University of Washington
Geography 349 (Professor Harrington)
International Trade Theory
Optimally, a trade theory would help us explain or predict
  • what nations export and import what goods
  • with what other nations
  • under which economic, geographic, and political circumstances,
  • with what consequences.
This would allow us to predict and prescribe the content, direction, and size of multilateral trade flows.


  •  not a full-blown trade theory (see above);  rather,
  •  an economic policy of governmental accumulation of wealth, in the form of gold bouillon for:
    •  domestic control
    •  investment
    •  international expansion
  •  reflects the era of nation-building and the shifting of European political power from feudal lords and the church to national sovereigns;  also reflects and supports the principal source of wealth -- trading
The trade-policy implication of this economic policy was the generation of a national trade surplus, paid for by accumulation of gold reserves.
Before fully developed financial systems, there was little international credit.  Therefore, a current-account surplus was not matched by net capital outflow (net loans or investment overseas);  rather, it was matched by a net inflow of gold to pay for the excess of goods exported from the country.  Some of this gold found its way to overseas investment by the sovereign.

1776:  Adam Smith's The Wealth of Nations

Political and economic liberalism found their expression in Smith's argument that the wealth of nations depends upon the goods and services available to their citizens, rather than the gold reserves held by the sovereign.

Maximizing this availability depends, first, on putting all resources to use, and then, on the ability --

  •  to obtain goods and services from where they are produced most cheaply (because of “natural” or “acquired” advantages), and
  •  to pay for them by production of the goods and services produced most cheaply in the country,
  •  with costs measured in terms of direct and "embedded" labor inputs.
This principle fit the development of capitalist economies based on production via wage labor (rather than trading commodities for profit);  reflects the manufacturing dominance of Britain;  reflects manufacturing economies of scale based on:
     - development of specialized equipment;
     - labor training and specialization;
     - long "runs" of one product.
 These are sources of acquired advantage.

The consequent trade policy is relatively free trade, so that a country should import goods that would be produced more expensively internally, where expense is measured according to the labor theory of value.
These imports are to be paid for by the production of goods that the country can produce with less use of labor per unit.
Exports flow from the country that can produce a product most cheaply.

• 1817:  David Ricardo's On the Principles of Political Economy and Taxation
• The possibility of system-wide gains from trade persists, even when a given country has an absolute advantage in the production of no product.
• Specialization and trade should occur according to the relative opportunity costs of production in each country, measured in terms of the alternative production given up to produce a tradable good.


resources required per unit output:

Tea Wheat
Sri Lanka 10 10
United States 5 4
  • The opportunity cost of tea in Sri Lanka is 1 unit wheat;  the opportunity cost of tea in the US is 1.25 units wheat.  Sri Lanka has comparative advantage in tea production, despite its absolute disadvantage in the production of each commodity.
  • To test for comparative advantage in the production of commodity A in a 2X2 model:  for which country is the opportunity cost of A smaller -- which is to say, for which country is the ratio of  resources required for a unit of A to the resources required for a unit of B smaller?
  • 10/10 < 5/4, thus, the comparative advantage of Sri Lanka is commodity 1, tea.
  • Although the US has an absolute advantage in the production of both tea and wheat, the US has a comparative advantage only in the production of wheat.  This is because its advantage in wheat is comparatively greater than its advantage in tea [Daniels and Radebaugh, 8th ed., p. 202].
• In Root's phrasing, "gainful trade will occur between countries when their pretrade relative price structures are different" [F.R. Root, 1990, International Trade and Investment, p.45].

