HI-POD CONFERENCE
Brussels, 10 September, 2012
The views expressed in this paper are those of the author(s) and not those of the funding organisation(s), which take no institutional policy positions.
This conference is funded under the grant: -Historical Patterns of Development and
Underdevelopment, HI-POD”-Collaborative Research Project under the European
Commission’s Seventh Framework Programme: SSH7-CT-2008-225342

Abstract
The widespread view that English Industrial Revolution was driven by labour substituting technical progress caused by high wages suffers from a basic defect: although English wages were high, so too was English productivity. We argue instead that England’s high wages and Industrial Revolution stemmed from a common source: the superior human capability of ordinary English workers, who were taller, better fed, and longer lived than their European contemporaries; possessing the technical skill or competence to implement the technology of the Industrial Revolution. While the expensive
English labour thesis predicts that labour should have flowed into England from France, in fact the flow of workers was in the opposite direction, with skilled artisans moving to France, showing that for the skilled mechanics on which early industrialization relied, England was a low wage economy.
Although central to our understanding of modern economic growth, the concept of human capital is largely absent from existing explanations of the Industrial Revolution. One recent answer to the question of why the Industrial Revolution happened first in England is that high wages there suffers from a basic flaw: while wages in England were high compared with Continental Europe, so too was English productivity. English workers in 1780 earned about 80 per cent more than their
French counterparts but, when it came to manual tasks like reaping and threshing wheat, we show that they were about 80 per cent more productive.
We argue instead, that England’s high wages and Industrial Revolution both flowed from a common source: the superior human capability of English workers. This capability has two dimensions: one is that the average British worker was superior in definable dimensions to others.
The other is that the very best and most skilled British workers were more numerous and better than their continental competitors. The eighteenth century saw an increasing number of scientific and technological ideas appear across Europe with potential economic applications, and the binding
1
1This approach dates back to Rothbarth (1946) and Habakkuk (1962), and has recently been revived by Allen
(2009a) and modelled by Acemoglu (2010). For a more detailed critique of the hypothesis, see Kelly, M okyr and Ó
Gráda 2012.
1

constraint on economies came in applying these ideas in a practical form. England had no particular advantage over continental Europe when it came to generating novel technological ideas: its decisive advantage lay instead in its ability to implement and make practical improvements to these ideas, and this advantage came from its supply of skilled artisans and superior workmen. This supply of skilled labour reflected in turn the superior human capability of English workers — using Heckman’s (2007) term for the triad of cognitive skills, non-cognitive skills and health-resulting from better childhood nutrition and its well developed and flexible system of apprenticeship.
Our view of the Industrial Revolution, in other words, closely resembles that of Nelson and
Phelps (1966) where a country’s ability to adopt a potential technology depends on the competence of workers. We allow this competence in turn to depend on parental investment through a
Ben-Porath (1967) process, and show that this model leads naturally to a sudden take-off from a low technology equilibrium to a high technology one of the sort experienced by England.
What evidence do we have for our claims about the superior human capability of English workers? A growing literature, surveyed below, shows that human cognitive and other capacities are strongly influenced by nutrition and the exposure to infectious illness from conception through adolescence, and particularly in early childhood. Naturally, we have little evidence about the developmental environment or human capability of children during the eighteenth century, but we do have systematic measurements of two variables known to be strong indicators of the conditions of childhood development: height and life expectancy. The differences between France and England in the late eighteenth and early nineteenth centuries that these reveal are startling. As we will show below, British men were substantially taller, had longer life expectancy, and were more skilled and better trained than French workers (which will be our standard of comparison).
Fogel (2004) has emphasized the synergy between income and health: rising living standards improve nutrition, and the taller and stronger workers that result can work harder and generate more output. Here, we extend Fogel’s focus on the physiological impact of nutrition to emphasize the connection between the higher cognitive capacity of better fed workers, and their ability to adopt and improve technology. In addition, however, Britain’s institutions were more conducive to the kind of skill formation that may have mattered most in the Industrial Revolution.
Underlying the superior human capability of English workers in both dimensions was the fact that eighteenth century England was already a fairly prosperous place, able to support two unique social institutions: the Old Poor Law, and its system of apprenticeship. The poor law was a national system of public charity funded from local taxes that went mostly to subsidize the elderly and families with young children as well as the unemployed. This reduced malnutrition and gave landlords, who paid the tax, an incentive to restrict population growth by discouraging large families, for instance through limiting the availability of housing (Solar, 1995). It also aided in some human capital formation through so-called pauper apprenticeships.
While English school enrolment and literacy were mediocre by contemporary standards, most
2

boys above farm labourers were trained as apprentices. In contrast to continental Europe, where occupational guilds enforced apprenticeship rules by regulatory power, English apprenticeship was based on private contracts between parents and masters, making it highly flexible in response to increased employment opportunities in new industrial sectors. In addition, some of the leading sectors of eighteenth century England – such as clock-and instrument-making, shipbuilding, and mining – turned out to be useful sources of the skills needed to implement the technology of early industrialization.
The rest of the paper is as follows. Section 1 outlines a model where ability to adopt technology depends on the human capability of workers, and shows how it leads to a sudden technological take-off. Section 2 shows that while English wages were almost twice French levels, so too was English productivity in manual agricultural tasks. Section 3 outlines the large gap in heights, nutrition, and life expectancy between France and England, while Section 4 looks at the comparative advantage of the two economies revealed by labour flows between them. Evidence for the connection between height, wages, and human capital has hitherto been lacking before the twentieth century, but in Section 5 we show the strong connection between these variables for a new dataset on France during its own period of rapid industrialization in the second quarter of the nineteenth century. Finally, Section 6 outlines the reasons for England’s supply of skilled artisans in its distinctive apprenticeship system, and its national Poor Law.
1
In this section we outline a model of growth where an economy’s ability to implement new technology depends on the human capability of its workers, which is determined in turn by the investment of their parents in feeding and educating them. We do not model here the act of invention itself but focus on diffusion and adoption, because Britain’s precocity was not so much in the act of invention itself as in the rapid utilization and improvement of new techniques wherever their origin.
2
The adoption of a new technique usually required a fair amount of tweaking and microinvention to adapt a technique to local circumstances. To adopt a new technology, a country thus needs skilled workers with some minimum level of “competence,” which is acquired by investment in human capability in part through childhood nutrition and in part through training in the form of master-apprentice contact.
Output is a function of both the quantity of workers and their “quality” (human capability).
The model combines the Nelson and Phelps (1966) model in which productivity growth depends on the difference between the actual techniques in use and the productivity of best-practice techniques with the Ben-Porath (1967) model, which analyses the growth of human capability as a function of parental investment decisions.
2
This certainly was perceived by informed contemporaries. David Hume, for instance, noted that
“every improvement which we have made [in the past two centuries] has arisen from our imitation of foreigners ...
Notwithstanding the advanced state of our manufacturers, we daily adopt, in every art, the inventions and improvements of our neighbours” (Hume, [1777], 1985, p. 328).
3

