports. 26 TABLE 3 Summary of Results of Studies Relating Exports Growth To Output Growth (Group B) Study Sam le/Period Argued Improvement Estimated Equation Tyler -55 middle income -Less restrictive - P + pft + p2t + (1981) LDCs. Eliminating sample of countries. 1 those with income per No selection bias + P3X capita (Y/N) < 300. with respect to -Period: 1960-77 Balassa's studies. Feder -2 samples of LDCs -Distinguishes theor- 2 - p1(I/Y) + p2t + (1983) 1) 19 middle income etically and empiric- countries ally between: factor + X* X/Y (1) 2) 31 middle income productivity differ- + low income ential and external- Y - 71(I/Y) + -2t + countries ity effects of EP policies in a 2-sector + X* X/Y + model (export and non- export sectors). + 4k (2) Kavoussi -73 low and middle -Examine whether the Y - 0 + ft + 02 + (1984) income LDCs, charac- positive correlation 1 2 terized by whether between X and t + P3 (1) (Y/N)60 $ $360 and also holds for low Y- a + at + t+ excluding oil income countries. + a3X + a4km (2) exporters. (Xm - share of manu- -Period 1960-78. factured Xs on total exports). Ram -73 LDCs, low and -Avoid selection 0- + pii/Y + 0 t + (1985) middle income dist- bias. 2 inguished by -Test for absence of + 3X + P4D (low (Y/N)77 g $300 simultaneity bias. income dummy). -2 subper'.ods: 1960- 70, 1970-77. 27 Table 3 (continued) Study Results Other Tests Tyler - 0.57 for the whole sample. Pearson and Spearman rank cor- (1981) Drops to 3 - 0.055 when 6 OPEC relations between 9 and countries are excluded. a) manufacturing output b) domestic investment c) is growth rate d) manufactured X8 growth rate (all were positive and significant). Feder 1) 3 - 0.4 and significant, Performs regression of (1) for (1983) introduction of exports as 17 developed countries and explanatory variable increased found that z3 was significant the R . but 74 wasn't, i.e., product- 2) A - 0.13 statistically ivity benefits of export expan- significant in both samples. sion were exhausted in those Evidence of substantial prod- countries. uctivity differentials in addi- tion to the usual externalities effect. Kavoussi 1) A3 > 0 and significant for Spearman rank correlations (1984) the whole sample and for both between 2 and GAP for subsamples. But was twice as a) The whole sample (positive large for the middle-income and significant). than for the low-income b) The 2 subsamples (higher countries. for the middle income than for 2) a4 is not significant for the low income group). the sample as a whole. c) The sample excluding For the middle-income group: countries where Xm > 44% A a3 < 0 not significant (reduce the significance of A a4 > 0 significant. the middle income group). For low income groups: opposite results. Ram -? > 0 and significant in all -White's test is performed to (1985) the regressions. check for homoskedasticity of -A3 larger in the second disturbances and simultaneity and lower for the low-income bias (favorable result). countries. 28 each study, their distinctive features with respect to previous works, the functional form postulated, the econometric results obtained and the statistical procedures employed to test their hypotheses. As can be seen the main concern of these studies has been to analyze the magnitude, significance and sensitivity of the parameter associated with exports growth. In general a large and significant value for this coefficient has been found. This has been interpreted as supportive evidence in favor of the benefits of export-promoting policies. A secondary concern of this line of research has been to determine whether the positive relation between exports and output growth is robust to the inclusion of low-income LDCs in the regression sample; the results in this respect, however, have not been conclusive. A major shortcoming of these empirical findings is the fact that they r.re obtained without controlling for the actual trade orientation (bias) of the countries included in the sample.21 Furthermore, these studies tend to derive policy recommendations without having tested the effectiveness of (some of) the policy instruments they advocate to use. In this sense, then, the results obtained by them should be considered, as Bhagwati (1986) points out, as only a very indirect evidence in favor of the adoption of export promoting policies. Another criticism to these studies has been focused on the issue of causality. The empirical findings obtained from standard regression techniques do not permit us to determine whether the growth of exports "cause" the growth of output -- as these studies claim -- or if the 2Two recent papers by Balassa (1981, 1985) have tried to overcome this deficiency by including policy related indicators as explanatory variables in the regression. However, the variables chosen as proxies for the trade orientation of the regime raise some doubts regarding the relevance of the results. 29 causality runs in the opposite direction -- as has been claimed, fore instance, by Findlay (1984). Some recent empirical works have addressed this issue by applying Granger "causality" tests to the relation between exports and output.22 Table 4 summarizes the methodology and results of two of these works. As can be seen the evidence reported in these papers is quite mixed; the causality from exports to output is accepted only for a small number of countries, and in the majority of cases the hypothesis of two-way causality cannot be rejected. These results seem to confirm the need for a more detailed and careful specification of the link between trade orientation, exports growth and economic performance. Studies Based on Kravis' "Decomposition Method" and on the Heckscher- Ohlin Framework The general issue of the effects of alternative trade regimes on the evolution of exports and on the overall economic performance of LDCs has been addressed rather differently by the studies of group C in Table 2. The objective of these studies has been to identify the relative importance of the different factors that explain the rate of growth of exports in develop- ing countries. In particular, these papers have followed the methodology established by Kravis (1970) and have assumed that the evolution of exports of a given country is determined by three factors: i) the evolution of external demand; ii) the change in the degree of competitiveness (measured as the change in the country's world market share); and iii) the change in the degree of export diversification (proxied by the share of non-tradition- al exports in total exports). While the first of these determinants is ' 22There is, of course, a great deal of controversy on the real meaning of "causality' and on the significance of the Granger-type tests. See, for example, Leamer (1985). Notice, also, that Sections 3 and 4 of Balassa (1985) are not subject to the causality criticism. 30 TABLE 4 "Causality" Tests on the Effects of Exports on Growth Study Saple/Period Test Jung & 37 LDCs for Granger "causality" tests and F-tests for the Marshall the period signs of coefficients in: (1985) 1950-1981 Rt Wiai Qt-i + j b t-j +t it-Eici It -i + Ed1X t-j + vt (Box-Pierce Q to test for autocorrelation) -Lag length generally r3stricted to 2. Hutchinson 34 LDCs for Granger "causality" tests and F-tests to & Singh the period capture export externality effects in: (1987) 1950-1985 A~~~~~~~~ nt -Zai 8- + Ebjfit- + Zckt- + P It i Zdi at-i + Ze nt j + Zf a t-s + vt where: n - (t-k) non-export sector growth St - AI/Y, change on Investment/GDP ratio. (continued) 31 Table 4 (continued) Study InterDretation of Outcome Results Jung If and Then 4 countries classified as EP & 5 countries classified as ERG Marshall i. E > O EP 3 countries classified as IGE (1985) i 2 countries classified as GRE (export promoted growth) -22 countries couldn't be properly classified. t - 2 Es > 0 IGE Including countries always 8 a characterized as EP examples like (internally generated X8) South Korea, Taiwan, Brazil. I - 2 EG < 0 ERG (export reducinp, growth) 2 - Q Ea < O GRE (growth reducing exports) if c<-> I or 4 not significant, no interpreta- tion is given. Hutchinson .Lf= and mhmn -10 countries (mainly primary & producing) showed evidence of Singh X.n f 6 > 0 Export exter- positive export externalities. (1987) nalities -3 countries showed evidence of A n - I ESe > 0 Growth exter- growth externalities. nalities -3 countries showed evidence of X - fi E8 < 0 Negative negative externalities. or externalities 18 countries couldn't be classi- A n E 2 a < O fied, including Brazil and South Korea. 