There was an Industrial Revolution, and it was slow.
The debate’s been over for two decades. The gradualists, armed with new growth accounting techniques, have won, exorcising forever the Ashtonian vision of an eighteenth-century leap into exponential growth. Where historians once thought the spinning jenny, steam engine, and factory system the immediate preconditions of England’s dramatic transformation, they now accept that these advances emerged in a trickle in a few tiny sectors. Nicholas Crafts, a primary exponent of this “limited-gradualist” view, opens a recent survey on productivity growth during the Industrial Revolution on a triumphalist note:
The British industrial revolution was a landmark event in world economic history. While it was once seen as a period of dramatic ‘take-off’, it is now generally accepted that it was an episode of gradual acceleration in labour productivity growth which eventually led to sustained increases in living standards (Crafts 2021, p. 1).
Crafts suggests that the research frontier on the 1750-1850 expansion has shifted. Economic historians should now be asking the following questions: “why did productivity growth accelerate?”, “how soon did workers benefit from faster productivity growth?”, and “why did productivity growth experience a major slowdown as soon as the 1870s?” Having exorcised the ghost of Rostow and established the empirical facts of Britain’s output trajectory, causality and distribution necessarily become the chief problems for the field. We can move on.
The macroeconomic figures are largely settled, for the moment; a substantial revision by Broadberry et al (2013) points to faster growth, but not much. That said, Crafts still tries to justify the low numbers—re-revised substantially since his 1985 “revision”—in light of qualitatively obvious increase in productivity-enhancing inventive activity. But what are the consequences for the history of English economic development in the eighteenth century if the decisive changes came after the middle of the nineteenth? Indeed, given that the latest statistics show sluggish growth until 1850 and stagnation after 1870, we appear to have shortened the Industrial Revolution out of existence.
Yet Britain did eventually start climbing onto the “inverted hockey curve” of rising living standards. Today, Manchester and Birmingham are modern cities with cars, helicopters, and health care—a far cry from the sooty slums of two centuries ago, or the nascent pre-industrial towns of a century earlier. If the transformation didn’t happen between 1750 and 1850, and if growth was stumbling after 1870, when did Britain become modern?
These smaller questions fold into a bigger one: what is the right meta-narrative1 for the Industrial Revolution? Tragedy, triumph, nascent triumph, or nothing at all?
In his 1982 essay “The Industrial Revolution: A Misnomer,” Rondo Cameron famously denied that the event had happened at all. “Probably no term from the economic historians’ lexicon,” he wrote, “has been more widely accepted by the public at large than ‘industrial revolution.’ That is unfortunate, because the term itself has no scientific standing and conveys a grossly misleading impression of the nature of economic change” (Cameron 1982, p. 377). He concludes that the process of expansion allotted to the time frame was “a long, drawn-out process, in no sense inevitable, which scarcely deserves the epithet ‘revolutionary’”—a discontinuity neither in time nor in space.
Cameron was reacting against two strands of research: the qualitative narratives of T.S. Ashton2 and David Landes3 and the foundational statistical work of Walther Hoffman, Phyllis Deane, and W. A. Cole. Hoffman, in his 1955 British Industry, 1700-1950, proposed an industrial output series that showed 1770—the year that Hargreaves patented the spinning jenny and the one after Arkwright invented the water frame—was the point at which growth accelerated. Deane and Cole, working from eighteenth and nineteenth-century data, laid down the statistical foundations of the classic view by showing that both industrial output and per capita income began to rise quickly from the end of the eighteenth century.4
The Deane and Cole vision rankled with historians for two reasons. First, it cut against the “gradualist” tradition of Max Hartwell, who had pushed back against the excessive enthusiasm of the “catastrophic” view (positive and negative). Explosive change never sits well with neoclassical economics, in any event, which struggles to this day to incorporate dramatic growth with natural equilibrating tendencies. Second, the statistical representation of discontinuity and acceleration resembled the vision of “take-off” presented by Walt Rostow in his 1960 book The Stages of Economic Growth: A non-Communist Manifesto, which presented a theoretical formula according to which nations—Britain as the exemplar—modernized through industrialization. Rostow’s intellectual work was unsurprisingly unpopular in the postmodern academy, and was further tarnished by the author’s controversial service as National Security Advisor during the Johnson administration’s descent into Vietnam.
