«I am mightily impressed with efforts to more tightly measure the development of industry in the British industrial revolution. However, after reading ...»
Thoughts On The Industrial Revolution
Jack A. Goldstone
I am mightily impressed with efforts to more tightly measure the
development of industry in the British industrial revolution. However, after
reading the following essays:
Nicholas Crafts, “Productivity Growth in the Industrial Revolution: A new
Growth Accounting Perspective”
C. Knick Harley, “Cotton Textiles and the Industrial Revolution”
Ian Inkster, “Potentially Global: A Story of Useful and Reliable Knowledge and Material Progress in Europe circa 1474-1914” Margaret Jacob and Larry Stewart, Practical Matter: Newton’s Science in the Service of Industry and Empire 1687-1851 Joel Mokyr, “The Great Synergy: the European Enlightenment as a factor in Modern Economic Growth.” Peer Vries, “Is California the Measure of all things Global? A rejoinder to Ricardo Duschesne, ‘Peer Vries, the Great Divergence, and the California School.’”,
I feel I should respond to several issues:
(1) “New industries” vs. “general improvement” and Total Factor Productivity (TFP) Crafts and Harley both examine the extent to which productivity gains in ‘new sectors’ and cotton in particular contributed to overall economic growth in from 1760 to 1860. Harley argues that the spectacular gains in 1 cotton had few spillovers, in that the technological improvements in cotton production were quite specific to that industry. He thus argues one must look elsewhere than technological change to explain the overall growth in the English economy in this period. Crafts looks more broadly at several ‘modern sectors’ – cotton, wool, shipping, railways, iron, and canals – and at the impact of steam power. He too finds that the impact of modern industries on overall economic growth was quite small before 1830, and that steam power in particular contributed little to overall growth until after that date.
Crafts does, however, find that the modern industries contributed significantly to such improvements in labour productivity as did occur in the industrial revolution. Decomposing total labour productivity growth of.78% per year in Britain from 1780 to 1860, he finds that capital deepening and TFP gains in the modern sectors contributed.46%, while corresponding gains in the ‘other’ sectors contributed.32%. Thus although the modern sectors were only a small portion of the total economy, they contributed well over half of total gains in labour productivity. Thus Crafts concludes that it is ‘perfectly feasible … to regard technological innovation as responsible for the acceleration in labour productivity growth that marked … the industrial revolution.” Nonetheless, much of the productivity growth during this period, and most of the economic growth, came from ‘other’ sectors. In addition, according to Crafts’ own calculations, increases in the rate of TFP growth did not become substantial until after 1830 (.75% after that date vs..3% earlier), and prior to 1830 the largest contributions to overall economic growth came neither from capital deepening nor increases in TFP but from labour force growth.
While I endorse Crafts’ conclusion on the importance of technological innovation to gains in labour productivity, I nonetheless believe these 2 analyses are wrongly conceived. Crafts, Harley, and others have demonstrated that the overall impact of the ‘modern sectors’ on total output remained quite low up to 1830. Thus if cheaper cotton, cheaper iron, and cheaper coal (which were the main effects of increased steam power and other technological improvements) had any effects on the aggregate economy, it was not through increased output of those commodities or expansion of those industries, but through the effect of lower prices of these items on the rest of the economy. To the extent that new industries contributed important inputs (in higher quality and quantity at constant or lower marginal prices) to older industries, their effect was more generally dispersed. Thus cheaper wrought iron helped all industries in which metal tools were used (ploughs and sickles to saws and machine tools) or metal was an input (nails, cutlery, buckles, buttons). Expanding supplies of coal at constant prices supplied metal working, residential heating, but also pottery, breweries and brick and tile. Cheaper and plentiful cotton yarn and cloth stimulated the traditional trades for dying, finishing, tailoring, printing, warehousing, merchandising, frame-knitting, lacemaking, etc. of cotton. And the expanding import of raw cotton and the export of cotton and metal manufactures had impacts on shipping, insurance, brokerage, and information (newpapers, coffeehouses) among other trades.
Part of this is simply arithmetic. Consider the expansion of the iron industry from 1800 to 1830. Improvements in technology between 1800 and 1830 allowed the iron industry to triple its output of pig iron from 250,000 tons in 1800 to 750,00 tons by 1830 while the price of bar iron fell in this same period by 60%. Two main technical improvements – the widespread use of coke for smelting and the use of steam engines to power the blast in furnaces – combined with lower costs of inputs (mainly coal) to produce this increased volume of output at reduced prices. However, if we examine the 3 total value in pounds sterling that iron output itself contributed to economic growth in this period, it is tiny – total growth in the value of iron output in this period was only 20% (3 times the output at 0.4 times the price = 1.2 times total value), for an average growth rate of 0.6% per year. Not bad, but clearly not a figure that would lift an economy even if it was a large sector, which it was not. So the real impact of the changes in the iron industry would in no way appear in changes in the value of output in that industry, which was small, since increases in output and declines in price offset each other in growth accounting. The contribution to growth would instead appear in the impact on the rest of the economy of having three times as much iron available at forty percent of its previous price. The effect of this cannot be calculated simply by using constant elasticity, because the greater availability of iron at much lower prices in this period spurred a transformation in the uses and demand for iron, such that iron output tripled again in the next twenty years to over 2 million tons, while growing demand for iron was such that prices remained stable.
