See Thomas Sowell's doctoral thesis: Say's Law. An Historical Analysis. Princeton University Press, 1972.
I know I use robots continuously. I start with the iPhone that wakes me up in the morning, then read the mail I receive on my laptop, chase down on the internet the information I need to write my stuff (after all, Wikipedia is half run by machines!), send my 'copy' through the ether, so to speak, and read the newspaper where my article will be 'printed' also from far away. I well remember when I used to dictate my articles over the phone to patient tachygraphers at a cost perhaps higher than what I was paid for the piece. Still, like many others, I feel a vague disquiet at the as yet benign invasion of the robots.
To allay my fears I remind myself, first, of Say's Law and second, of the fate of horses. Let me start with Jean Baptiste Say (1767-1832). He was the French economist who introduced Adam Smith to French public opinion and to the learned classes of the European continent, most of whom read French rather than English. Say was born in a Protestant family that had taken refuge in the Swiss city of Geneva, fleeing persecution in France. The family returned to the textile city of Lyon, where Jean Baptiste was born. As a young man, he was sent by his father to England to see for himself how England was industrialising. It was not his only visit to the land of economic progress. Later in life he met and befriended the great economists of the time—David Ricardo, Thomas Malthus, James Mill, and John Ramsay McCulloch—and corrected and bettered some of the classical theories of English political economy. He published his successful Treatise of Political Economy in 1803. Napoleon invited him to dinner to try to make him correct some anti-protectionist and anti-statist passages in the book. Say, a moderate republican and a defender of individual liberty, did not comply. At different moments in his life, Say tried his hand at business—banking, insurance, cotton spinning, and sugar—variously succeeding and failing as an entrepreneur. In 1828-29 he published an ambitious Course of Political Economy in no less than six volumes. Say made signal contributions to economics. In price theory, he introduced utility into the demand side of price formation. He corrected Adam Smith by including services in the definition of wealth. Notably, he introduced the figure of the entrepreneur in the explanation of economic progress. Finally and in parallel with James Mill, he codified what is known as 'Say's Law', to which I now turn in search of hope and consolation in a world beset with robots.1
One of the fears caused by robots is that total production in the economy will so increase that demand will not be sufficient to take it up, especially if people have become unemployed due to competition from non-humans. Could this bring about a general glut of the whole system? Would there then a crisis of overproduction, which might cause more unemployment? What Say's law sets down is that the very fact of bringing goods to the market amounts to demanding goods in exchange. In the case of overproduction, the economy will tend to return to equilibrium if supply prices are flexible and are allowed to fall. This may take time, since the alarm caused by falling prices may lead to a period of money hoarding. In the long run, however and as Ricardo said, "there is no limit to demand," surprising or even shocking though this may sound to people who hate our consumptionist civilisation. This whole idea was summed up by Say in the following phrase: "Supply creates its own demand". 2
Say and his classical companions of course accepted that there could temporarily be excess production in one industry and shortages in another whose products were more demanded. This would be the case with path-breaking technological advances, such as steam threshing machines or mechanical knitting frames in the early 19th century (or robots today). Though not leading to general under-consumption, those advances would for a time bring technological unemployment in their wake, as workers were displaced by the new, more productive methods or by free trade. What is called 'sunk capital' can also be affected, so there may also be resistance put up by the owners of capital or land endangered by shifts in supply through new productive methods or through trade.
The revolts of workers who saw their wages fall because of competition from new machines are now a part of the social history of capitalism. In the years around 1810, summers were very wet in England, badly affecting crops. Also, the general blockade of trade with Britain decreed by Napoleon in Berlin and Milan in 1806 and 1807 throttled the import of grain and made bread unwontedly expensive. The wages got by threshing wheat from November to February in England gave employment in the lean winter months. Hostility to the new-fangled steam threshing machines erupted in an epidemic of machine breaking by crowds allegedly led by a legendary 'Captain Ludd'—hence the name 'Luddites' applied to workers who wanted the competition of new machinery stopped. A repeat movement happened in the summer of 1830, dubbed the 'Swing Letters'. Again, times were harsh in the hungry thirties. It was a time of political turmoil, what with the campaign for Parliamentary reform in England and the Revolution in France. Landlords who were known to use threshing machines in the Southern English counties received letters signed by a certain Captain Swing, threatening to break the threshers up and demanding higher wages.
This was not the only example of harsh technological competition suffered by manual workers. By 1830 the livelihood of some 400,000 handloom weavers was being threatened by the power-looms and knitting machines in the cotton mills. Those weavers were whole families who worked at home for merchants who 'put out' work at growingly extortionate rates, given the higher productivity of large factories. Petitions rained on Parliament. By 1840 the number of handloom weavers had fallen to 100,000, mainly specialised in silk fabrics and the more expensive cloths. The recourse to the 1834 Poor Law was felt to be a paltry remedy, since Poor Houses were organised on the principle that their standards should be less attractive than the least paid employment. A Royal Commission was formed to examine the plight of these workers, chaired by the economist Nassau William Senior. Famed for his understanding of the rapidly increasing productivity of the factory system when other classical economists saw England "within a hands breadth of the stationary state," Senior saw there was not much that could be done. Attrition had a human cost, but in the end an open door to radical technological change was the only answer. It is thus that the hungry 1830s and 40s turned into the more prosperous second half of the 19th century—helped by Robert Peel's decision in 1844 to repeal the Corn Laws. The social safety net of today is much more comfortable than Dickensian poor houses. The opportunities of gainful employment, much greater.
