The History of Economic History: Part I
The Cliometric Revolution
In the beginning, there was nothing. Then, there was candlelight.
Economic history was born alongside academic economics in the late nineteenth century, a small part of a broader process: the formation of the modern university. Harvard’s first chair in economic history was William J. Ashley, a British scholar who received the post in 1892, a few years before his colleagues formed the school’s economics department. In England, the LSE became the first British university to offer an economics degree, and economic history was one of the offered specializations. Early tensions emerged between classical theorists and German-trained historical economists, who eschewed deduction for induction. Following Gustav Schmoller in Germany, John H. Clapham in England, and Richard T. Ely in the US, the early economic historians felt that cultural and temporal circumstances prevented the generalizations required to construct theory. The first articles were cross-disciplinary: those appearing in the Journal of Political Economy (JPE) or the Quarterly Journal of Economics (QJE) were methodologically indistinguishable from those appearing in the American Historical Review (AHR). But the avoidance of general theory began to alienate economic historians from their economist allies, symbolized by the absence of historical work from the early issues of the American Economic Review (AER) after 1911.
Convergence forces emerged during the 1920s, with the formation of institutions like the National Bureau of Economic Research (NBER), which—under the guidance of Wesley C. Mitchell—sought to build long-run historical datasets for policy advising. At the prodding of Clapham, who decried past scholars’ tendency to extrapolate from faulty data, economic historians were increasingly pressed to construct and incorporate longer, more rigorous time-series during the interwar period. For the first time, research began to focus primarily on “such questions as How much? How many? How quickly? Or How representative?” Economic history began to re-embrace formal theory, and while historians were prevalent in the formation of the Economic History Association (EHA) in 1941, the charge was led by economists like Earl Hamilton. In the first issue of the Journal of Economic History, the EHA’s publication, Simon Kuznets, an NBER pioneer and former Mitchell student, called for collaboration between economic historians and “statistical economists” to achieve “the final goal of economic study.” Kuznets’ students, including Robert Gallman and Richard Easterlin, would help to transform the discipline by their meticulous collection of national income and capital stock data. Quantitative studies began to take on old historical dogmas; C. M. Thompson, for example, argued that the South was not over-specialized in cotton because the region produced sufficient foodstuffs, while Carter Goodrich and Sol Davison attacked the Turner thesis by demonstrating the high costs of westward migration and the small numbers of actual migrants from the industrial sector.
The new stock of quantitative knowledge and the drive for theory proved transformative for economic history. In 1957, a Conference on Research in Income and Wealth held jointly by NBER and the EHA produced a famous volume containing a range of fundamental statistical series for the nineteenth-century United States. These included Gallman's estimates of commodity output, Towne and Rasmussen's farm gross product and investment series, Richard Easterlin's regional income estimates, Stanley Lebergott's wage series, Ethel Hoover's price index, Edward Budd's factor shares, and estimates of the balance of payments by a young Douglass C. North. Formalization, however, was immediately controversial. Claudia Goldin (1995) suggests that the main point of contention was the “huge fossilized stock of accepted wisdom concerning major projects, figures, and events of the past” which was absent in the other fields of empirical economics. “If one generation claimed labor supply functions were backward bending, but the next estimated a positive supply elasticity using new data and methods,” she writes, “their results could be reconciled by merely supposing that in the meanwhile the underlying parameters had changed.” But economic history’s knowledge base was both old and cumulative, and the findings derived from exploration of the new data frequently threatened the entire structure.
In 1958, John Meyer and Alfred Conrad published a paper that symbolized the consensus-breaking potential of leveraging theory and data in historical study. “The Economics of Slavery in the Antebellum South” had initially been presented at the joint NBER-EHA meeting of the previous year; in it, Conrad and Meyer challenged the prevailing view that American slavery was an economically moribund institution that was fading out prior to the Civil War. They estimated production functions for male and female field hands and, using historical price series, found that the rate of return on the former ranged from 4.5 to 6.5 percent in typical cases to over 10 percent on the best Mississippi alluvial lands. Women, meanwhile, earned 7 to 8 percent. Since comparable investments in other sectors of the economy might return 6 to 8 percent or less, they concluded that slavery was at least as profitable, and certainly not on the verge of dying out. Conrad and Meyer’s scientific project precipitated a series of studies testing the reproducibility of their findings, using different assumptions, data, and production specifications. The fact of their vindication is less relevant than the sheer novelty of their enterprise and the reaction inspired.
