Introduction

A Trade Myth 

As early as 1999, Roy J. Ruffin of the University of Houston and Federal Reserve Bank of Dallas pointed out that, contrary to popular belief, the top U.S. imports from Mexico were not clothing, fruits, and vegetables. These represented only 10 percent of the U.S. imports. Electrical machinery and equipment (and related parts) ranked first, representing 27 percent of U.S. imports from Mexico. Vehicles ranked second, and nuclear reactors, boilers, and related items were third. Interestingly, the United States’ top three exports to Mexico were these same three categories. Not only were Mexico’s exports to the United States quite similar to its imports, but Mexico’s exports were more concentrated in those big items.1 Current trade data show this same phenomenon of the United States importing and exporting in the same product categories, thus challenging the predicted trade patterns of traditional models of international trade. The chapter reviews several theories on the causes and consequences of trade. These newer approaches depart from trade theory as presented earlier by relaxing several assumptions employed in the basic trade model. Some of the implications of this relaxation will be presented in theories that incorporate differences in technology across countries, an active role for demand conditions, economies of scale, imperfect competition, and a time dimension to comparative advantage. Finally, intra-industry trade (IIT)—a common element in several theories—will be discussed in detail, because it is a prominent feature in international trade of manufactured goods. It is important to recognise that the causes of trade are more complex than those portrayed in the basic Heckscher-Ohlin model. 

Early Post-Heckscher-Ohlin Theories of Trade

The Imitation Lag Hypothesis

The imitation lag hypothesis in international trade theory was formally introduced in 1961 by Michael V. Posner. Discussion of this theory paves the way for a better-known theory—the product cycle theory. The imitation lag theory relaxes the assumption in the Heckscher-Ohlin analysis that the same technology is available everywhere. It assumes that the same technology is not always available in all countries and that there is a delay in the transmission or diffusion of technology from one country to another. Consider countries I and II. Suppose that a new product appears in country I due to the successful efforts of research and development teams. According to the imitation lag theory, this new product will not be produced immediately by firms in country II. Incorporating a time dimension, the imitation lag is defined as the length of time (e.g., 15 months) that elapses between the product’s introduction in country I and the appearance of the version produced by firms in country II. The imitation lag includes a learning period during which the firms in country II must acquire technology and know-how in order to produce the product. In addition, it takes time to purchase inputs, install equipment, process the inputs, bring the finished product to market, and so on. In this approach, a second adjustment lag is the demand lag, which is the length of time between the product’s appearance in country I and its acceptance by consumers in country II as a good substitute for the products they are currently consuming. This lag may arise from loyalty to the existing consumption bundle, inertia, and delays in information flows. This demand lag also can be expressed in a number of months, say, four months. A key feature in the Posner theory is the comparison of the length of the imitation lag with the length of the demand lag. For example, if the imitation lag is 15 months, the net lag page 175 is 11 months, that is, 15 months less 4 months (demand lag). During this 11-month period, country I will export the product to country II. Before this period, country II had no real demand for the product; after this period, firms in country II are also producing and supplying the product so the demand for country I’s product diminishes. Thus, the central point of importance in the imitation lag hypothesis is that trade focuses on new manufactured products.2 How can a country become a continually successful exporter? By continually innovating! This theory has considerable relevance for present-day concerns about the global competitiveness of U.S. firms. Further, it seems to be more capable of handling “dynamic” comparative advantage than are the Heckscher-Ohlin and Ricardo models.

The Product Cycle Theory

The product cycle theory (PCT) of trade builds on the imitation lag hypothesis in its treatment of the role of time in the diffusion of technology. The PCT also relaxes several other assumptions of traditional trade theory and is more complete in its treatment of trade patterns. This theory was developed in 1966 by Raymond Vernon. The PCT is concerned with the life cycle of a typical “new product” and its impact on international trade. Vernon developed the theory in response to the failure of the United States—the main country to do so—to conform empirically to the Heckscher-Ohlin model. Vernon emphasizes manufactured goods, and the theory begins with the development of a new product in the United States. The new product will have two principal characteristics: (1) it will cater to high-income demands because the United States is a high-income country; and (2) it promises, in its production process, to be labor-saving and capital-using in nature. (It is also possible that the product itself—e.g., a consumer durable such as a microwave oven—will be labor saving for the consumer.) The reason for including the potential labor-saving nature of the production process is that the United States is widely regarded as a labor-scarce country. Thus, technological change will emphasize production processes with the potential to conserve this scarce factor of production. The PCT divides the life cycle of this new product into three stages. In the first stage, the new-product stage, the product is produced and consumed only in the United States. Firms produce in the United States because that is where demand is located, and these firms wish to stay close to the market to detect consumer response to the product. The characteristics of the product and the production process are in a state of change during this stage as firms seek to familiarize themselves with the product and the market. No international trade takes place. The second stage of the life cycle is called the maturing-product stage. In this stage, some general standards for the product and its characteristics begin to emerge, and mass production techniques start to be adopted. With more standardization in the production process, economies of scale start to be realized. This feature contrasts with Heckscher-Ohlin and Ricardo, whose theories assumed constant returns to scale. In addition, foreign demand for the product grows, but it is associated particularly with other developed countries, because the product is catering to high-income demands. This rise in foreign demand (assisted by economies of scale) leads to a trade pattern whereby the United States exports the product to other high-income countries. Other developments also occur in the maturing-product stage. Once U.S. firms are selling to other high-income countries, they may begin to assess the possibilities of producing abroad in addition to producing in the United States. If the cost picture is favorable (meaning that production abroad costs less than production at home plus transportation costs), page 176 then U.S. firms will tend to invest in production facilities in the other developed countries. If this is done, export displacement of U.S.-produced output occurs. With a plant in France, for example, not only France but other European countries can be supplied from the French facility rather than from the U.S. plant. Thus, an initial export surge by the United States is followed by a fall in U.S. exports and a likely fall in U.S. production of the good. This relocation-of-production aspect of the PCT is a useful step because it recognizes—in contrast to H-O and Ricardo—that capital and management are not immobile internationally. This feature also is consistent with the growing amount of direct investment abroad by firms in many countries, such as by U.S. and western European firms in China and Southeast Asia and by Chinese investment in Africa. Vernon also suggested that, in this maturing-product stage, the product might now begin to flow from western Europe to the United States because, with capital more mobile internationally than labor, the price of capital across countries was unlikely to diverge as much as the price of labor. With relative commodity prices thus heavily influenced by labor costs, and where labor costs were lower in Europe than in the United States, Europe might be able to undersell the United States in this product. Relative factor endowments and factor prices, which played such a large role in Heckscher-Ohlin, have not been completely ignored in the PCT. The final stage is the standardized-product stage. By this time in the product’s life cycle, the characteristics of the product itself and of the production process are well known; the product is familiar to consumers and the production process to producers. Vernon hypothesized that production may shift to the developing countries. Labor costs again play an important role, and the developed countries are busy introducing other products. Thus, the trade pattern is that the United States and other developed countries may import the product from the developing countries. Figure 1 summarizes the production, consumption, and trade pattern for the originating country, the United States. In summary, the PCT postulates a dynamic comparative advantage because the country source of exports shifts throughout the life cycle of the product. Early on, the innovating country exports the good but then it is displaced by other developed countries—which in turn are ultimately displaced by the developing countries. A casual glance at product history yields this kind of pattern in a general way. For example, electronic products such as television receivers were for many years a prominent export of the United States, but eventually Japan, China, and the Republic of Korea emerged as competitors, causing the U.S. share of the market to diminish dramatically. The textile and apparel industry is another example where developing countries (especially China, Taiwan, Malaysia, and Bangladesh) have become major suppliers on the world market, displacing in particular the United States and Japan. Automobile production and export location also shifted relatively from the United States and Europe to Japan and later still to countries such as South Korea and Malaysia. This dynamic comparative advantage, together with factor mobility and economies of scale, makes the product cycle theory an appealing alternative to the Heckscher-Ohlin model. There is no single all-encompassing test (such as the Leontief test of Heckscher-Ohlin) to verify empirically the product cycle theory. Instead, researchers have examined particular features of the PCT to see if they are consistent with real-world experience. For example, new product development is critical to the PCT, and it is often the result of research and development (R&D) expenditures. Therefore, economists hypothesize that, in the U.S. manufacturing sector, there should be a positive correlation between R&D expenditures and successful export performance by industry. A number of early tests indicated this result, including those by Donald Keesing (1967) and William Gruber, Dileep Mehta, and Vernon (1967). Kravis and Lipsey (1992) found that high R&D intensity page 177 was positively associated with large shares of exports by U.S. multinational companies (MNCs). Furthermore, in recent years exports from U.S. MNCs abroad have displaced sales from U.S. plant locations, which is consistent with the direct-investment and export-displacement features of the PCT. In addition, in 1969, Louis Wells examined the income elasticity of demand of the fastest-growing U.S. exports and found that trade in “high-income”-type products indeed grew more rapidly than other products—again, an occurrence consistent with the PCT. Early empirical work also found that the United States and other developed countries tended to export new products whereas developing countries tended to export older products, and that research-intensive U.S. industries had a high propensity to invest abroad.3

