Human capital and modern economic growth
The role of education and human capital in the Industrial Revolution is more ambiguous than much of the New Growth literature would suggest. Britain, the most advanced industrial nation in 1850, was far from being the best educated, the most literate, or in some other way the best-endowed in traditional human capital.
Increases in male literacy in Britain during the Industrial Revolution were in fact comparatively modest and its educational system as a whole lagging behind [Mitch (1998)]. The Lutheran nations of the Continent - Germany and the Scandinavian nations - were far more literate and, in one formulation, “impoverished sophisticates”.[96] Jewish minorities throughout European history were unusually well-endowed in human capital [Botticini and Eckstein (2003)], yet contributed little or nothing to the Industrial Revolution before 1850. Clearly human capital is indispensable as a concept, but we need to be far more specific as to what kind of human capital was produced, for and by whom, what was the source of the demand for it, and how it was distributed over the population. In his recent survey, the social historian Peter Kirby (2003, p. 118) concludes that the idea that nineteenth century education and literacy emerged as a response to a need for a trained labor force is misleading. There was a significant gap between formal ‘education’ and ‘occupational training’, the latter remaining embedded in the workplace in the form of apprenticeships and trainee positions. Before 1870, at least, the rate of return on formal education in his view was so low that its benefits did not outweigh the costs. That is not to say that being literate did not convey advantages in terms of social and occupational mobility [Long (2003)], but many of the skills that we associate with formal schooling could be attained informally.The historical role of human capital in economic growth must then be re-examined with some care.
In terms of the framework delineated here its primary importance was in reducing access costs: literate and educated innovators could and did read articles, books and personal letters from scientist, as well as familiarize themselves with techniques used elsewhere. They could understand mathematical and chemical notation, interpret figures, read blueprints, and follow computations and mechanical arguments. Moreover, by knowing more, the cost of verification fell: some obviously bogus and ineffective pieces of propositional knowledge could be rejected offhand. Secondly, a more literate and better educated labor force is assumed to be more competent, that is, be able to execute instructions contained in more and more complex techniques. Yet because the total set of useful knowledge could be divided up more and more thanks to better access, the actual amount of such knowledge that a single worker had to control may not have increased, it may have just changed, becoming more specialized, a smaller slice of a bigger whole. Human capital may have been more important in learning new instructions than in executing more complex and difficult techniques: as technology changed more rapidly, technical tricks had to be learned and unlearned at more rapid rates.71
Above all, investment in human capital is supposed to have created the conditions for faster innovation. It made for the prepared minds that, as Pasteur famously said, are favored by Fortune. Much technological progress consisted of fumbling and stumbling into some lucky find - but only systematic training allowed inventors to recognize what they found and how to apply it most fruitfully. Yet it is a fair question to ask of all economists who draw links between demographic change and human capital on the one hand and technological progress on the other - whether through the quality-quantity trade-off or otherwise - how many inventors and technically competent people were needed to generate sustained technological progress.
The answer, of course, depends, on what we mean by “competent”. Eighteenth century Britain did have a cadre of highly skilled technicians and mechanics, almost all of whom were trained in the apprenticeship system rather than in formal academies, and these contributed materially to its technological development. The Continent, too, had its share of skilled and well-trained craftsmen, although if we are to judge from the net migration flow of talent, Britain may have had an edge, especially in coal-using industries.[97] But the process of training apprentices did not always correspond to the neoclassical depiction of human capital formation. In addition to imparting skills, it was a selection process in which naturally gifted mechanics taught themselves from whatever source was available as much as they learned from their masters. Such sources multiplied as a direct result of the Industrial Enlightenment. In the eighteenth century the publishing industry supplied a large flow of popular science books, encyclopedias, technical dictionaries and similar “teach-yourself” kind of books.[98] These mechanics and technicians were the ones that made the Industrial Revolution possible. They generated a stream of microinventions that accounted for the actual productivity gains when the great breakthroughs or macroinventions created the opportunities to do so. They were also the people who provided the competence to carry out the new instructions, that is, to build and operate the new devices according to specifications.[99]
How many such people were necessary? Better not teach the peasants how to read, Voltaire reputedly said, for someone has to plow the fields.[100] Technological change in the era of the Industrial Revolution, based on invention, innovation, and implementation, did not necessarily require that the entire labor force, or even most of it (much less the population at large), be highly educated; the effects of education depended on whether the relation between innovation and the growth of competence was strong and positive.