The model concerns the diffusion of best-practice blueprints from the original inventor to the level of artisans where it can be incorporated into everyday production. We therefore define the state
~ of the art level of international scientific knowledge by A and assume that it grows exogenously.
Within a given country, the level of technology in use is A. This level of technology evolves according to a Nelson-Phelps process, depending on the gap between scientific knowledge and the country’s own technology; and on the quality and level of skill of ordinary workers H:
(1)
~ where 0 < ä, å < 1. The technology in use in an economy cannot exceed the frontier value A ,and cannot fall below a minimum level A. To simplify notation in what follows, we assume that the minimum technological level is unity: A = 1.
The human capability of each artisan evolves according to the Ben-Porath equation:
(2) form of nutrition, basic schooling and training as an apprentice, and needs to equal H t-1
ì/ë to maintain the existing level of human capability of the workforce.
To close this simple model we need to specify what determines the level of investment in the next generation of workers. We suppose that individuals have two periods in their lives: when young they receive investment from their parents; and when old they receive income as workers that they use to maximize utility which comes from their own consumption Ct and investment in their child
(we assume constant population for now). The utility function is thus:
(3) where 0 < ã < 1. Workers supply one unit of labour inelastically, and live hand to mouth, making and receiving no bequests. It follows that parents invest a fraction ã of their income in their children.
Output in this economy comes from the production function
4

(4) t worker pays a share of his income to the government or landlords, receiving nothing in return. In addition, however, the government may tax the landlord class and redistribute this money to workers in the form of a Poor Law. It follows that the disposable income of workers is (1 - ô)Y/N where ô is the net rate of tax and rent after subtracting Poor Law transfers.
To analyse the evolution of useful knowledge A and human capital H it will be simpler, both for intuition and for drawing phase diagrams, to adopt the trick of using the inverse of human capability
(5) that we will refer to as misery, remembering that it refers to low levels of nutrition, health, schooling and other outcomes of childhood deprivation.
It follows that useful knowledge and misery evolve according to the log-linear system of
(6) of this system has a sign that is not immediately obvious: the (µ-ëá) term multiplying log M in the misery equation. If µ < ëá, the real wage at the technological minimum rises as population N or net taxes ô increase as (9) below shows. In addition, the misery process is unstable: a rise in misery lowers output and human capital investment, increasing misery next period in a self-reinforcing process so long as population remains stable. We therefore assume that (µ - ëá) is positive.
3
The analysis that follows is an adaptation of a linearized Lotka-Volterra dynamic system of two competing species. See e.g., Hofbauer and Sigmund, 1998, pp. 22-28.
5

6

The isoclines of the system, between the minimum and maximum levels of technology, are
(7)
The knowledge-misery system has four possible steady states depending on the relative
á position and slope of these isoclines. In the phase diagrams we denote N /ã(1-ô) by N*
In panel (a) of Figure 1, the misery isocline lies everywhere above the knowledge isocline so that misery dominates. The only equilibrium is at point B, with the log of useful knowledge equal to its lower bound, which we have set at zero.
In panels (b) and (c) the isoclines intersect at C = (
— — log A , l
— o g
—
M ) where:
(8)
In panel (b), the own effect terms in (6) dominate the cross effect terms ä(ì-ëá) > ëå so the technology isocline is steeper than the misery curve. As a result, the intersection point C is globally stable.
In panel (c), cross effects dominate own effects ä(ì-ëá) < ëå so the misery curve is steeper.
As a result, the intersection point at C = (
— — log A , l
— o g
—
M ) is a saddle dividing the space into two basins of attraction, one converging on point D with technology at its lower bound, the other at point
~
E where the misery isocline cuts the upper bound of technology log A .
Finally, in panel (d) of Figure 1, the knowledge isocline lies everywhere above the misery
4
There is also the case where the isoclines exactly coincide, making every point along them a steady state.
7

isocline, so the system converges to a steady state at point F where the misery isocline cuts the upper
~ bound of technology log A .
Intuitively, the evolution of useful knowledge and misery in (6) resembles an eco-system with two competing species: the growth of each species is retarded by the presence of the other. In panel
(a) of Figure 1, conditions are so favourable 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; with the converse holding in panel (d) where the species “knowledge” dominates. In panel (b), the species have little impact on each other and both co-exist at positive levels, while in panel (c) they have a strong impact on each other but the outcome depends on which species initially has a sufficiently large population to dominate the system and drive the other one out. We now show how this simple interaction of knowledge and misery leads to a sudden take-off in knowledge: an
Industrial Revolution.
There are two economies that we shall call France and England. 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 England than in France: w > w where subscripts E and F denote England and France respectively. This
E F
F) income even if it left the mean unaffected.
5
This implies that, at the technological minimum where the system begins, workers in each economy i = E, F have log disposable income
(9)
In this pre-industrial world, both economies start in a Malthusian equilibrium where births equal deaths. For this to be possible in a purely Malthusian world with English living standards and life expectancies above French ones, it must be that England has a lower marriage rate and fertility of marriage or a higher mortality rate due to a higher level of urbanization (Voth and Voigtländer,
2012). The difference in fertility can also stem from the incentives created by a poor law for landlords to minimize the number of mouths they have to subsidize as the disposable income
5
W e can also allow England to have a higher value of ë in the Ben-Porath equation (2), to reflect the greater efficiency of its apprenticeship system in creating useful skills; but we will focus here solely on the impact of greater transfers to workers.
8

equation (9) shows. If landlords are required to provide workers with a higher level of disposable directly. We therefore expect that, in otherwise identical regions of France and England, the English requirement to provide a higher level of subsistence to workers implies that English regions would
F below the French one. and the equilibrium of each economy by E and F. We want to see how these change as the
~ technological frontier A gradually rises through time.
~
Our starting point, in panel (a), is a stark Malthusian world with little knowledge: log A is arbitrarily small and the knowledge isocline A1 lies completely below the two misery isoclines. 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. the English misery isocline. We suppose that misery and useful knowledge strongly affect each other
ä(ì-ëá) < ëå so that the misery isocline is steeper: when the opposite holds the evolution of the system is broadly similar as we will see below. While a steady state exists at the knowledge frontier, as in panel (c) of Figure 1, because the English economy is starting in the basin of attraction of
9

