In 1817 a business man working in London, David Ricardo, published a book, On thePrinciples of Political Economy and Taxation, a collection of his pamphlets. The many topics discussed by Ricardo are clearly worthy of a 200 year celebration. Chapter 7 of his book, On Foreign Trade, is the focus of what I wish to talk about in this paper. The concept of Comparative Advantage that he described has since been thought of as The Ricardian Trade Model. My own view, shared by many, is that the Ricardian model still has an important place in International Trade Theory two centuries after the book’s appearance.
The Ricardian Trade Model
Imagine a world in which initially there are two countries, each consuming and producing the same two commodities. If one country is better at producing one commodity, and the other country is better at producing the other, it seems obvious that both countries could benefit if free trade between the countries becomes possible. By contrast, suppose one country is better at producing both commodities than is the other. It may seem that in such a setting there is no basis for mutual gains from international trade. Ricardo disagreed. He argued that both countries could benefit from free trade as long as the pattern of trade reflects each country exporting the commodity in which it has a comparative advantage in production, even if one country does not possess an absolute advantage in producing the commodity it is exporting compared with the other country. Ricardo emphasized that such a result rests on an important assumption: International trade takes place in these commodities, but labor, the factor of production, does not move from country to country. Phrased differently, free trade takes place in commodities but not in the factors that produce the commodities. If a country’s labor force is less productive in the production of both goods than its trading partner, this country can nonetheless be competitive in production of the commodity in which it has a comparative advantage (i.e. compared with its costs of producing the other commodity) if the wage rate is sufficiently lower than the rate in the other country. The kind of international trade that is considered by economists typically is not assumed to be found in a world in which all markets are international. Countries are countries for a reason – they do not want all items (such as most factors of production) to be tradeable or mobile in world markets.
In working with economic models in international trade theory it is often the case that much can be learned by starting with a model that has a small number of commodities and countries. The smallest case, of course, is one in which only two countries produce and trade in only two commodities, and this I will discuss first. However, the following Table shows the technology for a 3x3 case with three countries, A (America), B (Britain) and country C (Continental Europe), and there are three commodities that each country would initially produce if no international trade takes place: Corn (Co), Linen (Lin) and Cloth (Cl).1 The numbers indicate the quantity of labor required to produce each commodity. Ricardo assumed that labor is the only factor of production, and the quantity of labor required for each commodity is assumed constant, unless technical change is explicitly assumed.
America Britain Continental Europe
Corn 10 10 10
Linen 5 7 3
Cloth 4 3 2
First consider the 2x2 setting that leaves Continental Europe aside and with America and Britain only producing corn and linen. Neither country has an absolute advantage in the production of corn, but America has an absolute advantage in linen production since 5 is lower than 7. However, international trade between this pair of countries is possible because Britain has a comparative advantage in producing corn since 10/7 is lower than 10/5.2 As in Ricardo assume that labor is the only factor of production, and that there are constant returns to scale, so that doubling input of labor doubles the output of the commodity being produced. The concept of the world production possibility curve for this case of two countries and two commodities is well known, and illustrated in Figure 1. Point 1 shows the world production of Linen if both countries produce nothing but Linen. That would only be the case if the price of Linen were sufficiently high. If the relative price of Corn should sufficiently increase, which country would be the first to produce some units of Corn? Answer: the country with a comparative advantage in the production of Corn, and that is country B since 10/7 is less than 10/5. Comparative advantage reflects a comparison between two cost rankings in one country with a similar comparison of two production costs in the other country. If the relative price of Corn when trade takes place should rise above 10/5, country B has become completely specialized (at point 2), but country A would start shifting some of its labor towards producing some Corn only if the relative price of Corn reaches the slope of line 2-3 in the diagram. Point 4 shows outcomes if country A produces nothing but Corn and country B is producing only Linen, but this is not a combination of production that could be found if competitive markets are in equilibrium. In a competitive equilibrium no firm produces a commodity that has a lower market price than the cost of production for that firm. Furthermore, there cannot be an equilibrium if a firm does not produce a commodity which has a higher price than the cost for that firm to produce the commodity. These two conditions are required for equilibrium in markets if they are purely competitive, an assumption basic to the Ricardian trade model.