Raymond Vernon (1979) later suggested that the PCT might need to be modified. The main alteration concerns the location of the production of the good when the good is first introduced. Multinational firms today have subsidiaries and branches worldwide, and knowledge of production conditions outside the United States is more complete than it was at the time of Vernon’s original writing in 1966. Thus, the new product may be produced first not in the United States but outside the country. In addition, per capita income differences between the United States and other developed countries are not as great now as in 1966, so catering to high-income demands no longer implies catering to U.S. demands alone. Even with this modification, the salient features of scale economies, page 178 direct investment overseas, and dynamic comparative advantage still distinguish the product cycle theory from the Heckscher-Ohlin model. One hesitates, however, to distinguish the product cycle theory so clearly from the Heckscher-Ohlin model. Elias Dinopoulos, James Oehmke, and Paul Segerstrom (1993) constructed a theoretical model that has PCT-type trade emerging as a result of differing factor endowments across countries. The model utilizes three production sectors in each country: an innovating high-technology sector, an “outside-goods” sector that engages in no product innovation, and a sector that supplies R&D services to the high-technology sector. Like H-O, there are only two factors (capital and labor), identical production functions across countries, and constant returns to scale. Assuming that the R&D sector is the most capital-intensive sector, a capital-abundant country produces a great deal of R&D. This enables a firm in the high-technology sector in that country to obtain a temporary monopoly in a new product—with patent protection—and then to export the product. After the patent expires, production occurs abroad with some export from that location. While a complete explanation is beyond the scope of this book, Dinopoulos, Oehmke, and Segerstrom’s model generates PCT-type trade as well as IIT (a concept discussed later) and a role for MNCs. Thus, Heckscher-Ohlin and the product cycle theory may well be complementary, not competing, theories. In similar fashion, James Markusen, James Melvin, William Kaempfer, and Keith Maskus (1995, p. 209) introduced the idea of a life cycle for new technologies containing elements of both the Dinopoulos, Oehmke, and Segerstrom model and the product cycle model. Noting the growing importance of technology in the trade of industrialized countries, Markusen et al. suggest that, just as there is a product cycle for consumer goods, there increasingly appears to be a cycle for techniques of production and machinery, as techniques and machines developed in industrialized countries eventually find their way into labor-abundant developing countries. This technology cycle is driven by the capital-abundant, high-wage countries where there is both a cost incentive and a sufficient market demand to warrant new labor-saving technology and new-product development. The capital-abundant countries thus produce a flow of new products and innovations, with firms often protected by a temporary monopoly via patents to produce for the home market. Because the new labor-saving technologies are not consistent with the relative factor abundances in the labor-abundant developing countries, those countries initially have little economic incentive to acquire the innovations. Consequently, capital-abundant countries export the new products utilizing the new technology. Eventually, however, as incomes start to rise in developing countries and even newer technologies are produced in the developed countries, the machines embodying the original “new” technology are exported by capital-abundant countries and the final products start being produced in the labor-abundant countries. Later, as in the product cycle theory, the machines themselves may be produced in the developing countries and exported from them.