32 taken as exogenous for the developing countries, the last two are usually interpreted as being the result of domestic trade orientation policies. The empirical analysis of these studies has consisted, basically, on computing correlation coefficients between the value of exports and the three explanatory factors for a particular group of countries in a given period. Depending on the size and significance of these coefficients the studies have determined whether the observed rate of growth of exports can be considered primarily a result of the adoption of export-promoting policies or of the conditions prevailing in the world market. Overall, the results obtained by these studies are not fully conclusive. The relative importance of domestic as opposed to external factors seems to be very sensitive to the group of countries selected and to the period chosen for the estimation. In general, though, it appears that, as would have been expected, the exports performance of LDCs is strongly dependent, but not deterministically so, on the existence of favorable conditions in the world market. It is not clear, however, what practical and new implications can be derived from this finding; especially since the link between the competitiveness and diversification factors and the changes in the trade regime appears to be extremely indirect and weak. An alternative methodology for investigating the consequences of a change in the trade regime of LDCs has been used by the studies of group D in Table 2. Taking the Heckscher-Ohlin model as a benchmark, these two works performed different tests that confirmed that the factors content of exports of developing countries is consistent with these countries' factors endowment. Having determined that LDCs exports are relatively labor intens- ive, these studies relate the actual distortions in factors markets (Krueger 1983) and in the pattern of trade across regions (Havrylyshyn 1985) with an 33 excessive anti-export bias. Thus they emphasize on the potential gains in terms of efficiency, employment and exports growth of the adoption of a more outward oriented trade regime. However, despite their sound theoretical base, these studies' empirical analysis is somewhat incomplete. In particular, the fact that they do not provide an adequate and comparable measure of the degree of distortions created by the trade regimes prevailing in the countries analyzed undermines their policy recommendation. Summary The evidence discussed here -- plus a number of other contributions not included because of space considerations -- provides a broad picture in support of the joint hypotheses that domestic policies affect exports and that exports expansion has been associated with more sustained output growth. It is indeed very difficult to dismiss the massiveness of these results. It is fair to say then, that a firmly established emirical finding, is that a greater participation in international trade helps growth. However, from a policy perspective this is not an overly helpful finding, since it does not add too much to the current debas.e on openness, laissez faire, export promo- tion and trade liberalization. We would be hard pressed to say which set of policies would b*.tter achieve an efficient process of integration into the world economy or to give precise prescriptions on what is the "optimum" degree of openness. Two basic problems remain unsolved in this regard. The first is the lack of clear theoretical underpinnings that would support these empirical results. The second refers to the repeatedly mentioned measurement problem. Further development in this line of research would require, at least, the use of an appropriate index of trade orientation in attempts to explain cross-country differences in economic performance. 34 IV. Recent Theoretical Developments on the Relation Between Trade and Growth As pointed out in Section III the traditional policy and empirical literature on trade and growth have not been based on firm theoretical grounds. In fact, as Helpman (1988, p. 3) has recently pointed out "current theory is not suitable to deal in a satisfactory way with these alternative views (on the relation between trade policy and growth)".23 The most serious problem of traditional growth theory -- that is, of Solow's neoclassical growth model and its many variants -- is that in equilibrium (i.e., in the steady state) the rate of growth of output is independent of policy related parameters. This means that two identical countries that only differ with respect to their vectors of policies will still grow, in equilibrium, at the same jate. An additional and well-known property of these models is that the equilibrium rate of growth of an economy is independent of its savings rate. Thus, if two otherwise identical countries have different savings ratios, they will still grow at the same rate in the steady state. The model predicts, however, that they will have different equilibrium levels of income per capita. These limitations of traditional growth models can be illustrated with a simple aggregate production function. Let Y be total output, F( ) be a constant returns to scale production function, K the capital stock, L the labor force and A a parameter that represents "technology". Further, let's assume that technological progress is of the "labor augmenting" type. Aggregate output is then determined in the following way: 23Findlay (1984) and Smith (1984) provide excellent surveys on the relation between growth and trade. They don't deal, however, with the role of domestic commercial policy; their main interest is, in fact, to explain how growth will affect trade patterns. 24See Dixit (1976). 35 Yt F(Kt LtAt) (3) Denoting the rate of growth of output by g, the rate of growth of the labor force by n, and (net) investment by I, we have: g . q)t + en + e( )4 which is the equation that has been used extensively in accounting empirical studies of -he determinants of growth. These studies have found that, on average, tectnological progress accounts for between 1/2 and 1/3 of real growth in the developing countries. The presence of the investment ratio in equation (4) suggests that a higher level of investment will be translated into a higher rate of output growth. This, however, is nt an equilibrium situation. In the steady state the output capital ratio should be constant and equation (4) collapses to: g - n + (A) (5) In many ways this expression is disappointing, since it states that the equilibrium rats of growth is determined by the sum of two exogenous vari- ables -- the rate of growth of the labor force and the rate of technological progress. Furthermore, this setting not only predicts that the equilibrium rate of growth of an economy will be completely independent of policy measures, but also that the rates of output growth will converge across countries 25 As a consequence of the recently renewed interest in growth theory, a number of authors have tried to devise simple models that are free of some of the limitations of the Solow-type framework. Lucas (1988) provides a 25This last statement assumes, of course, that countries face the same exogenously determined n and (A/A). 