10

the low knowledge equilibrium, it stays at this point. The rising technological frontier has no impact on production technology because the level of human capability is too low to absorb it: we might say that the technological enlightenment has no impact down on the farm. This is the preindustrial world of Malthusian stagnation of Hansen and Prescott (2002) and Clark (2008).
In panel (c) of Figure 2, the continued gradual rise in scientific knowledge causes the technology isocline to move above the English misery isocline, but still to cut the French one from above. As a result, while France stays at the minimum technology steady state at F, England jumps to the technological frontier at E. A small rise in the knowledge frontier thus causes a sudden divergence between the economies to occur. Because English human capability is initially slightly higher than French, England can start to apply technological knowledge to production, giving rise to a cumulative process of rising living standards, rising human capital, and rising 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.
~
This divergence is not permanent however. As the knowledge frontier log A continues to causing France to converge to the same technological frontier as England. While technology in both economies is the same, living standards in England will remain higher so long as it continues to enjoy higher transfers and a lower population. However, to the extent that rising living standards erode political support for a generous poor law, while urbanization removes social restrictions on early marriage while advances in medical knowledge reduce the “urban penalty” if higher mortality,
English and French living standards will converge.
If, on the other hand, misery and technology interact weakly ä(ì-ëá) > ëå so that the knowledge isocline is steeper, the evolution of the system is slightly different. When the technology isocline first rises above the English misery isocline, England moves to a steady state where the two intersect, and France will follow some time afterward. Both economies move steadily down along their misery isoclines as the knowledge isocline rises, until they reach the technological frontier.
2 Wages and Productivity in France and England.
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 labourers were 80 per cent higher. It is, however, invalid to conclude that English labour was therefore expensive, until we compare productivity and can thus infer unit labor costs.
We can impute labour productivity by comparing piece rates and day rates in French and
English agriculture. The average time needed to reap an acre of wheat in early nineteenth century
11

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).
Clark (1991, 449) similarly estimates the average output of a day’s threshing at 0.24
man-days per bushel or 0.65 man-days per hectolitre. For France, threshing took from 0.9 man-days to 1.25 man-days per hectolitre, or nearly double the English rate (Grantham, 1991, 363).
Reaping and threshing were manual activities with almost no capital input and fairly little skill. Even 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 labour. We shall see below in Section 5 that variation was higher still within mid-nineteenth century France, with the lowest departments having one third the wage, and one quarter the agricultural productivity of the highest.
Having shown that high wages may have reflected more productive labour rather than expensive labour, we now look at what made English workers more productive, and argue that the answer lies in their higher human capability.
6
This estimate is consistent with Arthur Young’s eighteenth century observations. Dividing the median costs 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 (Young 1771, IV, 293-296; 1772). Young wrote that "Strength depends on nourishment; and if this dfference be admitted, an English workman ought to be able to do half as much work again as a Frenchman" (Young, 1793, II, 315-316).
12

3
The traditional approach to human capital in the industrial revolution is to observe that literacy was not high in pre-industrial England, and to conclude that human capital therefore could not have mattered. For instance, Reis (2005) estimates adult literacy in Britain around 1800 at 60 percent for males and 40 for females, slightly below Northern France (71 and 41 percent respectively) but above
Southern France (44 and 17 percent). It is, however, open to question whether literacy at this early stage of the Industrial Revolution mattered a lot to technological development. Indeed, Mitch (1992,
1999) has argued that for its industrial needs, Britain, despite its mediocre literacy rates, may have been over educated. The point is that the standard human capital measures do not reflect England’s advantage, or explain its precocity.
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 capability in turn are strongly determined by the individual’s nutritional and disease environment from conception to adolescence. Naturally we have little direct evidence on either human capability or its environmental determinants for England or France before the Industrial
Revolution, but we do have systematic data on two variables that are strongly correlated with these: height and life expectancy.
For France, England and Sweden during the nineteenth century, Crimmins and Finch (2006) find that childhood mortality rates, which reflect nutrition and environmental conditions, are strongly inversely related with subsequent old age mortality and adult heights, which they attribute to differing burdens of childhood infection and inflammation; while Bozzoli, Deaton and
Quintana-Domeque (2009) show that postneonatal (between one month and one year after birth) mortality is a strong predictor of adult height in the US and Europe from 1950 to 1980. Case and
Paxson (2008) find a strong connection between height and cognitive ability in the US and Britain since the 1950s, with the strong impact of adult height on earning disappearing when childhood cognitive ability is controlled for; while for industrializing England Baten, Crayen and Voth (2012) find that rising food prices are associated with subsequent falls in numeracy, measured by ability to recall one’s age. In the public health literature for developing economies, Victora et al. (2008) find a strong connection between childhood malnutrition and shorter adult height, reduced schooling, and lowered economic productivity; while Walker et al. (2011) survey the biological mechanisms through which stunting, iodine deficiency, and iron-deficiency anaemia inhibit brain development in children.
As well as being linked to cognitive function, height is closely correlated with physical strength. The strength of a muscle is proportional to its cross section, which increases as the square of height as empirical studies show. For example, a study of Indian female labourers implies an
13

elasticity of about two between height and grip strength, while a study of champion weight lifters found that weight lifted "varied almost exactly with height squared" (Koley, Kaur and Sandhu 2009;
Forde et al. 2000).
We have systematic data on the height of young men in the late eighteenth and early nineteenth centuries for both France and England, shown in Figure 3. The English data, from Nicholas and
Steckel (1991) are the heights of convicts transported to Australia; while the French data, from Weir
(1997), are the heights of accepted conscripts into the French army. The available evidence uniformly suggests that English young men were substantially taller than their French counterparts.
English records based on convicts come disproportionately from the poorest strata of society which were probably shorter than average. Using records of recruits into the East India Company,
Mokyr and Ó Gráda (1996) find that literate recruits were around 0.5 inches (1 cm.) taller on average than illiterate ones; while those from weaving backgrounds, a notoriously impoverished group in early Victorian Britain, were up to an inch (2.5 cm.) shorter than other recruits. Using slightly later data, for men born between 1817 and 1841, Riley (1994, Table 1) finds that the average height of the shortest occupational group, manufacturing, was nearly one centimetre lower than the national average of 169.9 cm. The consensus is that heights were static or, more likely, fell somewhat during the early nineteenth century (Floud, Wachter and Gregory 1990, 288-290; Johnson and Nicholas
1995) so it would appear plausible that average English height around 1800 was about 170 cm.
The French data are the average for 20 year old conscripts who exceeded the minimum height requirement, requiring the estimated average height for the total population to be corrected for truncation. However, for 1817, complete data for all recruits is provided by Villermé (1829) who reports the mean height of 20 year olds as 161.5 cm, with 28.6 per cent below the minimum height same time, full adult height was attained only around age 23, so, based on English evidence
(Nicholas and Steckel 1991, Figure 1; Johnson and Nicholas 1995, Figure 1) these 20 year olds were probably a centimetre below full adult height. This suggests average male height in 1817 France of around 162.5 cm, or roughly 7.5 cm shorter than the English average.
For 1825-1829, Angeville (1836, Table 3) reports the average height of accepted recruits by administrative department, and the proportions rejected because of height or fitness. Under normality, Angeville’s national average of 165.1 cm for accepted recruits, and 16 per cent rejected
7
5This coefficient of variation predicts that 27.9 per cent should be below 157 cm; that 20.6 per cent should be above 167.8, while Villermé reports 20.0 per cent; and that 12 per cent should be over 170.5 while Villermé reports 11.2
per cent.
14

