Precocious Albion: A New Interpretation of the British Industrial Revolution.

Prepared for Annual Reviews of Economics, July 2013

Morgan Kelly, Department of Economics University College, Dublin, email: Morgan Kelly

Joel Mokyr Departments of Economics and History Northwestern University, Evanston, IL 60208 email: j-mokyr@northwestern.edu

Cormac Ó Gráda Department of Economics University College Dublin email: Cormac O Grada

mailto:j-mokyr@northwestern.edu
1

Introduction

Why was Britain the cradle of the Industrial Revolution? The literature on the topic is quite substantial, and very little of a consensus has been reached since the survey in Mokyr (1999). The dominant schools are divided between those who focus on geographic endowments (such as coal), those who focus on politics and institutions (including the Glorious Revolution), and those who stress Empire and Britain’s colonial successes.

In what follows we present an argument that focuses on the quality of the British labor force. While in the past claims for human capital as an explanation of Britain’s leadership have been dismissed because of its mediocre schooling and literacy rates (Mitch, 1998), we argue that this focuses on the wrong variables. Instead, we highlight two very different dimensions of human capital. One is the physical condition of the average British worker. We will argue that better nutrition made British males grew up on average to be healthier and taller than their continental counterparts. Health and height meant both more physical strength and in all likelihood higher cognitive ability, and hence higher labor productivity. As nutrition was costly, better health can be seen as investment by parents in their children’s human capital. The other is that the distribution of ability and dexterity in Britain was more skewed, so that there was a much larger density in the right tail, that is, a relatively large contingent of highly-skilled and capable technical workers. That contingent may have contained a higher endowment of skills, through a more flexible and effective system of training young men in the apprenticeship system, but what counted above all was its highly skilled mechanics and engineers, who may not have been a large proportion of the labor force.

The net result is that on the eve of the Industrial Revolution, while Britain may have had more accessible coal and a larger overseas empire, the main reason for its precocity was its higher level of human capital. We do not dismiss other explanations, such as institutions. The contrast between institutions and human capital suggested by Glaeser et al. (2004) only exists if we employ a very narrow definition of institutions. If we include a wider definition of both formal and informal institutions, as well as distributional arrangements such as the Poor Law, we will show how in fact part of the comparatively high level of human capital was rooted in Britain’s institutions.

Induced Innovation and the Industrial Revolution

An obvious implication of a higher productivity of British workers is that the observed wage level in Britain was higher than elsewhere. Robert Allen (2009a, 2009b, 2010), ), the leading collector and organizer of the wage data on which this observation is based, has suggested that the higher wages themselves may have been instrumental in bringing about the Industrial Revolution.

2

Allen (2009, pp. 52-56, 143) recognizes the difference in human capital levels between B ritain and the rest of1

the Continent, but does not explore the implications for his interpretation.

O ne implication of Acemoglu’s model is that, because of the growth in the supply of unskilled labor in B ritain2

(migrants, as well as women and children), technological progress might have been what he calls “skill-replacing.” B ut

while the machines may have in some cases (especially in textiles) have done that, it also increased the demand for very

highly specialized skilled labor that could build, install, and maintain the new equipment. An increase in the supply of

skills, he shows, can under certain conditions actually increase its price (that is, the skill premium). Elsewhere,

Allen has resuscitated the idea of induced innovation and re-introduced it into the literature of the origins of the Industrial Revolution. The argument builds upon the literature that flourished in economic history in the 1960s on the effects of high wages on labor-saving innovation. He submits that the British Industrial Revolution was driven by a set of labor-saving and coal-using innovations, stimulated by the high cost ratio of labor to energy in Britain, relative to France (which is taken to be representative of the rest of Europe). The high level of British wages is attributed by Allen primarily to labor demand: the growing commercial and maritime sector and the growth of urban centers raised real wages in Britain, as it did elsewhere. This argument has obviously resonated with other scholars. For instance, Jean-Laurent Rosenthal and Bin Wong (2011) adopt it wholesale in their account of the difference between nineteenth-century technological progress in China and Britain.

