Part VII. Foreign Trade Was Not an Engine of Growth

Trade, then, was important as a context for British growth. But it was not an engine of growth. For the period in question Mokyr makes the clearest case.²⁹ The underlying argument is that domestic demand could have taken the place of foreign demand (Mokyr earlier [1977] had shown likewise that the shuffling of domestic demand was no more promising). To be sure, Britons could not have worn the amount of cotton textiles produced by Lancashire at its most productive: cotton dhotis designed for the working people of Calcutta would not have become fashionable at Marks and Sparks on the High Street of Salisbury or Aberdeen. But in that case the Lancastrians would have done something else with the labor and capital and resources and ingenuity employed in cotton textiles. As Hume put it in the 1740s, “if strangers will not take any particular commodity of ours, we must cease to labor in it.” Of course. But, he continued, in another of his astonishing anticipations of modern economics, “the same hands will turn themselves towards some refinement in other commodities, which may be wanted at home.”³⁰ Or rather, will be wanted at home, since that is how the alternative employment will be guided, as though by an invisible hand. The exporting of cotton cloth is not sheer gain. It comes at the cost of something else that its makers could have done, such as building more houses in Cheshire or making more wool cloth in Yorkshire.

That is, nothing like all the income received from exports is a net gain. Think of the opportunity costs of producing American medical equipment for exports. Pittsburg doesn’t produce such things out of the air. To make the magnetic resonance machine sold to, say, Finland the Pittsburgians divert labor, capital, natural resources from other potential employments, local or elsewhere, such as making more education at the University of Pittsburg, or moving to Philadelphia and making more candy. Exports are not the same thing as new income. They are new markets — which is to say new ways of getting importable things — not new income. They are a way of acquiring Nokia cell phones by showering the Finns with American machinery, telecommunications equipment and parts, aircraft and aircraft parts, computers, peripherals and software, electronic components, chemicals, medical equipment, agricultural products, bonds, and engraved pieces of paper (costing 4 cents to make) marked “dollars.”

The alternative of making the cell phones in America for Americans (“Buy American”) is a rather worse deal for Americans. But it is no catastrophe. American national income would not deflate to zero like a balloon if we did not trade for Nokia cell phones. (Motorola will be glad to explain that point to you.) Given innovation (a big, big given), the source of wealth is specialization and trade within a country, regardless of whether the country then sells snowmobiles to the Eskimo or TV sets to Nebraskans. Domestic efficiency is what gets us out to the production possibility curve, as economists put it (as innovation pushes it out). Your nation, or town, or even in the extreme your own household, does not have to trade with outsiders to live. Each can be an innovative and alert Crusoe on his island and survive without exporting or importing. The point is obvious for big, innovative countries like France or the United States, which can do much better than “survive” without foreign trade. They can achieve very high incomes by attending to innovation, trading merely with other Americans or Frenchmen within their borders, if persuaded by protectionists to do so (as both were to a greater or lesser degree in earlier eras).

In other words, the primitive conviction most non-economists have that foreign trade is the only source of wealth, that money must somehow come from outside to puff up the economy and make us rich, is wrong. You see it in the claim that subsidizing a new sports stadium will “bring dollars into the community.” The dollars are good only for a local owner of land. They have no effect on the rewards to mobile labor or capital. But public opinion gets fooled into voting for the stadium, because of “multiplier effects” (it sounds like technical economics, but only a bad economist thinks that multiplier effects work in anything but conditions of mass, nationwide unemployment). You can see the power of the conviction that a foreign, outside trade is the only source of wealth in the role of fish exports in the political economy of Iceland or of exports generally in that of Japan. The conviction is imprudent and unjust, good for a few exporters and bad for everyone else. “Export or die” is a foolish economic policy, which has undermined domestic policies for growth in poor countries. Imports and the exports to get them are a shift of attention, not consciousness itself. Trade as an essential engine: it seems not.