36 good survey of the analytical issues involved in this line of research and suggests that the explicit introduction of human capital accumulation in a two-good open economy model can indeed explain the nonconvergence of rates of growth across countries. Becker and Barro (1988), on the other hand, have looked at the economics of population in order to understand the deterlminants of the rate of growth of the labor force and output. Increasing REturns to Scale and Endogenous Technological Proeress In general, the recent efforts to construct more satisfactory theories of growth have focused on relaxing two key assumptions of the neoclassical model: (1) the constant returns to scale technology; and (2) the exogen- eity of technological progress. The first assumption has been replaced by the introduction of increasing returns to scale. Work in this area has been pioneered by Romer (1986a,b, 1988), who has suggested a number of growth models with economies of scale that are either exogenous or endogenous to the firm.26 Although in both settings the presence of increasing returns makes it possible to generate an equilibrium situation with nonconvergent rates of growth across countries, the models with internal economies of scale provide a more appealing explanation. These models use an Ethier (1982)-type production function which has tbe property that a greater degree of specialization in the productive process increases efficiency. This pioneer work does not deal directly, however, with the question of the relation between trade policy and growth. With respect to the er.dogenous technological progress, Helpman and Grossman (1988) have recently build an elegant two-country model that emphasizes the role of R&D in the growth process. This model assumes the 26 Allyn Young (1928) was an early proponent of explanations of growth based on increasing returns. 37 existence of differentiated intermediate inputs produced by oligopolistic firms, and of country-specific final goods. Investment in R&D allows each country to produce "blueprints" for new intermediate inputs. The production function for final goods is, as in Romer (1988), of the Ethier (l982) variety, where total factor productivity is enhanced by an increase in the variety of intermediate inputs. In this framework, investment in R&D is translated into "acquired" comparative advantages. If there are cross- country differences in the efficiency with which R&D takes place, relative to efficiency in manufacturing, the model generates very rich dynamics. For example, if changes in commercial policy reallocate spending towards the final commodity produced by the country with comparative advantage in R&D, the equilibrium world rate of growth will decline. An important feature of this model is that in the steady state both countries grow at the same (equilibrium) rate. Boldrin and Scheinkman (1988) have developed a simple two-countries, two final goods model that is able to generate different rates of growth across countries. In this model there are infinite intermediate inputs, and each country is subject to an economy wide learning-by-doing process a la Arrow (1962). It is further assumed that the efficiency of learning-by- doing depends on the distribution of the labor force across sectors; the higher the proportion of workers employed in the modern (high technology) sector, the more efficient will be the learning-by-doing. As a result of this assumption the externality will affect the dynamics of comparative advantage. Krugman (1988) has also focused on endogenous technological progress and has developed a model based on Schumpeterian premises, where firms undertake innovations based on expected future monopoly rents. The model 38 considers a three period economy with N goods produced per period, and one factor of production -- labor. In period 1 firms have to decide whether to invest in improving the production process. If they do, they obtain monop- oly rents in period 2; in period 3, however, the new technology becomes common knowledge and the monopoly rents disappear. A crucially important feature of this model is that the existence of monopoly rents imply a tradeoff between static welfare losses (i.e., the traditional triangles) and dynamic innovative gains. For our purposes, however, the most interesting result is that "the presence of endogenous technological development increases the gains from international integration." Recent work by Shleifer (1986) and Murphy, Shleifer and Vishny (1988) also make use of the Schumpeterian framework to develop interesting models of growth. In particular, this latter paper exploits the existence of multiple equilibria to give formal underpinnings to Rosenstein-Rodan's (1961) "big-push" hypothesis. Once again, however, these models do not deal with the narrower question of the relation between trade policy and growth. There is little doubt that these new developments in growth theory have opened up a highly promising area of inquiry that will help answer a number of questions that could not be tackled satisfactorily by the traditional approach.27 However, although many of these authors refer to their research as dealing with fundamental development issues, most of these papers have been derived with the structures of the advanced countries in mind. For instance, the models of endogenous growth focus on the determinants of the rate at which innovations are create rather than on the rate at which existing innovations are r . While, undoubtedly, the development of 27Other important papers include Rebelo (1987), Kohn and Marion (1987), Barro (1988), and Jones and Manuelli (1988). 39 new inventions should be central in any attempt to explain the historical patterns of growth in today's mature economies, the question of the rate of absorption of technology is the more important one for the developing countries. A second problem with this line of research is that very little effort has been made to provide empirical support for these theories. As Krugman (1988) has recently indicated, "the priority is really not how to construct cleverer models, but how to build a bridge to reality" (p. 29). V. New Evidence on the Relation Between Trade Regime and Economic Performance The purpose of this section is to present new tests and evidence on the relation between trade regimes and economic performance. These tests differ from previous work in two respects. First, an attempt has been made to find an index of trade regime orientation that, at least partially, satisfies the three requirements discussed in Sections II and III: (i) objectivity; (ii) continuity; and (iii) comparability. Second, for the first time (at least to my knowledge) an attempt is made to test the liberal trade regime hypothesis. In effect, we make use of an index of trade intervention that penalizes equally the presence of import taxes and export subsidies. The section starts with the presentation of a minimal model on the relation between trade orientation and growth. As most of the literature reviewed in Section IV, the model emphasizes the role of technological innovations. However, contrary to that literature we focus on the rate of absorption of technological advances. V.1 Trade Orientation, the Absorption of Technological Progress and Growth In this subsection a minimal model on the relation between trade orientation and growth is derived. This