How are the gains from trade divided between two trading partners?
If xi refers to the domestic cost of producing x (a or b) within country i, and Ci/Cj  refers to the exchange rate between currencies i and j (how many units of Currency i equals one unit of Currency j), then:
  • trade is gainful when ai/bi (the ratio of the costs of producing a and b in country i) is not equal to aj/bj (the ratio of the costs of producing a and b in country j)
  • if  ai/bi < aj/bj, then i will export a and import b, so long as exchange rates allow:
  • ai/aj < Ci/Cj < bi/bj  (which is to say that the exchange rate is not so extreme as to wipe out the differences in opportunity costs for each item across the two countries).
 For example:
  US beef costs  $10 in the US
  US cameras cost $20 in the US
  Japanese beef costs ¥5000 in Japan
  Japanese cameras cost ¥1000 in Japan
 · Without a common numeraire (such as labor-value), we don't know the absolute advantages.  However, the relative values within the two countries, 1/2 and 5/1, tells us that the US comparative advantage is beef and the Japanese comparative advantage is cameras.  Why?  Because if costs are accurately measured, then the opportunity costs of beef (in terms of cameras foregone, since there are only two possible products) are lower in the US than in Japan:  1/2 a camera foregone in the US;  5 cameras foregone in Japan.
 · Despite this clear comparative advantage, trade will only occur if
  10/5000 < $/¥ < 20/1000, which is the same as
  0.002 < $/¥ < 0.02, which is the same as
  1/500 < $/¥ < 1/50.

Q:  What would make the dollar fall so much against the yen that a dollar would buy fewer than 50 yen, and this bilateral trade would end?  (That is, in a world where neither currency was used for other international purposes, and where there was no monetary policy).
A:   Such a low Japanese demand for US beef and such a high American demand for Japanese cameras that the exchange rate fell this low.

This relative demand for products from trading partners, expressed via its effect on exchange rates, determines the division of gains from trade.

  • At $1 = ¥200, the US will import 2 cameras for each unit of beef it exports;  this is 1.5 more cameras than it could make domestically in lieu of one unit of beef;  the US gain from trade is 1.5 cameras per unit of beef.
  • At $1 = ¥183.33, the US will import 1.83 cameras for each unit of beef it exports;  the gain from trade becomes 1.33 cameras per unit of beef.
  • At $1 = ¥120, the US will get $10 = ¥1200 for each unit of beef it exports to Japan;  this would allow the US to import 1.2 cameras (at ¥1000 each);  this is 0.7 more camera than the US could make domestically at the opportunity cost of 1 unit beef, so the US gains from trade is 0.7 camera/beef.
The equation below notes that the gains to country 1 from exporting commodity a is the amount of commodity b that can be imported from country 2 (per unit of commodity a that is exported) minus the cost of producing commodity a for export (expressed as the amount of commodity b that is foregone).

 Simplified,  country 1's gains from trade = G1 = (A1/B2)(C2/C1) - a1/b1

  • The first ratio (A1/B2) is the ratio of (i) the unit price that Country 1's exports fetch on the world market (in terms of Country 1's currency) to (ii) the unit price that Country 2's exports fetch on the world market (in its currency).
  • The second ratio (C2/C1)is the exchange rate:  the "price" of the exporting country's currency, which (without capital or money markets) is also a measure of the relative demand for that country's exports.
  • The third ratio (a1/b1) is the opportunity cost of producing the exported product in the exporting country.
  • (A1/B2)(C2/C1) is what Country 1 can get in return for exporting a unit of a
  • a1/b1 is what Country 1 gives up as it produces a unit of a
  • So can you see why (A1/B2)(C2/C1) - a1/b1 is the "gains from trade"?  Wait, ask yourself -- can you see why?
Note, then, that G1 (the exporting country's gains from trade):
  • increases with the relative unit price of Country 1's vs. Country 2's export items (this reflects the global supply/demand balance for these items),
  • increases with strength of the Country 1's currency, and
  • decreases with the opportunity costs (in Country 1) of producing the exported item.

Let's run through the Japan/US example, above, using this formula. 
a1 = the cost of producing the US export item, beef, in the US = $10/beef.  For now, let's assume that costs equal prices: a1 = A1 .
b2 = the cost of producing the Japanese export item, cameras, in Japan = ¥1000/camera.  For now, let's assume that costs equal prices: b2 = B2 .
C2/C1 = the exchange rate, varied;  let's take the example where it's ¥200/$1.
b1 = the cost of producing the alternative product, cameras, in the US = $20/camera.