This hostility ensured that a new “revisionist” agenda launched in the early 1980s found a receptive audience. In 1982, C. Knick Harley published “British Industrialization Before 1841: Evidence of Slower Growth During the Industrial Revolution” in the Journal of Economic History. As the title suggests, Harley produced evidence—a new index of industrial production—that the rate of expansion during the 1760-1830 period had been far more gradual than Deane and Cole had suggested. Growth was 1/3 slower, and thus incomes during the mid-eighteenth century had been two times higher. Why had Hoffman, Deane, and Cole overshot? Harley argued that they had overemphasized the importance of the dynamic sectors: Hoffman inflated cotton, while Deane and Cole overweighted overseas trade. England’s advanced industries were a modern island in a backward sea:
E. J. Hobsbawm has recently written “whoever says Industrial Revolution says cotton.” One might want to add iron and steam. But these three industries were only a minority-less than a quarter of British manufacturing even in the 1840s-and while their technological transformations have received due attention, they must be kept in proper perspective in a study of British industrialization in the aggregate. The older textile industries of wool, linen, and silk and the agricultural processing industries of milling and baking, brewing and distilling, and leather processing certainly generated more income than the technological leaders. The diverse, dispersed, and unspectacular industries must all be given proper weight, even though the evidence is scant and fragmentary, if British industrialization is to be understood (Harley 1982, p. 268).
The rate of growth of Britain’s industrial output index was largely contingent on the weights allotted to cotton and iron; the former two sectors grew by 2200 and 350 percent respectively during the period 1770 to 1815, whereas most others were in the range 40 to 200 percent. Hoffman had further erred in assigning sectors with no data the weighted mean of growth rates in other industries; this had the effect of doubling the implied weights of cotton and iron from 6 percent each to 12 percent. Deane and Cole, meanwhile, had attributed export booms resulting from military conflict and colonial population growth to domestic technical change and lower prices, which led them to overrate the speed of change after 1770.
Harley’s paper marked a turning point. His figures implied that industrial production grew at a rate of 1.5 to 1.7 percent from 1770 to 1815 and 2.9 to 3.3 percent from 1815 to 1841. Since manufacturing was England’s most dynamic occupational category, national income would have had to move even more slowly—1.31 and 2.23 percent (and a measly .33 and .86 per capita) per year respectively. Positive and accelerating, no doubt, but a far cry from the third-world industrial revolutions taking place around him in the 1980s. But his pessimism was backed up by the 1985 publication of Nicholas Crafts’s British Economic Growth During the Industrial Revolution, which offered a slightly faster rate of 2.15 percent for 1770-1815. Crafts’s “optimism” was short-lived; yet another revision by Crafts, Leybourne, and Mills (CLM) in 1989 dropped the growth rate below Harley’s estimate, to just 1.4 percent per annum. Table 1 (Jackson 1992) compares the first and second generations of estimates:
The dour outlook of the “Crafts-Harley” view aroused a virulent reaction, particularly among the historically-minded economic historians. The first came from Maxine Berg and Pat Hudson, who—in the tellingly entitled “Rehabilitating the Industrial Revolution” (1992)—scorned the econometric advances of the previous decade. Growth accounting came in for particular vitriol, for statistically smothering qualitative discontinuities in the eighteenth-century British economy. Since the aggregated figures moved slowly; since the factories, engines, and machines predominated in isolated sectors, the “Industrial Revolution” had been reduced to an impressionistic illusion—a useless, misleading name for one part of a smooth upward curve.
In the last decade the gradualist perspective has appeared to triumph. In economic history it has done so largely because of a preoccupation with growth accounting at the expense of more broadly based conceptualizations of economic change. New statistics have been produced which illustrate the slow growth of industrial output and gross domestic product. Productivity grew slowly; fixed capital proportions, savings, and investment changed only gradually; workers' living standards and their personal consumption remained largely unaffected before 1830 and were certainly not squeezed. The macroeconomic indicators of industrial and social transformation were not present and so the notion of industrial revolution has been dethroned almost entirely leaving instead only a long process of structural change in employment from agrarian to non-agrarian occupations. (Berg and Hudson 1992, pp. 24-5).
Berg and Hudson attacked on two fronts. First, they contested the accuracy of the Crafts-Harley calculations as being fraught with estimation errors and unfounded assumptions about the rate of change in the “uncounted” sectors, particularly in services, which Crafts had limited to a pace no faster than industry. His conservative view ignored wholesale changes in transportation, finance, and the legal professions obvious to any observer at the time. Other omitted sectors included glass, metalworking, distilling, pottery, and chemicals—all of which saw genuinely revolutionary advances during the period. Changes in the composition and quality of inputs and outputs, they claimed, would introduce and leave “major sources of innovation… neglected.” Furthermore, low entry costs may have meant that much of the innovation was being done by inchoate firms with few or no written records. Together, these biases would not “cancel out,” as growth accounters assumed, but lead to the systematic underestimation of the rate of change. In a final and devastating aside, they add:
Crafts's recent statistical analysis of industrial output series for Britain, Italy, Hungary, Germany, France, Russia, and Austria shows that Britain and Hungary were the only countries to exhibit a prolonged period of increase of trend rate of growth in industrial production during the process of industrialization. In the light of the qualitative evidence of the extent and speed of change in Germany and Russia in particular, this finding suggests either that the macro estimates are far from accurate and/or that paying undue attention to changes in the trend rates of growth at the national level is not a helpful starting point for identifying or understanding economic transformation (Berg and Hudson 1992, p. 30).