An economy that is growing technologically more sophisticated and more productive may end up using techniques that are more difficult to invent and artefacts that are more complex in design and construction, but may actually be easier to use and run on the shop floor. Production techniques became more modular and standardized, meaning that labor might become more specialized and that each worker had to know less rather than more. If much of the new technology introduced after 1825 was like the selfactor - simpler to use if more complex to build - it may well be that the best models to explain technological progress (in the sense of inventing new techniques rather than implementing existing ones) should focus not on the mean level of human capital (or, as model-builders have it, the level of human capital of a representative agent), but just on the density in the upper tail of the distribution. In other words, what mattered above all was the level of education and sophistication of a small and pivotal elite of engineers, mechanics, and chemists. Dexterous, motivated, well-trained technically, and imaginative, with some understanding of the science involved, these workers turned the ideas of the “Great Men” into a more productive technology. The new technological system depended on the increased skills of low-level technicians, supervisors, foremen, and skilled artisans who introduced and operated new techniques on the shop floor and made the necessary adjustments to specific tasks and usages. What knowledge the firms could not supply from its own workforce, it purchased from the outside in the form of consulting engineers.[101]Technical education for the masses might have been beneficial because among the working classes there might have been “diamonds in the rough”, technically gifted lads who, with the proper training, could become part of the creative elite. The sample of 316 industrialists assembled by Crouzet (1985) - admittedly only the tip of a largely unknown pyramid - contained only 31 persons whose occupations were “unskilled workmen” and only 16 fathers out of 226 “founders of large industrial undertakings” were working class.
The bulk of the labor force consisted of rank-and-file workers whose ex post technical skills may have mattered but little, and thus any model that relates human capital to demographic behavior runs into a serious dilemma. Technological progress and competence had a complex relation with one another because ingenuity and detailed propositional knowledge could be frontloaded in the instructions or artefacts, thus reducing the competence needed to carry out the actual production.[102]It stands to reason that the ratio of competence to knowledge was higher in agriculture than in manufacturing and in services, since a great deal of competence involved uncodified knowledge about very local and time-specific conditions of soil and weather. The share of agriculture in the labor force and total output declined, and this may be one reason why the relative importance of this form of human capital has declined in the twentieth century. It has also been suggested [Harris (1992a)] that the importance of tacit skills was especially prominent in coal-using industries such as glass and iron, which explains Britain’s initial advantage in these industries and the need for Continental Europe to import British skilled workers after 1800 during the years of “catching-up”.
The human capital argument can be tested, at a rudimentary level, by looking at the ratio between skilled and unskilled wages (or “wage premium”). The problem is of course that without estimating a complete model of the market for skills, the historical course of that ratio cannot be assigned to demand or supply factors. If, however, we assume that technology is the prime mover in this market and we keep in mind that the supply of skills will lag considerably behind a rise in wages (since the acquisition of skills takes time), it would stand to reason that if the Industrial Revolution led to a net increase in the demand for skilled labor, we should observe some increase in the skill premium during the Industrial Revolution.
No such change can be observed. Indeed, recent research into the wage premium has established that it changed little between 1450 and 1900, yet it was much lower in Western Europe than in either Southern and Eastern Europe or Asia, indicating perhaps that Europe was more capable of generating the kind of skills and abilities we associate with human capital in an age in which literacy mattered less [Van Zanden (2004)]. It is even more surprising that this skill ratio declined precipitously in the twentieth century [Knowles and Robertson (1951)]. This could be caused by an (otherwise unexplained) increase in supply, but it is at least consistent with a story that stresses the ability of unskilled labor to operate effectively in a sophisticated technological environment.The argument I propose, that technological progress is driven by a relatively small number of pivotal people, is not a call for a return to the long-defunct “heroic inventor” interpretation of the Industrial Revolution. The great British inventors stood on the shoulders of those who provided them with the wherewithal of tools and workmanship. John Wilkinson, it is often remarked, was indispensable for the success of James Watt, because his Bradley works had the skilled workers and equipment to bore the cylinders exactly according to specification. Mechanics and instrument makers such as Jesse Ramsden, Edward Nairn, Joseph Bramah, and Henry Maudslay; clock-makers such as Henry Hindley, Benjamin Huntsman (the inventor of the crucible technique in making high-quality steel), John Whitehurst (a member of the Lunar Society), and John Kay of Warrington (not to be confused with his namesake, the inventor of the flying shuttle, who was trained as a reed- and comb-maker), engineers such as John Smeaton, Richard Roberts, and Marc I. Brunel; ironmasters such as the Darbys, the Crowleys, and the Crawshays; steam engine specialists such as William Murdoch and Richard Trevithick; chemists such as John Roebuck, Alexander Chisholm, and James Keir were as much part of the story as the “textbook superstars” Arkwright, Cort, Crompton, Hargreaves, Cartwright, Trevithick, and Watt.[103] These were obviously men who could squeeze a great deal out of a narrow epistemic base and who could recognize more effective useful knowledge and base better techniques on them. Eventually, however, there was no escaping a more formal and analytical approach, in which a widening reliance on physics and mathematics was inevitable. Oddly enough, this approach originated in France more than in Britain.[104] Over the nineteenth century, the importance of advantages in competence (tacit skills and dexterity) declined, and that of formal codified useful knowledge increased, thus eroding the advantages Britain may have had in its skilled craftsmen that other nations envied and coveted in the years before 1815.