Of particular interest are cases in which every country is completely specialized to a single commodity. In the 2x2 case there are three “classes of complete specialization”: (i): both countries produce only corn; (ii): both countries produce only linen; (iii): one country is specialized to corn and the other to linen. There is only one possibility in (i) and only one possibility in (ii). However, in (iii) there are 2 possibilities, with one country producing corn and the other country linen. In this class of complete specialization each country is producing the commodity in which that country has a comparative advantage, A in Linen and B in Corn.
Now use the level of aijto capture the “man-hours” required to produce a unit of commodity j in country i.The statement that countries A and B have these comparative advantages can be made by the following comparison of cost ratios of Corn relative to Linen:
aBco/aBLin < aAco/aALin
An alternative way of stating this is by cross-multiplying:
(1’) aBco aALin < aAco aBLin
In words: The assignment of country A in linen and country B in corn is the minimumproduct of the pair of possible assignments of one country in linen and the other in corn, and this assignment reflects comparative advantage.
Consider, now, the situation in which there are three countries and three different commodities as shown in the Table. Note that the kind of comparison of ratios shown in equation (1) for the 2x2 case seems not appropriate for higher dimensions. However, the kind of comparison used in equation (1’) can be used. Unlike the 2x2 case the 3x3 case has more than four classes of complete specialization: it has ten such classes. In three of them all three countries are producing the same commodity, so that there are no alternatives. Another set of possibilities is that a single country produces Corn and the other two countries produce the same commodity – either Linen or Cloth. Or a single country produces Linen and the other two produce the same other commodity. Or a single country produces Cloth and the other two countries produce in common one of the other two commodities. Finally, there is a tenth possibility: each commodity is only produced by a single unique country. This latter class of complete specializations has a full six number of possibilities. In considering what is the only assignment (out of the six) that could be supported in a competitive equilibrium, note that country C has an absolute advantage over the other two countries in producing Cloth, and recall that in the two-commodity previous case it was country A that had the comparative advantage in producing Linen and country B in Corn. In the 3x3 setting this assignment (A in linen, B in corn, and C in cloth), is an assignment for all three countries that satisfies bilateral comparisons and it has a product of 100. However, an alternative assignment, with country A in Corn, B in Cloth and C in Linen has a product of 90. For any class of assignments, the winner has the minimum value of the product of labor requirements.3
The argument that leads to this result can be easily explained. Suppose the assignment of countries to commodities for the class of complete specialization is indeed the assignment supported by a set of commodity prices. Let pj be the price of commodity j in equilibrium faced by all three countries, and the wage rates in each country by wi. Therefore, in equilibrium,
wAaACo = pCo
wBaBCl = pCl
wCaCLin = pLin
Furthermore, if that is indeed an equilibrium there must not be any country that could produce a commodity at a lower cost than its market price. For example, consider the one that put country A in Linen, country B in Corn and Country C in Cloth (the assignment discussed above). If that assignment is ruled out as a contender at given prices, cost of production must be greater than price for at least one commodity. That is,
wAaALin ≥ pLin
wBaBCo ≥ pCo
wCaCCl ≥ pCl
with a strict inequality for at least one industry. Multiply the first set of equalities to get:
wAwBwCaACoaBClaCLin = pCopClpLin
Multiply the three inequalities by comparison to get:
wAwBwCaALinaBCoaCCl > pCopClpLin
A comparison reveals that the product of labor coefficients that is consistent with equilibrium, aACo aBCl aCLin, is indeed less than all other products of labor coefficients in the class in which each country is specialized to a different commodity. However, it is important to note that Comparative Advantage does not by itself reveal which is the overall equilibrium pattern of production, since this depends not only on production technology, but also on world demand patterns. What comparativeadvantage does reveal is which trading patterns cannot be consistent with equilibrium (and this fraction grows substantially with more countries and commodities).
Figure 1 shows the world production possibilities when there are two countries and two commodities. Note that the diagram not only shows that country B has a comparative advantage in producing Corn, it shows the extent of B’s advantage. In the 3x3 case the world transformation surface is three dimensional. Figure 2 is only two-dimensional, and thus does not show the extent of differences between countries, but it does show the ordering according to comparative advantage. For simplicity, the effective productive relative size of each country is arbitrarily assumed roughly the same. Consider the extreme case in which all three countries are completely specialized in producing Corn. The diagram shows that country C is the relatively worst producer of Corn, and will be the first country to release labor to produce Linen if it’s relative price, pLi/pCo, should increase sufficiently or, if it is the relative price of Cloth that increases, labor moves out of Corn into Cloth.4 There are 10 classes of specialization in Figure 2. Nine of these have only one or two commodities produced and are simple to explain, like in Figure 1. The only class that is different is the one in which each country is specialized to a different commodity. This class has six contenders. The only contender that could be supported in competitive markets has country A producing Corn, country B Cloth and C Linen. As shown in the preceding paragraph, the assignment of A into Linen, B into Corn and C into Cloth does indeed satisfy the bilateral conditions for each pair of commodities, but does not satisfy the requirement that the product of all three assignments be minimal, since 90 < 100.