model provides the basis for the cross-country empirical results reported below, and differs from the works 40 discussed in Section IV in two respects. First, it concentrates on out of steady state situations and, second, it assumes that the country of interest is a small developing country that faces given prices and, more important for our purposes, is inserted in a world where innovations take place in the advanced covxntries. That is, our economy faces exogenous technological progress; the key question, then, is how fast, and how efficiently it absorbs this technological progress. Of course, the idea of focusing on out of steady state situations when attempting to explain growth is not new. For instance, Corden (1971) follows this approach in his two-stage production model of trade and justif- ies it by stating that: "It may be uninteresting to describe a state which is many years ahead and which may indeed never be reached ... [Flocusing on a theoretical ultimate state is thus purely an expositional device ... [A] concern only with steady states would have obscured significant aspects of the trade and growth process" (p. 219). Moreover, Lucas (1988) has recently recognized that one possible way of overcoming some of the limitations of the traditional growth theory is, indeed, to explore the characteristics of out-of-steady state situations.28 Our model is based on an important insight developed by W. Arthur Lewis in his monumental work Theory of Economic Growth. In this study Lewis argues that those developing countries that are more integrated to the rest of the world will have an advantage in absorbing technological innovations generated in the advanced nations. In Lewis' words: "New ideas will be accepted more rapidly in those societies where people are accustomed to ... 28He says: "Off steady state behavior would open up some new possibilities, possibly bringing the theory into better conformity with observation ..." (p. 14). However, he then goes on to say: I do not view this route at all promising". 41 change ... [A] country which is isolated is ... by contrast unlikely to absorb new ideas quickly ..." (1955, p. 178). There are a number of ways in which Lewis' insight can be formalized at the microeconomic level. One possibility is to postulate a "learning-by- looking" process of the type proposed by Findlay (197?).29 According to that framework the mere association with newer commodities and technologies increases the efficiency with which innovations are absorbed. In this section, however, I will not try to provide the microfoundations of this absorption process, but I will derive a simple aggregative that captures Lewis' insight regarding the connection between trade orientation and growth.30 The model itself is a simple adaptation of the Nelson-Phelps (1965) study on the relation between human capital and growth. Let's start with an aggregate production function of the form assumed in equation (3) (which is reproduced here for convenience): yt -F(KtLtAt) (6) from where we have already shown that the rate of growth of real output can be expressed as: g + i(Y)ten + e()t (7) (where the notation used in Section IV has been maintained). Contrary to traditional growth models we will consider the rate of technological change as endogenous. More specifically, we will assume that 29A 1984 study by Dalham, Ross-Larson and Westphal provide microeconomic empirical support to this general view. Rodrick (1988) develops optimizing models that yield to opposite results. 30The formulation of such a micro model would have to be based on some of the new and rich literature on North-South trade. 42 in every period the change in the technology parameter A is proportional to the difference between the technological level achieved by "the" world economy, W, and the degree of technological process in the domestic country: A - O(W-A); O < O 1 (8) where O is a parameter that measures the speed at which technological progress generated in the rest of the world is absorbed by the domestic country.31 Lewis' effect is captAred by assuming that 9 is a function of the country's trade orientation. In particular, let r be a continuous index of a country's trade regime whose value achieves a minimum under free- trade (laissez-faire) and increases as the extent of trade intervention 32 grows.3 In such a case the relation between 9 and r will be given by: 9 - O(T); ,'(*) < 0. (9) Then, if it is assumed that the stock of world's technological knowledge, W grows at the rate w: Wt W0e w, (10) it is easy to show that the trajectory of the domestic stock of technology will be given by: 31An alternative would be to assume that the O index determines the proportion of innovation2 that is actually absorbed. An undesirable characteristics of this formulation, however, is that the gap between the domestic and "world" stocks of knowledge would grow unboundedly. 32Lewis is not clear on whether he considers gRennes as the key determinant of the rate of absorption of innovations, or whether the crucial variable is how distorted the external sector is. This, of course, is related to the debate on export-promotion vs. liberal regimes. In this model we interpret r as a measure of distortions. Whether this is appropriate is, of course, an empirical question. 43 At - (AO - w )eWOt + ( )W t (11) The steady state of this model, then, is characterized by three important properties. First, in the steady state the stock of domestic knowledge will be: t 90 0t At - o Woew (12) Second, the steady state will imply the existence of an equilibrium technological gap (G): G W-A w (13) A - which, in turn, will depend positively on the degree of trade intervention: aG/B? > 0. Third, the eauilibrium level of domestic knowledge, A, will be negatively affected by increases in the trade orientation index: 8A , 2 W e < 0 (14) B1r O+ 2 9 This means that a country with a higher degree of trade intervention will have a lower equilibrium level of technological knowledge, and a larger technological gap, than a. otherwise identical economy. This fact is depicted in Figure 1 where the two p schedules represent different rates of growth of A; i.e., p - A/A - O((W-A)/A]. p(tL) is the A/A schedule corresponding to a low value of the intervention index, while p(rH) cor- responds to a high value of r anid, consequently, to a larger equilibrium technological gap. This result not only has implications for the levels of output in the steady state but also suggests that as an economy goes through a trade liberalization (lowering r), it will grow faster than an otherwise 44 identical economy that maintains its degree of intervention. On the other hand, the out of steady state effects of changes in r on the actual rate of accumulation of knowledge is: a (W-A) °' < 0 (15) This expression states that, for a given gap (W-A/A,, a higher degree of trade intervention r will reduce the rate of growth of knowledge accumula- tion in the domestic country. In terms of our growth equation (7) this implies that after taking into account the contributions of capital accumulation and labor force growth, and after controlling for the techno- logical gap, a higher degree of trade intervention will reduce the rate of growth during the transition towards a more regulated regime. The empirical analysis presented below makes wide use of this result. A difficulty with the empirical application of this model, however, is that the relation between the technological gap and the 9 coefficient is nonlinear. To overcome this problem, in the empirical analysis we have mainly concentrated on a linear specification that allows us to isolate the role of trade intervention. Equation (16) provides