Trade policy in developing countries.

By Edward F. Buffie.

Cambridge, UK: Cambridge University Press, 2001. Pp. vi, 400. $27.95 (paperback).

This is a very interesting and useful text for the "post-Seattle" discussion of trade policies in developing countries. If the protests associated with the World Trade Organization (WTO)

meetings there have taught us anything, it is that a very vocal opposition has emerged to the concept of trade liberalization among developing countries. The dichotomy in debate is stark: The proponents of trade liberalization speak of the static gains from trade and the power of markets in enticing disciplined economic activity, while the opponents decry the profits accruing to the rich at the expense of the poor that this liberalization will facilitate. The beliefs on both sides are passionate and strongly held, but the grounding in economic theory is limited. Professor Buffie has taken on the task of analyzing the theoretical costs and benefits of trade liberalization in this monograph. As he puts it, "It can be said with some assurance that extreme protectionist policies are economically harmful. But the case for free tra de is not airtight, and there is no general theoretical presumption that the market failures common in LDCs (less developed countries) favor an export-oriented rather than an import-substituting trade strategy" (p. 5). He succeeds handsomely in illuminating these questions with a logic embedded in theory. Let me begin with a disclaimer. Buffie and I seem to have a similar way of framing research questions, and thus I am willing to accept a number of maintained hypotheses in his work to which others might object. These include the following:

(i) modeling of economic activity as the outcome of optimizing behavior,

(ii) a presumption that developing countries are characterized to a greater extent than developed countries by certain forms of market failure,

(iii) a presumption that mathematical modeling can be used to represent the economic interactions of individuals in developing countries, and

(iv) the absence of explicit exploitation of developing countries by developed countries in the international marketplace.

If your own beliefs run counter to these, then Buffie's logic--and my review--will leave you dissatisfied.

Buffie sets an ambitious agenda in this monograph. He analyzes a number of specific policy questions and does so within a consistent theoretical framework that is both intertemporal in nature and has sufficient complexity to capture the features he associates with developing countries. As he says, "I am trying to sell not only a particular set of conclusions but also a general approach to the analysis of policy issues" (p. 3).

His general approach can be characterized as follows:

(i) He builds a static mathematical model of the developing economy. Equilibrium in the economy at each moment is characterized by constant returns to scale in production, optimizing individual agents, and preexisting factor-market based distortions that hold the economy away from its first-best outcome. The equilibrium is summarized in the duality approach of revenue and expenditure functions. There are typically more than two goods and more than two factors of production. Buffie's focus is supply side in nature, and he provides a careful discussion of substitutability of factors in production.

(ii) He describes the dynamic processes that carry the economy from period to period. The analysis is based on optimal control solution of a representative agent. The agent allocates saving to investment in the various productive sectors; a convex investment-cost function, of the Abel (1979) and Hayashi (1982) variety, ensures that adjustment to the optimal capital stock is gradual. The future reversibility of government policy choices provides a second rationale for intertemporal reallocation.

(iii) He draws welfare inferences from the comparative dynamics associated with policy shocks to the model. Buffie's preferred description of a developing country is a 3 X 3 trade model (three goods, three factors) in the steady state and of even greater dimension along the dynamic path. Because of the complexity of the model, he relies heavily on numerical simulation to illustrate the implications for policy choice.

This is not a book of equations, nor is each chapter devoted to inverting 6 X 6 matrices. Buffie is quite ingenious in introducing simplifications when appropriate and in teasing analytical results from the model. In an early chapter, he examines the difference between import substitution and export promotion within a static trade model; in later chapters, he focuses on dynamics and on credibility effects through judicious simplification of the basic model. Each chapter is well crafted to be complete enough to answer the question at hand.

Buffie's "general approach" requires the use of duality theory and the solution of systems of linear differential equations. He is determined that the reader understand these--so determined that he devotes three chapters in the monographs to "tools and tricks of the trade," proving therein a careful explication of the mathematical techniques involved. Those who have not yet mastered the applications of duality theory or of linear differential-equations solutions in economics will find this an excellent and detailed explanation of the mathematics involved. This feature also makes the monograph quite valuable as a text or reference work for graduate courses in international trade theory.

Developing countries in the real world differ substantially in their economic structure, of course, and this makes difficult the definition of a single economic model. Buffie's modeling choices suggest that he has a Central American, African, or Caribbean economy in mind. The developing country he models is small in world markets. It produces an export good, an import-competing good, and a nontraded good. It imports capital goods and intermediate inputs. Land, labor, and capital are used in the production process, but land is a factor specific to the export good.

In addition to this economic structure, there are two market distortions that characterize this developing country. First, the equilibrium real return to capital is greater than the social discount rate on capital, leading to too little investment. Buffie posits a "public good" aspect to investment in developing countries--while investors internalize some of the gains from capital formation, other gains to the economy as a whole are not captured by the investor. Second, there is an artificially high wage in the import-competing sector (e.g., because of union activity or minimum-wage requirements) that leads to a gap between the wage in that sector and the wage in the other (export and nontraded) sectors.

Buffie considers the impact of changes in trade policy on the equilibrium in that economy. Trade policy instruments include tariffs on consumption imports, tariffs on intermediate inputs, quotas on either of these imports, and a subsidy to exports. He derives optimal settings for these instruments and compares them to settings suggested in the current trade policy debate. Questions he answers in this context include (but are not limited to) the following:

* Is free trade always first best?

* What is the impact of trade liberalization on employment and wages?

* Does the credibility of the government's commitment to trade liberalization affect the welfare gains and losses of the policy?

* Should foreign investment into the developing country be regulated?

