If there’s a topic which has generated debate in economic history, that is the Industrial Revolution and its effects on living standards. Fierce academic debates have developed around the issue of how the early stages of modern capitalist development led to an improvement or a decline in worker’s niveau de vie (Voth, 2004). Marxist historians as Hobsbawm argued that in the first century of the Industrial Revolution in England, the working class saw no improvement in their standard of living due mainly to longer working hours, devastating sanitary conditions due to overcrowding in factories and greater inequalities between capital and labour. However, some economic historians have taken a more optimistic view of the effects on living standards of the early stages of the Industrial Revolution and have tried to demonstrate improvements in those by measuring variations of real wage levels and even changes in welfare through alternative indicators to income. Since the 1970s income as a measure of living standards has been heavily criticized in academia, mainly due to income being just an input for welfare and not an output by itself, with its declining marginal utility playing an essential role in giving greater relevance to alternative indicators. Innovation in cliometrics and the adaptation of research techniques in economic history to it brought to the centre stage anthropometric evidence as a valuable resource to establish trends in living standards (Voth, 2004). Several studies have used height as a measure of net nutritional status and as a variable closely correlated to living standards from birth to age 25, in attempts to analyse the standard of living of the working class from 1750 to 1850, which can be interpreted as the first century of the Biritsh Industrial Revolution. However, even after decades of research, conclusions from these analyses are quite divergent. Even though the original intention was to construct reliable techniques to analyse living standard trends through anthropometric evidence’s analysis, this has proven to present several flaws and inconsistencies, due mainly to the scarce, biased and sometimes inconsistent data available from that epoch. Even though conclusions from this evidence are not robust, if analysis is done taking into account several biases of the data and implementing modern data analysis techniques, as the introduction of data dummies to give greater consistency to data series, we can obtain certain robust trends about living standards at the time and present some conclusions.
In this essay I will briefly review, analyse and sometimes criticize some very relevant works on living standards during the early stages of the Industrial Revolution, based on anthropometric evidence. Firstly, I’ll try to answer the question regarding if anthropometric evidence is at all valid as a measurement of living standards, presenting some of its flaws and how economic historians as Cinnirella (2008), Oxley and Horrell (2009) or Bodenhorn et al. (2017) have tried to compensate for these flaws and present some of their conclusions, which sometimes diverge. Finally, I will put all this research in perspective and analyse if we can obtain some king of general conclusion from these works, regarding living standard trends during the early phases of the Industrial revolution.
Firstly, Cinnirella (2008) finds anthropometric evidence more valuable than trends in real wages to analyse living standards at the time due mainly to the lack of data about income and the unreliability of some of that information. Cinnirella (2008) gives great relevance to height due to being a measure of the net nutritional status of a person throughout its development period, with external events as pandemics, wars or work stress affecting this development and being reflected in final height data. However, we can’t completely reject income data when employing anthropometric evidence to analyse living standards, as the relationship between income and height is many times positive and non-linear, apart from hard to disentangle, which causes a serious sample-bias when selecting height data to analyse. However, in some cases, the relationship between income and height data can be invalidated when the effect of a certain pandemic or a general decline in food quality affects all the population, as Cinnirella (2008) shows. As surprising as it may seem, this fact has even led to some studies pointing to an inverse relationship between height and income. As none of these conclusions is definite and unique, this puzzling evidence has driven to the “industrial growth puzzle”, where despite rising income per capita, average height declined in several European countries at the time. Other authors as Bodenhorn, Guinnane and Mroz (2017) have tried to solve this puzzle, or at least provide some logical consistency to it by questioning the reliability of the data that presents an apparent decline in height for several European countries in the 1750-1850 period, as is the case with Great Britain, Sweden and most of central Europe. The coincidence in height data collection between all these countries is that they all collected height data from volunteer military ranks rather than conscripts. A volunteer sample entails that those measured for height are those individuals who personally chose to enlist in the army, which can lead to severe sample-biases when analysing. One of the problems comes from incentives to join the military, because as the economy develops and incomes rise, historically, the portion of the population willing to join the army becomes smaller, as military service becomes a less attractive option to the most productive people. So, one justification Bodenhorn et al. (2017) give for questioning the reliability of the conclusions presented by researchers analysing height data from countries with armies formed by voluntaries is that military heights declined mainly because tall people, who normally had better economic and educational status at the time, increasingly opted for other career paths different to the military. This is supported by the fact that “height puzzles” are less frequently observed in those nations that filled their ranks through conscription at the end of the XVIIIth century, from which researchers can obtain more diverse and less income or class biased height data.