the linear specification that captures this out of steady state property of the model: -1(I)i + en + yvi + uGi + Si (16) where gj is the growth rate of real output in country J, n is the rate of growth of the labor force in country J, * is a trade intervention index in country J, G is the technological gap in that country, and ej is an error term. The interest of our empirical study is to determine whether the coefficient of r is indeed negative as suggested by the model. A serious difficulty in 'mplementing this equation, however, refers to finding an 44a / / I1 1 /H~~~~ I -k (h I- , 45 adequate index of trade orientation. The subsection that follows deals with this issue. V.2 An Index of Trade Intervention There is little, if any, hope that researchers will ever be able to construct the ideal index of trade intervention that will capture all the subtleties of reality. Empirical research, then, has to proceed on the basis of finding proxies for this ideal index. If better, more reliable, and less controveraial proxies to this index are developed, the quality of that research will tend to improve. In a recent paper on openness Edward Leamer (1987) has used a large cross country data set to compute patterns of trade according to comparative advantage. He then takes "[t]he difference between the 'predicted' and actual trade intensity ratios ... as an indicator of trade barriers" (p. 7). In this paper we use these intervention indexes as a proxy for the extent of trade intervention. Leamer uses a traditional Heckscher-Ohlin general equilibrium model of trade as his theoretical framework. "Predicted" comparative advantage trade ratios are computed using a regression analysis that considers three goods aggregates and seven factor of production. Leamer then defines a rate of intervention which "measures the extent to which trade is distorted by pol- icy, positively or negatively" (p. 26). For every country this intervention index is defined as the ratio of the sum of the absolute value of the residuals from the regression to GNP.33 33Leamer also constructs an alternative intervention index where the denominator is the sum of the absolute value of predicted net trade. In private communication, however, Leamer has argued that for the purpose at hand using GNP as the denominator is more reasonable. 46 The Leamer index has a number of attractive features. First it is, as far as any index can possibly be, objective. No attempt whatsoever has been made a .riori to classify the countries' trade regimes. A debatable point, of course, is whether the three goods, seven factors empirical model used to predict trade is adequate to capture comparative advantage. In this respect, recent massive empirical evidence provided by Leamer (1984) himself indeed indicates that this is a reasonable model. The second attractive feature of this index Js that it reflects all types of trade interventions, independently of whether they are export promoting or import substituting. In that regard, the export promoting activities of a number of countries will be reflected in a very high intervention index. As a consequence of this property, this index can be useful to test the 'liberal trade regime" hypothesis. A third desirable property of this index is that it is comparable across countries. However, the index has some limitations. First, it has been computed for only one year (1982). It is not possible, then, to analyze the evolu- tion through time of the degree of trade intervention for any particular country. Second, the index is only as good as is the model used to predict "comparative advantage trade". Another important variable of the model derived above is the technological gap. Obviously, however, there are no direct measures of this gap. Ir this study a measure of the level of education achievement was used as a proxy for it. The specific variable was defined as the percentage of people enrolled in secondary schools as a fraction of each age group. An advantage of this proxy is that it focuses on labor augmenting technological progress, which is the type of technological knowledge captured by the model. Additionally, this variable is available for all countries for 1982 47 and for all but two in 1970.34 When other proxies for the gap were used -- such as income per capita -- the re, vs obtained were not altered in any significant way. V.3 Trade Restrictions and Cross Countra Differences in Growth: Some New Evidence The question we want to address is whether there is a relation between trade regimes and the rate of growth of output in the developing countries. More specifically we want to know if differences in trade orientation help to explain the observed differences in economic performance across countries. Ideally, an analysis on this type should capture the evolution of the relevant variables during several years, however, our index of trade orientation r is only available for one year (1982). This problem was tackled in the following way. First, it was assumed that r captured the cross-country differences in trade orientation for a period longer than 1982. Regressions were then estimated using that index togeth-er with 1970- 82 averages for output growth, the investment ratio, labor force growth and the level of educational attainment. These data were obtained from Worl Development Report 1984. Since the assumption that the intervention index applies to the whole 1970-82 period is not fully satisfactory, regressions were also estimated for the year 1982 only.35 In this case, the data on labor force and education still came from the World Development Report 1984, while those for growth and investment ratios were obtained from the IMFs 34Alternatively one could use the.Harbison-Myers index, that include enrollment rates in higher education. A limitation of this index, however, is that it is a flow index only. 35Notice, however, that what is required is that the distribution of the r's across countries captured by the 1982 index reflects its behavior along the whole period and = that the absolute values of the index are maintained. 48 International Financial Statistics. Given the likely presence of heteroskedasticity in cross-country regressions of this type, equation (10) was estimated using a weighted regression procedure, where each country's income per capita was used as the weight. The results obtained are reported in Table 5. As can be seen from equations (14.1) and (14.2), the coefficients of the investment ratio and of the rate of growth of the labor force have the expected signs and are signi- ficant in every regression. More importantly for our purposes, the coefficients of the intervention index were always negative and highly significant. These results provide evidence supporting the hypothesis that higher trade intervention affects negatively economic performance. In equa- tiona (15.3) and (15.4) we have included the product of the intervcntion index and the technological gap as an independent variable. This was done because equation (15) suggested a nonlinear relation between growth and trade intervention is nonlinear. As before the results support our hypothesis. An interesting question refers to whether the relation between growth and trade intervention is strictly linear or if we can detect some non- linearities. When a r2 term was incorporated the following result was obtained using weighted regressions: g- -1.742 + 0.286 (I/Y)j + 1.489 n - 0 039 G (-0.972) (6.353) (3.556) J (-1641) J - 3.758 r + 0.520 x2 Period: 1970-82 (-1.982) J (0.536) J N: 28 R2 0.706 As can be seen, the coefficient of the intervention index is negative and that of its square is positive. This last coefficient, however, is not significant. 