An air of excitement was sweeping the discipline. In 1963, Douglass North, recently appointed co-editor of the JEH, declared that “a revolution is taking place in economic history in the United States.” “Even a cursory examination of accepted ‘truths’ of U.S. economic history,” he proclaimed, “suggests that many of them are inconsistent with elementary economic analysis and have never been subjected to—and would not survive—testing with statistical data.” There was a sense that historical study had suddenly become amenable to great change—a desiccated set of long-cherished theories and assertions promised to collapse in spectacular fashion under the weight of formal analysis. “In those early years,” recalled Jonathan Hughes, “it seemed like you could hardly miss. Pick any topic in Economic History. Did it make sense as theory? If not, why not? Were there data available? If so, BINGO.”
The next of the series of studies challenging attacking the consensus on American economic history was produced by Robert W. Fogel, a former student of Kuznets. Born in 1926 to Ukrainian Jewish immigrants, Fogel had attended Cornell University as a history major (though minoring in economics), and on his graduation in 1948 worked for nearly a decade as an organizer for the Communist Party before rejecting the ideology as “unscientific.” Turning to economics, he received his MA from Columbia in 1960 (having studied under George Stigler) and his PhD from Johns Hopkins in 1963, where he worked with Kuznets. In the following year, he published his first major work: Railroads and American Economic Growth (1964), derived from his doctoral dissertation. Provoked both by the long-held view of the railroad as the “engine” of U.S. economic growth and by a burgeoning literature on the centrality of efficient transportation to development, Fogel sought to calculate the “social savings”—the addition to the social surplus from falling transport costs—of American infrastructure. His innovation here was the use of the counterfactual method of reasoning. By stressing the importance of the railroads to growth, old-school historians were hypothesizing that America would have grown sluggishly in their absence; thus Fogel’s book was “an extended thought experiment of what the U.S. economy would have looked like if the railroads had never been built.”
His most dramatic finding, the much-cited “interregional” social savings calculation, was that the abrupt disappearance of the large trunk lines linking the food-exporting Midwest with the importing Northeast would have decreased national product by less than 0.6 percent. Existing water routes would have carried the four main crops—wheat, corn, pork, and beef—with similar efficiency. The real controversy, however, was over “intraregional” social savings—what would have happened if the railroads connecting farmers with Midwest markets had vanished. Fogel made four sets of calculations, each based on different assumptions about the alternative mode of transport used—existing land and water routes or a hypothetical array of replacement canals—and the area cultivated (use all existing land, or only that within the feasible margin). The number was still small, around 1 percent of 1890 GNP, meaning that the contribution of railroads to the shipment of agricultural products—a quarter of all goods moved by rail—was less than 2 percent of national income. While this was a substantial figure to allocate to any single mechanical device, it was gallingly small by comparison with the expectations of the “big push” and “engine of growth” theorists.
Together with North, whose work on ocean shipping helped to demolish the view that technical change contributed most of the productivity gains of 1600-1860, Fogel inaugurated the “Cliometric Revolution” in economic history. This rapidly became a clash between the economic theorists and the “traditional” historians, who objected that historical models were too complex to be captured by mathematical relationships. The “young turks” held meetings at Purdue University and called their field the “new economic history,” a coinage of North’s. Quantification came slowly—labor economics was undergoing a similar transition at the time, which by comparison occurred practically “overnight”—but it came nevertheless. Fogel and North came to stand for two different perspectives on the possibilities of the new science; Fogel was the consummate empiricist, dogmatically proving facts over and over, while North was the “grand theorist,” developing narratives about the role of institutions in the broad sweep of growth. North’s ocean shipping study, which highlighted the reduction in piracy as a motive force behind efficient changes in ship design, presaged this methodological divergence.
The great arena of the Revolution was the slavery debate, which had been simmering unquietly since Conrad and Meyer had gone to press. North's The Economic Growth of the United States, 1790–1860 (1961), Fogel and Stanley Engerman's Time on the Cross (1974), and Fogel's later recapitulation Without Consent or Contract (1989) were the defining works. Both North and Fogel sought to explain the regional differences in American income that persisted after the Civil War, with the South lagging the US average by more than 40 percent as late as 1920—a collapse that neither could attribute to the consequences of the conflict itself. North saw the involvement of the slaveholding South in a complex network of export trade as the villain; the substantial receipts from sales of cotton to Britain and the Northeast paid for purchases of Midwestern foodstuffs, Northern industrial goods, and European luxuries. The inequality of Southern society prevented these profits from being invested in capital accumulation and internal improvements, as in the North, and channeled them into elite consumption instead. The existence of the slave-master institution (presaging North’s later work) prevented efficiency-improving trades between the two parties, even if those transactions would have raised elite incomes in the long run.