The induced innovation argument has a venerable pedigree in economic history as an explanation of why technological progress differed across economies. In the early stages of this literature, it was applied to explain the difference between American and British technology, with Britain cast in the role of the low wage economy (Rothbarth, 1946; Habakkuk,1962; Temin 1966, 1971). Paul David (1975) attempted to resolve the issue of substitution vs. technological progress and proposed a model that resembles Allen’s framework. He assumes that innovation was mostly “local” (that is, the product of learning by doing), and that this learning was faster in the more mechanized techniques. If that was the case, the choice among existing techniques (substitution) will have driven high-wage economies to choose labor-saving techniques, and these willhave generated further innovation in the neighborhood of high capital/labor ratios, leading to falling costs in those techniques. Eventually the unit costs of the mechanized techniques became so low that even the relatively low wage economies will have mechanized, and thus the British Industrial Revolution spread to the Continent. In Allen’s view, this model describes, roughly speaking, the history of the Industrial Revolution in Europe, although he is willing to leave some room to exogenous factors such as the Enlightenment and the Scientific Revolution.1

In more recent years , Daron Acemoglu (2001, 2002, and 2010) has shed further light on the economics of induced innovation (Acemoglu, 2001, 2002, 2010). His research has recast the literature in terms consistent with endogenous growth, by postulating that innovation is brought about by profit-maximizing individuals or firms, who then become monopolistic sellers of the new technique or good. These models provide a more rigorous foundation for the induced innovation literature, but in the final analysis this work has not helped cast much light on the role of high wages in the British Industrial Revolution. All the same, in some historical situations, it has been shown2

3

Acemoglu (2010) specifically mentions the H abakkuk and Allen work as examples of labor saving technological change

that might have been induced by labor scarcity. He concludes (2010, p. 1071) that “whether labor scarcity and high wages

may induce innovation and technology adoption in practice is an open empirical question and is likely to depend on the

specific application.” Furthermore, in Acemoglu’s model biased technological progress is triggered, if at all, by a rise

in wages (not a high level). T here is little evidence that wages actually rose sharply before the Industrial Revolution, and

after 1750 the growth in labor supply due to the acceleration in population growth makes this quite unlikely.

that when there is a strong unanticipated supply shock to factor prices, induced innovation can help to bring about adjustment. Good examples of such a phenomenon are the contributions of Walker Hanlon (2013) on the British response to the Cotton Famine and that of Richard Hornbeck and Suresh Naidu (2012) on technological response to the Mississippi Floods of 1927.

As an explanation for the British Industrial Revolution, an argument based on high wage levels relative to other economies needs to make strong assumptions, , the most basic one being that higher British wages entailed higher unit labor costs for British employers. As we will show below, it is far from clear that this assumption was met during the British Industrial Revolution. This is hardly a new idea. It was noted in a passage by Arthur Young commenting on the low cost of French labor: “labour is generally in reality the cheapest where it is nominally the dearest. The quality of the work, the skill and dexterity of performance come largely into account” (Young, 1790, p. 311). In 1824 Thomas Malthus made the same point: “Generally, my opinion is, that the efficiency of labour in France is less than in England, and that that is one of the principle causes why the money price of labour is lower in France than in England” (Great Britain, 1824, p. 600).