The Linder Theory

This theory explaining the composition and pattern of a country’s trade was proposed by the Swedish economist Staffan Burenstam Linder in 1961. The Linder theory is a dramatic departure from the Heckscher-Ohlin model because it is almost exclusively demand oriented. The H-O approach was primarily supply oriented because it focused on factor endowments and factor intensities. The Linder theory postulates that tastes of consumers are conditioned strongly by their income levels; the per capita income level of a country will yield a particular pattern of tastes. (Note that Linder is concerned only with manufactured goods; he regards Heckscher-Ohlin as fully capable of explaining trade page 179 in primary products.) These tastes of “representative consumers” in the country will in turn yield demands for products, and these demands will generate a production response by firms in that country. Hence, the kinds of goods produced in a country reflect the per capita income level of that country. This set of particular goods forms the base from which exports emerge. To illustrate the theory, suppose that country I has a per capita income level that yields demands for goods A, B, C, D, and E. These goods are arrayed in ascending order of product “quality” or sophistication, with goods A and B, for example, being low-quality clothing or sandals while goods C, D, and E are farther up the quality scale. Now suppose that country II has a slightly higher per capita income. Because of its higher income, it may demand and therefore produce goods C, D, E, F, and G. Goods F and G may be quality products (such as silks or fancy shoes) not purchased by country I’s lower-income consumers. Each country is therefore producing goods that cater to the demands and tastes of its own citizens. Given these patterns of production, what happens if the two countries trade with each other? Which goods will be traded between them? Trade will occur in goods that have overlapping demand, meaning that consumers in both countries are demanding the particular items. In our example, goods C, D, and E will be traded between countries I and II. The determination of the trading pattern by observing overlapping demands has an important implication for the types of countries that will trade with each other. Suppose that we introduce country III, which has an even higher per capita income than country II. Country III’s consumer demand may be for goods E, F, G, H, and J. Which goods will country III trade with the other two countries? It will trade goods E, F, and G with country II but will trade only good E with country I. For all three countries I, II, and III, Figure 2 portrays the income-trade relationships, recognizing that there is a representative range of individual incomes around each country’s per capita income level. Looking at the Linder model as a whole, the important implication is that international trade in manufactured goods will be more intense between countries with similar per capita income levels than between countries with dissimilar per capita income levels. The Linder conclusion is consistent with aspects of the product cycle theory and fits with the observation that, for much of the post-World War II period, rapid growth in trade in manufactured goods has taken place between developed countries. The Linder theory has been subjected to a number of empirical tests. A common type of test is formulated as follows: Suppose that we have figures on the absolute value of the per capita income differences between a given country I and its trading partners. Then we get information on the intensity of trade between country I and each of its trading partners. The Linder theory would hypothesize that the relationship between these two series is negative because the greater the difference between the per capita incomes of country I and a trading partner, the less intensely the two countries will trade with each other. Studies, such as that by Joel Sailors, Usman Qureshi, and Edward Cross (1973), have indeed found a negative correlation. However, a complicating factor is that countries with similar per capita incomes often tend to be near one another geographically, so that the intense trade may also reflect low transportation costs and cultural similarity. After correcting for geographic proximity and other shortcomings, studies by Hoftyzer (1975), Greytak and McHugh (1977), Qureshi et al. (1980), and Kennedy and McHugh (1980) found little or no evidence in support of the Linder theory. At this point, the primary tool employed in these studies was simple correlation analysis. More recently, gravity models in a multiple regression context have been used in the testing of the Linder theory. (The gravity model framework is discussed later in this chapter.) The gravity models focus on the interaction between the resistance (geographic distance) and attraction (similar demand patterns). The expectation is that controlling for geographic factors, countries with similar demand patterns will trade more intensively with each other. Two tests using gravity models found little or no evidence to support the Linder theory (Hoftyzer, 1984; Kennedy and McHugh, 1983). However, using a gravity model to control for distance between countries and other determinants of trade, Jerry and Marie Thursby (1987, p. 493) found that support for Linder’s hypothesis was overwhelming in their study of the manufactured goods trade of 13 European developed countries, Canada, Japan, the United States, and South Africa. Only Canada and South Africa failed to have a significantly negative regression coefficient for per capita income differences with a trading partner on the volume of trade with that given partner. Additional studies by Hanink (1988, 1990), Greytak and Tuchinda (1990), Bergstrand (1990), and McPherson et al. (2000) found evidence supporting Linder’s hypothesis using a gravity model. In addition, Bukhari et al. (2005), after controlling for size of trading partner and relative price levels, found that imports into each of the South Asian countries of Bangladesh, India, and Pakistan were, in a statistically significant manner, negatively related to the absolute value of the per capita income differences between the respective South Asian country and its trading partners. Rauh (2010) found similar results for Germany’s trade with its European partners. These results are consistent with the expectations from the Linder model. Several studies have also indicated that a problem with omitted countries may have created a bias rejecting the Linder hypothesis.

Finally, we need to make one very important point concerning the Linder theory. In our example of countries I, II, and III, the theory identified the goods that would be traded between any pair of countries. However, the Linder theory did not identify the direction in which any given good would flow. When we said that countries I and II would trade in goods C, D, and E, we did not say which good or goods would be exported by which country. This was not a slip in the model; Linder made it clear that a good might be sent in both directions—both exported and imported by the same country! This phenomenon was not possible in our previous models of trade, for how could a country have a comparative advantage and a comparative disadvantage in the same good? The answer to this question will be pursued later (see the final section of this chapter), but this type of trade could clearly occur, for example, because of product differentiation. This term refers to products that are seemingly the same good but which are perceived by the consumer to have real or imagined differences. Clearly two different makes of automobiles are not the same in the consumer’s mind. Nor does the consumer regard as equivalent two different brands of beer, tennis rackets, or word-processing programs. Linder’s theory can incorporate this notion of product differentiation, because country II might be exporting Hyundais to country III and country III might be exporting Ford Fusions to country II. Countries that export and import items in the same product classification are engaging in intra-industry trade (IIT). The topic of IIT, covered later in this chapter, is an aspect of Linder’s theory that has generated considerable theoretical and empirical work.

Economies of Scale

Some alternative trade theories are based on the existence of economies of scale. In several of these models, the economies of scale are external economies pertaining to the industry rather than the firm. In such industries, as output increases, firms experience cost reductions per unit of output because, for example, the industry growth is attracting a pool of qualified labor. In a two-country world where the countries have identical PPFs and demand conditions, there is normally no incentive to trade. If the two industries experience economies of scale, the model generates a potentially new reason for trade. In spite of the fact that both countries begin with identical autarky positions, a shock that results in each country moving to specialization in different goods and trading would lead both countries to experience gains from trade. See Appendix A at the end of this chapter for the complete development of this model. While the gains from trade are clearly a result of the cost reductions that come from specialization and the resulting economies of scale, there are a number of uncertainties in this model. First, there is no way to know which country will specialize in each good. Second, something unusual is needed to jolt production away from the autarky point, but there is no way to predict the cause of the shock. In spite of these uncertainties, the analysis opens up new possibilities for gains that do not exist in traditional models. Whether the introduction of increasing returns is more realistic than the constant-returns assumption is a debated question, but increasingly, economists do think that scale economies can be important.

More Recent Alternative Trade Theories

The Krugman Model This theory of trade represents a family of newer trade models that has emerged since Heckscher-Ohlin. While Paul Krugman has developed other models, we refer to this widely cited model (November 1979) as the Krugman model. This model rests on two features that are sharply distinct from those of traditional models: economies of scale and monopolistic competition. In the Krugman model, labor is assumed to be the only factor of production. The scale economies (which are internal to the firm) are incorporated in the equation for determining the amount of labor required to produce given levels of output by a firm, as shown here:

L stands for the amount of labor needed by the firm, a is a constant (technologically determined) number, Q represents the output level of the firm, and b specifies the relation at the margin between the output level and the amount of labor needed. The equation works as follows: If a = 10 and b = 2, this means that when the firm’s output level is 20 units, then the labor required to produce that level of output is L = 10 + (2)(20), or 50 units of labor. However, suppose that output doubles to 40 units. The labor required to produce 40 units is L = 10 + (2)(40), or 90 units. What does this equation imply? It means that a doubling of output requires less than a doubling of input; that is, economies of scale in production exist. All firms in the economy are assumed to have this type of labor requirement equation. It should be evident that this equation is not applicable to a Ricardian model, because constant costs of production would make the relevant labor-usage equation L = bQ; that is, the labor input has a constant relation to the amount of output. The second main characteristic of the Krugman model is the existence of the market structure of monopolistic competition. In monopolistic competition, there are many firms in the industry and easy entry and exit. In addition, there is zero profit for each firm in the long run. However, unlike the perfect competition of traditional trade theory, the output of firms in the industry is not a homogeneous product. The products differ from each other, and each firm’s product possesses a certain amount of consumer brand loyalty. Product differentiation leads to advertising and sales promotion as firms attempt to differentiate their products in the minds of consumers. (For a review of monopolistic competition and the effect of changes in the price elasticity of demand, see Appendix B at the end of this chapter.) The Krugman model is most easily portrayed through Krugman’s basic graph (see Figure 3). On the horizontal axis, we place consumption of a typical good by any representative consumer in the economy, that is, per capita consumption, c. The vertical axis page 183 indicates the ratio of the price of the good to the wage rate, P/W. The basic notions of the model are illustrated in this figure and through explanation of the PP and ZZ curves.