I will not try to summarize his answers. Not only is that bad form in general for a reviewer, but in this case it would not do justice to the care with which Buffie works through the possible permutations of the model. If your research or policy interests lead you to ponder these questions, you will benefit from his careful analysis of them.

Let me provide one teaser, however. Buffie will no doubt receive the most flak from his conclusion that complete trade liberalization may not be a first-best policy for his prototypical small open economy. This sounds counterintuitive, for trade theorists have traditionally found that the first-best commercial policy for a small open economy is free trade. Rodrik (1987) summarizes nicely the theoretical underpinning for this discrepancy. It is true that the optimal policy for a small open economy is always free trade if there are no other market distortions. However, with preexisting distortions, the theory of the second best comes into play. Protective commercial policy can raise welfare in this case because although it introduces welfare-reducing distortions, at the same time it reduces the negative effects of the preexisting distortions. In Buffie's model, the import-substitution policy has a negative direct effect on welfare by introducing deadweight losses. However, it has a positive effect by reducing t he welfare losses associated with underinvestment or the wage gap. On net, the positive effects dominate the negative effects for small levels of protection.

I have taken the liberty of creating in Figure 1 an illustration of the steady state in Buffie's model with preexisting distortions. (1) The two distortions in the model are indicated on the two axes: The vertical axis shows the wage gap for labor, while the horizontal axis shows the aggregated level of capital stock in the economy. (2) Buffie considers only steady states in the upper-left quadrant of the figure; he characterizes these as having underinvestment and under-employment. The steady-state welfare of the economy is measured by the iso-welfare ovals denoted. [U.sub.0] through [U.sub.3], with welfare rising with the subscript. Beginning from free trade, the first-best steady state is (1, [K.sup.*]) at point E. The preexisting distortions under free trade constrain the economy to a point along the "Free Trade" locus. (3)

Suppose that the initial distorted equilibrium is ([[omega].sub.FT], [K.sub.FT]) at point A. Introduction of an infinitesimal tariff on the imported consumer good has the effect illustrated in Figure 1 in the movement in the steady-state equilibrium from point A to point C. The deadweight loss from the tariff is negligible in size, but the impact on existing distortions is not. First, the policy raises the steady-state wage in the informal sector through increasing employment in the import-competing sector; this is the movement from point A to point B and is welfare increasing. Second, the policy lowers the relative price of capital and increases aggregate capital stock to [K.sub.TP]; this is the movement from point B to point C. The end result is a welfare increase from import protection. (4)

Buffie's analysis provides a compelling example of the "second-best" nature of import protection in this model. He is quite careful in developing the calibrated model, performs extensive sensitivity analyses on model parameters, and replicates the result time after time.

Buffie set himself the task of analyzing the impact of trade policy choices on developing countries. His monograph has a two-part comparative advantage in this analysis:

* He takes great care in constructing a realistic model on sound theoretical bases and in subjecting the parameters of that model to extensive sensitivity analysis.

* He provides an exhaustive introduction to the mathematical concepts used in deriving and solving the model.

In my opinion, its comparative disadvantage also takes two parts:

* There is relatively little attention to the results of empirical research. Empirical studies are mentioned, but the book typically does not provide a description of the methodology used or the generality of those results. Those teaching graduate courses in trade and development will probably wish to supplement this book with readings on empirical research.

* The conclusions drawn are model dependent. While there is extensive sensitivity analysis to the parameters used in the model, there is no consideration of the implications of changing the model structure. Three changes come to mind. First, what if the export sector were not treated as agrarian and labor intensive? Second, what if the relevant preexisting distortions in the developing country differ from the wage gap and investment distortion modeled here? Many critiques of activist trade policy are based on distortions such as corruption or inefficient use of productive inputs not modeled here. Third, what if there were another source of growth over time--for example, total factor productivity growth or human capital growth--endogenous to the choice of trade policy?

In sum, Buffie has written a provocative and well-reasoned book. You may not accept Buffie's conclusions. You may disagree with his assumptions. However, he has raised the debate on trade policy to the appropriate level: that of the specifics of economic processes in these developing countries. I have added readings from this book to our graduate syllabus in international trade and am sure that our graduate students will benefit from exposure to his careful work.

(1.) The equilibrium depicted is the steady state of the model in chapter 5. Buffie does not have this diagram in his text, and I apologize for any misrepresentation that it engenders.

(2.) [W.sub.M] is the wage offered in the import-competing (or manufacturing) sector, while w is the competitive wage offered in the remaining sectors. The diagram illustrates steady-state solutions after any differences in returns on capital across sectors are eliminated through capital reallocation.

(3.) The magnitude of [W.sub.M] determines the slope of the locus, while the size of the wedge between social discount rate and private interest rate determines the level of aggregate capital stock and thus the point on the locus.

(4.) Buffie notes two important qualifications to this result. First, the result could be modified for larger changes in tariffs because then these welfare-increasing effects would be weighed against the negative welfare effect of increasing the tariff distortion. Second, these are welfare changes from steady state to steady state and thus ignore the potential welfare losses during the transition. Those transition losses are associated with initial effects that increase the wage gap (above A in Figure 1) before eventually falling to C. The potential transition loss from a fall in aggregate capital stock before rising does not occur in this model.

References

Abel, Andrews, 1979. Investment and the value of capital. New York: Garland Press.

Hayashi, Fumio. 1982. Tobin's marginal q and average q: A neoclassical interpretation. Econometrica 50:213-24.

Rodrik, Dani. 1987 Trade and capital-account liberalization in a Keynesian economy. Journal of International Economics 23:113-29.

[Figure 1 omitted]