A number of other possible objections to the argument have been raised. For one thing, Allen focuses on process innovation (much of it focused on the textile industry), in which unit costs were reduced through mechanization. While this is an apt description of some of the innovations we associate with the Industrial Revolution, we must keep in mind that new techniques were emerging along a broad front of production, and that many are hard to classify as either labor- or capital saving, as they involve entirely new or vastly improved products or services, from canned food to marine chronometers to vaccination. It also must be shown that the London male wages relied upon by Allen are representative of what textile mill owners expected to pay their workers (something strongly objected to by Jane Humphries, 2013). It should also be stressed that Allen focuses primarily on the British cotton industry, by all accounts one of the most dynamic industries of the Industrial Revolution and often associated with it. Yet Mokyr and Ralf Meisenzahl (2012) demonstrate that in many respects the inventive processes in the cotton industry were atypical, and that in most other industries, such as engineering, the incentives and backgrounds of inventors were quite different. Hence cotton was in some sense an outlier in the Industrial Revolution. Equally worrisome for his argument is that he must show that the new capital-intensive techniques were profitable for Britain but that for a long period they could not be used in France because of its cheaper labor. At least one investigation has shown that even for the cotton industry, the story may be problematic. Ugo Gragnolati, Daniele Moschella, and Emmanuele Pugliese (2011) demonstrate that on Allens’s own numbers, the jenny, while more profitable in Britain than in France, would under reasonable

4

Even in textiles, anecdotal evidence that high wages and cheap coal were the prime mover in mechanization3

is mixed. Arthur Young reported in 1807 from W itney (Oxfordshire) that it was a low-wage area, suffering from “the

want of vicinity to coal” — yet it had introduced spinning jennies and “spring looms” (flying shuttles). T he labor-saving

innovations did not help raise local wages and most of the local poor were denied a share in the increasing prosperity

(Young, 1807, p. 326).

assumption also been profitable in France from the onset (for a similar view, see Charles Foster and Eric L. Jones, 2013, pp. 103-04). 3

Moreover, Allen focuses on the high cost of labor relative to the cost of energy. This is perfectly reasonable, given that in many of the British industrializing areas coal was abundant and cheap. It should be kept in mind, however, that steam power, the paradigmatic technology in which fossil energy supposedly replaced labor, was often used to replace horses or watermills. This indicates that the Industrial Revolution, rather than simply substituting resources for labor, replaced one form of resources by another. It is telling that in Cornwall, where coal was expensive, its high cost did not slow down technological progress, but simply re-oriented it into another direction. Indeed, the high cost of coal has been cited as the stimulus for the development of fuel-saving technology in Cornwall (Nuvolari and Verspagen, 2009, pp. 686-87). The success of Cornish engineers such as Arthur Woolf in developing fuel-saving engines wherever coal was expensive suggests that what was driving technological progress was something deeper and stronger than cheap coal and high wages, although the latter were affecting the direction into which innovation moved. Coal was important, but it was itself subject to technological progress, and its cost and availability were clearly endogenous to deeper forces. As E.L. Jones (2012, p. 7) remarks, “industry was growing in the North before any significant generation of power using coal, while trades vital for inventiveness — notably clock and watchmakers in South Lancashire — used little fuel.”

Finally, it may be added that the evidence for technological progress during the Industrial Revolution being on the whole labor-saving, as the induced innovation hypothesis would contend, is mixed at best. The macroeconomic record, questionable as the data are, is summarized by von Tunzelmann (1994, pp. 289-91). Apart from a short period during the Napoleonic Wars, he found little evidence that technological change in Britain as a whole was on balance labor-saving before 1830. Even after that year, in his view, when there was a clear-cut shift toward more labor-saving machinery, it was dampened by “the continuing labour-surplus of males” (ibid., p. 291). The microeconomic evidence from the patent records, assembled by Christine MacLeod (1988), is equally troubling for the labor-saving inventions hypothesis. She calculates that labor saving was a stated goal of patentees in only 4.2 percent of all patents taken out between 1660 and 1800, whereas capital saving was the goal in 30.8 percent of all patents. Looking at what patents actually achieved, only 21 percent of all inventions can be said to have saved labor, compared to the 30.8 percent that were said to save on capital (MacLeod, 1988, pp. 160-71).