The use of comparative advantage by Ricardo does not by itself reveal what each country produces. That depends upon demand patterns as well as supply, i.e. where in Figure 2 does equilibrium lie. What the Ricardo condition does is to reveal what assignments are ruled out in free trade conditions. In the 2x2 case there were a total of 4 possible assignments in which each country was specialized to a single productive activity. Only one assignment, that in which country A produced corn and B produced linen was ruled out, 25% of the total. In the 3x3 case there are 27 possible assignments but 17 are ruled out, i.e. 63%. (In the 4x4 case the percentage ruled out would be 86%). The importance of comparative advantage is that the percentage of possible production patterns (in the class of complete assignments) that are ruled out becomes very high.
Figure 2 attempts to show the important features of the 3-dimensional world production possibilities, but only has two dimensions in which to show this. The three corners show every country completely specialized in producing the same commodity. Along the edges Figure 2 shows the movements as in Figure 1. Now consider a point such as point x, which I assume is the point where a world indifference curve is tangent to the world production possibility surface. Country A produces both Linen and Corn, and country C produces both Linen and Cloth. What does country B produce? Only Cloth, with the relative value of the price of Cloth compared with Linen or with Corn already determined by countries A and C. Point x lies on a “flat” along which relative prices are all the same. For points along the lines only one relative price is constant, while at the point where each country is specialized to a different commodity no single relative price is completely determined.
The Heckscher-Ohlin Model
The Ricardian Model was challenged 102 years after its appearance by a Swedish economist, Eli Heckscher in 1919, and then by one of his students, Bertil Ohlin, in 1933.5 Their basic purpose was to examine the determinants of the pattern of trade, especially when there is more than one factor input. Whereas Ricardo in his model assumed that a country’s labor force was the only factor of production in any commodity, both Heckscher and Ohlin considered the possibility that in production there is more than just labor: they considered capital and/or land as two other production factors. 6
This setting raises several important possible differences between countries: (i) The overall relative quantities of factor endowments may differ between countries. For example, America has a higher ratio of land to labor than does Sweden. As well, America might be a relatively capital abundant country compared to Sweden. (ii) The relative ratio of factors in which a specific commodity is produced may differ significantly from that in other commodities. (iii) Furthermore, factor ratios used in producing a commodity might be sensitive to changes in the relative prices of land or capital compared with wage rates on labor. All these possibilities were absent in the Ricardian model.
Ricardo assumed that technologies differed from country to country, and that these differences were the only basis that determined the pattern of comparative advantages among countries. In order to focus on otherexplanations leading to trade patterns, Heckscher and Ohlin assumed that all countries shared the sametechnological knowledge. And, if relative factor prices should change, so would the techniques of production. For example, if labor costs are reduced, more labor-intensive means of production are apt to be undertaken.7
World Production Possibilities used in the Ricardian model (admittedly long after Ricardo’s time) can also be used in discussing Heckscher-Ohlin models. However, the linear features displayed in Figures 1 or 2 are not found in H-O models. Instead, small changes in relative commodity prices might lead to small changes in relative outputs produced by several countries. Nonetheless, the central features found in the Ricardian model survive as well in models in which there are many kinds of inputs used in each of several commodities and countries. Patterns of trade still reflect the difference among countries in comparative advantage if we assume that whereas commodities may have global markets, some factors of production are nonetheless exchanged only within countries so that factor prices need not be similar even if traded commodity prices are globally determined. Although the models of Heckscher and Ohlin may differ from the Ricardian model, the importance of comparative advantage comparisons among countries are still found. Comparativeadvantage becomes important when some factors of production have their returns determined in country markets. Price changes for commodities that have global markets do indeed affect local markets in which factor prices are determined, but do not require that factor price changes must be the same between countries.