49 TABLE 5 Growth and Trade Orientation in Developing Countries: Weighted Regressions Eg. (14.1) Eg. (42) Eg- (14.3) La. (14.4) Period (1960-82) (1982) (1960-82) (1982) Constant -1.924 -7.935 -4.225 -9.661 (-1.111) (-2.575) (-2.780) (-3.768) Investment Ratio 0.280 0.323 0.302 0.343 (6.533) (4.506) (5.755) (4.277) Labor Force Growth 1.461 1.992 1.496 1.916 (3.573) (2.966) (4.369) (3.142) Technology Gap -0.037 -0.020 - - (-1.611) (-0.575) Trade Intervention -2.697 -3.137 - - (-3.766) (-2.491) Trade Intervention - - -0.065 -0.052 Tech. Gap (-3.336) (-2.169) N 28* 30 28* 30 R2 0.702 0.503 0.656 0.473 The 1960-82 regressions exclude Ethiogia and Portugal because there are no data on early educational achievements. The countries included in the regressions are listed in the Appendix. Numbers in parentheses are t-statistics. R2 is the coefficient of determination. 50 It was also investigated whether the relation between trade regime and economic performance was dependent on the stage of development. Helleiner (19--), among other authors, has argued that the hypothesized positive rela- tion between openness and performance does not apply to very poor countries. According to this view only after a certain threshold of development is reached, outward orientation and export growth will have a positive impact on real income growth. In order to investigate this possibility, a new explanatory variable (INTPC) was addad to equation (16). This is an inter- active term constructed as the product of the intervention index and the level of income per capita. The following result was obtained in this case: * - -1.629 + 0.269 (I/Y) + 1.465 n- 0.038 G - 2.828 r (-0.784) (4.332) (3.507) J (-1.600) (-3.227) + 0.005 INTPC Period: 1970-82 (0.270) N: 28 R2: 0.703 Consistent with Helleiner's hypothesis the coefficient of INTPC is positive; however, it is not significant. The coefficient of the interven- tion index itself, on the other hand, remains significantly negative.36 In general, the regression analysis reported above provides support for the hypothesis that there exists a negative relationship between the degree of restrictions on international trade and economic performance in the developing countries. In all regressions the coefficient of the index on- trade impediments was negative and statistically significant. These regres- sions indicate that after taking into account the roles of capital accumulation, growth in the labor force, and technological gap, countries with higher degrees of trade intervention tend to grow, on average, slower 36When the nonlinearity and the Helleiner hypotheses were jointly tested for the year 1982 only, very similar results were obtained. 51 than countries with lower trade restrictions. Finally, it is important to notice that in the regressions reported in Table 5, as well as in the different variations of them, the R was considerably high for a cross country analysis. Those values for the coefficient of determination suggest that our model, that combines elements of the production function approach with the trade orientation literature, is capable of explaining approximately one half of the cross country variation in real growth. VI. Con&luinfRmrks This paper has discussed a number of important issues related to the literature on the relat!on between trade orientation and economic perform- ance in the developing countries. It was argued that a major shortcoming of the current policy debate is its increasingly dichotomizod and ideologized tone. In the last few years not only have positions become more rigid, but also we have witnessed an increasing confusion on what the key concepts, such as liberalization, outward orientation and so on, exactly mean. Three main conceptual shortcomings were detected in the current debate. First,ithe notion of a continuum of trade regimes (which was present in the early works) has been lost. Second, "liberalization' is no longer under- stood as a procesE that can have different intensities. And th'rd, a number of authors have confused liberalization with laissez faire. This last problem has resulted ih people advocating what, at least to some (including myself), seem to be contradictory positions, such as favoring openness and outward orientation at the same time as opposing "liberalization". 52 Defining Trade Regimes and Liberalization There is no doubt that the only way to put order back into the debate is to start by clearly defining the different concepts involved. In that regard Bradford and Branson's recent taxonomy of trade regimes, discussed in Section II, is a useful beginning. This taxonomy has a number of attractive features. First, it contemplates a continuum of trade regimes. Second, it places two dirigiste regimes at the extremes. And third, it recognizes explicitly that a neutral trade regime -- characterized by a bias ratio equal to one -- is different from laissez-faire. Undoubtedly the debate will be enriched if the concept of liberalization is used as referring to a process that can have different intensities. In that regard a definition close to that of the original Krueger-Bhagwati NBER project is possibly the most useful one: Trade Liberalization is a process that makes greater use of the 2rnce system. oaking the trade regime more transparent and bringing domestic prices closer to world Drices. This is a rather vague definition of liberalization, but the choice of it has been completely deliberate. This is because the vague- ness disappears as soon as we provide two key pieces of information: (1) the initial trade regime, and (2) the actual intensity of the liberalization process. In terms of the Bradford-Branson taxonomy an economy would be undertaking a liberalization process whenever it moves from either extreme towards the laissez faire position. Trade Regimes and Economic Performance Thus far we have discussed the definition of trade regimes and of liberalization. But, is there a relation between how liberalized a trade regime is (i.e., how close it is to laicsez faire) and economic performance? Is there an "optimal" trade regime? If so, is that "optimal" regime 53 independent of the country's specific characteristics? These of course, are the key policy questions. Their answer has both a theoretical and an empirical dimension. A well known, and disturbing property of traditional neoclassical growth models is that changes in policy affect only the steady state level of income per capita, and do not affect the equilibrium rate of growth of output. Although applied development economists have always been aware of this feature of the model they have continued using it in their empirical investigations, trying to determine whether there has been a sustained relation between trade (and other) policies and growth. The theoretical basis of these empirical analyses have usually been less than rigorous. Recently, however, we have witnessed a renewed interest in growth theory. A number of authors have tried to develop equilibrium models that are able to capture some of the more salient stylized facts of actual growth processes: First, rates of growth don't seem to converge across countries as the traditional theory suggests. Second, policy packages do seem to make a difference. Some of the important contributions to this literature were reviewed in Section IV. It was argued that an important shortcoming of these models is that they take the economic structure of the large mature economies as a starting point. In the meantime, and as the theoretical underpinnings are revised and greatly improved, the policy debate continues. In Section III the empirical literature on the relation between trade regimes and economic