15

for height is consistent with a population mean of 163.8 cm and a coefficient of variation of 4 per cent.
8
Weir (1997) reports estimates of heights from 1804 onwards based on the sample of accepted recruits but assumes a coefficient of variation of only 3.5 per cent: the same as in modern western populations where the heights of the poor have more or less converged on those of the rich. This low coeffcient of variation gives a 0.4 cm increase in average height for the late 1820s compared with our estimate; but a very large gap for 1817: Weir estimates an average height of 163.2 cm compared with Villerme’s reported value of 161.5 cm.
9
The average French figure disguises large regional variation within France. Using Angeville’s data, assuming a coeffcient of variation of 4 per cent, the gap between the tallest and shortest departments is 8 centimetres, with the tallest departments being only 1-2 cm below the English average. While 16 per cent of recruits nationally were below the 157 cm threshold, in the shortest departments this rises to 33 per cent.
As we noted above, modern studies suggest that height was correlated was a range of productivity- enhancing capabilities, both physical and cognitive. It stands to reason that the higher cognitive ability of British workers would also translate into a better capacity to adapt and apply new techniques. English wages and agricultural productivity were nearly twice those in France. We find a larger range of variation across French departments. Industrial wages in 1839-40 the highest paid department were 3.2 times those in the lowest; while agricultural productivity in 1852 (measured as man hours to produce a hectolitre of wheat) in the most efficient department was 4.2 times that in the lowest. As the maps in Figure 3 show, both are strongly correlated with height.
Estimates of average nutrition in eighteenth century England vary widely. 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 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 causes.
8
6For 1,000 accepted recruits, another 335 were rejected for reasons of height, and 765 for other physical
9
Assuming normality and a coeffcient of variation of 4.75 per cent implies that the average height of accepted recruits was 165.1 cm. Inferring average population height from this figure using W eir’s assumption of a 3.5 per cent coeffcient of variation gives a value of 163.8 cm, close to W eir’s reported estimate.
16

from id., pp. 114–15); while Broadberry et al. (2011) estimate 2,100 kcal, more or less unchanged between the mid-thirteenth and mid-nineteenth centuries. At the other extreme, Muldrew (2011, p.
156) estimates average calories per capita 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.
To assess the plausibility of these estimates, we can consider the income elasticity of calorie demand that they imply. Subramanian and Deaton (1996) estimate this elasticity at 0.3 to 0.5 for
India, while Logan (2009) estimates it at 0.5 to 0.7 for late nineteenth century Britain. Assuming that
English income in the late eighteenth century is 75 per cent higher than France, the Fogel and Flood et al finding that English calories were 25 per cent higher translates into an elasticity of one third, at the lower end of the Subramanian-Deaton range; while the Broadberry et al estimate is roughly zero; while the Allen estimate that English caloric intake was 50-90 per cent higher translates into an elasticity of two thirds to 1.2, at the upper end of the Logan range.
English workers were not only taller and better nourished; they lived longer. Longer life expectancy reflects better nutritional and health status; and contributes to higher investment in human capability to the extent that parents are more likely to invest in their children if these had a better chance to survive. While it may be the case that high infant mortality shifted average health up by trimming the lower tail and “cleansing the weakest” it is far more likely that the process that killed the weak also tended to weaken the survivors. Inflammations, which killed many children, also led to life-long reductions in height and increased morbidity (Crimmins and Finch, 2006).
Table 2 gives estimates of infant mortality and life expectancy for England and France, the
English data coming from Wrigley et al. (1997, 295, 224, 256, 291), and the French from Blayo
(1975). It can be seen that expectation of life at birth in England was 36.4 years in the first half of the eighteenth century and 40.3 years in the second half; whereas in France in the 1740s it was about
25 years; and between 1750 and 1789 around 28.1 years. Most of the English advantage was due to higher survival rates of infants and children, but a smaller gap for those surviving to adulthood still exists.
France’s high infant mortality reflected to some extent its lower standard of living and inferior diet. Matossian (1984) links higher French death rates to consumption of the wrong kind of food, whereas Fogel (2004) emphasizes the link between inadequate food consumption-hunger-and
“premature death.” Some of the English advantage may have been due to longer breast-feeding ; and possibly to a faster decline in smallpox due to the application of pre-Jenner inoculation techniques, especially after the adoption of improved methods by Robert Sutton in the 1760s Mokyr (2009, 285,
244).
17

Table 2: Infant mortality and life expectancies in England and France during the eighteenth century
The implementation of new techniques required a high-quality labor force, but of the most essence was a skilled and well-trained elite of technicians, engineers, mechanics, and similar personnel. That the quality of highly-skilled English artisans was a decisive factor in its industrial success was the opinion of many contemporary commentators, summarized by Mokyr (2009, 107-118) and
Meisenzahl and Mokyr (2012). For instance Jean-Baptiste Say noted in 1803 that “the enormous wealth of Britain is less owing to their own advances in scientific acquirements, high as she ranks in that department, as to the wonderful practical skills of her adventurers in the useful application of knowledge and the superiority of her workmen.” Josiah Tucker in 1758 wrote “the Number of
Workmen [in Britain] and their greater Experience excite the higher Emulation, and cause them to excel the Mechanics of other Countries in these Sorts of Manufactures.” Contemporary policymakers clearly viewed highly skilled laborers as a source of comparative advantage, as witnessed by the laws prohibiting the emigration of artisans and exportation of machinery (neither of which appear to have been particularly effective). Laws prohibiting the emigration of artisans and the exportation of machinery were first passed in 1696, and repeated through the eighteenth century, to be repealed in 1824.
Decisive evidence on the relative quality of English workers comes from the direction of labour migration. If the expensive English labour hypothesis were true, we would expect the flow to have been from France to England, in the form of unskilled workers taking advantage of higher
English wages. In fact, the flow was in the opposite direction, and composed of English and Scottish skilled artisans. Interestingly, this flow of artisans to France precedes the Industrial Revolution,
18