Second, they propose an alternative conceptualization of the Revolution itself, based on the increased use of female and child labor, growing regional specialization, and rapid population growth and structural transformation. Change in the economy was broad-based and swift; the “analytical distinction” between ancient and modern sectors was unfounded and unhelpful in accounting for the kind of movements that did occur. Pointing to paper, wool, and chemicals, Berg and Hudson argue that technical change was possible in slow-growing sectors, while some sectors with traditional methods—coal-mining and construction—expanded quickly. Assuming that high-investment, TFP5-focused innovation should be the “sole yardstick of industrial innovation” would obviously lead one to see a couple of advanced sectors pulling along a stagnant remainder.
Peter Temin offered a theoretical critique of the Crafts-Harley position five years later, in his excellent “Two Views of the British Industrial Revolution.” Using a Ricardian trade model, he showed that Britain’s success at exporting a wide range of goods was inconsistent with the proposition that technical change was limited to cotton and iron. He also produced compelling trade data in support, showing that Britain exported hundreds of thousands of pounds’ worth of everything from pottery and silverware to soap and stationary. If these industries had stagnated, comparative advantage dictated that Britain should have exported even more of its key goods in order to import their products from abroad.
When the smoke from these volleys had cleared, however, Crafts and Harley remained in control of the field. A second paper summarizing the “Crafts-Harley” view appeared in the fall of 1992 with even lower estimates of industrial growth (Harley’s figures falling to 1.5 and 3.0 percent for 1770-1815 and 1815-1841 respectively). Charles Feinstein’s wage estimates confirmed the per capita output data; real wages increased only modestly until the mid-1840s. Antras and Voth, meanwhile, marshalled new figures to confirm the slow-growth picture for 1770-1860. Capping off this string of successes, Crafts and Harley took on Temin using a Competitive General Equilibrium (CGE) trade model that demonstrated that a broad range of exports was consistent with an economy with diminishing returns to agriculture and a swiftly growing population (both of which clearly characterized Britain at the time). Thus Joel Mokyr, writing in 1999, could write that “for a country that undergoes structural change while it grows, very sudden accelerations in the growth rate of the kind that Rostow envisaged are simply impossible” (Mokyr 1999, p. 12).
So where does this leave us? The Industrial Revolution was not nearly as dramatic as the transformations of developing countries during the latter half of the twentieth century, and was dwarfed—even in Britain—by the postwar growth miracle. Living standards were stagnant until at least after the Battle of Waterloo, and maybe longer.6 Worse, a recent survey by Crafts has shortened the far end of the period, too. TFP change from 1873-1913 was a miserable .10 percent per annum after removing human capital accumulation, as GDP per capita growth fell to just one percent, slower than over 1830-56. And we know well that the First World War, the troubled twenties, and the Depression-wracked thirties all contributed to regression and stagnation.
Was the Industrial Revolution really just 1856-73, the sole period during which per capita GDP grew faster that two percent per year? A 40 percent increase in living standards is nothing to sniff at, but we’ve come to expect that and more from “steady state” growth in the modern world. England’s GDP per capita in 2013 prices was about 1,900 pounds in 1750 and exceeds 30,000 today—an increase of 1500 percent. When did all that happen? Clearly not by 1800, when per capita GDP was 2,300 pounds, nor even by 1900, when that number was nearly 5,000. It was only after 1945 that sustained growth—more than two percent—became a reality.
So the two centuries between 1750 and 1950, the first half of which was the “first” Industrial Revolution and the second half the “second,” really amounted to this: a sluggish acceleration that slowed again after 1870 and which only jolted into life after the war. The real period of economic dynamism appears to have been the postwar boom—a long but temporary episode at that.
Having destroyed the Industrial Revolution as growth miracle, what do we have left? Shall we follow Cameron and discard the concept entirely—as the fictitious and misleading nomenclature for part of a continuous process?
Surprisingly, Crafts and Harley didn’t think so. Here’s their now-canonical summary of “what the Industrial Revolution was” (1992):
We are happy to agree with O'Brien that… “over the period 1750-1850 the growth of the British economy was historically unique and internationally remarkable.” [But] industry overall grew much more slowly than was once thought. Revolutionary changes in industrial technology were not widespread and productivity improvements contributed only modestly to the growth of GDP before the second quarter of the nineteenth century, probably causing a growth of national income of about one-third of one per cent annually. To be sure, industrial change helped to alter social structure, demographic behaviour, and savings habits, all of which may have stimulated growth. Nevertheless, it seems impossible to sustain the view that British growth leapt spectacularly in one generation as a result of innovations in manufacturing (Crafts and Harley 1992, pp. 704-5).