When I say that the plight of working horses brings me some consolation as I think of the possible effect of robots on employment, I do not mean to belittle their centuries' long contribution to the productive or destructive efforts of humanity. They were used in peaceful transport. Also, in all wars up to the middle of the 20th century, pack animals were used and sacrificed on the battlefields by their cruel and warring masters. I myself did my peacetime military service in a mounted regiment, where I learnt to ride the friendly beasts. In Hyde Park in the centre of London there stands a most affecting monument to the horses, mules, and donkeys misused by all armies in history. Anybody visiting London should take time to stop before it.
What I mean by linking the case of humans menaced by robots with that of horses displaced by the internal combustion engine is that there you have an example of job destruction by technological advance. The point I want to make is that horses could not and did not learn to drive motorcars and lorries. We men and women can learn new trades, albeit with some difficulty when we get older. Luddites are a danger to the very people they want to help because in the end progress will prevail.
Here is the crux of the matter. Technical advance makes for growth. The increasing use of robots is but a continuation of the story of triumphant capitalism. The countries of what we call the West have experienced a singular and unexpected two centuries of growth that is pulling the entire population out of poverty and into prosperity as never before in history. Now the rest of the world is joining in the ride, which we hope will last. There have been periods of growth in history that did not extend to the entire population of the world and ended in ruination. The examples that come to mind are those of Mesopotamia, Egypt, Rome, China, and the Aztec and Inca empires in America. In all of them the flame of growth died or was snuffed out by military disaster. Something quite different, however, happened in Britain around the beginning of the 19th century and then spread to continental Europe and North America. Productivity increased by leaps and bounds and was not even stopped by two world wars.
The size of these developments in the capacity to produce more with less are difficult to measure precisely. William Nordhaus, whose first claim to fame with the general public was his joint authorship with Paul Samuelson in the later editions of Economics, the textbook which has been the initiation rite of so many into the queen of social sciences. More recently, he has become one of the high priests of the new religion of climate change. Whatever the disagreements one might have with that textbook and this alarm, he is an outstanding professional economist. I want to draw attention to the striking way in which he has presented the growth of real wages from 1800. To calculate real wages he has to take nominal wages and deflate them by a price index. He finds much to query about the price indices normally used to do this calculation. He shows that price increases over the years are much exaggerated by not properly accounting for improvements in the quality of goods due to technological progress. The case he uses is that of lighting. Prices per 'lumen' (the measure of light intensity) have fallen much more quickly than shown by ordinary price indices. If the same method he uses for the true price of light is applied to general consumption (so as to take account of the incredible improvement of goods and services over the last hundred years), then real economic growth will turn out to have been much greater in the United States and Western Europe than is generally thought. "In terms of living standards, the conventional growth of real wages has been a factor of 13 over the 1800-1992 period." If the downward bias is corrected as Nordhaus thinks it should be on the lines of the cost of lighting, "real wages have grown by a factor of 970." Some growth! 3
Let me show a graph I have used in a previous column, since it comes from so respected a pen as that of Angus Maddison. It shows the sudden increase in per capita production in the United States and the United Kingdom around 1840. The inflation adjustment is of the conventional kind: the productivity surge would be much larger if we used the Nordhaus method.
For more on these topics, see the EconTalk podcast episode Gary Marcus on the Future of Artificial Intelligence and the Brain, December 2014; and I'm becoming increasingly worried about AI, by Scott Sumner, EconLog, March 14, 2017. See also "Politics and Welfare: The Political Economy of the English Poor Laws", by George R. Boyer, Library of Economics and Liberty, December 2, 2002; Industrial Revolution and the Standard of Living by Clark Nardinelli in the Concise Encyclopedia of Economics; and "Ned Ludd, Handloom Weaving, and Franco-German Moral Banking", by Anthony de Jasay, Library of Economics and Liberty, January 4, 2010.
Recently, Nobel laureate Robert J. Shiller has taken up the idea, characteristically hatched in the European Parliament, that a moderate tax should be laid on robots so as to put a brake on robot use. This tax would "slow the adoption of disruptive technologies". He seems to be worried about the effects of robots in increasing inequality and considers the tax would "accord with our natural sense of justice". The produce of the tax could then be used to finance an insurance fund to help the displaced workers who need time to change their calling. Professor Shiller seems to be another of those Nobel laureates who speaks outside his field of speciality, which in this case financial economics, not economic growth.4
What Shiller does not do is underline the trade-off between technical progress and anti-market regulation. In his future pronouncements in favour of a tax on robots he should pay some attention to worries generally expressed about the trend of economic growth in the United States and the rest of the Western world, especially Europe. I recently watched a TED talk by Professor Robert J. Gordon of Northwestern University.5 As he has been doing for a number of years, he worried about the fall in the long term growth rate of the American and the world economy due to the lack of great technical innovations. He showed us pictures of 19th and 20th century transformative innovations that, he says, are not happening any more—water closets, electrical appliances such as washing machines and refrigerators, the internal combustion engine, airplanes or telephones—but he pictured no robots. Professor Gordon may turn out to be right in discounting robots and artificial intelligence as engines of growth, if the likes of Professor Shiller have their way. Luddites all!
See Thomas Sowell's doctoral thesis: Say's Law. An Historical Analysis. Princeton University Press, 1972.
One could not think of a view more contrary to the Keynesian idea that the main macroeconomic problem in a downturn is a failure of aggregate demand. Crises caused by the failure of misapplied investment suddenly revealed by a technical shift are characterised by a general thirst for cash or money hoarding. Capitalist growth is inevitably cyclical. It can be made less volatile by trying to keep the money supply stable, as Milton Friedman proposed. But this question is for another day.
William D. Nordhaus: "Do Real Output and Real Wages Measures Capture Reality? The History of Lighting Suggests Not." NBER. PDF file.
Robert J. Shiller: "Why robots should be taxed if they take people's jobs". The Guardian, March 27, 2017.