implying that the advantage that England enjoyed in the area of technical competence predated its
1815, a large number of British technicians found their way to the Continent, where they installed, maintained, and managed new equipment and instructed local workers how to use it.
11
Evidence that the mercantilist laws trying to protect this resource were effective is sparse, but they did result in a Parliamentary investigation that yielded a large quantity of anecdotal evidence on the quality of British labour. Thus it was maintained that an English engineer, turner, or iron founder, working in France, will make twice as much as a French one. “The English workmen, from their better methods, do more work and better than the French...and though their wages are higher, yet their work does not cost more money in France than when done by Frenchmen, though their confirmed by the flows of immigrants, as well as by the reverse flow of Continental engineers who
How large were the flows? Alexander estimated that in the years 1822 and 1823 alone,
16,000 artisans moved from England to France (ibid., 108). This seems exaggerated, but a year later
Galloway estimated the stock of English workers in France at 15,000-20,000 workers Great Britain
(1824, 37, 43).
10
John Holker, an English mechanic and political refugee in France, set up a textile manufacturing plant in 1752 and, despite the risks as a Jacobite refugee, returned to recruit many of his skilled workers in Britain. By 1754 he employed 20 English artisans who were allocated among French workers so that skills could be disseminated in the most effective fashion (Henderson, 1954, 16; Harris, 1998, 60).
11 The paradigmatic figure here was Aaron M anby (1776-1850) who set up a large engineering works in
Charenton that employed 200-250 Englishmen in the mid-1820s with the French providing only the manual labour
Henderson (1954, 54).
12 The great inventor Bryan Donkin noted that a worker in the paper industry who might have made 18-20 s. a week in B ritain, was hired at 50s. in France. An engineer named Alexander Galloway felt that a person of similar qualifications would make 22 s. in Paris and 36 s in London — but then added that English workmen in Paris would make twice what the locals would make (2 guineas), indicating the difference in the perceived quality of the workmen
(ibid., 24). John M artineau testified similarly that a French blacksmith would make in France 4 francs a day, while an
English smith in Paris would make 10- 11 francs (ibid., 7).
13
Among the many Germans who came to Britain to acquire technical expertise, we can mention W ilhem von
Reden, sent to study British coal mining techniques in 1776; Johann Gottfried B rügelmann who travelled to study
Arkwright’s famous Cromford mill in 1794, before setting up his own mill near Düsseldorf; F.A.J. Egells, a W estphalian locksmith sent by the Prussian government to England in 1819 to study machinery engineering; and Jacob M ayer, who worked for a time at Sheffield before opening a cast-steel mill near Cologne Henderson (1954, ch. 4).
19

The strong correlation between wages, height and measures of human capability such as patents, literacy and school attendance can be seen within a cross-section of France during its own episode of rapid industrialization, in the second quarter of the nineteenth century. Figure 3 shows the correlations across each of the 85 departments of mainland France of height and literacy of conscripts, school enrolment, industrial employment, number of patents, urban population, the crude death rate, and per capita wheat consumption for the late 1820s, industrial wages from 1839-40, and agricultural productivity (man days to produce a hectolitre of wheat) for 1852. Details of these variables are given in the Appendix. It can be seen that most of these variables are strongly correlated, with the exception of health of army recruits and crude death rates.
20

In part, the negative correlation between height and fitness of conscripts reflects recruitment quotas: in areas with many recruits rejected for short stature, the remainder would be declared fit for service irrespective of health (Marshall, 1834). However, height is also positively correlated with crude death rates and urbanization, consistent with the idea of Bozzoli, Deaton and
Quintana-Domeque (2009) that in areas of high mortality, small infants die at a disproportionate rate leading the survivors to be taller on average.
While most variables are correlated, as the second panel of Figure 3 shows, they also show strong spatial correlation: departments in northern and eastern France tend to have high levels of productivity, literacy, patents and so on, and central, southern and western departments low ones.
There is therefore the possibility, typically neglected in economic analyses of cross-sectional data, that these correlations simply reflect similar latent characteristics of neighbouring regions, rather than true relationships between individual variables.
Table 3 gives regressions of agricultural productivity, industrial wages, literacy of conscripts, and life expectancy on a variety of covariates. We again emphasize that these do not purport to be structural regressions: their purpose is to summarize correlations in the data. The first column of each gives standard OLS results. However carrying out a Moran test for spatial correlation of residuals, using a weighting matrix W where neighbours of each region are given equal weights which sum to one, shows strong spatial correlation in every case. The second column therefore gives a Kelejian and Prucha (2010) version of a Cliff-Ord spatial lag regression, implemented by Piras
(2010), where:
(10) and residuals follow
(11)
21

22

where residual structure of Kelejian and Prucha (2007) where, instead of the spatial AR1 in (11), residuals are required only to obey å = Rå + ç where R is a non-singular matrix. Estimating these regressions, using a variety of kernels to weigh the geographical distance between the centres of each department to estimate the covariance matrix, gave similar results to the Cliff-Ord model, and we do not report them here.
Starting with productivity, OLS suggests a strong association with height, industrial employment, and, especially, number of patents. However, controlling for spatial autocorrelation, none of these variables has a large coeffcient or is statistically significant at conventional levels.
Omitting all other variables, however, there remains a modest but significant relationship between height and productivity. Looking at wages, OLS suggests a strong relationship with height (the relationship with the closely related literacy is slightly higher), patents, and urbanization, and these relationships turn out to be robust to spatial correlation, although the strength of the effect diminishes in each case. Turning to literacy of conscripts in the late 1820s, we add life expectancy at birth (to capture returns to investment in human capital) and the ratio of annual births to marriages (to see if there is any quantity-quality trade-off where areas with fewer births have greater investment in education). The OLS regressions show a strong association with height and patents only; but adjusting for spatial autocorrelation in the second column, patents no longer have a large or significant effect, while life expectancy has a modest but significant impact. Finally, looking at life expectancy, only urbanization is significant in the spatial regression, with height, patents, and industry being individually insignificant. Excluding patents and industry, the regression shows a modest relationship between height on life expectancy.
Similar data for Britain for the nineteenth century are harder to come by, but the little there is confirms our hypothesis that there is a positive association between high wages and some measure of labor quality. One measure of human capital that seems to have withstood the test of time is an index of age-heaping, which tends to be higher among less numerate and educated populations. In the table below we produce some cross-sectional results for England at the county level. As our dependent variable we use the wage level at the county level as reported by Hunt (1986), though these are agricultural wages. The independent variables are a variety of measures of “human quality.”
One of them is the literacy rate by county adapted from convict data (Nicholas and Nicholas, 1992, p. 11); another are two measures of nineteenth century age-heaping from data generously provided to us by Professor Joachim Voth (see Baten, Crayen and Voth, 2012 for details.) We also utilize the county by country “quality of diet” developed recently by Horrell and Oxley (2012) and British army height data using the standard source based on the work of Floud, Wachter and Gregory. All those data are seriously flawed, yet they are consistent with the hypothesis that wages were associated with a higher “quality” of the average worker (although there is no possibility to disentangle the exact causal relation with the data at hand).
23

