Not Even Coal
Yet four impressive scholars recently have insisted on coal: Anthony Wrigley (1962, 1988), Kenneth Pomeranz (2000), Robert Allen (2006), and John Harris (1998). The historical demographer Wrigley has long claimed that the substitution of mineral fuel for wood and animal power made the Industrial Revolution. In one sense Wrigley is obviously correct, since wood could not have fueled the steam engines and blast furnaces of England—though observe that the United States used wood to power steamboats on the Mississippi and charcoal to refine iron in Pennsylvania well into the nineteenth century. But coal deposits do in fact correlate with early industrialization. The coal-bearing swath of Europe from Midlothian to the Ruhr started early on industrial growth. As Jones observes, however, a capability of exploiting an endowment may matter more. English coal was important from an early date in heating London’s homes, blackening the Black Country, eventually running Manchester’s steam engines for cotton mills—though Manchester, New Hampshire’s cotton mills kept using falling water. It is hard to imagine big electricity generating stations running on logs. Eventually hydro-electric and especially atomic power do something to replace coal, and we all hope that wind and solar and geothermal power will prevail. But coal still matters a lot.
Yet the sheer availability of coal does not seem, at least on static grounds, to be important enough for the factor of sixteen, or even a doubling 1780-1860. Economically speaking a coal theory, or any other one-step geographical theory, has an appointment with Harberger. The share in national income of land was much higher in the eighteenth century than now (a third then as against 2 or 3 percent now), but the share of coal land within all land was small. The calculations would be worth doing, but they probably would turn out like the others. Gregory Clark and David Jacks have recently argued that substitutes for coal meant that an upper bound on the loss from a coal-less Britain would have been a mere 2% of national income—when what is to be explained is a 100% increase down to the mid-nineteenth century and much larger increases afterwards.28
Especially, of course coal, could be moved, and was—it went to Amsterdam and London, like iron and lumber from Sweden, or French salt, or Irish cattle. The presence of coal somewhere reachable at low cost may have been important for the steam stage of industrialization, say 1800-1950. And before the railway a transport route by sea would have been very important. But the point is that the coal didn’t need to be on the spot. As Goldstone notes, if the coal fields had been located in Normandy, then London and the Cornish mines would have imported their coal from France, and we would have no sage talk about the necessity of British coal inside the legal confines of Britain. Yet Normandy would not necessarily have industrialized, if lacking the requisite dignity and liberty of the bourgeoisie (whose standing there may be inferred from Madame Bovary). The place where steam engines were most used was Cornwall, with no coal. Norrland in Sweden exported lumber and paper pulp, but did not make the house frames or the paper.
The coal advocates are right, however, to emphasize that any argument about industrialization needs to be made comparatively. The Chinese in the seventeenth century had long been using coal on a big scale to get the high temperatures to fire ceramics, exporting the result westward.29 Kenneth Pomeranz argues for the importance of the accident that in Europe, especially in Britain, cheap coal existed close to populations. China’s coal was far away from the Yangzi Valley—the Valley being until the nineteenth century a place which was in other ways, he argues, comparable to Britain in wealth, at the high end of the $3 ± $2 a day of our ancestors. The Valley was where the demanders of coal and in particular the skilled craftsmen were. China used coal (and natural gas, of all things) early, but its coal was inland, with no cheap water routes like London’s “sea coal” from Newcastle, used in English lime kilns and glassmaking from the thirteenth century and increasingly for house fuel (the local price of firewood had sharply increased) around 1600.
Yet one might object that a more vigorous proto-innovation (“vigorously exploiting its endowment”) would have moved the industry to, say, Manchuria (not entirely unnaturally, perhaps, under the rule of Manchus after 1644), or at any rate to some other coal-bearing lands of the gradually widening Central Kingdom, exporting the finished products instead of the raw coal. After all, eventually China did just that, as on a smaller scale the British did in the (newly) industrial northwest and northeast, or the Germans in Silesia, or on a larger scale the Europeans did in exporting finished products to the world. You do not have to move coal, even before the railway made moving it cheap. You can move people and move finished goods.
Coal as merely a new source of heating, in short, does not work very well for explaining our riches. Robert Allen, who would disagree, has noted that coal was anyway relatively cheap in England. By the end the eighteenth century, certainly in London, and even the once-poor north, English people enjoyed higher real wages than most of the Continent, except the Netherlands: “Craftsmen in London or Amsterdam earned six times what was required to purchase the subsistence basket [of goods], while their counterparts in Germany or Italy only 50% more than that standard.”30 His argument is that cheap coal relative to scarce labor led to innovation. That is, he attributes the scale of British innovation to the pattern of factor scarcities. Labor was scarce relative to coal fuel in Britain, and so innovations would be labor-savings. And so Britain would have a large volume of innovations.
