The myth of technological unemployment

By Gerard Jackson
web posted November 1998

Any period of prolonged unemployment or rapid technological change invariably conjures up the myth of technological unemployment, even though history and economics has long since discredited it. What is particularly worrying about the current resurrection of this economic fallacy is that it is so-called academics who are struggling to give it respectability. Greg Callaghan's article Digital Downsizing in Employment is Creating a New Social Divide (The Australian 14/2/98) epitomises all too well what now tends to pass for serious economic and social comment in the Australian media. At least it demonstrated the utter poverty of left-wing economic thinking on technology and its basic intellectual dishonesty.

The article concentrated on the alleged job-destroying evils of information technology. Callaghan claimed that there is a school of economists who are convinced that the new technology "is triggering a work-force implosion". In support of this statement Callaghan devoted most of his article to the anti-capitalist views of Professor John Shields, Paul James, John Buchanan and Jocelyn Pixley -- none of whom are economists. Shield lectures in industrial relation, James teaches politics, Pixley is a sociologist while Buchanan is the deputy director of the Australian Centre for Industrial Relations, Research and Training. Now anyone who lecturers in any of these subjects is not precluded from being an economist. But these lecturers opinions on unemployment and technology make it abundantly plain that they know little, if any, economics or economic history.

James claimed, without a shred of historical evidence or theoretical support, that the next generation will experience the "80/20" work force where only 20 per cent will be fully employed, the remainder suffering unemployment and underemployment. (This is the kind of economic illiteracy that Callaghan called "prophetic".) His only proof was the rapid rise in service sector employment in recent years. James further asserted that our "swelling army of cast-offs" were made redundant by technology, suggesting that lengthening hours of employment supported this view. He compounded his economic nonsense with the Marxist assertion that productivity gains are largely going to "corporate elites and the owners of capital".

Professor Shields showed himself to be just as dogmatic as James, categorically asserting that "the information revolution is...a net destroyer of jobs. His idea of proof was tell us that the new technology had "already devastated banking, steel, cars and manufacturing in general. Shields also thinks part-time employment and work "intensification" supports his anti-capitalist and technophobic assertions. Pixley was just as bad, implying that most new jobs in America were for "car-park attendants, trolley-pushers and waiters." (No statistical support, of course, just the usual smug ideological assertions.) When she did use statistics it was with the kind of ideological thoughtlessness that characteristise much left-wing thinking.

When Callaghan pointed out that America's unemployment rate had fallen to 4.6 per cent her response was to state that only 34 per cent of America's jobless claim unemployment benefits compared with 82 per cent in Australia; in addition, there are 1.2 million people in US prisons and 8 million on parole. Callaghan was so struck with the latter figure that he italicised it. What Pixley was obviously trying to convey the left-wing fiction that America's unemployment figures are rigged. As expected, Callaghan played along with her. What readers were not told is that America's participation rate (that part of the working-age population who are part of the labour force) is 74 per cent compared with Australia's 69 per cent. The same participation rate in Australia would see its unemployment rate rise to about 1.5 million. It should also be mentioned that America's working-age population is 4 per centage points higher than Australia's. These facts make the American employment achievement all the more remarkable. Obviously Pixley's figures on unemployment claims are totally irrelevant, nothing more than an ideological red herring and a desperate attempt to try and rescue what is basically another untenable Marxist-inspired criticism of capitalism.

As for America's prison population, there are about 1.2 million prisoners under federal and state jurisdiction. This figure not only includes those under the legal authority of the prison system but also those who are held outside its facilities, many of whom are on work release -- including parolees. In other words, parolees are included in work force figures despite Pixley and Callaghan's attempt to imply otherwise. Unfortunately Pixley seems to evince that state of mind that causes so many on the Left to treat statistics and economic reasoning with contempt. Buchanan's only apparent contribution to the article was to attack downsizing (a dreadful euphemism for sacking) for unemployment rather than technology.

Putting it simply, these people are really giving us a variation of the long-discredited Marxist thesis that predicted capitalism would bring about the increasing immiseration of the proletariat, destroy the bourgeoisie (the middle class), create a massive reserve army of the unemployed while concentrating wealth, power and industry in the hands of the capitalist few.