Thus, our formula G1 = (A1/B2)(C2/C1) - a1/b1 suggests that the US gain from trade =
($10/beef / ¥1000/camera)  (¥200/$1) - $10/beef / $20/camera =
($10/beef) (camera/¥1000)  (¥200/$1) - ($10/beef) (camera/$20) =
                                   2 cameras/beef  -  .5 camera/beef  =  1.5 cameras/beef, or 
1.5 cameras gained by the US in return for each unit of beef exported.

At the same time, Japan has a gain from this trade:
(¥1000/camera / $10/beef)  ($1 / ¥200) -  ¥1000/camera / ¥5000/beef =
(¥1000/camera) (beef/$10)  ($1 / ¥200) - (¥1000/camera) (beef/¥5000) =
                                      .5 camera/beef - .2 camera/beef, or
.3 unit of beef gained by Japan in return for each camera exported.

As you think about this, and view the graphs in the textbook (Figs. 6.2 and 6.3), you'll recognize that we're making some assumptions:

  1. We assume that national production is on the production possibility frontier (PPF), with no un-used factors (e.g., no unemployment).  If some available and relevant factors are not being used (e.g., labor unemployment in the export sector), then we can't clearly say that producing more of the export item would mean producing less of other items. 
  2. We assume that factors are homogeneous within a country (e.g., L is perfectly substitutable across individuals and across sectors), so that workers, capital, and resources not used in one sector can work in another.
  3. To make the numerical example simpler, I've assumed that prices equal costs (including a standard rate of return to capital and entrepreneurship).  That's not always the case:  sometimes prices are much higher than costs (e.g., in a monopoly), and sometimes international prices are lower than domestic costs (that's called "dumping").  We can deal with this by recognizing that a1 might differ from A1 and b2 might differ from B2.
  4. We've ignored transportation costs -- but we could implicitly include them in A1 and B2 in the first term of our gains-from-trade equation -- or we could add them in as a variable in the first term -- this would allow us to show what happens with transportation costs fall (e.g., with containerization) or rise (e.g., with higher fuel prices).
  5. We assume that factor prices reflect the value of marginal product attributable to the factors.  In reality, this is affected by the competitive structure of the industry (higher but declining in a monopolistic or oligopolistic industry), the nature of the factor markets (supply is restricted in the case of unionized L or the guild professions), and finally, wages/salaries are not solely determined by economic forces.

These are all problems with using traditional trade theory to understand and to prescribe trade flows in the current economy.

1935:  Bertil Ohlin's Interregional and International Trade, based on earlier work by Eli Heckscher

What explains the differences in opportunity costs for producing the same product in different countries?  Possibilities include skill or technology (including a preference for producing in different ways), availability of materials or resources, or the pricing of inputs.

1) ubiquitous technology (the same everewhere),
2) general preferences to use the best available methods,
3) perfect competition in domestic factor markets (which should push factor prices to reflect their opportunity costs),
4) input requirements that differ across different products, but
5) only one particular mix of inputs for each different product, and
5) immobility of factors,
comparative advantage should depend on relative factor availability.

• "A country has a comparative advantage in the production of goods that use relatively large amounts of its abundant factors of production and a comparative disadvantage in the production of goods that use relatively large amounts of its scarce factors of production." Goods trade can be considered the indirect trade of factor services [Root, p.69].
 --> What is meant by abundant and scarce?  Measured by relative prices received by an additional unit of two factors, in one country versus that relationship in another country -- in other words, geography.

1) heterogeneous factors -- of differing quality for specific production processes
2) substitution within a production function (a given product can be made with different proportions of inputs, depending on the relative prices of inputs in different places or at different times)
3) raw-material extraction is based on absolute advantage

The principles of comparative advantage and factor proportions form the basis of the traditional, neoclassical theory of international trade.  Note that this is a normative theory, in that it asks the question "If we had a goal of maximizing world production (the goods and services available to citizens of each country), how would we proceed?"  If we assume that
  • resources (a.k.a. factors of production) are immobile, but that
  • goods are mobile and
  • technology is stable and ubiquitous,
then we maximize world production and the goods and services available to each country by using resources for the production of goods that face the lowest opportunity cost within each country, and trading the locally unneeded products for other goods, produced with the lowest opportunity cost for resources in other countries.