Neither “and so” makes much economic sense. The economic historian H. J. Habakkuk in 1962 made tentatively the same argument about the United States in the nineteenth century: labor was scarce relative to capital, and so America innovated by saving labor. Allen himself accurately summarizes one crushing point against such an argument, following critics such as Peter Temin and other economic historians reacting to Habakkuk: “one problem is that businesses are only concerned about costs in toto—and not about labor costs or energy costs in particular—so all cost reductions are equally welcome.”31 Well put. As another leading student of technology, Tunzelmann, remarks, “In truth, it is extremely difficult to make a logical theoretical argument for the seemingly self-evident proposition that scarce labor should induce labor-saving bias in technology.”32 A shilling got from saving not labor but coal—which was in fact the obsession of early users of steam engines, as Margaret Jacob has shown from their writings—is the same shilling that one got from saving labor (which Jacob notes was seldom mentioned in the writings of the engineers she has examined).33 Later, in the nineteenth century, as Allen and I discovered some time ago, the British iron- and steel-making made advances mainly by saving coal, as for example Neilson’s recycling of hot gases from the blast furnace to cut coke usage by two-thirds, or the hard driving later in the century with similar results.34 By that time Britain had even higher wages, and the real price of coal had not much changed. What happened, one may ask, to the alleged labor-saving bias between the late eighteenth and the late nineteenth centuries?
If wages relative to coal prices were all that mattered, Jacob has also noted, Belgium and the extreme south of the high-wage Netherlands, both of which had coal, and in any case could import it very cheaply from Northumberland across the North Sea, would have been the Birminghams and Manchesters of the late eighteenth century. And to look at the point from the opposite side, why did not industry on the low-wage parts of the Continent away from the Netherlands therefore explode with coal-saving innovations? You can see the underlying illogic: something is always relatively scarce, “and so” innovation in saving the scarce input will be high. And so every age and place has an incentive to innovate in great volume. Something is wrong in the logic.
Cheap coal can indeed explain the location of power-hungry industries in Lancashire vs. Wiltshire, or Birmingham vs. Bordeaux (though, by the way, Allen does not sufficiently acknowledge the importance of water power). If one is willing to glide by the point that a shilling is a shilling, as Allen does so glide, after tipping his hat to the critics of Habbakuk, then the high ratio of wages to coal might be supposed, illogically, to affect the composition of innovations. But the matter to be explained in the Industrial Revolution is not the composition of innovation, but its magnitude. Patrick O’Brien and Calgar Keyder recognized the point long ago, arguing that France took “another path” than Britain did to the twentieth century. One could ask therefore why in eighteenth-century Italy or indeed China there was not a labor-using path to the modern world. That British innovations were biased (as the economists put it) towards labor saving, if they were (though in iron making, as I said, they definitely were not), says nothing at all about how many innovations in total the British would make. If spaghetti is cheap relative to rice in Italy compared with Japan you can expect Italians to eat relatively more spaghetti than rice. But such an expectation does not say anything about how much food in total the two countries will consume, one sort of food aggregated with another. In explaining modern innovation the aggregate is what matters, not the pattern.
It is easy to get confused about the economics here. China did use labor-intensive methods of all kinds. But doing so is merely using old technology (not innovating new technology) in a way determined by the abundance of labor relative to, say, land. In such matters Allen properly affirms that relative prices matter. Yet using people to hoe the fields by hand instead of using capital-intensive methods such as great iron plows is not an advance of the sort that made us rich compared to our great-great-great-great-great grandparents. It is not an “advance” at all, in fact, but a choice of different routines from existing plans of business, different paths on the same map. Allen cites Rainer Fremdling, who has persuasively shown that the non-use of coke for iron on the Continent before the 1850s—it had been in use in Britain for a century by then—was not an entreprenurial failure (as Landes for example had argued) but a matter of relative prices.35 Peter Temin had argued earlier, likewise, that the use of charcoal for blast furnaces in the U.S. at the time was another case in point: wood for charcoal was cheap relative to coal there.36 And I had done the same sort of research on British iron makers about a claimed “failure” to use now Continental techniques of by-product coking later in the century, or a “failure” to have in other ways the same pattern of use of technology as the Americans or Germans (David Landes again made the claim I was criticizing; Landes does tend to leap to scolding for sloth and incompetence whomever was not using whatever he asserts without quantitative inquiry was the best technique; it is a corollary of his race-to-the-swiftest, élan-vital theory of world history).37
Splendid though such quantitative researches in historical economics are, however, they are not the same as explaining the innovativeness of British vs. Continental economies in the eighteenth and early nineteenth centuries, or the innovativeness of Europe generally 1700 to 1900. To explain the size as against the composition of innovativeness you need factors like a lead in the practical side of the Enlightenment (Jacob, Mokyr, Israel) or in entrepreneurial élan vital (Landes: though note how poorly the hypothesis does in the late nineteenth century) or—to come to the One True Explanation—in the extent to which a rhetoric of dignified and liberated business had been adopted (McCloskey). One needs, to put it again in economic jargon, an explanation of absolute, not comparative, advantage. Relative prices of the sort economists usually concern themselves with, in other words, have a highly doubtful connection with the amount of innovativeness in total. As Allen argues, the scale of Britain’s mining of coal and lead explains “why steam engine research was carried out in England.”38 That sounds reasonable. Margaret Jacob for example would probably agree. For the same reasons, as Alan Olmstead and Paul Rhode have recently argued, biological innovation in crops and livestock took place in the United States during the nineteenth century—this against still another version of the scarce-labor hypothesis (which claims that mechanization was the key to American agriculture).39 But economies of scale to innovation in a leading industry is not a theory of the amount of innovation of all sorts, in banking and insurance and cotton and wool and glassmaking and printing. The total amount of innovation is what is to be explained.