The upward-sloping PP curve reflects the relationship of the price of the good to marginal cost. As consumption increases, demand becomes less elastic. This is like the ordinary straight-line demand curve studied in the introductory economics course—at lower prices and larger quantities, demand is less elastic than at higher prices and smaller quantities. Thus, the expression [eD/(eD + 1)], which is developed in Appendix B, increases and, with constant marginal cost, profit maximization dictates a higher price. Thus, P/W rises as c increases, and the PP curve is upward sloping. The ZZ curve in Figure 3 reflects the phenomenon in monopolistic competition that economic profit for the firm is zero in long-run equilibrium. (Ignore the Z′Z′ curve for the moment.) To arrive at the downward slope, remember that zero profit means that price is equal to average cost at all points on the ZZ curve. From any given point on the curve, if per capita consumption (c) increases (a horizontal movement to the right), average cost is reduced because of the economies-of-scale phenomenon specified in this model. Hence, to maintain the zero profit and to move back to the ZZ curve, price must be reduced (a vertically downward movement); this yields a downward slope for the curve. Clearly, when the downward-sloping ZZ curve is put together with the upward-sloping PP curve in Figure 3, there is an equilibrium position. (We’re assuming that you’ve been page 184 able to figure out where it is!) At point E, the representative monopolistically competitive firm is in equilibrium because it is charging its profit-maximizing price (and hence is on the PP curve), and this is a long-run equilibrium position because economic profit is zero (because the firm is on the ZZ curve). In Figure 3, the firm thus settles at (P/W)1, and the per capita consumption level of the product is c1. To introduce international trade, suppose that we designate the home country of this representative firm as country I. Let’s now introduce another country, country II, which is identical to country I in tastes, technology, and characteristics of the factors of production. (Country II could also be identical in size, although that is not necessary.) Traditional trade theory would conclude that, with these same general supply and demand conditions (and hence relative prices), the two countries would have no incentive to trade with each other. However, Krugman (and Linder) would disagree. When the two countries are opened to trade, the important point to note is that the market size is being enlarged for each representative firm in each country, because there are now more potential buyers of any firm’s good. And, when market size is enlarged, economies of scale can come into play and production costs can be reduced for all goods. In Figure 3, if the firm being considered is in country I, the opening of the country to trade with country II means that consumers in both countries are now consuming this product (as well as all other products)—country II’s consumers now add country I products to their consumption bundle, just as country I’s consumers add country II products to their consumption bundle. If the firm’s total output is momentarily held constant, there is thus, with the larger consuming population but with the spreading out of consumption to other, newly available products, less per capita consumption of this firm’s product at each P/W than was previously the case. This is equivalent to a leftward shift of the ZZ curve, represented by the Z′Z′ curve in Figure 3. [For example, if country II’s population is identical in size to country I’s population, then the size of the consuming population has doubled—this would lead to a shift to the left of ZZ by 50 percent (per capita consumption would be one-half of its previous value); if country II’s population were 25 percent of that of country I, the ZZ curve would shift to the left by 20 percent, because per capita consumption would now be 80 percent (=1/1.25) of its previous value.] Given the shift of ZZ to Z′Z′, there is thus disequilibrium at the old equilibrium point E, and movement takes place to the new equilibrium position E′. As movement from E to E′ occurs, P/W falls from (P/W)1 to (P/W)2 and per capita consumption of this firm’s good falls from c1 to c2. Notice that, although per capita consumption at the new equilibrium has declined in comparison with per capita consumption at the old equilibrium, it has not declined proportionately to the extent by which the size of the consuming population increased. [If it had, per capita consumption would be at the level on the Z′Z′ curve that is associated with (P/W)1.] Because per capita consumption has not decreased proportionately with the increase in the size of the consuming population, this means that total consumption of the firm’s product has increased; with this increased output by the firm, the scale economies have come into play and have reduced unit costs (and hence the price of the firm’s output). As noted, the opening of trade has reduced P/W in this country (and has done so in the other country, too) because economies of scale have been realized. However, if P/W has decreased, obviously its reciprocal W/P has increased. The significance of an increase in W/P is that real income has risen because W/P is the real wage of workers. Thus, trade causes an improvement in real income and a corresponding increase in output of all goods. Also, there is a less tangible but nevertheless very real additional benefit from trade—consumers now have foreign products available to them as well as home-produced products. This increase in product variety and consumer choice should also be counted as a page 185 gain from trade. Further, the trade between the countries in this model is trade of similar but differentiated products. As in the Linder theory, an exchange of similar goods emerges, that is, IIT, a result that conforms with the nature of much of international trade in the present-day world. Consequently, this model can explain trade between similar countries. Krugman has noted elsewhere (1983) that factor endowments can determine the broad range of types of goods a country will export and import; within that broad range, however, product differentiation and scale economies play a very important role in generating trade and the gains from trade. Finally, another potential result of the Krugman model’s trade is also important. The increased well-being from trade is available to all consumers. Thus, even if a person is a “scarce factor of production” in a Heckscher-Ohlin context and would tend to lose from trade, the gains for that person in the Krugman model both from a higher real wage due to the scale economies and from the increased variety of goods due to product differentiation can more than offset the loss from being a scarce factor. Hence, the “gainer-loser” income distribution aspects of trade do not necessarily occur if trade consists of an exchange of differentiated manufactured goods produced under conditions of economies of scale.