performance was selectively reviewed. Although this literature has provided a persuasive general case supporting the hypothesis that there is a relation between policy measures, export expansion and output growth, it has two major limita- tions. First, it lacks a firm theoretical basis. Second, it provides 54 indirect tests that don't shed much light on the issue of trade regimes. The problem is mainly one of measurement. These studies have not been able to construct a convincing objective and continuous index of trade intervention. In Section V we derived a modified neoclassical growth model and presented new cross section results that do use an index of intervention that satisfies most of the requirements discussed in Sections II and III. The model is based on Arthur Lewis' proposition that a more open economy and less distorted trade regime will result in a faster rate of absorption of technological progress origincting in the advanced countries. The model is very simple and uses an aggregative neoclassical framework to analyze the out of steady state relation between trade intervention and growth. The empirical analysis ui d a 30 country cross section data set. The results obtained provided strong support for the model indicating that, after con- trolling for the relevant variables, a higher degree of trade intervention will imply a decline in the rate of growth of output. Although these results are very encouraging, there are a number of unresolved issues that chould be tackled by future research. First, more detailed micromodels of the process of technology absorption should be explored. Second, the search for an even better index of intervention should continue. 55 t-A TABJE A. 1 Countries Included in Regression Analysis of Section V Bangladesh Ethiopia Pakistan Sri Lanka Turkey Dominican Republic Ecuador Indonesia Philippines Nicaragua Colombia Egypt Cameroon Thailand El Salvador Peru Horocco Ivory Coast Costa Rica Brazil Yugoslavia Israel Greece Argentina Panama Malaysia Portugal Trinidad-Tobago Hong Kong Singapore 56 REFERENCES Arrow, K. (1962), "The Economic Implications of Learning by Doing," Review of Economic Studies, 29: 155-73. Balassa, B. (1965), "Tariff Protection in Industrial Countries: An Evaluation," Journal of Political Economy, 73: 573-94. (1971), The Structure of Protection in Developing Countries, Baltimore: Johns Hopkins University Press. (1978), "Export Incentives and Export Performance in Developing Economies," Weltwirtschaftlichos Archiv, 114: 24-61. (1981), "The Newly Industrialized Developing Economies After the Oil Crisis," Weltwirtachaftlichos Archiv, 117: 142-94. (ed.) (1982), "Development Strategies and Economic Performance," in Development Strategies in Semi-Industrialized Economies. London: Oxford University Press. (1985), "Exports, Policy Choices and Economic Growth in Developing Countries After the 1973 Oil Shock," Journal of DeveloRment Economics, 18: 23-35. (1988), Changing Trade Patterns: An Econometric Investigation (Amsterdam: North-Holland). _________ and D. Schydlowsky (1972), "Domestic Resource Cost and Effective Protection Once Again," Journal of Political Economy: 63-69. Barro, R. (1988), "Government Spending in a Simple Model of Endogenous Growth," unpublished ms. Harvard University, May. Becker, G. and R. Barro (1988), "A Reformulation of the Economic Theory of Fertility," Ouarterly Journal of Economics, 103: 1-25. Bhagwati, J. (1978), Anatomy and Conseguences of Exchange Control Regimes, Cambridge, MA: Ballinger Publishing Co. for NBER. (1986), "Export Promoting Strategy: Issues and Evidence," World Bank Discussion Paper UPERS7, October. and T. Srinivasan.(1983), Lectures on International Trade, Cambridge, MA: MIT Press. Bianchi, A., et al. (1987), "The Adjustment Process in Latin America, 1981- 1986," in V. Corbo, M. Goldstein and M. Khan (eds.) Growth-Oriented Adiustment Programs, Washington, DC: International Monetary Fund and World Bank. 57 Boldrin, M. and J. Scheinkman (1988), "Learning-By-Doing, International Trade and Growth: A Note," Working Paper #462, UCLA Dept. of Economics. Bradford, C. and W. Branson (1987), Trade and Structural Change in Pacific Asia, Chicago: University of Chicago Press for NBER. Collins, S. and W. Park (1988), "External Debt and Macroeconomic Performance in South Korea," NBER Working Paper #2596. Cooper, R. (1987), "Round Table Discussion," in V. Corbo, M. Goldstein and M. Khan (eds.) Growth-Oriented Adiustment Programs, Washington, DC: International Monetary Fund and the World Bank. Corden, W. (1966), "The Structure of a Tariff System and the Effective Protection Rate," Journal of Political Economy, 74: 221-37. (1971), "The Effects of Trade on the Rate of Growth," in J. Bhagwati (ed.), Trade. Balance of Payments and Growth: Essays in Honor of Charles P. Kindleberger, Amsterdam: North-Holland. Dalham, C.B., Ross-Larson, R and L. Westphal (1984) "Managing Technological Development: Lerssons From the Newly Industrialized Countries," World Bank Staff working Paper #717. Dixit, A. (1976), The Theory of Equilibrium Growth, New Delhi: Oxford University Press. (1986), "Tax Policy in Open Economies," in A. Auerbach and M. Feldstein (eds.) Handbook of Public Economics, Amsterdam: North-Holland. and V. Norman (1980), Theory of International Trade, Cambridge University Press. Edwards, S. (1988), "Structural Adjustment Policies in Highly Indebted Countries," in J. Sachs (ed.) The DeveloRinE Countries' Debt Crisis, Chicago: University of Chicago Press. Ethier, W. (1982), "National and International Returns to Scale in the Modern Theory of International Trade," American Economic Review, 72: 389-405. Feder, G. (1983), 'On Exports and Economic Growth," Journal of DeveloRment Eonomics, 12: 59-74. Findlay, R. (1984), "Growth and Development in Trade Models," in R. Jones and P. Kenen (eds.) Handbook of International Economics Vol. I, Amsterdam: North-Holland. G. Grossman and E. Helpman (1988), "Comparative Advantage and Long Run Growth," unpublished ms., August. Harberger, A. (ed.) (1984), "Economic Policy and Economic Growth," in World Economic Growth, San Francisco: Institute for Contemporary Studies. 58 Havrylyshyn, 0. (1985), "The Direction of Developing Country Trade: Empirical Evidence," Journal of Development Economics, 19: 255-81. Helpman, E. (1988), "Growth, Technological Progress and Trade," NBER Working Paper #2592. Hutchinson, H. and N. Singh (1987), "Exports and Economic Growth in Developing Countries: Identifying Externality Effects," UCSB Working Paper, March. Johnson, H. (1965), "The Theory of Tariff Structure with Special Reference to World Trade and Development," in Trade and Development, Geneva: Institute of International Studies. Jones, L. and R. Manuelli (1988), "A Model of Optimal Equilibrium Growth," enpublished ms., Northwestern University, April. Jung, W. and P. Marshall (1985), "Exports, Growth and Causality in Developing Countries," Journal of DeveloRment Economics, 18: 1-12. Kavoussi, R. (1984), "Export Expansion and Economic Growth: Further Empirical Evidence," Journal of Development Economics, 14: 241-50. (1985), "International Trade and Economic Development: Recent Experience of Developing Countries," Journal of Develoning Areas, 19: 379-92. Kohn, M. and N. Marion (1987), "The Implications of Knowledge-Based Growth for the Optimality of Open Capital Markets," unpublished ms. Kravis, I. (1970), "Trade as a Handmaiden of Growth: Similarities Between the Nineteenth and Twentieth Centuries," Economic Journal, 80: 850-72. Krueger, A. (1978), Foreign Trade Rezimes and Economic Develonment: LiberalizAtion Attempts and Consequences, Cambridge, MA: Ballinger Publishing Co. for NBER. (1981), "Interaction Between Inflation and Trade Regime Objactives in Stabilization Programs," in W. Cline and S. Weintraub (eds.) Economic Stabilization in Deve,loning Countries, Washington, DC: Brookings Institution. (1983), Trade and Emploment in Developing Countries: Synthesis and Conclusions, Chicago: University of Chicago Press. (1984), "Problems of Liberalization," in A. Harberger (ed.) Wdrl Economic Growth, San Francisco: Institute for Contemporary Studies. (1986), "Problems of Liberalization," in A. Choski and D. Papageorgious (eds.) Economic Liberalization in Develoning Countries, Oxford: Blackwell. 