The historian John Harris argued for coal in a way that makes more sense than the static arguments favored by the economists. He wrote that “the move [in Britain in the seventeenth century and before] to general use of a cheaper mineral fuel. . . . . nearly always necessitated important technical change in order to accommodate the use of the equipment of the relevant industry,” such as glass-making or salt-making. “The long success with this change of fuel . . . over a couple of centuries was a major reason for a willingness to try new methods in other industrial fields and to be prised away from traditional practices.”40 Yes: the accident of easy coal and expensive forests could lead to a tinkering mentality (say) about applications of heat. (Though again the Chinese were in such matters many centuries ahead.) But in this case the Coal Effect works through habits of the mind, not through relative prices directly, as the economist would wish. I stand instead with the admirable Tocqueville: “Looking at the turn given to the human spirit in England by political life; seeing the Englishman. . . inspired by the sense that he can do anything. . . I am in no hurry to inquire whether nature has scooped out ports for him, or given him coal or iron.”41
How far have we gotten?
The claim is that the economist’s static model does not explain the factor of sixteen. The static model and its quasi-dynamic extensions can tell what did not cause the Industrial Revolution and its sequel, correctives to popular fable and sharpeners of serious hypotheses. But the kind of growth contemplated in the classical models, embedded nowadays deep within modern economics as a system of thought, was not the kind of growth that overtook Britain and in the late eighteenth century and then was gloriously continued in the nineteenth century and then in the wide world.
One might reply that many small effects, static and dynamic, could add up to the doubling of income per head to be explained: trade, coal, education, canals, peace, investment, reallocation. The late Charles Feinstein suggested this to me at a conference bringing the “new” economic history to Britain in the 1980s. I honor the liberal impulse to avoid unicausal explanations. On the other hand, the purpose of a science is to uncover causes, and if one cause such as gravity explains most of the phenomenon, such as a falling stone, then there can’t be a reasonable complaint that “unicausal explanations are always wrong in (physics or) history.” Sometimes they are right, or right enough for scientific purposes.
And another trouble—the historical trouble emphasized before—is that many of the suggested effects, whether in the first or the second century of modern economic growth, were available for the taking in earlier centuries. The mystery inside the enigma of modern economic growth is why it is so very modern. If canals, say, are to explain some major part of the growth of income it must be explained why a technology available since the beginnings of settled society, and used with increasing sophistication in many of them from the third millennium B.C.E. on, was suddenly so very useful as to cause an epochal rise in productivity. The Chinese invented the pound lock in 984 C.E. (it got to Europe in 1373) and in 1327 C.E. completed the Grand Canal of 1100 miles (the pride of French rationalistic engineering, the Canal du Midi, completed in 1681 C.E. from the Atlantic to the Mediterranean, was a mere 149 miles). China had elaborate systems of lockless transport canals many centuries earlier, as of course did ancient Mesopotamia and the Indus Valley civilization.42 The Iranians dug long tunnels through mountains to water their plains. The Romans led water for scores of miles on arches and through tunnels. What is so special about the Bridgewater Canal (1776) bringing coal to Manchester?
In any case, adding up the material causes proposed for the Industrial Revolution doesn’t seem to work, either. One trouble is that adding up a dozen effects shown to be individually on the order of 1 or 2 percent still does not come close to the 100 percent in the first century of the Industrial Revolution. But the deeper trouble is that the doubling is not enough, since in short order the result of modern economic growth was not a factor of 2 or even 3 but a factor of sixteen—not 100 percent but 1,500 percent—and greatly larger if the better quality of goods and services like lighting and health care and education could be properly accounted for.