The Reciprocal Dumping Model

The reciprocal dumping model was first developed by James Brander (1981) and was then extended by Brander and Paul Krugman (1983). (See also Krugman, 1995, pp. 1268–71.) As discussed in Chapter 8, a monopoly firm may charge a different (lower) price in the export market than in the home market. This price discrimination phenomenon in the context of international trade is called dumping, and it usually arises because demand is more elastic in the export market than in the domestic market. (As indicated earlier, with price discrimination, markets are separated and the same good is sold in the different markets at different prices, with the lower price being charged in the market where demand is more elastic.) In the Brander-Krugman model, there are two countries (a home country and a foreign country) and two firms (a home firm and a foreign firm) producing a homogeneous (standardized) good. An important feature is that there is a transportation cost of moving the good (in either direction) between the home country and the foreign country, so this is a barrier that can keep the markets separated. Suppose first that the transportation cost of moving the good between the countries is very high—each firm then may well produce only for its own home market, and each will have a monopoly position in that market. In such a situation, each firm will follow the usual criterion of producing at the output level where marginal revenue equals marginal cost, and of marking up price over marginal cost in accordance with the markup formula P = MC [eD/(eD + 1)] developed in Appendix B. Thus, each of the two firms is maximizing profit and is selling in only one market (the home firm in the home country and the foreign firm in the foreign country). Suppose, however, that the transportation cost is not so high, and the home firm notes that the price charged by the foreign firm in the foreign market exceeds the home firm’s marginal cost of producing a unit of output plus the transportation cost of moving that unit of home output to the foreign market. If this is so, then the home firm will want to sell in the foreign market as well as in the home market, since there will be some additional profit that can be made by doing so. Analogously, if the foreign firm notices that the price in the home market exceeds the foreign firm’s marginal cost plus the transportation cost of sending the good from the foreign market to the home market, then it will want to sell in the home country’s market, too. Thus, there clearly are possibilities for international trade to emerge in this model. Once the two firms start selling in each other’s country as well as in their own home countries, we enter a market structure of duopoly (two sellers in a market), and the price in each country will change from the previous monopoly situation because of the new rivalry. page 186 This duopoly structure gives rise to a situation where each firm must take account of the behavior of the other firm when choosing its own price and output. The recognized interdependence between the firms and the manner in which price and output decisions are made in this context are best addressed by game theory, a complicated subject about which we will have more to say in Chapter 15. In general, each firm determines an array of various profit-maximizing positions, one for each possible output level of the other firm. The two firms then interact with each other in accordance with these arrays; for the purposes of this present chapter, suffice it to say that an equilibrium price and output level will be determined in each market for each firm. There will be this one outcome which is satisfactory to both firms, and, in this equilibrium position, each firm will be maximizing its profit for the given output level of the other firm. An additional aspect of the situation can also be usefully mentioned. Each firm is maximizing profit in each market (MR = MC in each market), but note the nature of the marginal cost of selling in the foreign country’s (home country’s) market for the home (foreign) firm. That cost will include not only production cost but also the transportation cost. If demand structures are similar in the two countries, then, using the home firm as an example, the net price received in the foreign market (price in the foreign market less transportation cost) is likely to be lower than the price received in the home market (where there is no transportation cost). It is in this rather particular sense that “dumping” is occurring in the model—the price received in the foreign market is less than the price received in the home market. Note, however, what the main point of the model is from the standpoint of trade theory. International trade in a homogeneous product is occurring, with each country both exporting and importing the product. This result emerges importantly because of the imperfectly competitive market structure, and it could never emerge with perfect competition. We clearly have IIT here, and the model may help us to understand real-world trade flows consisting of the movement of similar products between countries. Finally, Brander and Krugman discuss welfare implications of the model. On one hand, welfare tends to increase for each country and for the world because previously monopolistic sellers in each country are now faced with a rival and this pro-competitive effect will put downward pressure on price. On the other hand, a negative welfare aspect exists in that there clearly is waste involved in sending identical products past each other on transportation routes! (It would be better from a transportation cost perspective by itself to have each country supplying exclusively its own market.) Hence, in general, we cannot say whether this two-way trade in a homogeneous product is on balance welfare enhancing or welfare reducing—the result will depend on the particulars of the actual situation being considered.

Vertical Specialization-Based Trade

Another avenue of research that has been explored concerns vertical specialization-based trade (see Hummels, Rapaport, and Yi, 1998). In this research it is recognized that, increasingly, various stages in the production process of any particular good are taking place in different countries. Thus, some components of an automobile engine may be made in Germany and then sent to the United Kingdom, where the engine is assembled; the engine itself is then sent to the United States for placement in the automobile. Such offshoring4 by the U.S. auto firm, rather than making or having the components made domestically, has been growing noticeably in recent years. While there is no new “theory” of trade here, the approach recognizes that comparative advantage may pertain to trade in intermediate goods as well as in final goods (as in traditional theory). page 187 IN THE REAL WORLD: 

New Venture Internalisation

Stage theory (Johanson and Vahlne, 1977) incorporates the idea that internationalization is a gradual process that occurs as firms establish themselves in the domestic market and then grow to the point where they have the managerial expertise and economies of scale needed to enter the international arena. However, since the late 1980s, researchers in international entrepreneurship have been observing an increasing number of new ventures in countries around the world that seek to operate internationally at or near inception (McDougall, 1989). The acceleration of international activity by firms should be seen as a result of several changes on the global scene. Many of the changes have resulted in a reduction in the transaction costs associated with conducting business internationally. Globally, barriers to trade and investments have been reduced through the World Trade Organization. In addition, regional agreements such as the European Union (EU) and the North American Free Trade Agreement (NAFTA) have further removed barriers to new ventures becoming internationally active. Technological advances (including the Internet and e-mail) and falling transportation costs have resulted in enhanced information flows between countries which facilitate new venture internationalization (Autio, 2005; Reynolds, 1997). As the economy becomes increasingly global, information about foreign markets, potential clients, and possible foreign partners is more easily available. While there are many reasons to expect more international new ventures, there are also some special challenges. New ventures are typically more resource-constrained than larger, more established firms in terms of both financial and human capital (Coviello and McAuley, 1999). Also, new ventures are more likely to suffer the liabilities of newness (greater risk of failure and particular difficulties of being young) and liabilities of foreignness (a disadvantage relative to local firms when operating in foreign markets) as compared to larger, more established firms (Hessels, 2008). New venture internationalization is not easily explained by the traditional trade theories that were developed to examine the countries involved in trade rather than the firms. Even the models that focus on firm characteristics were developed to explain internationalization among large firms (McDougall et al., 1994). As a result, cross-border entrepreneurship has become a growing area of research. Sources: Erkko Autio, “Creative Tension: The Significance of Ben Oviatt’s and Patricia McDougall’s Article ‘Toward a Theory of International New Ventures,’ ” Journal of International Business Studies 36, no. 1 (January 2005), pp. 9–19; Nicole E. Coviello and Andrew McAuley, “Internationalisation and the Smaller Firm: A Review of Contemporary Empirical Research,” Management International Review 39, no. 3 (1999), pp. 223–56; S.J.A. Hessels, International Entrepreneurship: Value Creation across National Borders (Rotterdam: Erasmus Research Institute of Management, 2008); Jan J. Johanson and Jan-Eric Vahlne, “The Internationalization Process of the Firm—A Model of Knowledge Development and Increasing Foreign Market Commitments,” Journal of International Business Studies 8, no. 1 (Spring–Summer 1977), pp. 23–32; Patricia P. McDougall, “International versus Domestic Entrepreneurship: New Venture Strategic Behaviour and Industry Structure,” Journal of Business Venturing 4, no. 6 (November 1989), pp. 387–400; Patricia P. McDougall, Jeffrey G. Covin, Richard B. Robinson, Jr., and Lanny Herron, “The Effects of Industry Growth and Strategic Breadth on New Venture Performance and Strategy Content,” Strategic Management Journal 15, no. 5 (September 1994), pp. 537–54; Paul D. Reynolds, “Who Starts New Firms? Preliminary Explorations of Firms-in-Gestation,” Small Business Economics 9, no. 5 (October 1997), pp. 449–62. This vertical focus on the production process has provided a new aspect of internationalization and is often discussed under the concept of global logistics/supply chain management. While the product cycle theory examined international movement of consumption and production for an entire product, this newer approach focuses on the division of the production of the product into distinct stages. The distribution of different parts of the production process to many different countries results in a global network of producers that must be coordinated to produce and distribute the final product. If different aspects of the production vary in terms of capital and labor intensity, the production process is thus often spread across a variety of developed and developing countries. The key decisions for the firm in this model include identifying countries that possess comparative advantages in the stages of the production process, developing a logistic system to coordinate page 188 the multiple inputs in the global supply chain, and maximizing the efficiency of production and sales globally. This internationalization of the production process has led to the formulation of newer models that focus on the characteristics of firms as opposed to the characteristics of countries.