In short, if we look beyond the economic statistics—if we follow Berg and Hudson in discarding output as our yardstick—we can find out how the Revolution was revolutionary. I have three propositions. First, we should remember that this was an “industrial,” not an “economic” revolution. Industry was transformed, slowly but inexorably, from small-scale handicraft (which persisted, no doubt, but residually) to mechanized factory production. And a growing share of the population moved out of agriculture and into that increasingly-recognizable branch of manufacturing—from 33.9 percent in 1759 to 45.6 percent in 1851. The Industrial Revolution was, in this sense, a rhetorical analog of the French Revolution—industry seized control from the economic ancien regime.
Second, the Industrial Revolution had an unmistakeable technological legacy. British technology from the “classic period,” from the power loom to the steam engine, really did become the basis of global production in the following century. That the economic ramifications were slow to arrive should be a lesson in patience for the historian. New technologies start by diffusing slowly. From 1760 to 1800, the contribution of the steam engine was .004 percent per year to capital deepening and .005 percent to TFP growth. Not until after 1850 had the high-pressure engine become widespread and efficient enough to be deployed in factories and on rail engines that these numbers each rose to .2 percent. A century passed between James Watt’s patent—the first revolutionary “general purpose technology”—and its maximum realization in TFP growth. In terms of sheer invention, meanwhile, 1757 really was a trend break in patenting volume, which rose at .54 percent before and 3.63 percent thereafter. 1760-1860 was an “Age of Invention,” and those inventions did prove the basis for sustained growth. Britain simply needed time. The postwar boom was founded on advances in electrification and metal inputs initially made before the assassination of Archduke Ferdinand. And remember the Solow paradox? We may not even yet be reaping the benefits of the “IT revolution” that began forty years ago.7 Britain became modern, and then it got rich.
Finally, we need not discard the immediate welfare consequences of early British industrialization. Soaring population growth in the late eighteenth and early nineteenth centuries threatened the island with a Malthusian demographic catastrophe. Crafts believes that even the slow growth of the eighteenth century may have had significant positive consequences: “if the acceleration of population growth was exogenous rather than a consequence of the industrial revolution, then it would be reasonable to argue that the industrial revolution averted a collapse of living standards. In that case, the working class benefited much sooner than Feinstein allowed” (Crafts 2021, p. 20). Without an Industrial Revolution, Mokyr reasons that GDP per capita in Britain could have been twenty percent lower in 1830 than in 1760. The thin grey line of modern industry did just enough to maintain Britons at a tolerable standard of living, and provided a launchpad for future improvement.
In closing, it’s worth quoting an immortal passage from T. S. Ashton’s The Industrial Revolution, 1760-1830. Writing in 1948, of a global economy at best only half-transformed, he observed that
There are to-day in the plains of India and China men and women, plague-ridden and hungry, living lives little better ... than those of the cattle that toil with them ... Such Asiatic standards, and such unmechanized horrors, are the lot of those who increase their numbers without passing through an Industrial Revolution (Ashton 1948, p. 111).
The Britain of 1860 was a world away from that of 1760—not much more prosperous, but endowed with the technologies, capital, and entrepreneurship unknown a century before.
Borrowed from postmodern philosopher Jean-Francois Lyotard, the term describes an “overarching account or interpretation of events and circumstances that provides a pattern or structure for people’s beliefs and gives meaning to their experiences.”
Ashton’s succinct The Industrial Revolution, 1760-1830 (1948) is a canonical statement of the classic view of the late eighteenth and early nineteenth centuries as both transformation and forward leap—the “wave of gadgets” view, in a sense.
The Unbound Prometheus (1969), which remains an essential survey of Britain’s progress from 1750 onward, following the continuation of the Industrial Revolution into the nineteenth century.
Their combined work, British Economic Growth, 1688-1959 (1962), was recognized to have “brought explicit economic models and statistical techniques to the study of British economic history, methodologically joining the ‘cliometricians’ in the United States.”
Total Factory Productivity (TFP): the ratio of aggregate output to aggregate inputs. Productivity, in short, is the material gain derived from a given set of resources. Most of economic growth has resulted from increases in this measure, also referred to as the “Solow residual.”
We’ve discussed Allen’s 2009 paper “Engels’ Pause” here before; this form of pessimism is ameliorated by the role that capital accumulation played in stimulating later growth.
I’ve argued elsewhere that we may never. But in any event, Solow’s pithy remark is telling: “You can see the computer age everywhere but in the productivity statistics.”