Yet the trade, of course, benefits the traders on both sides, some, or else it wouldn’t have happened. But again — here’s the nub of the issue — the benefit can be shown in static terms to be small. One of the chief findings of the “new” economic history, with its conspicuous use of economic rhetoric, is that static gains, as I have said, are very often small. Robert Fogel’s startling calculation in 1964 of the social savings from American railways in 1890 is the leading case.³¹ Replicated by Hawke in 1970 for Britain with broadly similar results (though higher on account of denser passenger traffic), in countries unlike Britain or the United States without easily navigable rivers, such as Mexico (Coatsworth 1979) and Italy (Fenoaltea 1971-72), the impact of railways turned out to be greater. But it was never enough to account for any but a small portion of modern economic growth. Fogel’s finding, with Harberger’s, were part of the gradual realization by economists in the 1960s that their beloved supple-and-demand framework did not explain The Great Fact. However essential one may be inclined to think railways were, or how crucial foreign trade to British prosperity, or how necessary the cotton mill to industrial change, the calculations lead to small figures, far below the factor of sixteen, or even a doubling.

For trade, how so? Think of British foreign trade around 1841, like railroads or whatever, as an industry for making consumable imports of wheat and lumber by selling exports of iron and cotton textiles made with Britain’s inputs of labor, land, and capital. From an economist’s point of view that is all it is, a machine for making imported sausage for consumption out of domestic ingredients. In 1841 the mighty United Kingdom exported some 13 percent of its national product. The terms of trade is the “productivity” of the industry that “makes” wheat out of cotton textiles sacrificed (that is, exported for the use of foreigners). The terms of trade tells how many bushels of wheat the British got for each yard of textiles. From 1698 to 1803 the range up and down of the three year moving averages of the gross barter terms of trade was a ratio of 1.96, highest divided by lowest; Imlah’s net barter terms range over a ratio of 2.32, highest divided by lowest.³² So the variation of the terms on which Britain traded was about 100 percent over century-long spans like these. Only 13 percent of any change in income, then, can be explained by foreign trade, statically speaking, under full employment: 100 x 0.13 = 13. Apparently we have another popular cause that doesn’t work very well, quantitatively speaking.

One might be tempted to see growth of sheer output sent abroad as an engine of growth. As has long been realized, however, to do so assumes that massive portions of the economy were idle (in contrast to the full-employment assumption that I just made tacitly). And no historical evidence has been marshaled to make plausible an assumption of massive unemployment — no evidence, for example, that real wages were unresponsive to changes in the relative scarcity of labor. The economist Theodore Schultz decades ago confronted the assumption of idle hands in India (“underemployment, surplus labor”) by noting that in the 1919 influenza epidemic there, which killed an appalling 5 percent of the population, agricultural output did not stay constant — as it should have if the marginal productivity of additional labor in the countryside were in fact zero.³³ If surplus labor does not apply to India in 1919, then surely not in Britain in 1719.

The so-called “vent-for-surplus” model boldly supposes that any sales abroad puts formerly idle, zero-product people to work. (But why don’t sales at home have the same “job-creating” effect? In which case, why would foreign trade matter?) Exports to French colonies in the eighteenth century, for example, are said to have put to work previous idle French workers. (I repeat: why did not domestic demand for carriages and servants have the same effect?) But in the 1780s the share of colonial exports in French manufacturing was only 2.5 percent.³⁴ And as Prados de la Escosura argued for the parallel case of the Spanish Empire, the loss of even that enormous empire resulted in little if any loss to the metropolis.³⁵ Again: trade doesn’t seem to work.

Trade, then, cannot be an engine of growth — not in the simple way envisioned by non-economists, at any rate, and anyway not on the scale necessary to explain much of the 1500 percent growth per capita in Britain from 1700 to 2000. The deepest economic reasoning is that the borders of countries cannot be important, or at any rate not important enough to make flows of exchange over one of them into an engine producing results on the scale of modern economic growth. Trade, after all, is trade, and it shouldn’t much matter whether you trade with someone down the street or with someone on the other side of the world. There’s nothing magic about goods crossing borders, as the Swedish economist Bertil Ohlin noted long ago. (Swedish and Canadian economists, used to living beside the great bears of the German Empire and the United States, tend to get this economic point right.) Your trade with the rest of the world is much of your consumption, but that is so merely because you are little relative to the wide world. Big countries like India or the United States tend to have lower shares of exports in national product than do little countries like Taiwan or the Netherlands. Thus among 20 major economies in 1992 a population 1 percent larger was associated with a ratio of exports to national product 1 percent lower.³⁶ Unsurprisingly.