59 Krugman, P. (1988), "Endogenous Innovations, International Trade and Growth," paper presented at the SUNY-Buffalo Conference on "The Problem of Development", May. Lal, D. and S. Rajapatirana (1987), "Foreign Trade Regimes and Economic Growth in Developing Countries," World Bank Research Observer, 2: 189-218. Leamer, E. (1984), Sources of International Comparative Advantage, Cambridge, MA: MIT Press. (1985), "Vector Autoregressions for Causal Inference?" Carnegie Rochester Conference Series on Public Policy, 22: 255-304. (1987), "Measures of Openness," Working Paper #447, UCLA Dept. of Economics. Lewis, A. (1955), The Theorv of Economic Growth, London: Allen & Unwin. Little, I., T. Scitovsky and M. Scott (1970), Industry and Trade in Some Developing Countries, London: Oxford University Press. Love, J. (1984), "External Market Conditions, Competitiveness and Diversification in LDCs Exports," Journal of Develooment Economics, 16. Lucas, R. (1988), "On the Mechanics of Economic Development," Journal of Monetary Economics, 22: 3-42. Hichaely, H. (1977), "Exports and Growth: An Empirical Investigation," Journal of Development Economics, 1: 49-53. (1988), "Trade Liberalization Policies: Lessons of Experience," paper presented at the Conference "For a New Policy Towards Foreign Trade," Brazil, April. Murphy, K., A. Schleifer and R. Vishny (1988), "Industrialization and the 'Big Push'," unpublished ms., Graduate School of Business, University of Chicago. Nelson, R. and E. Phelps (1965), "Investment in Humans, Technological Diffusion, and Economic Growth," American Economic Review, 55: 69-75. Ram, R. (1985), "Exports and Economic Growth: Some Additional Evidence," Economic DevelO2ment and Cultural chanfs, 33: 415-25. Rebelo, S. (1987), "Long Run Policy Analysis and Long Run Growth," unpublished ms., University of Rochester, March. Rodrik, D. (1988), "Closing the Technology Gap: Does Trade Liberalization Really Help?" NBER Working Paper #2654. Romer, P. (1986a), "Increasing Returns and Long Run Growth," Journal of Political Economy, 94: 1002-37. 60 -________ (1986b), "Increasing Returns, Specialization and External Economies: Growth as Described by Allyn Young," Rochester Center for Economic Research, Working Paper #64. (1988), "Endogenous Technological Change," paper presented at the SUNY-Buffalo Conference on "The Problem of Development," May. Rosenstein-Rodan, P. (1961), "Notes on the Theory of the 'Big Push'," in H. Ellis and H. Wallick (eds.) Economic Develogment for Latin America, New York: St. Martin's Press. Sachs, J. (1987), "Trade and Exchange Rate Policies ln Growth-Oriented Adjustment Programs," in V. Corbo, M. Goldstein and M. Khan (eds.) Growth-Oriented Adiustment Programs, Washington, DC: International Monetary Fund and the World Bank. _ (1988), "Conditionality, Debt Relief and the Developing Countries' Debt Crisis," NBER Working Paper #2644. Schleifer, A. (1986), "Implementation Cycles," Journal of Political EconogMy, 94: 1163-90. Singer, H. and P. Gray (1988), "Trade Policy and Growth of Developing Countries: Some New Data," World Development, 16. Smith, M.A. (1984), "Capital Theory and Trade Theory," in R. Jones and P. Kenen (eds.), Handbook of International Economics, Vol. I, Amsterdam: North-Holland. Tyler, W. (1981), "Growth and Export Expansion in Developing Countries," Journal of Develooment Economics, 9: 121-30. World Bank (1987), World Development Regort, London: Oxford University Press. Young, A. (1928), "Increasing Returns and Economic Progress," Econgmic Journal, 38: 527-42. PPR Workina Paper Series Contact Title Author Date for paper WPS164 Intersectoral Financial Flows in Developing Countries Patrick Honohan March 1989 W. Pitayatonakarn Izak Atiyas 60353 WPS165 Developing Countries' Exports of Manufactures: Past and Future Implications of Shifting Patterns of Comparative Advantage Alexander J. Yeats WPS166 Achieving and Sustaining Universal Primary Education: International Experience Relevant to India Nat J. Colletta March 1989 M. Philiph Margaret Sutton 75366 WPS167 Wage Determination in Rural Bangladesh: The Welfare Implications Martin Ravallion WPS168 Technological Change from Inside: A Review of Breakthroughsl Ashoka Mody March 1989 W. Young 33618 WPS169 Financial Sector Reforms in Adjustment Programs Alan Gelb Patrick Honohan WPS170 General Training Under Asymmetric Information Eliakim Katz April 1989 C. Cristobal Adrian Zidermar. 33640 WPS171 Cost-Effectiveness of National Training Systems in Developing Countries Christopher Dougherty March 1989 C. Cristobal 33640 WPS172 The Effects of Peru's Push to Improve Education Elizabeth M. King March 1989 C. Cristobal Rosemary T. Bellew 33640 WPS173 Staffing and Training Aspects of Hospital Management: Some Issues for Research Julo Frenk Enrique Ruelas WPS174 Trade Restrictions with Imported Intermediate Inputs: When Does the Trade Balance Improve? Ramon E. Lopez March 1989 M. Ameal Dani Rodrik 61466 WPS175 An Integrated Model of Perennial and Annual Crop Production for Sub-Saharan Countries Robert D. Weaver May 1989 A. Daruwala 33716 WPS176 Credit Rationing, Tenancy, Productivity, and the Dynamics of Inequality Avishay Braverman May 1989 C. Spooner Joseph E. Stiglitz 37570 WPS177 Cash-Flow or Income? The Choice of Base for Company Taxation Jack M. Mintz April 1989 A. Bhalla Jesus Seade 60359 PPR Workin. Paper Series Contact Title Author Date for paper WPS178 Tax Holidays and Investment Jack M. Mintz April 1989 A. Bhalla 60359 WPS179 Public Sector Pricing in a Fiscal Context Christopher Heady April 1989 A. Bhalla 60359 WPS180 Structural Changes in Metals Consumption: Evidence from U.S. Data Boum-Jong Choe April 1989 S. Llpscomb 33718 WPS181 Public Finance, Trade and Development: What Have We Learned? Johannes F. Unn April 1989 M. Colinet Deborah L Wetzel 33490 WPS182 The Experience of Latin America With Export Subsidies Julio Nogues April 1989 S. Torrivos 33709 WPS183 Private Investment in Mexico: An Empirical Analysis Alberto R. Musalem April 1989 L Spear 30081 WPS184 Women and Forestry Operational Issues Augusta Molnar May 1989 M. Villar Gotz Schreiber 33752 WPS185 Uniform Trade Taxes, Devaluation, and the Real Exchange Rate: A Theoretical Analysis Stephen A. O'Connell April 1989 A. Oropesa 61758 WPS186 The Uruguay Negotiations on Subsidies and Countervailing Measures: Past and Future Constraints Patrick A. Messerlin April 1989 S. Torrijos 33709 WPS187 The Flexibility of Wages When the Output Price Changes: An Empirical Study of 13 Industrial Countries Menahem Prywes April 1989 J. Israel 31285 WPS188 International Comparisons of Wage and Non-Wage Costs of Labor Luis A. Riveros April 1 R. Luz 61762 WPS189 The Treatment of Companies under Cash Flow Taxes: Some Administrative, Transitional, and International Issues Emil M. Sunily May 1989 A. Bhalla 60359 WPS190 Macro Performance Under Adjustment Lending Riccardo Faini April 1989 M. Ameal Jaime de Melo 61466 Abdel Senhadji-Semlali Julie Stanton WPS191 Openness, Outward Orientation, Trade Liberalization and Economic Perform- ance in Developing Countries Sebastian Edwards June 1989 M. Ameal 61466