In what follows, we will show that there is good reason to believe that there were far-reaching differences in the quality of labor between Britain and France on the eve of the Industrial Revolution that all point in the direction of British workers being more productive than their French colleagues. High wages were little more than a symptom of much deeper differences between Britain and the

5

rest of Europe. The very factors that made Britain’s workers more productive may well have also been important in generating the inventions, and (equally important), in disseminating and absorbing new knowledge and putting it to good use. We will show a number of things. One is that the differences in productivity between British and French workers were sufficient to cast doubt on the assumption that unit labor costs in Britain were higher than in France. Another is that this higher quality of labor helps explain the British Industrial Revolution without having to rely on induced innovation. In this case the high wage is not the cause of invention, but a symptom of deeper factors that drive both wages and technological creativity.

A Simple Model of Human Capital and the Industrial Revolution

In what follows, we present the verbal and graphical outline of a simple dynamic model that captures the main features of our view of the Industrial Revolution. The formal model may be found in the Appendix to this paper, retrievable at http://www.ucd.ie/t4cms/WP13_12.pdf. The historical reality the model reflected in the model is that technological ideas were generated by the Industrial Enlightenment, which redirected research efforts toward more pragmatic purposes, and re-organized useful knowledge to make it more accessible (Mokyr, 2009). But turning this into an Industrial Revolution required skilled and capable artisans who could build the new devices from blueprints, install, operate, and maintain them. These abilities took a large amount of training and adeptness, and we will refer to them as competence.

The importance of competence can be incorporated into a standard Phelps-Nelson growth model of human capital, in which productivity evolves as a function of human capital: specifically, there is some “maximum” level of attainable, and in each period the economy gets closer to it as a function of its competence. Competence in the next period itself is a function of the existing stock of competence (reflecting the fact that artisans were trained by other artisans) and the investment their parents make in their training (reflecting the facts that apprentices had to pay a fee to their master and that health depended on food consumption). The growth in productivity A is a function of past productivity and the level of competence in the economy. We then define a variable we term M (for “misery”), defined as the reciprocal of both health and competence.

The model then solves for two log-linear difference equations that follow the trajectories of both A and M over time. This will be recognized as adaptation of a linearized Lotka-Volterra dynamic system of two competing species (Hofbauer and Sigmund, 1998, pp. 22-28). The two state variables M and A can be seen as two competing “species” in a finite environment. The growth of each species is retarded by the presence of the other. Four possible equilibria are possible. One is that conditions are so favorable to the first “species”, misery in our case, that its “population” will be high regardless of the second species, which it always drives to its minimum level. This corresponds to the “Malthusian” (dismal) equilibrium, with population at the minimum of subsistence (maximum M), and A at its very minimum. The converse holds when A drives out M meaning that M is minimized (high levels of health and human capital). Two other equilibria are possible: if the species have little impact on each other both co-exist at positive levels; If they have a strong effect

6

on one another, it becomes indeterminate which of the two drives out the other, the outcome depends on initial conditions: whichever species initially has a sufficiently large population to dominate the system will drive the other out.

The model then posits two economies that we shall call France and Britain. Each faces the same best-practice technological frontier A that rises through time, reflecting the progress of

~

Enlightenment scientific knowledge in Europe, which easily crossed national boundaries. France and England differ in only one way: initially disposable income is higher in Britain than in France. This historical development is described in fig. 1. Our starting point, in panel (a), is a stark Malthusian

1world with little knowledge: log A is arbitrarily small and the knowledge isocline A lies completely ~

below the two misery isoclines pertaining to both countries. As a result, both economies are at an equilibrium at the lower bound of knowledge. As time passes, best-practice technology A will rise

~

exogenously, reflecting the progress of Enlightenment useful knowledge, and this will be the driving force behind the model. At some point, the A-isocline rises enough to intersect with the British misery-isocline but is still below the French one. This is depicted in panel (b). What happens then depends on the parameters: under weak interaction Britain may start moving to an interior solution at the intersection of the two; under strong interaction, the higher level of A only starts kicking in

3when the A isocline is entirely above it, as A in panel (c). At that time Britain jumps to a new at which it misery index is at minimum and its productivity is high, while the

~ 3equilibrium at log A

French remain in a Malthusian equilibrium until the A isocline has advanced sufficiently to make the transition possible there as well (panel 4). Because English human capability is initially slightly higher than French, England can start to apply technological knowledge to production earlier, giving rise to a cumulative process of rising living standards, rising human capital, and improving production technology. A gradual rise in knowledge above a critical level causes England to experience an industrial revolution, while France for a while appears mired in age-old backwardness.

