Whether the Industrial Revolution made ordinary people better or worse off is the oldest argument in the subject, and it has stayed alive because the four best kinds of evidence point in different directions.
The previous lesson put people inside the factory and noted that they walked there, which suggests it paid. This lesson asks what happened to the people who lived through the change, and it is the one place in the course where the honest answer is a set of numbers that will not reconcile. Real wages say one thing, the bodies of the people who earned them say another, and death rates say a third. Learning to hold all three at once is the point.
Why the question is hard
Start with what a standard of living measurement has to do, because most of the disagreement is buried in the method rather than in the archives.
To say a wage rose you need the money wage, which survives in wage books and contracts, and a price index to deflate it by, which has to be built from the prices of things people actually bought. The index is where the trouble is. Weight it towards bread and it says one thing; weight it towards tea, sugar and cotton cloth, which were falling in price, and it says another. The basket also changed: a family in 1850 bought goods a family in 1780 could not have bought at any price.
Three further problems have no clean solution. The money wage is a rate, not an income, so a series of daily wages tells you nothing about how many days were worked, and unemployment in this period was heavy, seasonal and unrecorded. Second, much of what a rural family consumed never passed through a market at all, so a labourer who lost common rights at enclosure and gained a cash wage may show up in the series as better off while eating worse. Third, a wage is paid to a person and a standard of living is enjoyed by a household, so anything that changed how many members of a household earned changes the answer without changing any wage.
Two traditions
The argument has a shape worth knowing, because it is one of the few historical debates where both sides were largely right about different things.
The pessimist tradition runs from Engels, whose Condition of the Working Class in England of 1845 described Manchester from the inside, through the Hammonds and to Eric Hobsbawm, who reopened the question in the Economic History Review in 1957. Its evidence is mortality, housing, food riots, and the testimony collected by parliamentary commissions. The optimist tradition, argued by Max Hartwell against Hobsbawm in the same journal in 1961 and later by Peter Lindert and Jeffrey Williamson, is built on wage series and consumption, and its strongest version, published in 1983, had blue-collar real wages roughly doubling between 1820 and 1850.
Charles Feinstein settled the wage half of it in 1998, in an article whose title, "Pessimism Perpetuated", says what he found. Rebuilding both the money wages and the cost of living index with more care, and allowing for unemployment, he got average real earnings rising by about 30 per cent across the seventy years from 1780 to 1850, with almost nothing before 1820. That is a real improvement and a slow one: compounded, is 0.38 per cent a year, a rate at which forty years of work leave a household about a sixth better off than it started.
Example. Money wages in a district rise by 50 per cent over a period while the cost of living rises by 40 per cent. What happens to the real wage, and how much can the answer be moved by the choice of price index?
The real wage index is , so real wages rose about 7 per cent. Now consider how firm that is. Suppose bread is 60 per cent of the basket and rose 55 per cent, while everything else is 40 per cent and rose 17.5 per cent, giving the 40 per cent average used above. A statistician who thought bread was only 40 per cent of spending would compute per cent inflation and report a real wage rise of per cent, nearly double. A statistician who thought bread was 75 per cent of spending would get 45.6 per cent and report 3 per cent. The same wages, the same prices, and three different answers, none of them wrong. This is why the standard of living debate outlived every attempt to settle it with a single series, and why Feinstein's contribution was as much about the deflator as about the wages.
Now you. During the French wars, from about 1790 to 1810, British money wages rose roughly 40 per cent while the cost of living rose roughly 70 per cent. What happened to real wages, and what does that do to the dating of any improvement?
Answer
Real wages fell: , a fall of about 18 per cent. So the two decades in which the cotton industry was growing fastest were decades in which the working population got poorer, because war inflation, bad harvests and the blockade of Baltic and continental grain pushed food prices up faster than wages could follow. The consequence for the dating is that whatever improvement the machinery eventually delivered arrives after 1820 at the earliest, which is why Feinstein finds almost nothing before then and why the classic period of the Industrial Revolution and the period of rising living standards barely overlap. It is worth being careful with the causation here: the war is a separate shock from industrialisation, so this is not evidence that machinery lowered wages. It is evidence that the two processes ran at once and that contemporaries had no way of telling them apart, which is part of why the political reaction described in a later lesson was aimed at machines.
Engels' pause
The most useful modern framing of the whole argument is Robert Allen's, and it turns the disagreement into a measurable gap rather than a difference of opinion.
Allen set the growth of output per worker beside the growth of real wages over the same years. Between 1780 and 1840, output per British worker rose by about 46 per cent while real wages rose by about 12 per cent. Both series go up, so both the optimists and the pessimists can quote one of them, but the interesting quantity is the distance between them. If workers produce 46 per cent more and are paid 12 per cent more, the share of output going to labour must fall, and the rest accrues to profits, which were reinvested. Allen named the period Engels' pause, after the man who described it while it was happening.
Example. Output per worker rises 46 per cent and the real wage rises 12 per cent over the same sixty years. What happens to labour's share of output, and what is the annual growth rate of each?
Labour's share is the wage divided by output per worker, so it changes by : a fall of 23 per cent in the fraction of what is produced that reaches the people producing it. Annualised over sixty years, output per worker grows at per cent a year and the real wage at 0.19 per cent. Neither rate would be visible to anyone living through it, which is exactly the difficulty: over a single working life of forty years the wage improves by about 8 per cent, less than the difference between a good and a bad year, while the country visibly fills with mills, ships and railways. That gap between what people could see being built and what they could see in their own household is a better description of the politics of the 1830s than any wage index on its own.