Two particular journal articles published in mid-twentieth century had profound effects on the basic Heckscher-Ohlin properties of small-scale models. The first paper was that written by Wolfgang Stolper and Paul Samuelson and published in 1941.8 Stolper and Samuelson assumed that a country that imports a labor-intensively-produced commodity from another country does so because the foreign country’s endowments are more labor-abundant. Their paper concerns the consequences for Home wage rates should the Home country protect production at Home by levying a tariff on imports of this commodity (call It commodity X) from the more relatively labor-abundant country. The surprising result: the Home wage rate will rise – indeed it will increase so much that the tariff actually increases the real wage rate at Home. The reason? Because the wage rate will increase relatively more than the commodity price (assuming a simple 2x2 model). The reasoning behind this result was even more simple than the argument they provided (although that was also clear). The authors assumed that commodity markets were sufficiently competitive so that if a tariff is imposed the relative change in the local commodity price must be an average of the two relative factor price changes (for labor and capital).9 If the price increase is only for the more labor-intensive commodity (and Y is the other commodity produced), and if x is defined as the relative change in the local price of commodity X, the Home wage rate relative change ( is the relative change in the return to capital) is >X> Y > so that Home’s real wage rises, regardless of taste patterns of laborers.
The second paper (1953) is one that suggested that in the United States the trade pattern comparing exports with import-competing industries seemed to contradict the expectations of the Heckscher-Ohlin model. The author was Prof. Wassily Leontief (Harvard University), and the Input-Output figures he used were for the United States. U.S. exports seemed to be produced at Home with more labor-intensive techniques than found in America’s import-competing sector. This surprising result (because all agreed that America was the most capital-abundant country) quickly was called the “Leontief Paradox” and many new articles (and Ph.D. theses) from other economists not surprisingly followed. Years later I suggested a simple argument that the kind of result that Leontief obtained is not that unusual.10 To keep the argument simple, suppose that only two factors are used in production, capital and labor. Figure 3 illustrates the technology for four different commodities, 1 through 4, shown by the four positively sloped schedules that illustrate that the capital/labor endowment ratio (which increases with growth) would drive up the relative wage/rent ratio if the economy produces only a single commodity. Each of the three dark horizontal lines illustrate that for some K/L endowment ratios the economy produces a pair of commodities. There are three regions in each horizontal line, regions I, II, and III, and, because two commodities are being produced with two productive factors, costs in equilibrium are equal to commodity prices and factor prices are determined, (but not necessarily the same in different countries).
Consider, now, two countries (among many): Home’s endowment is relatively labor abundant and suppose it produces commodities 1 and 2 in region III of the first horizontal locus. That is, Home produces mostly the second commodity but does also still produce some of the first commodity in its import-competing sector. By contrast, suppose Foreign is significantly more capital abundant, and produces commodities 3 and 4 in its region I, with exports of commodity 3 and local production in import-competing sector 4. Region II in each country has not been discussed because in those regions each country produces a sufficient amount of two commodities that it exports both of them and has no need to produce a commodity it also imports. That is, in thatregion there is no import-competing produced commodity to be compared with the exporting commodities. Note that Home, a more labor-abundant country, is exporting itsmore capital-intensivecommodity2, and Foreign, a more capital-abundant country, is exporting itsmore labor-intensive commodity 3. This pattern of trade seems to contradict expectations of the Heckscher-Ohlin theory. However, comparing factor-intensities for different commodities produced within a single country is not the same as comparing factor intensities between countries with different factor endowments. Ricardo’s emphasis on comparative advantage requires a comparison of technology between countries, and such data were probably at that time not available except for the United States. In his 1953 paper, Leontief compared factor intensities in both its export and import good in the same country. If both countries shared the same technologies between countries but the capital/labor ratios actually used depended on relative endowments, the more capital abundant country would use higher capital/labor proportions in all its commodities produced than in any commodity produced in the other more labor abundant country. (That sounds more like the results desired by Heckscher and Ohlin.)
Fragmentation and the Role of Services in Production
Recall that in the Ricardian model labor is the only factor of production. Attention was paid to situations in which the only production situations are those in which final consumption goods are produced. However, many items that get produced are parts and components that need to be assembled before the finished product emerges. Think of two kinds of productive activity. Labor produces a variety of what we can call “Production Blocks”, but these “parts” need to be “linked” together, and this we call “Service Links”, also requiring labor.11 Do all production blocks that are required to produce a particular final consumption commodity get produced in the same country that exports the final commodity to consumers? Not necessarily. Service Links may bring together Parts that are produced in different countries. It turns out that in the past few years international trade in parts and components have been growing more rapidly than other items, such as final goods. As Ricardo assumed, Labor is often not traded in international markets. However, some of the items that labor producesmay enter global markets. That is, the objects that labor produces may be traded on world markets without requiring that labor itself do so. Do these remarks suggest that David Ricardo’s emphasis on comparative advantage as the key explanation of the gains from trade becomes dated, no longer useful? Just the opposite. With greater possibilities for parts and components to be traded, who should be producing such parts and components? The answer: countries (and their firms) that have a comparative advantage in producing the individual parts and components.
