“HAS the machine in its last furious manifestation begun to eliminate workers faster than new tasks can be found for them?” wonders Stuart Chase, an American writer. “Mechanical devices are already ousting skilled clerical workers and replacing them with operators...Opportunity in the white-collar services is being steadily undermined.” The anxiety sounds thoroughly contemporary. But Mr Chase's publisher, MacMillan, “set up and electrotyped” his book, “Men and Machines”, in 1929.

The worry about “exporting” jobs that currently grips America, Germany and Japan is essentially the same as Mr Chase's worry about mechanisation 75 years ago. When companies move manufacturing plants from Japan to China, or call-centre workers from America to India, they are changing the way they produce things. This change in production technology has the same effect as automation: some workers in America, Germany and Japan lose their jobs as machines or foreign workers take over. This fans fears of rising unemployment.

What the worriers always forget is that the same changes in production technology that destroy jobs also create new ones. Because machines and foreign workers can perform the same work more cheaply, the cost of production falls. That means higher profits and lower prices, lifting demand for new goods and services. Entrepreneurs set up new businesses to meet demand for these new necessities of life, creating new jobs.

As Alan Greenspan, chairman of America's Federal Reserve Bank, has pointed out, there is always likely to be anxiety about the jobs of the future, because in the long run most of them will involve producing goods and services that have not yet been invented. William Nordhaus, an economist at Yale University, has calculated that under 30% of the goods and services consumed at the end of the 20th century were variants of the goods and services produced 100 years earlier. “We travel in vehicles that were not yet invented that are powered by fuels not yet produced, communicate through devices not yet manufactured, enjoy cool air on the hottest days, are entertained by electronic wizardry that was not dreamed of and receive medical treatments that were unheard of,” writes Mr Nordhaus. What hardy late 19th-century American pioneer would have guessed that, barely more than a century later, his country would find employment for (by the government's latest count) 139,000 psychologists, 104,000 floral designers and 51,000 manicurists and pedicurists?

Even relatively short-term labour-market predictions can be hazardous. In 1988, government experts at the Bureau of Labour Statistics confidently predicted strong demand in America over the next 12 years for, among others, travel agents and petrol-station attendants. But by 2000, the number of travel agents had fallen by 6% because more travellers booked online, and the number of pump attendants was down to little more than half because drivers were filling up their cars themselves. Of the 20 occupations that the government predicted would suffer the most job losses between 1988 and 2000, half actually gained jobs. Travel agents have now joined the government's list of endangered occupations for 2012. Maybe they are due for a modest revival.

The bureau's statisticians are now forecasting a large rise in the number of nurses, teachers, salespeople, “combined food preparation and serving workers, including fast food” (a fancy way of saying burger flippers), waiters, truck drivers and security guards over the next eight years. If that list fails to strike a chord with recent Stanford graduates, the bureau also expects America to create an extra 179,000 software-engineering jobs and 185,000 more places for computer-systems analysts over the same period.

Has the bureau forgotten about Bangalore? Probably not. Catherine Mann of the Institute for International Economics points out that the widely quoted number of half a million for IT jobs “lost” to India in the past couple of years takes as its starting point the year 2001, the top of the industry's cycle. Most of the subsequent job losses were due to the recession in the industry rather than to an exodus to India. Measured from 1999 to 2003, the number of IT-related white-collar jobs in America has risen (see chart 6).

Ms Mann thinks that demand will continue to grow as falling prices help to spread IT more widely through the economy, and as American companies demand more tailored software and services. Azim Premji, the boss of Wipro, is currently trying to expand his business in America. “IT professionals are in short supply in America,” says Mr Premji. “Within the next few months, we will have a labour shortage.”

If that seems surprising, it illustrates a larger confusion about jobs and work. Those who worry about the migration of white-collar work abroad like to talk about “lost jobs” or “jobs at risk”. Ashok Bardhan, an economist at the University of California at Berkeley, thinks that 14m Americans, a whopping 11% of the workforce, are in jobs “at risk to outsourcing”. The list includes computer operators, computer professionals, paralegals and legal assistants. But what Mr Bardhan is really saying is that some of this work can now also be done elsewhere.

What effect this has on jobs and pay will depend on supply and demand in the labour market and on the opportunity, willingness and ability of workers to retrain. American computer professionals, for instance, have been finding recently that certain skills, such as maintaining standard business-software packages, are no longer in such demand in America, because there are plenty of Indian programmers willing to do this work more cheaply. On the other hand, IT firms in America face a shortage of skills in areas such as tailored business software and services. There is a limited supply of fresh IT graduates to recruit and train in America, so companies such as IBM and Accenture are having to retrain their employees in these sought-after skills.

Moreover, Mr Bardhan's list of 14m jobs at risk features many that face automation anyway, regardless of whether the work is first shipped abroad. Medical transcriptionists, data-entry clerks and a large category of 8.6m miscellaneous “office support” workers may face the chop as companies find new ways of mechanising paperwork and capturing information.

Indeed, the definition of the sort of work that Indian outsourcing firms are good at doing remotely—repetitive and bound tightly by rules—sounds just like the sort of work that could also be delegated to machines. If offshoring is to be blamed for this “lost” work, then mechanical diggers should be blamed for usurping the work of men with shovels. In reality, shedding such lower-value tasks enables economies to redeploy the workers concerned to jobs that create more value.

Stuart Chase understood the virtuous economics of technological change, but he still could not stop himself from fretting. “An uneasy suspicion has gathered that the saturation point has at last been reached,” he reflected darkly. Could it be that, with the invention of the automobile, central heating, the phonograph and the electric refrigerator, entrepreneurs had at long last emptied the reservoir of human desires? He need not have worried. Today's list of human desires includes instant messaging, online role-playing games and internet dating services, all unknown in the 1920s. And there will be many more tomorrow.

Copyright © 2004 The Economist Newspaper and The Economist Group. All rights reserved.