1953:  Wasily Leontief published "Domestic production and foreign trade: the American capital position re-examined" in Proceedings of the American Philosophical Society (v.97).
1956:  Wasily Leontief published "Factor proportions and the structure of American trade: further theoretical and empirical analyses" in Review of Economics and Statistics  (v.38):

Using data available from the 1947 input-output (I-O) model of the US economy, Leontief calculated the capital and labor requirements for the production of $1 million of US exports and $1 million of US production in import-competing industries.  He found that

  • US export sectors (industries) required a higher proportion of labor than
  • US production in those sectors in which the US had substantial imports.

Why is this a paradox?

  1. The US is capital-abundant and labor-scarce, relative to the rest of the world.  (That's why US wages were the highest in the world in 1947.)
  2. The principle (or theory) of factor proportions says that a country should have a comparative advantage in the production of goods and services that require large proportions of the country's abundant factor and smaller proportions of the country's scarce factor(s).  If a country has an open (free) trade policy, market forces will result in exports of the goods and services in which the country has comparative advantage.  (This is the same as saying "Free trade results in specialization and trade according to comparative advantage.")
  3. US export sectors required a higher proportion of labor (the value of labor used in production as a proportion of total production costs) than did US sectors in which the US had substantial imports.  You can rephrase this to say "US exports are more labor-intensive than US imports."
  • The paradox:  why is a capital-rich country exporting relatively labor-intensive products? 
  • Analogous results were found in Japan, which was then L-rich and K-poor, yet Japanese exports were more K-intensive than Japanese production in sectors that faced import competition.

Possible explanations of the Leontief paradox (see a similar exposition by E.K. Choi)  Focus on the boldfaced explanations

1.  The US was not really capital-abundant and labor-scarce.  Perhaps all the troops returning from World War II flooded the labor market.  [It turns out that this is not a good explanation.]

2.  The principle of factor proportions is wrong.  [This would be so upsetting to trade theory that economists searched for other explanations.  See explanation 3.]

3.  US sectors that faced import competition became more capital-intensive than the competing imports.  This makes sense -- if you're a manufacturer in the US, and you have to compete with imports, you'd start using more sophisticated equipment and fewer workers, because your workers are probably more expensive than your foreign competitors' workers.) Notice that Leontief did not have data on how these imports were produced overseas:  he had data on how these sectors operated in the US.  Perhaps these sectors were actually labor-intensive in the countries that were exporting to the US.  [This is indeed a partial explanation of the paradox.  HOWEVER, this suggests that the same product can be made in different ways.  If THAT'S the case, how can you have a principle of factor proportions -- which relies on an assumption that some products are labor-intensive (everywhere), some products are capital-intensive (everywhere), and some are land-or resource-intensive (everywhere).

4.  There are different kinds of labor, and each should be considered a separate factor.  The most common (though simplistic) way to distinguish labor is to separate "unskilled," "moderately skilled," and "highly skilled" labor.  In 1947 the US may have been the most well-endowed country in the world in highly skilled and moderately skilled labor.  Its exports were indeed skilled-labor intensive relative to its imports -- PRESERVING THE PRINCIPLE OF FACTOR PROPORTIONS.  Furthermore, skilled labor is more expensive than unskilled labor (because it is scarcer than unskilled labor, everywhere in the world), so skilled-labor requirements "inflate" the proportion of labor costs within total costs. [This is a partial explanation of the paradox -- and see the product life cycle model, below.]

5.  The US imported natural-resource commodities whose extraction is capital-intensive, but in which other nations have an absolute advantage.   In other words, some trade occurs based on where the products are found, not on their labor- or capital-intensity of production.  [Empirically, this a partial explanation;  the paradox was more apparent in US bilateral trade with resources-rich countries (e.g., Canada), and was less strong when natural-resource sectors were excluded.] 

6.  Technology itself is a nation-specific factor of production, rather than being a universal attribute of production.  Furthermore, technology is a factor that is produced within a given nation (much like a commodity), but is not perfectly mobile or tradable.  This kind of thinking has led to "neo-technology theories of trade" (see below).  [This is potentially an explanation of the paradox;  US exports -- including cotton -- were and are more technology intensive than its imports.]