Firm-Focused Theories

There are several theories related to the characteristics of firms and their relationship to international trading activity. The first of these is stage theory, originally advocated by Johanson and Vahlne (1977). Stage theory suggests that internationalization is a gradual process that requires the acquisition, integration, and use of knowledge about foreign markets. As the firm grows, it accumulates resources, exploits economies of scale, and builds excess capacity. These resources enable management to direct greater efforts to exports when compared to smaller, younger firms (Bonaccorsi, 1992). Stage theory contains an element of “entrepreneurial learning” that takes place over time as owners and managers develop intellectual capital used in the development of internationalization strategies and resource allocation (Orser et al., 2008). If this theory holds, then firms that export are larger, more established, and run by older, more experienced managers. An alternative firm-based explanation for involvement in the international market is resource-exchange theory. Zacharakis (1997) argues that exporting is a result of “organizations entering into international transactional relationships because they cannot generate all necessary resources internally.” The resources that must be accumulated include firm-level tangible and financial assets as well as intellectual resources, including human attributes like growth orientation, management experience and knowledge, networks and command of foreign languages (Oviatt and McDougall, 1994; Dhanaraj and Beamish, 2003, as cited in Orser et al., 2008; Oviatt and McDougall, 1994). Resource-exchange theory differs from stage theory in that it does not assume that acquisition of organizational resources proceeds in a linear manner. The resource-exchange theory also helps to explain the differences in export propensity associated with owners’ managerial or entrepreneurial experience (Dhanaraj and Beamish, 2003). A third firm-level theoretical approach is referred to as network theory. This model differs from its predecessors in that it is often cited as an explanation of international new ventures (Orser et al., 2008). The firms may access strategic resources externally through interdependencies among network players. While the direct relationships stressed in the two previous models provide control over resources through ownership, network theory focuses on the ability of collaborations to speed the internationalization by firms through synergistic relationships among partners at various stages in the supply chain (Dana et al., 2004; Jones, 1999, as cited in Orser et al., 2008). Network theory challenges the stage theory contentions that firms need to be large and well established with experienced management to trade internationally. Rather, network theory provides a model in which new firms can find the experience and international expertise they need externally through networking. These new theories are useful in understanding the trend in the international entrepreneurship literature of focusing on firms, known as international new ventures, that operate internationally from their inception.5

The Gravity Model

An older model that has recently begun to be widely applied to issues in international trade is the gravity model of trade. This model has a relatively long history. (For early formulations, see Tinbergen, 1962, Pöyhönen, 1963, and Linnemann, 1966; for more recent page 189 discussion, see Disdier and Head, 2008, and Wang, Wei, and Liu, 2010.) It differs from most other theories (including traditional theory) in that it is trying to explain the volume of trade and does not focus on the composition of that trade. The model itself uses an equation framework to predict the volume of trade on a bilateral basis between any two countries. (The particular equation form need not concern us—the form has some similarity to the law of gravity in physics, which has resulted in the term gravity model being applied.) It is concerned with selecting economic variables that will produce a “good fit,” that is, that will explain at least in a statistical sense a substantial portion of the size of trade that occurs. The variables that are nearly always used in the equation as causes of, say, the flow of exports from a country I to a country II are: A national income variable for country II (GNP or GDP), which is expected to have a positive relationship with the volume of exports from I to II because higher income in II would cause II’s consumers to buy more of all goods, including goods from country I. A national income variable for country I (GNP or GDP), reflecting that greater income in I means a greater capacity to produce and hence to supply exports from I to II. Some measure of distance between country I and country II (as a proxy for transportation costs), with the expectation being that greater distance (greater transportation costs) would reduce the volume of exports from country I to country II. Sometimes other variables are introduced, such as population size in the exporting and/or importing country (to get at large market size and thus perhaps to economies of scale) or a variable to reflect an economic integration arrangement (such as a free-trade area) between the two countries. Empirical tests using the gravity model have often been remarkably successful, meaning that the volume of trade between pairs of countries has been rather well “explained.” In addition, by selecting different pairs of countries, other interesting questions can be addressed. For example, Helpman (1999, p. 138) discussed work in the literature that indirectly sought to distinguish between factor endowments and product differentiation as underlying causes of trade. The gravity equation worked best for similar countries that had considerable IIT with each other, better than it did for countries with different factor endowments and a predominance of traditional trade rather than IIT. At the minimum, these findings suggest that product differentiation is indeed a phenomenon to be considered above and beyond factor endowments. Although the theoretical underpinnings of the gravity model have been debated, the theory has proved empirically useful for helping us understand influences on the volume of trade. These econometric analyses have also linked the volume of trade to important economic variables. Such work is important if we are to make headway in understanding the world economy, and the volume of trade is not considered by many trade theories.