Figure 4 may seem to have a strange title, ending with a question mark. Most discussions of competitive markets in international trade rely on an assumption of the way in which improvements in technology take place. If we assume that markets are highly competitive, as is often assumed in discussing international trade, production functions are typically assumed to exhibit constant returns to scale. In Figure 4 a positively sloped line from the origin where production costs are shown on the vertical axis and levels of output by the horizontal axis is line 1. Such a line would indeed exhibit constant returns to scale. Line 2 suggests that a different technique might be possible, but this entails a set of fixed costs such as shown by line 2 emanating from point A on the vertical axis. If such costs are made it is possible that marginal costs of production can be lowered (once fixed costs are paid). It would not pay producers to switch to such techniques unless output is large enough, higher output than shown by the intersection of lines 2 and 1. A further change of techniques is shown by line 3, requiring further types of fixed costs shown by point B, after which marginal costs are reduced. The dark portion of three sets of lines 1, 2, and 3 suggest “increasing returns to scale”. Sufficiently large firms will have marginal costs of production as shown by line 3. Also shown in Figure 4 are dashed lines 2’ and 3’. Compared with lines 2 and 3, these suggest a significant fall in the costs of service link activities that reduce fixed costs A and B to A’ and B’ so that lines 2 and 3 drop to 2’ and 3’, the new heavy dashed lines. This is meant to capture the significant lowering of the cost of service link activities (and arbitrarily not to illustrate any reduction in marginal costs). The question mark in the title of Figure 4 is meant to suggest that the remark could be made that the recent lowering in costs are due to increasing returns to scale. I would argue that in a sense that is the case. However, I would prefer that in the lowering of costs that are brought about by the great lowering in service link costs (the lowering of A and B to A’ and B’), the heavy dashed parts should be distinguished from any severe technological lowering of marginal costs of activity “on the factory floor”.
International Trade Theory was a key attention when David Ricardo and friends talked and wrote about economic issues two centuries ago. In the past century much attention has been paid to economic models that extended the Ricardian model in allowing for more than a single input (labor in the Ricardian model) to emphasize that inputs such as capital and land may well be important in production. These more recent trade models typically agreed with the Ricardian emphasis that when considering trade among countries it frequently makes sense to assume that with more firms and countries engaged in production the economic model of perfect competition in global markets often remains quite appropriate even in more complicated and realistic models than the Ricardian model. However, although interest in the Ricardian model seemed to fade over time, it became clear that Ricardo’s emphasis on the importance of comparative advantage remained central to the analysis of global trade. The reason? Some items or productive factors are not exchanged in world markets, but are exchanged instead in country markets or even more local markets. This is especially the case often found in labor. Countries are countries for a reason, and many countries prefer to have control over movements of labor (as in the Ricardian model) as well as in many drugs or forms of capital. International trade is clearly important, but the world is far from being “One Country”.
Is Ricardo’s assumption that labor is the only factor of production still important and useful after 200 years? My vote is yes. Clearly, it is important because his assumption that an important factor of production still remains in a country market supports the basic role of comparative advantage. Furthermore, the fact that it is a very simple model should not be held against it. Suppose that the question that is asked is what is the effect on a country’s overall wellbeing if a traded commodity goes up in price because of a change in technology or taste patterns abroad. The Ricardian model suggests that the easy to obtain change in the wage rate suffices to answer the question, since labor is the only factor of production.
