The origins of the Industrial Revolution come second perhaps only to the origins of the First World War as a topic attracting contributors who are both numerous and distinguished. As with the Great War, Nicolas Baumard's enjoyable paper faces the challenge of persuading us not only that his hypothesis is right, but also that we understand precisely what distinguishes it from those of rival scholars. In this field, there is (relatively) little disagreement about the various phenomena that occurred in England in the eighteenth century: the growth in incomes, the development of various institutions increasing the protection given to private property, the increase in trade, the decline in violence, the increase in innovation.
The disagreement between scholars is partly about timing: Recent evidence from Humphries and Weisdorf (Reference Humphries and Weisdorf2016) based on calculations of annual rather than daily wages places the growth of labor incomes well before the mid-eighteenth century favored by the previous consensus. But mainly it is about causality: Was the increase in innovation an independent cause of increasing incomes, or was it (as claimed by Baumard) mainly the consequence of prior high incomes? It is extremely hard to say anything about causality, either in the numerous studies cited here about life history or in the studies showing that various apparently growth-reinforcing attitudes occurred in England both before and during the eighteenth century.
To take the former studies first, I doubt anyone would dispute that “Individuals living in conditions of affluence tend to have lower rates of time discounting, to be more optimistic, and to be more conscientious and trustful” (sect. 3.6, para. 2). The issue is to what extent priori affluence causes these other characteristics; it seems plausible that the causality runs in both (indeed, in all) directions. Correlational studies do not become evidence about causation through being cited in large numbers.
Nor are twin studies the magic bullet Baumard seems to think they are: even if a study (sect. 3.6, para. 1) “found that individuals with higher birth weight (within pairs of identical twins) are more likely to participate in the stock market (a proxy of risk-taking preference),” we do not know whether the higher birth weight caused later participation in the stock market through changing risk preferences or through other mechanisms (such as inducing parents to give preferential treatment to the heavier child, causing that child to become more prosperous than its twin).
This example prompts the question, valid for the latter studies, as well, of which proxies to pick for risk preference and time discounting among the hundreds potentially available. Why use participation in the stock market when this is correlated with many other things that are not risk preferences? Similarly, why use literacy rates in early modern Europe as evidence of diminished time discounting rather than, say, pointing to the reduced rate of cathedral building as evidence in the other direction? The willingness to design and build cathedrals has as good a claim as any other to be a proxy for long-term thinking. It is hard to avoid the suspicion that, if cathedral building had appeared to favor the Baumard hypothesis, it would have been cited, but because it does not, it was not.
The very wealth of possible correlates of the unobserved variables of Life History Theory means that Baumard several times loses sight of the aim of the exercise, which is to explain why England had its Industrial Revolution before any other country. In using literacy rates as evidence of lower time discounting, he cites the fact that England had higher literacy than France, while treating as of minor importance that England had lower literacy than Germany or Scandinavia. Similarly, he makes a great deal of Manuel Eisner's (Reference Eisner2001) evidence that English homicide rates began falling in the later Middle Ages, and of the fact that “on the eve of the Industrial Revolution … England was still ahead of the rest of Europe” (sect. 5.3, para. 1) – but he cannot have it both ways. What matters for his hypothesis is the change in homicide rates over time, and this occurred spectacularly across Europe (including in Italy), even if in England these were lower than elsewhere from the sixteenth to the nineteenth centuries.
This is not to dispute the plausibility of the hypothesis, but if rates of time discounting were falling everywhere in Europe, as they probably were, this phenomenon will not explain the English origins of the Industrial Revolution. And several things make me doubt their relevance to the question of English origins. One is that innovation is the fruit of a tiny subset of the population, whose affluence had been little correlated with the general living conditions of the population prior to the eighteenth century. The French upper classes were probably more affluent than their English counterparts in the sixteenth through eighteenth centuries, so it is not a lack of affluence that explains their unwillingness to engage in innovation.
A second reason for skepticism is that England's greater rate of urbanization produced affluence for those who survived childhood but at the cost of horrendous rates of infant mortality for those who grew up in the cities. It is not clear why Life History Theory should focus only on the affluence of those who survive to adulthood and not on the ex ante improbable fact of their survival.
Most importantly, the dominant fact about the modern world is that sooner or later the Industrial Revolution took place everywhere. As Baumard points out, quoting Deirdre McCloskey, the increase in prosperity it brought about was massive, far greater than it seems reasonable to attribute to any single cause. Even if increases in risk taking preceded increases in innovation, a claim that is at best unproven, the extent of innovation seems out of proportion to the alleged cause. This very strongly suggests that we are seeing the mutual and cumulative reinforcement of several different causes. That one such cause was the change in attitudes predicted by Life History Theory is a reasonable conjecture, and Baumard's paper, although always stimulating, would have seemed more convincing to me if it had been framed in this light.