Table 4: cross sectional correlations, England (s.e.’s in parentheses)
Height
Nutrition
Whipple
Literacy wage 1760
-3.78
a
(4.02)
-2.72
c
(.903)
-.966
(.41)
-.05
(.20) constant 472
(276) wage 1790
6.92
a
(5.07)
1.71
-.669
.13
(.25)
-295 c
(1.14)
(.514)
(348) wage 1840
6.48
b
(5.56)
1.41
d
(1.36)
-1.59
(.58)
-.05
(.27)
-149
(402) n
R 2
35
.42
35
.30
37
.25
a 1788-1805 average b 1824-44 average c 1797 d 1834
The results for the 1760 wage data are obviously weak, with both the height and the nutrition variables having the wrong signs. For the 1790 and 1840 wages, the signs are more reasonable and perhaps most encouraging, the wages are thoroughly negatively correlated with the degree of ageheaping (which reflects to some degree numeracy education and to some degree innate cognitive ability). All in all, the county-level regressions need more work, and in particular a better measure of the dependent variable.
24

We have seen how England benefited from a self-reinforcing nexus of high wages and high labour capability. Underlying this process were two social institutions unique to England: the Old Poor
Law, and an informal system of apprenticeship.
The idea that the English Poor Law, a unique institution, had some unintended positive effect on industrialization has been proposed before (Solar, 1995; Greif, Iyigun, and Sassoon, 2012).The
main argument made is that the Poor Law functioned as a social insurance scheme inducing people to take more risks than they would have otherwise. Starting in the early seventeenth century, English parishes were required to levy a property tax to fund local poor relief, which was directly largely to the elderly and families with many small children. By the end of the seventeenth century, Poor Law expenditure was about 1 per cent of national income, sufficient to provide half subsistence for 10 per cent of the population; and increased to around 2 per cent of national income by the end of the eighteenth century. Kelly and Ó Gráda (2011) argue that the weakening of the Malthusian positive check in the early seventeenth century was a result of this Old Poor Law, by noting that a strong positive check endured for a century in London, despite its higher wages, where systematic poor relief did not start until the 1720s; and that parishes with generous poor relief experienced weaker positive checks in the eighteenth century than those with lower spending. As well as improving living standards, especially of children, directly, the Old Poor Law probably maintained higher living standards indirectly by acting as a brake on population growth. As equation (9) indicates, areas dominated by a few landowners, so called “closed parishes”, had a strong incentive to restrict the supply of housing to minimise their liability for poor relief. In short, by being a progressive transfer, the Old Poor Law may have raised the nutritional standard of living of the median English person., and thus contributed to the higher quality of the labor force.
The second source of English human capability lay in its system of professional training through apprenticeship: in 1700 over a quarter of males aged 21 had completed an apprenticeship
(Mokyr, 2009, 118). As noted earlier, the English school system was not impressive by contemporary standards. However, the decisive group during the Industrial Revolution was artisans, and nearly all artisans were trained as apprentices by other artisans. The question is why the English system of apprenticeship worked better than elsewhere.
In her work on childhood labour and training, Humphries (2003; 2010, 282- 283) stresses that, unlike much of the Continent, apprenticeship in England was not normally enforced and
(2008).
14
The traditional, negative view of European guilds is disputed by Epstein (1998) but defended by Ogilvie
25

a repeated-interaction framework, relying on the capability of local networks based on kin, religion, and personal connections to create reputation effects that made the majority of both masters and pupils cooperate at a reasonable level even when the contracts themselves were less than complete.
The 1563 Statute that formally prohibited craftsmen to carry out their trade without completing their apprenticeship was not uniformly enforced, and after its repeal in 1814, apprenticeship remained the main form for boys to acquire a professional training. Moreover, an examination of a large sample of indentures (formal contracts between masters and apprentices) reveals a substantial increase in the number of apprentices in mechanical and machine-related occupations in the early stages of the Industrial Revolution (Van der Beek, 2010). As Humphries
(2010, 2) emphasizes, the Poor Law contributed substantially to the accumulation of human capital, through the funding of pauper apprenticeships. Contrary to common belief, for much of the eighteenth century these apprenticeships provided real human capital (especially non-cognitive skills such as docility and punctuality) to the children of the destitute who otherwise would have remained unskilled, as well as providing labour to new factories in remote areas.
This system of producing human capital was adaptive and worked well as long as the importance of formal schooling remained limited. It remained largely a private order institution, very much part of a British institutional structure that stood at the center of its success in the early stages of the Industrial Revolution This institutional structure represents a “civil economy,” one in which cooperative arrangements between individuals based on shared cultural norms and reputation mechanisms led to outcomes that elsewhere required direct state intervention.
Of course, it could be argued that eighteenth century England could only maintain institutions like the Old Poor Law and an extensive apprenticeship system because it was a prosperous place, with continuous demand for the products of industries that generated a supply of new artisans.
Among the industries of eighteenth century England that proved particularly important as sources of skilled craftsmen who drove the Industrial Revolution, Mokyr (2009, 114-115) highlights three: clock and instrument making (with particular contributions from Protestant Huguenots driven out of France); the shipping and shipbuilding industry; and coal mining with its demand for precisely made pumps, and iron rails for moving carts. There was also an element of demand: by the early eighteenth century Britain had a substantial middle class consisting of such people as merchants, professionals, well-to-do farmers, and artisans. These “middling sort” consumers demanded luxury consumer durables that required a high level of precision skills: clocks, harpsichords, tapestries, and fine china- and tableware (Berg, 2005). This market helped create the cadres of craftsmen
(themselves middle class), whose competence was essential if the innovative ideas of inventors were craftsmen that gave Britain the unique ability to implement and improve technologies developed across Europe that made her the leader of the Industrial Revolution.
15
De Vries has argued that the demand for consumer durables in the century before the Industrial Revolution shifted from an emphasis on the quality of the materials to an emphasis on workmanship (De Vries, 2008, p. 146).
26