Fogel famously took an entirely different tack. Time on the Cross was designed as a systematic refutation of the “traditional view” that Conrad and Meyer had once assailed: that slaves were inept and incompetent, plantation agriculture was inefficient, and that planters themselves were irrational for perpetuating the system. On the contrary: slavery was profitable, the antebellum South was prosperous, and Southern slaves and farms were more productive than their white and Northern counterparts. He observed that in 1860, hypothetically reducing enslaved persons to a minimum subsistence income raised white consumption above Midwest and on par with Northern levels. Together with Engerman, Fogel sought to demonstrate how control of this captive labor force undergirded the South’s economic power during the antebellum period. They asserted that planters were rational profit-maximizers who treated their human property relatively well—positive rather than negative incentives, keeping families together, providing adequate nourishment—and organized their workforces to exploit economies of scale. Plantations with 15 slaves were shown to have produced 40 percent more per unit input than those with fewer. Gang labor, Fogel surmised, was combined with efficient management practice, including the seasonal employment of slaves (especially the corn and cotton growing periods) and the strategic deployment of the elderly, infirm, and women on associated tasks such as tending livestock and buildings and producing manufactured goods. North’s “triangular trade” was a mirage—the South was actually self-sufficient in foodstuffs. Urban and industrial growth had been suppressed not by the moribund extractive institution of the planter elite, but by the South’s exploitation of a slave-based comparative advantage in cotton agriculture.
All of the book’s claims received intense scrutiny and criticism, both from Fogel’s fellow economic historians and from the standard historical profession. The gang labor hypothesis was assailed on multiple fronts, first for ignoring the fact that large plantations only appeared to be more productive because they grew proportionately more cotton than small farmers, and more recently for failing to recognize that picking rates were invariant to farm employment over time. Others attacked the “benign working conditions” claim, suggesting that Fogel and Engerman had misread source material on physical punishment and ignored nutritional and anthropometric evidence on well-being. But this is beside the point. Time on the Cross had, for the first time, placed neoclassical economic analysis at the forefront of the historical debate on a controversial question. Price signals, market processes, and above all rational profit- and utility-maximization were at the core of the Fogel-Engerman mechanism, and these assumptions and emphases proved nearly as contentious as the econometrics with which they handled data. Arguments about the book were really concerned with whether neoclassical models relying heavily upon these features were an appropriate tool for historical analysis; ultimately, they marked a fork in the road along which economists and their historian counterparts had hitherto traveled together. While the cliometricians challenged Fogel and Engerman on their own terms—Gavin Wright, for example, developing the “safety first” model to explain why smallholders grew less cotton and more corn—the historians increasingly ignored the Time on the Cross debate, and economic history, altogether. As mainstream economics offered up models with information costs and bounded rationality with which to expand upon (and supersede) the neoclassical method, economic historians became ever more deeply wedded to the program that Fogel had pioneered. Historians, feeling alienated, began to turn away.
Having fully re-embraced economics, economic history began to branch outward beyond the critique-based bounds of the Cliometric Revolution. Paul David, inspired by the retention of the ineffective QWERTY keyboard, developed the concept of technological path dependence, and, in partnership with Gavin Wright, sought to demonstrate that American resource abundance was a product of institutional and cultural factors, not merely endowments and incentives. Wright foregrounded the role of racism in preserving an economically sub-optimal apartheid in the South, while Joel Mokyr began to explore the role of the Enlightenment in kickstarting the Industrial Revolution. By the 1990s, however, there was a sense that the Cliometric Revolution had “eaten its own.” Christina Romer, in a 1993 talk titled “The End of Economic History?,” blamed the cliometricians for making economic history too familiar to the mainstream economist. “[A]s economic history began to use the same tools as other specialties,” Goldin observed, “many believed the economic history requirement for the doctorate could be abolished without consequence.” That the end of economic history did not arrive was in large part thanks to the efforts of North, who had been working on a grand theory of historical change whose assimilation would mark a turning point in the discipline’s history. Explaining this theory is deserving of a subsequent post, however—so for the moment, we stop here.
Eagerly waiting to your subsequent parts!
"Paul David, inspired by the retention of the ineffective QWERTY keyboard"
But the QWERTY keyboard is not "ineffective." It looks ridiculous, and wonderful arguments can be made about how other designs (I'm lookin' at you, Dvorak) would be so much better. But they actually aren't much better (the studies claiming they are, which David cites at second hand, have never been found). QWERTY became a standard because it consistently beat other keyboards in speed competitions when typewriters were new. Now, even though you can get a Dvorak keyboard for $22 from Amazon, just about no one switches over because the gain in speed or comfort just isn't enough to be worth switching.
(I got much of the above from THE FABLE OF THE KEYS by S. J. LIEBOWITZ and STEPHEN E. MARGOLIS, published in the Journal of Law and Economics back in 1990. A quick search will find non-gated copies of the article.)