You have
Access
The origins of the Industrial Revolution come second perhaps only to the origins of the First World War as a topic attracting contributors who are both numerous and distinguished. As with the Great War, Nicolas Baumard's enjoyable paper faces the challenge of persuading us not only that his hypothesis is right, but also that we understand precisely what distinguishes it from those of rival scholars. In this field, there is (relatively) little disagreement about the various phenomena that occurred in England in the eighteenth century: the growth in incomes, the development of various institutions increasing the protection given to private property, the increase in trade, the decline in violence, the increase in innovation.
The disagreement between scholars is partly about timing: Recent evidence from Humphries and Weisdorf (Reference Humphries and Weisdorf2016) based on calculations of annual rather than daily wages places the growth of labor incomes well before the mid-eighteenth century favored by the previous consensus. But mainly it is about causality: Was the increase in innovation an independent cause of increasing incomes, or was it (as claimed by Baumard) mainly the consequence of prior high incomes? It is extremely hard to say anything about causality, either in the numerous studies cited here about life history or in the studies showing that various apparently growth-reinforcing attitudes occurred in England both before and during the eighteenth century.
To take the former studies first, I doubt anyone would dispute that “Individuals living in conditions of affluence tend to have lower rates of time discounting, to be more optimistic, and to be more conscientious and trustful” (sect. 3.6, para. 2). The issue is to what extent priori affluence causes these other characteristics; it seems plausible that the causality runs in both (indeed, in all) directions. Correlational studies do not become evidence about causation through being cited in large numbers.
Nor are twin studies the magic bullet Baumard seems to think they are: even if a study (sect. 3.6, para. 1) “found that individuals with higher birth weight (within pairs of identical twins) are more likely to participate in the stock market (a proxy of risk-taking preference),” we do not know whether the higher birth weight caused later participation in the stock market through changing risk preferences or through other mechanisms (such as inducing parents to give preferential treatment to the heavier child, causing that child to become more prosperous than its twin).
This example prompts the question, valid for the latter studies, as well, of which proxies to pick for risk preference and time discounting among the hundreds potentially available. Why use participation in the stock market when this is correlated with many other things that are not risk preferences? Similarly, why use literacy rates in early modern Europe as evidence of diminished time discounting rather than, say, pointing to the reduced rate of cathedral building as evidence in the other direction? The willingness to design and build cathedrals has as good a claim as any other to be a proxy for long-term thinking. It is hard to avoid the suspicion that, if cathedral building had appeared to favor the Baumard hypothesis, it would have been cited, but because it does not, it was not.
The very wealth of possible correlates of the unobserved variables of Life History Theory means that Baumard several times loses sight of the aim of the exercise, which is to explain why England had its Industrial Revolution before any other country. In using literacy rates as evidence of lower time discounting, he cites the fact that England had higher literacy than France, while treating as of minor importance that England had lower literacy than Germany or Scandinavia. Similarly, he makes a great deal of Manuel Eisner's (Reference Eisner2001) evidence that English homicide rates began falling in the later Middle Ages, and of the fact that “on the eve of the Industrial Revolution … England was still ahead of the rest of Europe” (sect. 5.3, para. 1) – but he cannot have it both ways. What matters for his hypothesis is the change in homicide rates over time, and this occurred spectacularly across Europe (including in Italy), even if in England these were lower than elsewhere from the sixteenth to the nineteenth centuries.
This is not to dispute the plausibility of the hypothesis, but if rates of time discounting were falling everywhere in Europe, as they probably were, this phenomenon will not explain the English origins of the Industrial Revolution. And several things make me doubt their relevance to the question of English origins. One is that innovation is the fruit of a tiny subset of the population, whose affluence had been little correlated with the general living conditions of the population prior to the eighteenth century. The French upper classes were probably more affluent than their English counterparts in the sixteenth through eighteenth centuries, so it is not a lack of affluence that explains their unwillingness to engage in innovation.
A second reason for skepticism is that England's greater rate of urbanization produced affluence for those who survived childhood but at the cost of horrendous rates of infant mortality for those who grew up in the cities. It is not clear why Life History Theory should focus only on the affluence of those who survive to adulthood and not on the ex ante improbable fact of their survival.
Most importantly, the dominant fact about the modern world is that sooner or later the Industrial Revolution took place everywhere. As Baumard points out, quoting Deirdre McCloskey, the increase in prosperity it brought about was massive, far greater than it seems reasonable to attribute to any single cause. Even if increases in risk taking preceded increases in innovation, a claim that is at best unproven, the extent of innovation seems out of proportion to the alleged cause. This very strongly suggests that we are seeing the mutual and cumulative reinforcement of several different causes. That one such cause was the change in attitudes predicted by Life History Theory is a reasonable conjecture, and Baumard's paper, although always stimulating, would have seemed more convincing to me if it had been framed in this light.