Against the widely accepted view that high wages induced the British industrial revolution as a way of saving on labour, we show that after adjusting for productivity, English labor costs were no higher than those in France. We argued instead, that England’s high wages and industrialization both flowed from a common source: the superior human capability of English workers. We have argued that this capability has two closely related dimensions: a physiological and a cognitive. We have also argued that while much of the literature still thinks of the Industrial Revolution as a stream of inventions, the crucial factor was the highly skilled labor that implemented it and the strong, disciplined and docile workers who manned the machines.
The underlying idea is that technology, much like the performing arts, is an implementable form of knowledge: it takes one person to write the original, but if it is to be successful it needs to be implemented by competent individuals, whose skills, typically, do not necessarily include comparable creativity and originality. Britain and France both could count on a considerable supply of original genius, but when it came to the skilled artisans needed to implement and operate the new techniques and introduce incremental improvements to new technological ideas, England’s advantage was decisive.
While we do not have data on the cognitive advantage of English workers in the eighteenth century, we do have systematic information on two closely associated variables: height and life expectancy, both of which show that English children suffered from a less adverse developmental environment than their French counterparts.
Excepting wages and productivity, all variables are taken from the tables in Angeville (1836) with column numbers listed. All figures for military recruits are the average for 1825-1829; births, marriages, and deaths are average of 1825, 1826, 1828, 1830, 1832. o o o
Height: average height in cm of accepted conscripts. V37.
Literate: One minus percentage conscripts ignorant. V69.
Pupils: Number of schoolchildren per 1,000 inhabitants 1830-33. V71. o o
Industry. Percentage military recruits from industrial occupations. V84.
Patents. Number of patents per 1000 inhabitants 1829. V85. o
+ V43).
Unfit. Percentage of recruits rejected for physical causes other than height. V43/ (1000 + V39 o o
Urban. Percentage of population in towns with over 1500 inhabitants 1831. V7.
Crude death rate: Estimated from V5 natural increase and V9, mean of the inverses of crude birth and death rates. o Births per marriage: Ratio of births relative to male population aged 21 (V13) and marriages
27

relative to male population aged 21 (V15). o Wheat: Annual wheat consumption, hectolitres per 100 inhabitants, 1825-28. V129. o Wage: Average industrial wage, 1839-40 used by Chanut et al. (1995) and kindly provided by Gilles Postel-Vinay. o Productivity: Wheat yield per hectare divided by man days per hectare, Enquete agricole 1852 from Demonet (1990).
28

Acemoglu, Daron. 2010. "When Does Labor Scarcity Encourage Innovation?" Journal of Political
Economy 118(6):pp. 1037-1078.
Allen, Robert. 2005. “English and Welsh Agriculture, 1300-1850: Output, Inputs, and Income.”
Working paper Nuffield College.
Allen, Robert C. 2009a. The British Industrial Revolution in Global Perspective. Cambridge:
Cambridge University Press.
Allen, Robert C. 2009b. The British Industrial Revolution in Global Perspective. Cambridge:
Cambridge University Press.
Angeville, Adolph d’. 1836. Essai sur la Statistique de la Population francaise. Paris: F. Darfour.
Baten, Jörg, Dorothee Crayen and Joachim Voth. 2012. “ Poor, Hungry and Ignorant: Numeracy and the Impact of High Food Prices in Industrializing Britain, 1780-1850.” Economics Working Papers
1120 Department of Economics and Business, Universitat Pompeu Fabra.
Ben-Porath, Y. 1967. "The production of human capital and the life cycle of earnings." Journal of
Political Economy 75(4):352-365.
Blayo, Yves. 1975. "La Mortalité en France de 1740 a 1829." Population 30:123-142.
Bozzoli, Carlos, Angus Deaton and Climent Quintana-Domeque. 2009. "Adult Height and
Childhood Disease." Demography 46:647-669.
Broadberry, Stephen, Bruce Campbell, Alexander Klein, Mark Overton and Bas van Leeuwen. 2011.
“British Economic Growth, 1270-1870: an Output-based Approach.” Working paper London School of Economics.
Case, Anne and Christina Paxson. 2008. "Stature and Status: Height, Ability, and Labor Market
Outcomes." Journal of Political Economy 116:499-532.
Chanut, Jean-Marie, Jacques Mairesse, Jean Heffer and Gilles Postel-Vinay. 1995. "La disparité des salaires en France au XIXe siècle." Histoire et Mésure 10:381-409.
Clark, Gregory. 1991. "Yields per Acre in English Agriculture, 1250-1860: Evidence from Labour
Inputs." Economic History Review 44:445-60.
29

Clark, Gregory. 2008. A Farewell to Alms: A Brief Economic History of the World. Princeton:
Princeton University Press.
Crimmins, Eileen M. and Caleb E. Finch. 2006. "Infection, Inflammation, Height, and Longevity."
Proceedings of the National Academy of Sciences (USA) 103(2):498-503.
Demonet, Michel. 1990. Tableau de l’agriculture francaise au milieu du 19e siecle: l’enquête de
1852. Paris: Editions de l’EHSS.
Epstein, S. R. 1998. "Craft Guilds, Apprenticeship, and Technological Change in Preindustrial
Europe." The Journal of Economic History 58(3):pp. 684-713.
Floud, R., K.W. Wachter and A. Gregory. 1990. Height, Health and History: Nutritional Status in the United Kingdom, 1750-1980. Cambridge: Cambridge University Press.
Floud, Roderick, Robert W. Fogel, Bernard Harris and Sok Chul Hong. 2011. The Changing Body:
Health, Nutrition, and Human Development in the Western World since 1700. Cambridge:
Cambridge University Press.
Fogel, Robert W. 2004. The Escape from Hunger and Premature Death: Europe, America, and the
Third World. Cambridge: Cambridge University Press.
Forde, Lincoln E., Alvin J. Detterline, Kevin K. Ho and Cao Wenyuan. 2000. "Gender-and
Height-related limits of Muscle Strength in World Weightlifting Champions." Journal of Applied
Physiology 89:1061-1064.
Grantham, George. 1991. Land, Labour and Livestock: Historical Studies in European Agriculture.
Manchester: Manchester University Press chapter The Growth of Labour Productivity in the
Production of Wheat in the Cinq Grosses Fermes of France, 1750-1929, pp. 340-363.
Great Britain. 1824. "First Report from Select Committee on Artisans and Machinery." British
Sessional Papers, House of Commons (5).
Habakkuk, H.J. 1962. American and British Technology in the Nineteenth Century. Cambridge:
Cambridge University Press.
Hansen, Gary D. and Prescott, Edward C. 2002. “Malthus to Solow, ” American Economic Review
Vol. 92, No. 4 (Sept.), pp. 1205-1217.
Harris, J.R. 1998. Industrial espionage and technology transfer: Britain and France in the eighteenth century. Aldershot: Ashgate.
Heckman, James J. 2007. "The Economics, Technology, and Neuroscience of Human Capability
30