Now you. If real wages had instead kept pace with output per worker, what would they have done, and why might the eventual outcome have been worse?
Answer
They would have risen 46 per cent rather than 12, so the average worker in 1840 would have been about 30 per cent better off than they actually were, since . The reason this might not have been better in the long run is that the gap was not consumed by the rich in any simple sense: profits in this period were reinvested at rates that lifted Britain's investment from around 6 per cent of national income in the mid eighteenth century to around 12 per cent by the 1830s, and that investment is the mills, the railways and the iron works that eventually raised wages sharply after 1850. Allen's own reading is that the pause was the mechanism of accumulation, not an accident alongside it. Two cautions belong with that argument. It is a claim about aggregates and offers nothing to a particular family in 1830, and it is not a defence, since nothing shows the same accumulation could not have been achieved with a smaller gap. Notice also that the argument is testable in principle by looking at economies that industrialised later with different distributions, which is one reason the comparative work in the final lesson matters.
What the bodies say
Wages are a price. A height is an outcome, and it is one of the few measurements of this period that was taken on hundreds of thousands of ordinary people for reasons that had nothing to do with the argument.
Final adult height depends on nutrition and disease during growth, net of the work the body had to do at the same time. Armies, navies, prisons and charities all measured recruits, and those records let historians build a series by year of birth. Roderick Floud, Kenneth Wachter and Annabel Gregory published the major British reconstruction in 1990, and the finding that startles is the level rather than the trend: fourteen year old boys taken in by the Marine Society from the London poor in the 1770s averaged around 130 centimetres, more than 30 centimetres below a British fourteen year old today.
The trend is the part that bears on the argument, and it goes the wrong way for the optimists. Heights rose for cohorts born through the eighteenth century, then stalled and fell for those born roughly between 1820 and 1850, the very decades in which Feinstein's real wages were finally rising. The decline is of the order of an inch, it is contested in size, and the samples are awkward because armies impose minimum heights and recruit selectively from the poor. But the direction has survived a great deal of re-examination, and men born in towns were consistently shorter than men born in the countryside.
Reading the mortality evidence
The third body of evidence is death, and it comes with a trap that is worth walking into deliberately, because the same trap is set in a great deal of nineteenth century social statistics.
Edwin Chadwick's Report on the Sanitary Condition of the Labouring Population of Great Britain of 1842 tabulated the average age at death by class and place. In Manchester he gave 38 years for professional persons and gentry, 20 for tradesmen and their families, and 17 for mechanics and labourers. In rural Rutland the same three groups gave 52, 41 and 38. Liverpool was worse than Manchester, at 35, 22 and 15.
Those figures are real and the conclusion usually drawn from them is wrong. Average age at death is not life expectancy, and it is not a measure of how long an adult could expect to live.
Example. Take two populations. In the first, half of all deaths are of infants aged 1 and half are of adults aged 60. In the second, one fifth are infants and four fifths are adults at the same two ages. What is the average age at death in each, and what has changed about adult life?
The first gives years and the second years. The average age at death differs by nearly eighteen years between the two populations, and nothing whatever has changed about how long an adult lives, which is sixty in both. The entire difference is infant mortality. So when Chadwick reports 17 years for Manchester labourers, he is not saying that a Manchester labourer died at 17; he is reporting a figure dominated by the fact that a very large share of labourers' children died before their fifth birthday. The number is evidence about infant death, and it is powerful evidence, but it will mislead anyone who reads it as a lifespan.
Now you. Give a second reason why Chadwick's figure is lower for Manchester than for Rutland even if adults in the two places were equally healthy.
Answer
Because the average age at death depends on the age structure of the living population, and Manchester's was extraordinarily young. A town growing at nearly 3 per cent a year is filled with migrants in their teens and twenties who arrived, worked, and had children there, so the population contains few old people simply because it did not exist long enough to produce them, while Rutland's population was stable and therefore had a normal share of the elderly. Deaths are drawn from the living, so a young population produces young deaths whatever anyone's individual risk. The correct instrument is a life table, which asks about the probability of dying at each age separately and is therefore immune to both distortions, and William Farr was already building them at the General Register Office while Chadwick was publishing his averages. The general point is worth carrying beyond this subject: any statistic that averages over a population inherits the shape of that population, and comparisons between differently shaped populations need a measure that conditions on age.
Putting the three together
The three bodies of evidence look contradictory and are not, once each is asked the question it can answer.
Real wages measure the purchasing power of an hour of work. Heights and death rates measure what happened to a body in the place where that wage was spent. There is no contradiction in a worker being paid more, in a currency that buys more, while living in a town with no sewer, drinking water drawn downstream of a privy, working sixty-nine hours a week and losing two children in five before they were five years old. Wages rose. The environment in which they were spent deteriorated faster, for about two generations, and then improved.
That resolution has a consequence that shapes the rest of the course. If the losses were environmental rather than contractual, then no bargain between an employer and a worker could have fixed them, and neither could a higher wage. What was needed was drains, water, housing regulation and hours legislation, which is to say collective action of a kind that did not yet exist. The next lesson goes to the place where the environmental damage was worst and most measurable, which is the city, and the lesson after it to the political response.
One honest closing caution. Everything in this lesson is a national average built from unrepresentative samples, and averages hide the thing that mattered most to the people living through it, which is that the gains and the losses landed on different households. The mule spinner's family and the handloom weaver's family lived in the same town under the same wage index and had opposite experiences of the same fifty years, and E. P. Thompson's objection to the whole quantitative debate was that no index can capture what the weaver lost, which included a trade, a status and a way of ordering his own day.