Before doing what these critics have evidently refused to do and that is apply economic reasoning to the subject let us first examine a few facts. If what they say is true, then those countries with the greatest investment in information technology should have the highest unemployment rate. Yet Australia's unemployment rate is 100 per cent greater than American level even though American per capita investment in industrial robots and information technology vastly exceeds Australia's. And no amount of statistical abuse by the likes of Pixley can change that fact by one jot. Singapore's per capita investment in information technology and industrial robots also greatly exceeds Australia's. The result was not rising unemployment, as these academics predicted, but a rise in the demand for labour.

Whichever way one looks at it, there is no correlation between investment in information technology, or any other kind of technology and widespread persistent unemployment. The correlation is between market wage rates and non-market wage rates. In short, large-scale widespread persistent unemployment in Australia and other countries suffering the same phenomenon has been caused by pricing people out of work. Given the kinds of labour-market rigidities and wage-fixing arrangements that cause 'permanent' unemployment one would expect the emergence of a very high level of casual and part-time work -- even self-employment -- as people are driven into suboptimal employment. This is precisely what we find. Part-time work in Austalia is now up from 10 per cent in 1966 to 25 per cent today; in Holland it is 37 per cent while in America, with its much more flexible labour markets, it is 16 per cent.

Now if James is right and productivity gains are being largely confiscated by the "corporate elites" this would show up in government statistics in the form of significantly rising profits (though this would not actually support James' argument) and in falling real gross wages. Instead we find that average corporate profits are still below the average of more than 10 per cent of GDP that prevailed from 1940 to 1970 when they began to fall.

Moreover, James is clearly implying that wages as a proportion of corporate earnings have fallen. Not so. The Economic Report to the President 1996 showing the two-way division between the net profit of corporations and the total amount going to labour revealed that labour's per centage was 91 for 1994, 90.2 for 1995 and 89.8 per cent for the first quarter of 1996. The very opposite of what James et al asserts. Compare these figures with 1949 to 1959 when labour's per centage was 86 per cent, and 1960 to 1977 when it averaged 89 per cent. Labour's per centage was clearly lower during the 1950s, 1960s and 1970s than during (so far) the technology ravaged '90s. These figures also make nonsense of his other implication that gross wages have fallen significantly behind value productivity.

A study by Bank Credit Analyst, a Canadian group, showed that real wages for American workers have actually increased in line with value productivity. What has fallen in America is not real gross wages but take-home pay. Fiscal drag and wage increases mandated by government in the form of benefits is the culprit behind any falls in net pay. Furthermore, the American economist Leonard Nakamura calculated that inflation has been overstated for more than 20 years. If correct this would mean that even real take home pay had increased.

It is easily seen that James' claim that "corporate elites" have confiscated productivity gains is just Marxist hogwash. His claim that "capital owners" shared in the looting of workers' productivity gains has a delicious ring of unconscious irony about it. The reason is that these very same workers are, through pension funds, insurance companies, shares (nearly 60 million American workers own shares), etc, the biggest owners of capital in the country.

So far all that has been demonstrated is that current claims of technological unemployment are utter nonsense. What needs to be shown, however, is why this is so. No matter what these critics publicly claim or personally believe their assertions that technology destroys jobs is based on the fallacy that capital is a substitute for labour and not a complementary factor. This belief in turn springs from the fallacy of composition, confusing the part with the whole. This is not to deny that machines do not destroy jobs -- they do. But the process by which this is done expands real purchasing power, lifts living standards and raises the demand for labour. Therefore observers have confused the destruction of certain jobs with the destruction of employment, thinking they are the same thing.

Rather than being a substitute for labour capital is a means of economising labour, the scarcest factor because it is the least specific, and making it more efficient. Not only does capital economise on labour it makes it possible for labour to produce services that were once impossible (e.g., x-ray machines, jet flight telecommunications). Some might argue that we are not discussing capital but technology. This would be to miss the point. Technology can only be applied through capital. It would be pointless having the best programmers in the world if you cannot supply them with computers. Technology, including so-called information technology , is always applied through capital which in turn is fuelled by savings. It therefore follows that savings limit the extent to which investments in technology, no matter how highly advanced or desirable, can be made.