Note that we need not assume that Britain is originally richer than France: the French M- isocline would lie above the British one if Britain’s elasticity of converting knowledge into productivity was higher, or its ability to teach its apprentices was higher because the institutions governing apprenticeship worked better, or the elasticity of output w.r.t. human capital were higher, or some combination of those.

7

T his estimate is consistent with Arthur Y oung’s eighteenth century observations. D ividing the median costs4

of reaping an acre of wheat (60d) by the median harvest wage (20d-22d per day) on both Young’s southern and northern

circuits yields a rate just under three man days per acre (Y oung 1771, IV , 293-296; 1772). Y oung wrote that “Strength

depends on nourishment; and if this difference be admitted, an English workman ought to be able to do half as much

work again as a Frenchman” (Y oung, 1793, II, 315-316).

Fig. 1. Impact of rising technological frontier A on equilibrium in two economies with different ~

misery isoclines.

Wages and Productivity in Eighteenth-century Britain and France

There is no dispute on the main fact underlying this debate: English wages were considerably higher than French ones on the eve of the Industrial Revolution. Allen (2009b, ch. 2) calculates that the real wages of building craftsmen in London in 1780 were 83 per cent higher than those in Paris, while those of laborers were 80 per cent higher. It is, however, invalid to conclude that English labor was therefore expensive, until we compare productivity and can thus infer unit labor costs. Some indication of the differences in productivity can be attained from data in agriculture by comparing day rates and piece rates. The average time needed to reap an acre of wheat in early nineteenth century England was 2.9 man-days per acre, or 7.2 days per hectare (Clark, 1991, 449). This4

compares with France where the average cost ranged from 9.3 to 16.3 man-days per hectare, giving an average, weighted by regional output share, of 12.9 man-days per hectare (Grantham, 1991, 362). Reaping and threshing were manual activities with almost no capital input and fairly little skill. Even

8

allowing for considerable measurement error, the roughly 65-75 per cent productivity advantage for English workers suggests a real difference in the physical quality of labor.

England France

Real Wages, 1780

Craftsmen 1.82 1.00

Laborers 1.80 1.00

Agricultural Output per man-day

Reaping 0.14 0.08

Threshing 1.54 0.93

Table 1:

Sources: Clark (1991, 449); (Grantham, 1991, 363).

The exact reason why British workers were more productive that French ones is yet to be resolved. It might be pointed out, however, that if income per capita affected labor productivity (instead of just the other way around), we are in an efficiency -wage world, in which employers will find it in their interest to pay workers more than the lowest wage possible, because by paying a higher wage they increase their productivity. They will continue to do so until the increase in labor productivity is equal to the higher wage. A standard issue here is that coordination failure between employers may undermine this equilibrium. An employer may want to pay his workers a higher wage to elicit more work out of them. If this efficiency wage engenders personal loyalty to the employer and thus reduces the effects of asymmetrical information, this could work fine. However, if it works through a mechanism of improved worker strength and energy due to better nutrition or creates an inter-generational externality by improve the quality of workers available to the next generation of firms, other employers might free-ride on the higher wage, and a coordination failure would result in a low-wage equilibrium. Arguably the British Poor Law could be seen as a attempt to prevent local free-riding on improved worker quality.