Below the great engineers came a much larger contingent of skilled artisans and mechanics, upon whose dexterity and adroitness the top inventors and thus Britain’s technological success relied. These were the craftsmen, highly skilled clock- and instrument-makers, woodworkers, toymakers, glasscutters, and similar specialists, who could accurately produce the parts, using the correct dimensions and materials, who could read blueprints and compute velocities, understood tolerance, resistance, friction, and the interdependence of mechanical parts. These were the applied chemists who could manipulate laboratory equipment and acids, the doctors whose advice sometimes saved lives even if nobody yet quite understood why, agricultural specialists who experimented with new breeds of animals, fertilizers, drainage systems, and fodder crops. These anonymous but capable workers produced a cumulative torrent of small, incremental, but cumulatively indispensable microinventions, without which Britain would not have become the “workshop of world”. They were artisans, but they were the skilled aristocracy of trained craftsmen, not the average man in his workshop. It is perhaps premature to speak of an “invention industry” by this period, but technical knowledge at a level beyond the reach of the run-of-the-mill artisan became increasingly essential to creating the inventions associated with the Industrial Revolution.
The average “quality” of the majority of the labor force - in terms of their technical training - may thus be less relevant to the development and adoption of the new techniques than is commonly believed. The distribution of knowledge within society was highly skewed, but as long as access costs were sufficiently low, such a skewedness would not impede further technological progress. Rosenberg has pointed out that in Adam Smith’s view, though the modal level of knowledge may be low, the highest levels of scientific attainment were remarkable and the collective intelligence of a civilized society is great and presents unprecedented opportunities for further technological progress [Rosenberg (1965, p. 137)]. Avenerable tradition in economic history, in fact, has argued that technological progress in the first stages of the Industrial Revolution was “deskilling”, requiring workers who were able to carry out repetitive routine actions instead of the skilled labor of skilled craftsmen.[105] The “factory system” required workers to be supervised and assisted by skilled mechanics, and hence the variance of the skill level may have increased even if we cannot be sure what happened to average skills. Much innovation, both historically and in our time, has been deliberately aimed to be competence-reducing, that is made more user-friendly and requiring less skill and experience to use even if it took far more knowledge to design.[106] Human capital was instrumental in creating competence rather than useful knowledge itself, in teaching how to carry out instructions rather than writing them. Yet given that much of what I termed above competence consisted of tacit knowledge and experience, and given that much of the competence could be front-loaded into the equipment by a small number of brilliant designers, the role of the size of the population and its “mean” level of human capital should be questioned. It seems plausible that the degree of networking and the level of access costs within the relatively small community of highly trained engineers and scientists may have been of greater importance.
Furthermore, the term “skill” may be too confining. Human capital was in part produced in schools, but what future workers were taught in schools may have had as much to do withbehavior as with competence. Docility and punctuality were important characteristics that factory owners expected from their workers. “The concept of industrial discipline was new, and called for as much innovation as the technical inventions of the age”, writes Pollard (1968, p. 217). Early factories designed incentives to bring about the discipline, but they also preferred to hire women and children, who were believed to be more docile. Skill may have mattered less than drill. Some of the literature by economists on human capital acquisition may have to be reinterpreted in this fashion.
8.