Critics of technology, because that is what these people really are, implicitly assume that labour is specific, that it has no alternative use. Once it is dismissed from a particular line of production its services are rendered valueless. This is where a curious contradiction between some critics of capitalism emerges: there are those who argue the above and those who argue that a lowering of real wage rates reduce investment by inducing capitalists to hire more labour rather than invest in machinery. Though they appear to contradict each other the arguments are closely linked in the obvious sense that labour and capital are viewed as competing substitutes.

Let us imagine an overpopulated country in which all the land is fully utilised and capital is virtually nonexistent. Obviously the living standard of the general population would be at a bare subsistence level with little in the way of wages. According to both views it would not pay capitalists to invest. The first view claims that machines destroy jobs while the latter claims the cheapness of labour makes investment unprofitable. These views overlook the fact that machinery is primarily employed to raise output per unit of input, i.e., productivity, not to displace labour. A moment's reflection and one quickly realises that if either one were right there would never have been an industrial revolution.

Let us assume that in our imaginary country entrepreneurs calculate that by using textile machinery, let us say, can make handsome profits with little in the way of labour costs. Clearly the investment will be made in the machinery even though abundant labour is available and domestic weavers and spinners abound. Why? Because the machinery raises the value of the labourer's product by increasing his marginal product even though unit prices fall. As profits are maladjustments between supply and demand it follows that labour is undervalued in relation to the value of its product. This means the demand for labour will rise as more entrepreneurs move into the newly developing, though labour intensive, industry to compete away the profits. (I have assumed that the government does not interfere with markets.) This is basically what happened in eighteenth century England when its cotton industry took off. And the same thing happened to the Asian tigers. History demonstrates that though this process destroyed hundreds of thousands of jobs in the short term many more were created.

As I pointed out earlier on, much of the confusion stems from extending the experience of one company to that of the economy. Okay, a company decides to invest in the latest cost-cutting technology that will double output without raising total costs. If demand is elastic (sensitive to prices changes) a small cut in prices could see all of his output sold. No one is fired, prices have fallen somewhat and healthy profits are being earned. However, these profits act as a signal to other entrepreneurs who move in to compete them away. The demand for labour rises, output rises, as do costs while prices fall. Eventually stability is restored at a higher level of employment, output and investment, and profits, though not interest to the firm, disappear.

By buying the machines employment was created in their production and maintenance; more employment was directly created in the industry as it expanded to meet demand; purchasing power rose as the price of the product fell which in turn expanded the demand for other products. In addition, given unchanged time preferences, more savings would have been made available for other job-creating investments.

However, assuming that demand for the product was insufficiently elastic despite falling prices to maintain the industry's present level of employment then unemployment will appear. But note, whether labour is dismissed is determined by the shape of the demand curve for the product and not by the introduction of the labour-economising machinery. As the machinery (applied technology in reality) is of the cost-cutting type then the amount of labour in terms of payrolls that is dismissed will be greater than is used in the production and maintenance of the machinery. Hence there will be a net loss of employment in the industry.

But the effect of the new investment (and I must stress that this means new technology) is to expand total demand by expanding output at lower prices. This raises purchasing power which in turn expands the demand for the products of other industries. These industries respond by raising their demands for labour. Total demand is raised by raising total output. Therefore supplies are demands. (This is Say's Law). Money is the means by which we carry out indirect barter. So when the money price of any product falls this leaves consumers with more of their own products to exchange for the products of others. (This is basically all there is to Say's Law.) Hence as new technology expands demands it expands the demand for labour.

Obviously the central point that goods exchange against other goods, not money, and that supplies are demands is being neglected. When individuals engage in production they offer their products in exchange for the products of others. When investment raises the marginal productivity of labour more goods will be offered in exchange for other goods.