If a border was closed and is now opened there is a gain, the modest Harbergerian one of increased specialization. The most extreme cases in modern times are the substantial gains of income arising from the opening of Japan in the 1860s or the opening of Eastern Europe in the 1990s.³⁷ But the sheer tearing down of borders does not have the power to enrich us gigantically, and for example did not do so even for Japan and Eastern Europe — as by contrast Mokyr’s “macro inventions” in the making of cloth and surgeries and computers certainly do. Even the violent separation of East and West Germany left “only” a factor of, say, two or three on unification. Not sixteen.

If borders were such an engine, the economist points out, then one could draw an international border in England from Dover to Wroxeter, calling “foreign” all trade across the Watling Street border thus created, into and out of the ancient Danelaw, and thereby make trade within England into an engine of growth. Or you could call left-handed English people “foreigners,” and achieve the same result. The accounting reductio shows that there cannot be something special about foreign trade. If a demand by consumers that relocates production from one side of the English Channel to the other, or from one side of Watling Street to the other, or from left-handed people to right-handed, is enriching on anything but a modest Harbergerian scale, then one has an economic perpetual motion machine, by the mere words of the accounting. Words aren’t that powerful.

And historically, yet again, the problem is that if such a machine worked for Britain in the eighteenth century, why didn’t it work elsewhere and in earlier times? That is the central historical reason that something peculiar to the eighteenth century must explain the peculiarity of the eighteenth century and its denouement. Trade is ancient, as old at least as language. When people start talking in the full way we now call language, around it seems 50,000 B.C.E., they start trading, and we find evidence of the trade in their graves and trash dumps. Even stone for tools, as I have noted, came to be traded over hundreds of miles. Much later in the Bronze Age great trading empires with enriched metropolises were commonplace, and the tin to alloy with copper was shipped by Phoenicians to the Mediterranean from far-away Cornwall. “The light-hearted [Greek] master of the waves/ [sailed] to where the Atlantic raves/ outside the Western Straits,/. . . . and on the beach undid his corded bales.” Big cities and big trade have characterized many places from Mexico City to Hangchow. The Indian Ocean was a trading lake for a millennium before the Europeans got to it. The Northern Italian cities were traders, certainly, and they had even the European cultural traits that some historians believe made European success so inevitable from the Middle Ages on. But why didn’t the Florentines create an industrial revolution? “They did,” one might reply. No they didn’t, not on the scale of the Industrial Revolution. The same objection can be raised to modern growth theory among economists, which in parlor-trick fashion inserts economies of scale into the story just when it is needed to reproduce in the mathematics the rumblings of productivity in the eighteenth century and the innovation gone mad of the late nineteenth.

29. [back] Mokyr 1985b, pp. 22 23 and works cited there.
30. [back] Hume, "Of Commerce," in Essays, p. 284.***check volume.
31. [back] Fogel 1964.
32. [back] Deane and Cole 1962; Mitchell and Deane 1962, p. 330. Imlah 1958.
33. [back] Schultz 1964, p. 70; though see Sen's (not wholly persuasive) strictures in Sen 1967, and Dandekar 1966.
34. [back] O'Rourke, Prados de la Escosura, and Daudin 2008, p. 11.
35. [back] Prados de la Escosura 1993.
36. [back] Inferred from Foreman-Peck 2003, p. 375, who gives Maddison's figures. The scatter is a rectangular hyperbola, that is, a (negative) unit-elasticity curve.
37. [back] On the opening of Japan see Bernhofen and Brown 2009 and works cited there.