The classical model from Smith to Mill was one of reaching existing standards of efficiency and equipment. The model looked plausible until the late nineteenth century. To attach it to a place: the model was one of reaching Holland’s riches in 1700. And indeed as late as 1870 the Western European countries had merely done that, so far as average income per head was concerned. (They had prepared the technical and organization grounds for a growth gigantically beyond old Holland, but that is another and later matter). No wonder the classical economists imagined limits close to what they could see plainly in Holland, and had no idea that the $5.40 a day (in 1990 prices) that the average Western European earned in 1870—about what the average Dutch person had earned 170 years earlier—was to increase to an astounding $50 a day by the end of the twentieth century. According to Maddison’s figures, per capita income in the Netherlands was $2110 per capita in 1700 (expressed in 1990 dollars), about what in 1870 had been achieved in most western European countries—for example, France at $1876 and a collection of the twelve richest European countries at $2086.43
Holland was to the eighteenth century and even the early nineteenth century what America was to the twentieth, a standard for the wealth of nations. “The province of Holland,” wrote Adam Smith in 1776, speaking in precise terms about the west of the United Netherlands, “in proportion to the extent of its territory and the number of its people, is a richer country than England. The government there borrows at two percent., and private people of good credit at three. The wages of labor are said to be higher in Holland than in England, and the Dutch. . . trade upon lower profit than any people in Europe.”44 Smith’s emphasis on profit at the margin is characteristic of the classical school. The classical economists thought of economic growth as a set of investments which would, of course, decline in profit as the limit was reached. (The anxieties of stagnationism in the 1940s among economists such as Keynes and Alvin Hansen, as I’ve noted, were similar. They reckoned that opportunities had been exhausted, and that after the War the Great Depression would return. On the political left, Baran and Sweezy [1966] kept up the stagnationist argument for some decades after its time.)
Smith spoke a few pages later of “a country which had acquired that full complement of riches which the nature of its soil and climate, and its situation with respect to other countries allowed it to acquire.”45 He opined that China “neglects or despises foreign commerce” and “the owners of large capitals [there] enjoy a good deal of security, [but] the poor or the owners of small capitals . . . are liable, under the pretense of justice, to be pillaged and plundered at any time by the inferior mandarins.”46 In consequence the rate of interest in China, he claimed, is 12 rather than 2 percent. Not all the undertakings profitable in a better ordered country are in fact undertaken, says Smith, which explains why China is poor. Smith and his followers sought to explain why China and Russia were poorer than Britain and Holland, not why Britain and Holland were to become in the century or two after Smith so very much more rich (Smith, incidentally, was off in his facts about China here, as most Europeans were: not all of China was in fact poor). The revolution of spinning machines and locomotive machines and sewing machines and reaping machines and insurance companies and commodity exchanges and universities that was about to overtake north west Europe was not what Smith had in mind. He had in mind that every country, backward China and Russia, say, and the Highlands of his native Scotland, might soon achieve what the thrifty and orderly Dutch had achieved. He did not have in mind the factor of sixteen that was about to occur even in the places in 1776 with a “full complement of riches.”
In the event a vastly fuller complement of riches came from innovation in machines, both physical and social. Smith, of course, did mention innovation, in his discussion of the division of labor: “Men are much more likely to discover easier and readier methods of attaining any object, when the whole attention of their minds is directed towards the single object.”47 And he was eloquent on the need for sound institutions, such as public schools and sensible commercial policy. What is striking in his and subsequent discussions, however, is how much weight was placed on mere reallocations. But the reallocations, the reshufflings, the moving even of coal—mere efficiencies—we have found, were too small to explain what is to be explained.
-
[back] Clark and Jacks 2007.
[back] Goldstone 2009, p. 13.
[back] Allen 2006, p. 6.
[back] Allen 2006, p. 10. See Temin 1966, 1971.
[back] Tunzelmann 2003, p. 87.
[back] Jacob, personal correspondence, 2008.
[back] McCloskey 1973; Allen 1977.
[back] Fremdling 2000, referred to in Allen 2006, p. 18.
[back] Temin 1964.
[back] McCloskey 1973.
[back] Allen 2006, p. 27.
[back] Olmstead and Rhode 2008a, 2008b.
[back] Harris , p. 133.
[back] Tocqueville, 1835, p. 116.
[back] Temple 2007, pp. 218, 197.
[back] Maddison 2001, Appendix B, Table 21, p. 264.
[back] Smith 1776, I, ix, 10, p. 108.
[back] Smith 1776, I, ix, 14, p. 111.
[back] Smith 1776, I, ix, 15, p. 112; compare I, viii, 24, p. 89.
[back] Smith 1776, I, I, 8, p. 20.