The Melitz Model and Multiproduct Exporting

Marc J. Melitz (2003) set forth a model that has had widespread influence. In a framework similar in important ways to the Krugman model developed earlier in this chapter, the Melitz model specifies that each particular industry consists of monopolistically competitive firms with differentiated products and unrestricted entry and exit. Also, like Krugman, labor is the only factor of production, and each firm has a fixed overhead cost and a constant marginal cost. Krugman utilized a labor equation for the firm (see page 182 of this chapter) consisting of L = a + bQ, where L is the labor needed by the firm, a is a constant, Q is the amount of output, and b is the production relation between output and labor at page 190 the margin. Meltiz’s new feature is that his equation is written as L = a + Q/ϕ, where ϕ is an index of productivity for the firm, and ϕ varies across firms in the industry. With a higher ϕ, the labor needed by a firm would be lower and the marginal cost would therefore be lower. Alternatively, higher productivity (an increase in ϕ) for a first firm in comparison with a second firm can be visualized as the first firm producing a higher-quality product than the second firm but at the same cost. The level of productivity for any given firm in the Melitz model is simply treated as a draw from a distribution of productivity levels. With all of this in place and after introducing demand conditions, there is initially established an equilibrium in a closed economy, with an equilibrium number of firms and an equilibrium amount of output, by a process that by itself is not critical for our understanding of the basics of this trade model. However, it is to be noted that the equilibrium has a cutoff productivity level ϕ*—firms with lower than that level of productivity will have exited from the industry. From this initial situation of a closed-economy equilibrium, consider what happens when the possibility of trade is introduced. (As with the Krugman and Linder models, the sequence in the models is from firms producing only for the domestic market to firms producing for export as well.) Firms will consider exporting, but to engage in export certain costs need to be incurred. Not only will there be some per-unit or variable costs such as transport costs, but there will also be fixed export costs (costs that are independent of the size of exports). Firms must become knowledgeable about the foreign markets, must inform the potential foreign buyers about the products, must investigate foreign regulations, and must establish distribution channels, among other tasks. Only the most productive firms in an industry that can bear these costs (or reasonably obtain financing) will be in a position to pursue exporting. Then, once the fixed costs have been incurred, they can be spread over a larger volume as exports grow, and the exporting firm becomes even more profitable and productive. Overall, the average productivity level in the industry will have risen because of the exporting, and some lower-productivity firms will now be facing losses and will exit the industry. Further, the demand for labor by the exporting firms increases, with the consequence that the wage for the type of labor used in the industry rises. This wage increase will add to costs for all firms, and least productive firms may leave the industry for this reason as well. Of course, with the additional export costs, there will be an equilibrium cutoff productivity level ϕ*x that is higher than the cutoff productivity level for supplying the domestic market, so some firms will be successfully exporting while also selling at home while other firms in the industry will be producing for the domestic market but will be unable to engage in exporting. Overall, with trade, there has been an increase in aggregate productivity in this industry (and in every industry that is now engaged in exporting). Productivity in the industry will have increased because of the reallocation of production toward higher-productivity firms and away from lower-productivity firms (the least productive of whom have left the industry). The process has enhanced the country’s welfare through the increased productivity, a mechanism that was not included in a model such as Krugman’s (where firms in any given industry were identical). But, as with the Krugman model, the Melitz model realistically incorporates phenomena such as product differentiation and scale economies that did not exist in models that utilized perfect competition. In a later paper, Thierry Mayer, Melitz, and Gianmarco I.P. Ottaviano (2014) expanded on recent research that focuses on the real-world fact that many of the world’s exports originate from firms that are producing many different products or varieties of a product. Mayer, Melitz, and Ottaviano’s model spells out how, if there is an increase in competition in a particular export market, the existing exporting firms will lower the markup of price over cost on all goods exported to that market (thus providing welfare gains to consumers) page 191 because price elasticity of demand has increased. Importantly, though, the firms will shift their export product mix toward their “better performing products” or “core products”. This shift toward the better products will imply a rise in overall productivity of the firms because the better-performing products will be the products in which the firms are most productive or have the greatest comparative advantage. The product mix of any given exporting firm has been narrowed toward the high-productivity goods, which leads to an increase in the firm’s overall productivity. The authors empirically test their predictions using a large sample of French manufacturing firms, and they find support. The French firms do seem to emphasize their better-performing products in their large export markets where the firms have much competition. This basic heterogeneous-product firm model has been extended in different ways. As one example, Richard Baldwin and Rikard Forslid (2010) have studied the implications of a model of this general type for welfare in an importing country. If trade is liberalized, suppose that, following the Mayer-Melitz-Ottaviano hypothesis, exporting firms to the importing country then narrow their product mix to the better-performing varieties. In addition, if some domestic firms are forced out because of the liberalization of imports, then the overall effect of the liberalization can be a reduction in the number of types of goods available to consumers in the importing country. This is in contrast to other models, such as the Krugman model, where expansion of trade results in a greater number of varieties. The “anti-variety” effect can be injurious to welfare in and of itself, although Baldwin and Forslid conclude that, on balance, the importing country will experience an increase in welfare, because the general positive welfare effect of trade will outweigh this negative anti-variety effect. We are only scratching the surface here regarding the new multiproduct firm models, but it is clear that these models have various potential implications for supplementing (or displacing) features of the traditional trade models. And even with the development of these newer models, there are still further modifications or additions to consider. For example, J. Peter Neary (2010) has criticized these models because they utilize only the structure of monopolistic competition, a situation in which firms do not differ greatly from each other (if at all). Neary judges that what is needed is an examination of firms in a setting of oligopoly, a setting considerably different from that of monopolistic competition.

Concluding Comments on Post–Heckscher-Ohlin Trade Theories

From the discussion in this chapter of early and more recent analyses, it is clear that trade theory is moving in directions neglected by traditional trade theory. The newer approaches enhance our understanding of the causes and consequences of trade beyond the insights provided early on by the Heckscher-Ohlin model. We have looked principally at theories that allow for lags in diffusion of technology, demand considerations, supply chains, economies of scale, international capital mobility, dynamic comparative advantage, and imperfect competition. There is considerable theoretical analysis in this area, much of which we have not discussed. For example, there is a growing literature on government policy and how it can generate comparative advantage and alter the distribution of the gains from trade between countries. Examples from this literature will be given in Chapter 15. Further, yet another approach (by Paul Krugman) explores the role of location of production in the determination of comparative advantage and consequent trade patterns. Although the newer theories mainly focus on developed countries, what are their implications for the developing countries? The imitation lag hypothesis and the product cycle theory do not lead to particularly optimistic conclusions about the future export performance of developing countries because they suggest that developing countries may be confined to exporting older products rather than new high-technology goods. On the other hand, these theories suggest that a potential exists for moving away from exporting principally primary products toward exporting more manufactured goods, a process that is page 192 underway in many developing countries. However, theories such as those of Linder and Krugman imply that trade may increasingly take place between countries of similar income levels. This forecast may not bode so well for developing countries who wish to break into developed-country markets, although the analyses suggest that they may beneficially trade more among themselves in the future. Finally, economies-of-scale models indicate the difficulty of predicting future trade patterns but suggest potentially large gains from trade, as do the newer productivity models. 

Geography and Trade

In a series of lectures at the Catholic University of Leuven, Belgium, in 1990, later published under the title Geography and Trade (1991), Paul Krugman examined various economic issues that arise when firms make interdependent spatial decisions regarding the location of production. He dubbed this exercise “economic geography” because the concept of “location” seemed too narrow and restrictive. Assuming the presence of economies of scale, transportation costs, and imperfect competition, he examined possible reasons for the concentration of manufacturing production, the localization of production across a broad spectrum of goods within the United States, and the role played by nations in interregional and international trade. In these lectures, Krugman introduced a new perspective (although geographers might say “reintroduced an old perspective”) on the basis for trade in manufactured goods which rests on the observation that trade often takes place as a result of more-or-less “arbitrary specialization based on increasing returns, rather than an effort to take advantage of exogenous differences in resources or productivity” (p. 7). This strong accidental or serendipitous component of international specialization sets off cumulative processes that throughout history have tended to be pervasive. Thus whether one looks at Catherine Evans’s interest in tufting bedspreads in 1895 and the manner in which it generated a local handicraft industry that evolved into the center of the U.S. carpet industry in Dalton, Georgia, or the classic examples of Eastman Kodak in Rochester, New York, the giant Boeing Company in Seattle, or Silicon Valley in California, dynamic comparative advantage in manufactures often appears to have its roots in a quirk of fate which sets off important cumulative processes. Critical to these developments are, of course, economies of scale and transportation cost considerations. In addition, Krugman (p. 62) also points to Alfred Marshall’s belief that the pooling of labor and the readily available supply of specialized inputs also play a crucial role in promulgating specialized local production and regional and international comparative advantage. The existence of potential economies of scale is a crucial element in this approach in that firms have reason to concentrate their production activities as long as the cost advantages related to larger size are not offset by transportation costs associated with either inputs or final products. Thus, production locates both with regard to the size of the market and the availability of inputs, given the objective of concentrating production. Once established, production generates a dynamic of its own and tends to be self-sustaining. Nations are important from this perspective primarily because they adopt policies that influence economic decision making and the evolution of the critical cumulative processes. While policies are often enacted to inhibit the flow of goods and/or factors, government policies can also stimulate these critical processes, as was the case with the state government’s decision to provide financial support for the Research Triangle Park in North Carolina. 