Final Remarks (made by Paul Samuelson)
On many occasions Prof. Paul Samuelson had talked about the theory of international trade and its connection to the wider field of economic theory. One of these occasions was in 1968 in Montreal, where Samuelson was giving a Presidential Address to the Third World Congress of the International Economic Association.12 I was fortunate to have been at this occasion, and was struck by Samuelson’s remarks:
“Our subject puts its best foot forward when it speaks out on international trade. This was brought home to me years ago when I was in the Society of Fellows at Harvard along with the mathematician Stanislaw Ulam. Ulam, who was to become an originator of the Monte Carlo method and co-discoverer of the hydrogen bomb, was already at a tender age a world famous topologist. And he was a delightful conversationalist, wandering lazily over all domains of knowledge. He used to tease me by saying, ‘Name me one proposition in all of the social sciences which is both true and non-trivial’. This was a test that I always failed. But now, some thirty years later, on the staircase so to speak, an appropriate answer occurs to me: the Ricardian theory of comparative advantage; the demonstration that trade is mutually profitable even when one country is absolutely more – or less – productive in terms of every commodity. That it is logically true need not be argued before a mathematician; that it is not trivial is attested by the thousands of important and intelligent men who have never been able to grasp the doctrine for themselves or to believe it after it was explained to them.” (p.683)
If Ricardo could speak after 200 years he might say, “Thank you, this proves my case.”
Heckscher, Eli: “The Effect of Foreign Trade on the Distribution of Income,” (1919),
Ekonomisk Tidscrift, translated in Harry Flam and M. June Flanders, Heckscher-Ohlin Trade (Cambridge, Mass): The MIT Press, (1991).
Jones, Ronald W. “Comparative Advantage and the Theory of Trade: A Multi-Country Model,” Review of Economic Studies, June, 1961.
……………”Heckscher-Ohlin Trade Flows: A Re-appraisal”, (2008) Trade and Development Review, v.1, #1, pp. 1-6.
Jones, Ronald W. and Henrik Kierzkowski (1990): “The Role of Services in Production and International Trade: A Theoretical Framework”, ch. 3 in Jones and Krueger (eds.), The Political Economy of International Trade (Blackwell’s).
Leontief, W. : “Domestic Production and Foreign Trade: The American Capital Position Re-examined, Proceedings of the American Philosophical Society, 97, (4), September, 1953.
Ohlin, Bertil: Interregional and International Trade (Harvard University Press, Cambridge, Mass), 1933.
Ricardo, David: “On the Principles of Political Economy and Taxation”, 1817.
Stolper, Wolfgang, and Paul Samuelson, “Protection and Real Wages,” Review of Economic Studies, November, 1941.
1 This set of numbers for the 3-country, 3-commodity case was used in Jones, Review of Economic Studies, (1961), which also examined the case of many commodities and countries. An easy discussion of this paper was provided by Wilfred Ethier in Chapter 2 of Ethier, Helpman and Neary, Theory, Policy and Dynamics in International Trade, Cambridge University Press, 1993.
2 10/7 is the price of corn in Britain expressed not in units of labor but in units of linen if Britain were to produce both commodities. 10/5 would be the price of corn in units of linen in America if it were to produce both commodities.
3 As noted in the previous footnote the argument that supports comparative advantage as the significant feature of the Ricardian Model in higher dimensions is explained in Jones (1961).
4 Figure 2 also shows that countries A and B are the relatively worst producers of Cloth and Linen respectively. This is what leads to the triangles at each of the corners in Figure 2. This need not be the case with some other technologies for labor’s use.
5 Heckscher’s paper was not translated into English until 1949.
6 Ricardo, of course, was aware that in reality labor was not the only factor of production. He even used a four- letter word when talking of the price of another productive factor – rent. (See, for example, the second chapter of his 200-year old book.)
7 Some economists later allowed differences in technology as well as differences in factor endowments.
8 Wolfgang Stolper and Paul Samuelson: “Protection and Real Wages,” Review of Economic Studies,” v.9, pp. 365-79 had earlier been rejected by the American Economic Review. A 50th anniversary celebration led to a book entitled, The Stolper-Samuelson Theorem. Stolper was my first undergraduate professor (at Swarthmore College) and Samuelson was my professor for graduate work at M.I.T. This past history urged me to suggest to Alan Deardorff (at the University of Michigan along with Stolper) to have a 50th anniversary celebration there, especially since Alan was a student of mine when I taught a trade course at Cornell.
9 This presumes that there is no joint production. That is, every productive activity requires two inputs to produce a single output. This assumption was so commonly made that often little notice was paid to its importance in affecting factor prices.
1010. R. Jones: “Heckscher-Ohlin Trade Flows: A Re-appraisal”, Trade and Development Review, Volume 1, Issue 1, 2008.
11 See especially R. Jones and H. Kierzkowski, (1990).
12 The talk was entitled “The Way of an Economist” and is found in vol. 3 of The Collected Scientific Papers of Paul A. Samuelson.