Data selection problems when dealing with anthropometric evidence from the early Industrial Revolution period are also found in data obtained from prison samples, as these over-represent the poor and working classes at the time, due to unobserved characteristics that made them more prone to criminal activity (Bodernhorn et al., 2017). This is a problem when trying to derive a general trend of heights from the available data, as there’s no general height register for the time, and those registers available incur in severe sample-biases. However, from this data we can obtain certain conclusions for those groups who were notoriously represented in these samples (army and prisons): the poor working class. Bodenhorn et al. (2017) show that the industrialization “puzzle” is also present in the United States, where the pattern of declining heights from 1750 up to 1850 is puzzling because it reacts inversely to what conventional indicators implied at the time, which was that the American economy was growing and developing rapidly, a similar scenario to the one experienced in England, with the surprising inverse relationship at the time between economic growth and average stature.
Some explanations to the industrialization puzzle can be obtained from paying greater attention to basic factors. For example, declines in the availability of foodstuffs due to an increase in their relative price led to a downward trend of the net nutritional status of the population. Apart from it, a direct consequence of industrialization in the short term, as is widely known, was an increase in diseases and a worsening of basic sanitary conditions due to overcrowding of cities and ventilation issues in factories and house buildings, where workers lived. This affects negatively the average height measure, because sanitary conditions and higher relative prices of food had a greater negative effect on poor workers’ heights than the positive marginal effect that economic growth had on middle and upper classes heights. So, due to composition effect, the average height trend went decisively downwards at the time, irrespectively of rising income per capita. By carefully observing the data we can even perceive how height variations oscillate when analysing height trends by employment. For example, due to the extreme work intensity at industry at the time, the average height of young factory workers suffered much more than that of farmers or white-collar workers, which can be another clue to disentangle height data and eliminate certain biases when analysing it, providing us with more robust and maybe more conclusive anthropometric evidence from the time.
On the other hand, alternative explanations are given to the industrialization puzzle by pointing to severe measurement flaws. Ewout Depauw and Deborah Oxley (2019), in their paper Toddlers, teenagers, and terminal heights: the importance of puberty for male adult stature, Flanders, 1800-76, argue that adult stature doesn’t fully capture living standards at birth but is much better in signalling living conditions over adolescence growth years, due to being this period the most influential on terminal stature, specially from ages 11 to 18. Depauw and Oxley (2019) contradict the foetal origins hypothesis, which argues that nutritional status during pregnancy is the one that affects development in a greater way and is consequently reflected in adult terminal height. However, they believe that evidence points to disease environment, nutritional intake and sanitary conditions during the central puberty growth years being much more explicitly reflected on terminal height measurements than the standard of living of toddlers. Puberty is an essential period for determining terminal height, as it is a growth catch-up period, meaning that if growth was disrupted due to nutritional or health insults during early childhood, lost growth could be at least partly recovered if living standards improved during puberty years, with teenage boys in the late XVIIIth and early XIXth century being particularly sensitive to socioeconomic conditions for growth, as they had greater calory requirements than female teenagers (Depauw and Oxley, 2019). This is the main reason for the authors’ innovation in measuring height and living conditions at the time, by organizing data series differently in terms of how final height at different ages can relate to exposure to economic and health conditions at different moments throughout the growth period. They study this by collecting data from the prison of Bruges, justifying this as a suitable study source despite the already explained biases of prison registers, arguing that prisoners’ specific group reflected mainly conditions of the poor working class. To obtain long term results of health and welfare effects on growth and prevent temporary economic shock from affecting these results, Depauw and Oxley (2017) employ annual variations in prices and mortality rates to disentangle the more generalized connections with macroeconomic conditions.
Through this essay, I haven’t yet presented the various authors’ results and numerical conclusions, because they sometimes diverge and present different pictures of living standards at the time of the Industrial Revolution. These results aren’t valid for our analysis if before we don’t pause and employ some time in trying to understand and comprehend their different methodologies, and overall, the reasons they give for using their particular methodology and the flaws they present. Once this has been understood, we can now concentrate, at least partially, in analysing the results presented by the authors compiled in this essay’s bibliography, putting trends into context and observing the complexity and near impossibility of obtaining a single and solid conclusion of living standards at the time. However, this has never been the intention of these various studies, but to confront methodologies and lead to advances in economic history’s quantitative analysis.