Infra-Industry Trade

A characteristic of a country’s trade that has appeared in many new theories and is increasingly recognized as important in the real world is intra-industry trade. IIT occurs when a country is both exporting and importing items in the same product classification category. This trade differs from inter-industry trade, where a country’s exports and imports are in different product classification categories. Traditional trade theory dealt only with inter-industry trade, but IIT clearly constitutes an important segment of international trade. Table 1 indicates magnitudes and trends in the IIT in manufactured goods of selected countries. (IIT is more important in manufactured goods than in nonmanufactured goods.) It can also be noted that other data (not shown in Table 1) indicate that IIT is typically the highest for more sophisticated manufactured goods such as chemicals, machinery, transport equipment, and electronics, where scale economies and product differentiation can be important. Table 2 (see page 196) provides more recent additional data on IIT.

Reasons for Intra-Industry Trade in a Product Category

Unfortunately, comparative advantage based on factor endowments is of little or no help in predicting IIT. In fact, IIT will be relatively greater (compared with inter-industry trade) the more similar are the capital and labor endowments of the countries being examined. In view of this deficiency of the Heckscher-Ohlin model, we now look at several possible explanations for the occurrence of IIT. (For an extensive groundbreaking discussion of reasons for IIT, see Herbert Grubel and P. J. Lloyd, 1975.)

Product Differentiation

This explanation for IIT was outlined earlier. Briefly, many varieties of a product exist because producers attempt to distinguish their products in the minds of consumers to achieve brand loyalty or because consumers themselves want a broad range of characteristics in a product from which to choose. Thus, U.S. firms may produce large automobiles and non-U.S. producers may produce smaller automobiles. The consequence is that some foreign buyers preferring a large car may buy a U.S. product while some U.S. consumers may purchase a smaller, imported car. Because consumer tastes differ in innumerable ways, more so than the varieties of products manufactured by any given country, some IIT emerges because of product differentiation.

Transport Costs

In a physically large country such as the United States, transport costs for a product may play a role in causing IIT, especially if the product has large bulk relative to its value. Thus, if a given product is manufactured both in the eastern part of Canada and in California, a buyer in Maine may buy the Canadian product rather than the California product because the transport costs are lower. At the same time, a buyer in Mexico may purchase the California product. The United States is both exporting and importing the good. Another mechanism by which transport costs can lead to IIT is as specified in the reciprocal dumping model discussed earlier in this chapter.

Dynamic Economies of Scale

This reason is related to the product differentiation reason. If IIT has been established in two versions of a product, each producing firm (one in the home country, one in the foreign country) may experience “learning by doing” or what has been called dynamic economies of scale. This means that per-unit cost reductions occur because of experience in producing a particular good. Due to these cost reductions, sales of each version of the product may increase over time. Because one version was an export and the other an import for each country, IIT is enhanced over time because of this production experience.

Degree of Product Aggregation

This explanation rests on the observation that IIT can result merely because of the way trade data are recorded and analyzed. If the category is broad (such as beverages and tobacco), there will be greater IIT than would be the case if a narrower category is examined (such as beverages alone or, even more narrowly, wine of fresh grapes). Suppose a country is exporting beverages and importing tobacco. The broad category of “beverages and tobacco” [a category in the widely used Standard International Trade Classification (SITC) System of the United Nations] would show IIT, but the narrower categories of “beverages” and “tobacco” would not. Some economists think that finding IIT in the real world may be mainly a statistical artifact because of the degree of aggregation used, even though actual calculations use less broad categories than “beverages” and “tobacco.” Nevertheless, most trade analysts judge that IIT exists as an economic characteristic of trade and not primarily as a result of using aggregative classification categories.

Differing Income Distributions in Countries

This explanation for IIT was offered by Herbert Grubel (1970). Even if two countries have similar per capita incomes, differing distributions of total income in the two countries can lead to IIT. Consider the hypothetical income distributions plotted in Figure 4. page 195 Country I has a heavy concentration of households with lower incomes, while country II has a more “normal” or less skewed distribution. Producers in country I will be concerned primarily with satisfying the bulk of country I’s population, so they will produce a variety of the product that caters to consumers with incomes, for example, between y1 and y2. Producers in country II will cater to the bulk of country II’s households, say, those households between y3 and y4. Therefore, country II’s firms produce a variety of the good with characteristics that satisfy that group. What about a household in country I with a high income such as y6? And what about a household in country II with a low income such as y5? These consumers will purchase the good from the producers in the other country because their own home firms are not producing a variety of the good that satisfies these consumers. Hence, both countries have IIT in the product. This explanation can be applied in the context of the Linder model to help in predicting the pattern of IIT.

Differing Factor Endowments and Product Variety

In work that attempts to marry IIT with the Heckscher-Ohlin approach, Falvey (1981) and Falvey and Kierzkowski (1987) developed a model in which different varieties of a good are exported by countries with different relative factor endowments. Assuming that the higher-quality varieties of a good require more capital-intensive techniques, the model produces the result that higher-quality varieties are exported by capital-abundant countries and lower-quality varieties are exported by labor-abundant countries. Thus Heckscher-Ohlin can, in this framework, yield IIT. In related work, and building on the assumption that the higher-quality varieties require greater capital intensity in production, Jones, Beladi, and Marjit (1999) hypothesized that a labor-abundant country (such as India) may export capital-intensive varieties of a good to high-income countries (such as the United Kingdom or the United States) and keep the lower-quality, labor-intensive varieties for the home market. A further complication for trade theory! page 196

The Level of a Country’s Intra-Industry Trade Work

on IIT has gone beyond examination of reasons for IIT in any particular product. Attempts have been made to determine if the levels of IIT differ systematically by country, such as is suggested by Table 1. A study by Bela Balassa (1986) is an example of this research. Using a standard measure of IIT (see Appendix C to this chapter), Balassa examined a sample of 38 countries (18 developed and 20 developing countries) to test various hypotheses on factors associated with IIT. He hypothesized that a higher level of per capita income for a country is associated with a greater amount of IIT. Balassa’s reasoning followed Linder’s suggestion that, at higher levels of development, trade consists increasingly of differentiated products. Second, a positive association was postulated between IIT and total income of a country, because a larger national income permits greater realization of economies of scale. In his regression equations, Balassa also used independent variables representing items such as distance from trading partners, a common border with principal trading partners, and degree of “openness,” or degree of absence of trade restrictions, of countries. In general, the various hypotheses were essentially confirmed. Greater per capita income, greater national income, greater openness, and the existence of a common border with principal trading partners were positively correlated with the extent of IIT. Distance from trading partners (a proxy for transportation costs) was negatively associated with IIT. Thus, commonsense intuitions on IIT seem to be statistically supported. Another finding was that developing countries had their IIT better “explained” by the analysis than developed countries.