Formation." Proceedings of the National Academy of Sciences 104(33):13250-13255.
Henderson, W.O. 1954. Britain and Industrial Europe, 1750-1870: Studies in British Influence on the Industrial Revolution in Western Europe. Liverpool: Liverpool University Press.
Hume, David. [1777], 1985. Essays: Moral, Political, and Literary . Ed. by Eugene F. Miller.
Indianapolis: Liberty Fund.
Humphries, Jane. 2003. “English Apprenticeship: A Neglected Factor in the First Industrial
Revolution.” In The Economic Future in Historical Perspective, ed. Paul A. David and Mark
Thomas. Oxford: Oxford University Press pp. 73-102.
Humphries, Jane. 2010. Childhood and Child Labour in the British Industrial Revolution.
Cambridge: Cambridge University Press.
Johnson, Paul and Stephen Nicholas. 1995. "Male and Female Living Standards in England and
Wales, 1812-1857: Evidence From Criminal Height Records." The Economic History Review
48(3):pp. 470-481.
Kelejian, Harry H. and Ingmar R. Prucha. 2007. "HAC Estimation in a Spatial Framework." Journal of Econometrics 140:131-154.
Kelejian, Harry H. and Ingmar R. Prucha. 2010. "Specification and Estimation of Spatial
Autoregressive Models with Autoregressive and Heteroscedastic Disturbances." Journal of
Econometrics 157(1):53-67.
Kelly, Morgan and Cormac Ó Gráda. 2011. “Living Standards and Mortality in England since the
Middle Ages.” Working paper School of Economics, University College Dublin.
Kelly, Morgan, Joel Mokyr, and Cormac Ó Gráda. 2012. “Wages, Technology, and the Industrial
Revolution: a Reassessment.” In progress.
Koley, Shyamal, Navdeep Kaur and J.S. Sandhu. 2009. "A Study on Hand Grip Strength in Female
Labourers of Jalandhar, Punjab, India." Journal of Life Science 1:57-62.
Koller, Daphne and Nir Friedman. 2009. Probabilistic Graphical Models: Principles and
Techniques. Cambridge: MIT Press.
Logan, Trevon D. 2009. "The Transformation of Hunger: The Demand for Calories Past and
Present." Journal of Economic History 69:388-408.
Marshall, Henry. 1834. "On the Mortality of the Infantry of the French Army." Edinburgh Medical and Surgical Journal 42:34-49.
31

Matossian, Mary Kilbourne. 1984. "Mold Poisoning and Population Growth in England and France,
1750-1850." The Journal of Economic History 44(03):669-686.
Mitch, David. 1992. The Rise of Popular Literacy in Victorian England . Philadelphia, PA:
University of Pennsylvania Press.
Mitch, David. 1999. “The Role of Education and Skill in the British Industrial Revolution.” In Joel
Mokyr, ed., The British Industrial Revolution: An Economic Perspective, 2nd edn. Boulder, CO:
Westview Press, pp. 241–79.
Mokyr, Joel. 2009. The Enlightened Economy: An Economic History of Britain 1700-1850. New
Haven: Yale University Press.
Mokyr, Joel and Cormac Ó Gráda. 1996. "Height and Health in the United Kingdom 1815-1860:
Evidence from the East India Company Army." Explorations in Economic History 33:141-168.
Muldrew, Craig. 2011. Food, Energy and the Creation of Industriousness: Work and Material
Culture in Agrarian England, 1550-1780. Cambridge: Cambridge University Press.
Nelson, Richard R. and Edmund D. Phelps. 1966. "Investment in Humans, Technological Diffusion, and Economic Growth." American Economic Review 56:69-75.
Nicholas, S. and R. Steckel. 1991. "Heights and Living Standards of English Workers during the
Early Years of Industrialization." Journal of Economic History 51:937-957.
Ogilvie, Sheilagh. 2008. "Rehabilitating the Guilds: A Reply." Economic History Review
61(1):175-182.
Piras, Gianfranco. 2010. "sphet: Spatial Models with Heteroskedastic Innovations in R." Journal of
Statistical Software 35(1):1-21.
Reis, Jaume. 2005. “Economic Growth, Human Capital Formation and Consumption in Western
Europe before 1800.” In Living Standards in the Past: New Perspectives on Well-being in Asia and
Europe, ed. Robert C. Allen, Tommy Bengtsson and Martin Dribe. Oxford: Oxford University.
Riley, James C. 1994. "Height, Nutrition, and Mortality Risk Reconsidered." The Journal of
Interdisciplinary History 24(3):pp. 465-492.
Rothbarth, Erwin. 1946. "Causes of the Superior Effciency of the USA Industry As Compared with
British Industry." Economic Journal 56:383-390.
Scutari, Marco. 2010. "Learning Bayesian Networks with the bnlearn R Package." Journal of
Statistical Software 35(3):1-22.
32

Solar, Peter. 1995. “Poor Relief and English Economic Development Before the Industrial
Revolution.” Economic History Review 48:1-22.
Subramanian, Shankar and Angus S. Deaton. 1996. "The Demand for Food and Calories." Journal of Political Economy 104:133-62.
Victora, Cesar G, Linda Adair, Caroline Fall, Pedro C Hallal, Reynaldo Martorell, Linda Richter and
Harshpal Singh Sachdev. 2008. "Maternal and child undernutrition: consequences for adult health and human capital." The Lancet 371(9609):340-357.
Villermé, L. R. 1829. "Mémoire sur la taille de l’homme en France." Annales d’Hygiène Publique et de Médicine Légale pp. 351-396.
Walker, Susan P, Theodore D Wachs, Sally Grantham-McGregor, Maureen M Black, Charles A
Nelson, Sandra L Huffman, Helen Baker-Henningham, Susan M Chang, Jena D Hamadani, Betsy
Lozoff and et al. 2011. "Inequality in early childhood: risk and protective factors for early child development." Lance 378(11):1325-1338.
Weir, David. 1997. Health and Welfare during Industrialization. Chicago: University of Chicago
Press chapter “Economic welfare and physical wellbeing in France, 1570-1990,” pp. 161-200.
Wrigley, E. A., R. S. Davies, J. E. Oeppen and R. S. Schofield. 1997. English Population History from Family Reconstitution, 1580-1837. Cambridge: Cambridge University Press.
Young, Arthur. 1771. A Six Months Tour through the North of England. second ed. London: Strahan.
Young, Arthur. 1772.
A Six Weeks Tour through the Southern Counties of England and Wales. third ed. London: Strahan.
Young, Arthur. 1793. Travels in France. Dublin: Cross et al.
33