If the critics' reasoning is right then technological unemployment would go hand in hand with remarkable increases in productivity and masses of idle capital. (US productivity, incidentally, is twice that of Australia while the Australian unemployment rate is twice as high as the American rate.) This is because as companies invest in more labour- economising cost-cutting technology productivity rises and operations expand. But if the side-effect is mass unemployment then there would be a glut of goods and massive idle capacity would emerge. In other words, this kind of technology causes general over-production. But this could only happen if supplies are not demands. So long as there is sufficient land and capital to employment people then lasting mass unemployment is not possible in a free market. And as long as people's wants remain unsatisfied and the means (capital) exists to employ labour then our critics economic prediction of 80 per cent unemployment is nothing but a bogeyman.

Their argument that the above analysis only applies to what has passed because new technology is qualitatively different does not hold up. The same technophobic argument was used in the 1930s and 1950s without success. But their argument suggests that productivity increases during the Industrial Revolution must have been comparatively small otherwise the period would have been market by mass unemployment caused by rapid rises in productivity. A cluster of industrial innovations saw rapid strides productivity and employment that continued throughout the century. For example, virtually overnight Henry Cort's process raised the prooductivity of bar iron production by 1400 per cent. Between 1873 and 1886 the price of Bessemer steel fell by 75 per cent while the productivity of Bessemer furnaces rose by 400 per cent. These are remarkable productivity increases.

The English textile industry is another graphic example that should have caused mass unemployment. A variety of inventions and innovations led to a massive rise in productivity. It was estimated that by 1812 the productivity of a spinner was 2000 per cent greater than in 1770. So great were the increases in productivity by the 1800s that some workers, fearing technological unemployment, resorted to machine-breaking, even though the expansion of employment in the industry had been spectacular.

Without labour-economising technology the American telephone system would collapse. In 1972 it was estimated that using 1900 technology 20 million operators would have been needed to handle the volume of calls. Taken at face value technology had destroyed nearly 20 million jobs in this sector alone. Now it is estimated that current American telephone traffic uses so much computer power that if it were done manually the number of operators would exceed the numbers generating the traffic. So where did all the operators go? To other jobs, every one. That's where.

There is no point in critics asserting that information technology is qualitatively different when they do not even make the slightest attempt to explain why. Information is only useful to industry when it allows a more efficient allocation of resources. Any technology that allows such valuable information to be more efficiently and rapidly disseminated is to be welcomed, not condemned. On the other hand, I suspect that what is really being attacked is the microprocessor -- the very device that has given birth to a number of new industries and the hundreds of thousands of jobs it created. Only God knows how many future industries and jobs will owe their existence to this astonishing device. Finally we come to technology and income. That some groups can suffer in the short term from the introduction of technology is indisputable -- a similar thing happens with shifts in demand -- but the principle effect is to raise overall income in the longer term. However, it is not the existence of the technology per se that raises income but, as already stated, its application through investment in capital goods. These goods form an integrated heterogeneous structure consisting of stages of production. As the structure becomes longer, more complex and productive it raises the marginal productivity of labour. This phenomenon can conceal some surprising facts. Though only about 3.5 million Americans are directly employed in agriculture, 20 million or more are directly and indirectly employed in the food industry of which agriculture is the highest stage. Seeing the economy as integrated stages of production, as we should, rather than isolated sectors, as is frequently the case, helps put things like information technology in their proper perspective because it forces us to seek out economic linkages.

If an advanced economy is experiencing a significant shift in income from wages to capital, this suggests that the population is either expanding at a faster rate than investment or the production structure is shrinking. In either case the result would be falling wages. That real wage growth in America has slowed since the '70s should be cause for alarm but not surprise. It was in the '70s that capital investment as a ratio of the labour force began to fall. Though it would require another article, I believe that regulatory costs (estimated at $US688 billion for 1997) and the taxing, big-spending, Keynesian policies of successive American governments have combined to greatly retard the country's economic performance. But these are thoughts that our left-wing ideologues resolutely refuse to entertain.

Current Issue

Archive Main | 1998

E-mail ESR




1996-2020, Enter Stage Right and/or its creators. All rights reserved.