• Raymond Vernon, 1966, "International trade and investment in the product life cycle"

• Concepts of product cycles had been developed in industrial economics and in marketing since the 1920's.  Vernon, however, became concerned with the technological bases for PLCs in the late 1950s. with his work on the New York Regional Plan.  He later extended these concerns to the international realm.
• Concerned with only certain products:

     · manufactured goods (or services that can be produced remotely from consumption, like call centers)
     · income-elastic demand
     · L-saving in use
• His original formulation also assumed a ranking of nations by income and wage levels, with steadily rising income and wage levels of the ranked nations.
• Assumes that product innovation is market-led [define product versus process innovation;  technology-push versus market-pull models of innovation].
• Assumes that process technology undergoes stages distinguished by labor-intensity, standardization, unit cost, and ability for technology to be embodied in capital equipment and standard operating procedures.

 With these assumptions, Vernon built a story of these particular kinds of products facing
1) introduction in the highest-income, highest-wage country where the products found their first demand;  production is small-scale, changing, expensive, and uses highly skilled L;  demand is not very price-elastic, because of differentiation, and the nature of pioneering adopters
2) growth of demand and production in the original country, with declining costs and prices;  some export demand from countries with lower incomes and wages — the high-wage, L-scarce country is exporting L-intensive products
3) maturity of demand in the original country, with standardized and increasingly K-intensive production;  establishment of foreign operations in newer markets to serve them and to overcome real or threatened trade barriers
4) decline of product demand in the original country, with increasing competition from other suppliers and other products;  the cost pressure and the ability to embody the technology in K equipment and SOP pushes production "offshore" to low-wage countries, using financial K and K equipment from the originating country — the high-wage, L-scarce country is importing K-intensive products, because the K and technology are mobile within the MNC, while the less-expensive L is not mobile.

This helps resolve the Leontief paradox by explaining, for a limited class of goods, US exports of these goods while they are L-intensive and import of these same goods when they are K-intensive.

However, this model differs from the Heckscher-Ohlin theorem:

  •  technology changes over time, from L toward K;
  •  the mobility of technology changes over time
  •  standardized technology and financial K are mobile factors within the MNC
Today, of course, international differences in incomes and wages do not form a neat ranking, and improvements in communication make it easier to engage in new-product production anywhere within a MNC's global production network.

The “new growth theory” and “new trade theory”:  Romer, Krugman, Helpman:  1980s.
See The Royal Swedish Academy of Sciences descriptions of Paul Krugman's contributions to trade theory and economic geography:  generaltechnical.
See a short video on Paul Krugman and "new trade theory":  (here)

Six assumptions: 

1) Technology is an explicit factor of production, but one that is itself produced with inputs of capital and labor (and thus endogenous to the model of growth). 

2) However, the production of new technology reflects decreasing returns to the application of capital and labor (doubling the resources allocated to new technology does not immediately double the rate of technological advance). 

3) The production of new technology creates externalities:  all the benefits of new technology can't be appropriated by the entity (firm) that invests the resources to create the technology. 

4) There are increasing returns to scale in the use of technology (a little technology goes a long way -- can be used to improve quality or reduce cost of infinite number of units).  More generally, there are increasing returns (economies of scale) for the facility, organization (company), and country that specializes in a specific product.

5) While technology is mobile (across companies and countries), there is imperfect mobility of the ability to use technology, based on localized investments in infrastructure, institutions, and labor.

6) Most importantly, says Krugman, is the assumption of increasing returns to scale (unit costs of production fall with increased scale of output).  For advanced manufactures and for many services, this results in benefits from specialization (the more a company or country produces of a very specific item, the lower the opportunity costs of production).  These benefits include the ability to use and reuse specialized technology, equipment, and even reputation, at little additional cost.  This helps explain a country importing and exporting in the same industry -- the imports and exports can actually be different specific forms of the same type of product (automobiles, aircraft, financial services...).


 1) To the extent that individual companies cannot appropriate all the returns to their investment in new technology (because of externalities in the development and deployment of technology),

  •  companies may under-invest in technology (R&D) -- and
  •  are very likely to under-invest in "technology infrastructure" (e.g., education and communications).