There were also more general increases in the technology of transport, including Macadamizing road surfaces, and improved postal coaches that greatly reduced travel time for people and information that Crafts does not include in his ‘modern sectors.’ Improvements in tools for working wood and metal expanded the capabilities of traditional trades, ranging from a host of improvements in the accuracy and capabilities of lathes, most notably Ramsden’s screw-cutting lathe of 1770, later enlarged with unprecedented accuracy and uniformity to 1/10,000 of an inch by Maudslay in 1797, to the development of circular saws for cutting and milling lumber, to Naysmith’s development of the steam hammer in 1839 which made it possible to forge much larger metal items. There were also major improvements in the production of chemicals, such as sulphuric acid in 1746 4 (by Roebuck) and soda in 1780 (by Keir). Indeed, Crafts does not include mining, machine tools, road transport, chemicals, water works, potteries, milling, threshing, and a host of other areas in his ‘modern sectors,’ even though they all had been impacted by technological improvements by 1860;
thus his estimate of the portion of labour productivity gains produced by technological change is very conservative.
Finally, and perhaps most important, Crafts and others divided GDP growth into portions due to population growth, capital deepening, and TFP.
but population growth too was endogenous. Britain’s population growth from 1760 to 1830 was highly exceptional in Europe, being far faster than in any area other than land-rich Russia. I have shown (in my 1986 paper in Population Studies, and this was accepted by Wrigley & Schofield – see their intro to the paperback edition of English Population History if you wish) that England’s population growth in this period can be attributed with good precision to a large shift (5+ years) in the age at first marriage of about 20% of the population, and this appears to be very tightly linked to shifts of population out of agriculture and to opportunities for employment in regions that were home to the ‘new industries.” Combined with a slight decrease in the death rate, the increased fertility of this fraction of the population raised the Net Reproductive Rate to unprecedented levels and created the population boom. Recent research by the Cambridge Population group shows that most of the population increase in England in this period shows up as growth (including migration) in Lancashire, the areas around Birmingham, the Northumberland Coal region, and London. Without such employment opportunities, it is unlikely that marriage would have accelerated and birth rates would have grown in this fashion. So calculated rightly, the majority of the population increase during the industrial revolution was almost certainly an endogenous product of technological change. And 5 the growth of population centres in London, Lancashire and Birmingham further stimulated more intensive effort and investment in agriculture – again not involving new technology, but reconfigurations of stock fattening and grain farming and drainage and other investments to raise output to respond to a changing market.
Part of the answer to the riddle that has arisen around the Crafts/Harley results – why was economic growth during the industrial revolution so slow? – can be found by finding the proper point of reference.
Prior to the Industrial Revolution, if one examines economic history over the prior five centuries, growth was truly slow. Real wages in most European countries were no higher in 1750 than they had been five hundred years earlier, and population growth was nil as well – England and France appear to have had as many people (over 6 and 20 million, respectively) in 1250 as they did in 1750. In other words, for the five hundred years prior to the industrial revolution, net total output growth was effectively zero. Moreover, even in relatively ‘Golden Ages,’ such as the Dutch growth experience from 1630 to 1730, total output growth, including population growth and gains in per capita income, never exceeded 1% per annum. In China, where the Qing ‘economic miracle’ is lauded for producing an unprecedented tripling of population between 1620 and 1820 with no decline in living standards, this still implies an annual growth in output of only 0.56 percent per annum.
Thus for the English economy to grow by 1.7% per annum from 1780 to 1831, even if much of that was accomplished through population increase, was a major breakthrough in proper historical perspective (the only prior period of similar population growth, in 1550-1660, involved a doubling of population but a halving of real wages, so that total net output grew only slightly if at all). By comparison, economic growth from 1700 to 1789 in France (in constant prices), a typical mature late pre-industrial economy, 6 was only 0.34 % per year, and this in an economy participating in the Atlantic trade boom and the centre of the Enlightenment. British growth, at about double that rate during the same period, was already starting to come up against the limit of previously seen pre-industrial growth rates, so for growth to accelerate further after 1780 was quite surprising.
The big reason for the apparent paradox of ‘slow growth’ in the IR was the misguided expectation that the industrial revolution would somehow take a mature pre-industrial economy (for which we would expect long term zero growth net, and positive growth periods of 1% increase per annum at best) and immediately raise its growth rate to that of a mature industrial economy (3% per year or better). In fact, the transition from mature pre-industrial economy to mature industrial economy took a little over a hundred years, during which aggregate growth rates stepped up from an initial breakthrough rate of 1.7% in 1780-1831 to 2.4% in 1831-73 and then still higher in the late nineteenth and twentieth century. But that such a transition began and took place at all was certainly, in historical terms, an economic revolution.
In sum, I would argue that one of the main impacts of technological changes in the period up to 1830 was demographic -- to spur population growth and redistribute population by increasing opportunities for wage employment in certain regions. Yet since for this added population, only some will find employment in the new industries, and for all the bulk of their consumption demand will inevitably be for products of the ‘old’ economy (food and shelter and warm wool clothing), the old economy will initially be stimulated to grow along with the new.
It seems that behind the Crafts and Harley analyses are models of the impact of technological change that see either (1) technologically advanced sectors growing wholly separate and independently from the rest of the economy if their technology is only of use in that sector, or (2) 7 technologically advanced sectors growing by displacing or squeezing out other sectors. Neither of these, however, is accurate, as technologically advanced sectors usually are initially complementary to older sectors.
Workers and managers at new jobs still need to be housed and fed and served and supplied, and that means expansion of the old sector. In addition, technical change in sectors that provide inputs to other sectors (or potential inputs) leads to changes in demand and output in those sectors as well. Just as providing a railway spur may greatly increase the production of horse-drawn carts in that region to move goods to the railhead, growth in technologically-leading sectors may promote growth in other sectors that do not directly employ the technology or even the products of those sectors, but are responding to opportunities created by advances in the leading sectors – including in this case population growth.