Our argument, then, is that British workers were of higher quality than French ones. This would not only explain their higher wages, but also provide a critical link that explains why British workers were able to take advantage of the technological opportunities emerging in the eighteenth century. This is not a traditional human capital argument: as is well-known, in this period Britain led Europe neither in the quality and quantity of its educational system nor in observed literacy rates. Instead of human capital in its conventional, narrow sense of rates of literacy and schooling, we want to focus on the wider concept of what Heckman (2007) has termed human capability. Human capability is the triad of cognitive skills, non-cognitive skills (for example self-control, perseverance, time preference, risk aversion, preference for leisure), and health. These components of human

9

capability in turn are strongly determined by the individual’s nutritional and disease environment from conception to adolescence. Nutrition, Health and Physical Capability

Perhaps the most obvious source of difference between British workers and French workers is that the former seem to have been fed better. There is a fair amount of anecdotal evidence that suggests that for the bulk of the population, Britons were better fed than Frenchmen. That said, estimates of the exact gap differ quite a bit. Robert Fogel (2004) famously argued that English workers, by being better fed than their French counterparts, were capable of more work, and estimated that the median French worker consumed about 2200 kcal per day, considerably less than a median English diet of about 2600 kcal. The amount of energy available for work (after the needs of basal metabolic demand) per capita in his estimate was about one-third higher in England than in France: 600 kcals in France in 1785, against 812 kcals in England in 1750 and 858 kcals in 1800 (pp 9-11). The more recent calculations in Roderick Floud et al. (2011, pp. 99) are similar and put the English mean around 1800 at 2,456 kcal as opposed to the French mean of 1,847 (computed from id., pp. 114–15); while Stephen Broadberry et al. (2011) estimate 2,100 kcal, more or less unchanged between the mid-thirteenth and mid-nineteenth centuries. At the other extreme, Craig Muldrew (2011, p. 156) estimates average calories per capita in England at 5,047 in 1770, falling to 3,977 in 1800, although these estimates rely on implausibly high output of coarse grains. In the middle is Allen (2005) with an estimate of 3,800 kcal in 1750 falling to 2,900 in 1800.

The net effect of higher infant mortality on adult productivity depends, to some extent, on why infant mortality was so much higher in the first place. France’s high infant mortality reflected to some extent its lower standard of living and inferior diet probably had something to do with it. Mary Matossian (1984) has linked higher French death rates to greater consumption of the wrong kind of food, whereas Fogel (2004) and Bernard Harris et al. (2010) emphasize the link between inadequate food consumption — hunger — and “premature death.”

In terms of the quality of the diet, the data suggest a much higher percentage of animal protein consumption in Britain than on the Continent. Muldrew’s claim (2011, p. 153) that British adult laborers consumed about 0.7 lbs of meat a day around 1770 may err on the high side, but Floud et al.’s estimates (2011, p. 210) of a per capita consumption of 4.9 oz in 1750 and 4.4 oz in 1800 still compare favorably with Jean-Claude Toutain’s estimates of France’s per capita meat consumption of 0.1 – 0.13 lbs per capita on the eve of the French Revolution. In Germany, too, meat consumption was low (see Blum, 1974, pp. 413–15). Anecdotal evidence is abundant: many travelers visiting Britain commented on British carnivorous habits. Thus B.L. de Muralt, a Swiss traveler, in 1726 noted that “the pleasures of the table in this happy nation” contained much roast beef which is a favorite dish as well at the King’s table as at Tradesman”(Muralt, 1726, pp. 39-40). In 1748 Per Kalm, a Swede, similarly remarked that he does not believe that any Englishman “who is his own master, has ever eaten a meal without meat” (Kalm, 1892, p. 15).

10

A case could me made for assuming a bigger coefficient of variation. Villermé (18//) reports that in 1817 the5

average height of all those measured was 161.5 cm with 28 per cent below 157cm. T hat is consistent with a coefficient

of variation of 4.75 per cent.

Figure 2. English and …

Read more
Applied Sciences
Architecture and Design
Biology
Business & Finance
Chemistry
Computer Science
Geography
Geology
Education
Engineering
English
Environmental science
Spanish
Government
History
Human Resource Management
Information Systems
Law
Literature
Mathematics
Nursing
Physics
Political Science
Psychology
Reading
Science
Social Science
Home
Homework Answers
Blog
Archive
Tags
Reviews
Contact
twitterfacebook
Copyright © 2022 SweetStudy.com