  •   2) A country may gain comparative advantage in a product because it was quick to gain economies of scale in that product.  As a result, it can produce the product more efficiently, relative to other products, than can its trading partners -- not because of factor endowments, but because of the skilled labor, specialized infrastructure, networks of suppliers, and localized technology that have developed to support that industry.

     3) With the additional assumption (which is empirically based) that purchasers benefit from a choice among similar products, we can see how more than one country might gain comparative advantage in similar products, based on (a) similar factor endowments and (b) economies of scale in the production of particular variants of the products.  This helps explain the cross-trade in similar products, usually with national variations -- think of the typical variations among German, Japanese, Swedish, and US-based automobiles.

    Policy implications:

    Thus, there may be a role in government support of technology, via measures such as

      •  R&D tax credits
      •  trade policy to support key, technology-improving sectors
      • "technology infrastructure"
    1980s:  strategic trade policy (industrial policy) based on differential technological linkages among sectors (some sectors lead to development of skills, infrastructure, and institutions that foster further innovation across a range of related industries)
    c.f. the importance that Porter places on “related and supporting industries” to national competitiveness.

    This may all seem bewildering, but so is the prospect of trying to predict and/or prescribe the myriad international trade flows.  Trade in different products is likely to be best interpreted by different models, as I implied when I noted the characteristics of "product-cycle" goods.  Here's a stab at such a typology:

    extracted resources
    absolute advantage
    (1) natural advantage:  is the resource available, and can it be extracted and transported at reasonable cost?
    (2) K availability for extraction
    region-specific goods and services (products such as Champagne or  Roquefort, services such as tourism requiring luxury hotels on sunny, warm white-sand beaches) absolute advantage
    (1) natural advantage
    (2) K availability
    (3) trademarking and reputation

    commodity goods and services (agricultural products, basic manufactures)
    comparative advantage
    (1) "factor proportions":  factor intensity of production and factor endowments of countries
    (2) scale economies
    innovative products or services with income-elastic demand
    product life cycle
    (1) innovative capacity in early stages
    (2) standardized technology, K availability, scale economies in later stages
    ideas, innovation or technology itself
    new trade theory
    sustained application of resources (risk K, highly skilled L) toward creation of new technology or ideas

    What trade or development policies are implied by a desire to promote these different types of export capabilities?

    Specialization and trade according to comparative advantage may maximize world production, but under some circumstances it may not lead to increased wealth and living standards in poorer countries.  Two of these "trade traps":

    IMMISERATING  TRADE  (concept developed by Jagdish Baghwati):  If a country's imports depend on its export of basic, raw commodities, the country may face an inability to increase export earnings for three reasons, each related to the first term in our "gains from trade" model, above:

    1. Raising prices will reduce demand by a greater proportion (because most raw commodities can be accessed from other countries, and because most raw commodities have substitutes)
    2. Lowering prices will not increase demand by an equal proportion (because world demand for many commodities does not increase substantially because of a price decline:  think about world demand for copper)
    3. As the world economy increases, the world demand for many raw commodities increases at a slower rate (demand is income inelastic).
    OVER-ABUNDANCE  OF  POTENTIAL  LABOR:  If the supply of labor is limited, then an increase in labor demand (e.g., because of export growth) will yield an increase in wages (and if the general price level doesn't rise as fast because productivity is increasing, real wages and living standards will rise).  However, some countries have a seemingly inexhaustible supply of additional labor, because of a slow but steady shift of people from subsistence or agricultural activity to an industrial labor force.  Thus, export increases may increase national income but not wage levels -- leaving little incentive for businesses to increase labor productivity.

    This is a complicated and contentious question.  Let me propose this:  the level of skill required for a job relates to the amount of time a person with typical mental and physical ability would take to be very proficient at the job.  Imagine a physically capable 20-year-old with no formal education.  How long would it take for that person to be able to be a good fast-food worker?  Industrial assembler?  College professor?  Electrical engineer? 

    copyright James W. Harrington, Jr.
    revised 1 February 2013