By looking at results, Voth (2004) finds that average heights for the period 1760-1830 increased by 3.3 cm, from 167.4 cm up to 170.7cm, falling afterwards to 165.3 cm, which leads him to argue that it is impossible to obtain a historically meaningful conclusion about living standards at the time from looking at height data while sampling biases, truncation problems in relation to army samples or general historical data deficiencies persist, which is why he decides not to present any firm conclusion as definite from anthropometric data. Other authors as Cinnirella (2008), find a declining nutritional status throughout the entirety of the period, being consistent with the rising trend in food prices in relation to wage rates. The price trend for food items rises strongly in the first half of the analysed period, concretely from 1750 up to 1800, along with declining farm labour real wages. Cinnirella (2008) gives an alternative explanation to other authors. For him, parliamentary enclosures of open fields played a very relevant role in determining the nutritional status of the British population in the early phases of the Industrial Revolution. Enclosures, along with an increasing population and a process of urbanization caused a notorious inflation of food prices, due also to the loss of common rights and allotments that these enclosures lead to, which had a direct consequence on the value of arable land, causing it to rise and translating this effect to wheat prices, making agricultural labourers more dependent on wages and more sensitive to food prices variations. Thus, we could take the worsening of the net nutritional status at the time as an endogenous consequence of the land enclosures. Apart from that, the decline of cottage industry is pointed to as an adjacent cause of the nutritional status deterioration, with more than 50% of the population living in urban centres, which directly translated into lower quality of food, higher prices and extremely low levels of sanitation; all of them being insults to growth and development. Cinnirella (2008), therefore concludes that the height trend he presents along with all the above-mentioned evidence contributes to reinforce the pessimistic view about the standard of living of the working class during the Industrial Revolution.
An alternative case to Britain’s is that of Flanders’, which is studied by Deborah Oxley and Ewout Depauw (2019), as I explained before. In their paper, they show how the existence of two crises affecting the Flemish economy (1846-1849 and 1853-1856) means that prison data of heights can be employed to investigate the impact on height of reaching puberty during a crisis, and how this is a more accurate measure of the effect of insults to net nutritional status on adult height. Mean male height in the prison of Bruges was 167.5 cm around the year 1800, being the same in 1875, with a decline in average height between the two years, being notable during the downturn periods. For those born around the later 1840s, living standards seem to have been better for them during their puberty years (coinciding with the period after the two downturns), with average height increasing for this generation in line with changes in per capita GDP. These remain in stark contrast with prisoners born in 1838, which turned eight years old in 1846 and fifteen years old in 1853, having spent four growing years during the first crisis and entering adolescent growth during the second crisis, being this the main cause why they present declining growth trends in contrast with those who were born ten years later.
In conclusion, we can agree that the core issues the anthropometric literature discusses are extremely relevant to understanding the process of modern economic growth and its effects on living standards. However, the height literature has heavily relied on sources that present severe sample biases as forms of selective sampling. So, if we wish to solidly uncover the “industrialization puzzle”, we should be aware of the consequences of the sample selection process and introduce correction mechanism for them when analysing the data. The debate on the effects of the Industrial Revolution on living standards will probably continue for many decades, mainly because there is evidence of both, improvement and worsening of living standards at the time. However, if we want anthropometric evidence to solidly contribute to clearing several unknowns, researchers must bear in mind how sample selection biases affect conclusions and interpretations.
-Voth, H.-J. (2004). «Living Standards and the Urban Environment» in R. Floud and P. Johnson, eds., The Cambridge Economic History of Modern Britain. Cambridge, Cambridge University Press. 1: 268-294
-Ewout, D. and D. Oxley (2014). “Toddlers, teenagers, and terminalheights: the importance of puberty for male adult stature, Flanders, 1800-76.” Economic History Review, 72, 3 (2019), pp. 925-952.
-Bodenhorn, H., T. W. Guinnane and T. A. Mroz (2017). “Sample-Selection Biases and the Industrialization Puzzle.” Journal of Economic History 77(1): 171-207.
-Oxley and Horrell (2009), “Measuring Misery: Body mass, ageing and gender inequality in Victorian London”, Explorations in Economic History, 46 (1), pp.93-119
-Cinnirella, F. (2008). “Optimists or Pessimists? A Reconsideration of Nutritional Status in Britain, 1740–1865.” European Review of Economic History 12(3): 325-354.