To those that have shall be given

JUST ACROSS THE road from Gothenburg’s main railway station, at the foot of a pair of hotels, a line of taxis is waiting to pick up passengers. The drivers, all men, many of them immigrants, chat and lean against their vehicles, mostly Volvos. One of them, an older man with an immaculate cab, ferries your correspondent to Volvo’s headquarters on the other side of the river. Another car is waiting there, a gleaming new model with unusual antennae perched on two corners of its roof. An engineer gets in and drives the car onto a main commuter route. Then he takes his hands off the wheel.

Volvo, like many car manufacturers, is putting a lot of work into automated vehicle technology. Such efforts have been going on for some time and were responsible for the development of power steering, automatic transmissions and cruise control. In the 2000s carmakers added features such as automated parallel parking and smart cruise, which can maintain a steady distance between vehicles. In 2011 Google revealed it was developing fully autonomous cars, using its detailed street maps, an array of laser sensors and smart software. It recently unveiled a new prototype that can be configured to have no driver controls at all, save an on/off button. Traditional car manufacturers are taking things more slowly, but the trend is clear.

In many ways driverless cars would be a great improvement on the driven variety. Motoring accidents remain one of the leading causes of death in many countries. Automated driving promises huge improvements in both fuel efficiency and journey times and will give erstwhile drivers the chance to do other things, or nothing, during their trip.

Yet its effect on the labour market would be problematic. Only ten years ago driving a car was seen as the sort of complex task that was easy for humans but impossible for computers. Driving taxis, delivery vans or lorries has been one of the few occupations in which people without qualifications could earn a decent wage. Driverless vehicles could put an end to such work.

The apocalypse of the horsemen

Before the horseless carriage, drivers presided over horse-drawn vehicles. When cars became cheap enough, the horses and carriages had to go, which eliminated jobs such as breeding and tending horses and making carriages. But cars raised the productivity of the drivers, for whom the shift in technology was what economists call “labour-augmenting”. They were able to serve more customers, faster and over greater distances. The economic gains from the car were broadly shared by workers, consumers and owners of capital. Yet the economy no longer seems to work that way. The big losers have been workers without highly specialised skills.

The squeeze on workers has come from several directions, as the car industry clearly shows. Its territory is increasingly encroached upon by machines, including computers, which are getting cheaper and more versatile all the time. If cars and lorries do not need drivers, then both personal transport and shipping are likely to become more efficient. Motor vehicles can spend more time in use, with less human error, but there will be no human operator to share in the gains.

At the same time labour markets are hollowing out, polarising into high- and low-skill occupations, with very little employment in the middle. The engineers who design and test new vehicles are benefiting from technological change, but they are highly skilled and it takes remarkably few of them to do the job. At Volvo much of the development work is done virtually, from the design of the cars to the layout of the production line. Other workers, like the large numbers of modestly skilled labourers that might once have worked on the factory floor, are being squeezed out of such work and are now having to compete for low-skill and low-wage jobs.

Labour has been on the losing end of technological change for several decades. In 1957 Nicholas Kaldor, a renowned economist, set out six basic facts about economic growth, one of which was that the shares of national income flowing to labour and capital held roughly constant over time. Later research indicated that the respective shares of labour and capital fluctuate, but stability in the long run was seen as a good enough assumption to keep it in growth models and textbooks. Over the past 30 years or so, though, that has become ever harder to maintain as the share of income going to labour has fallen steadily the world over.

Recent work by Loukas Karabarbounis and Brent Neiman, of the University of Chicago, puts the global decline in labour’s share since the early 1980s at roughly five percentage points, to just over half of national income. This seems to hold good within sectors and across many countries, including fast-growing developing economies like China, suggesting that neither trade nor offshoring are primarily responsible. Instead, the two scholars argue, at least half of the global decline in the share of labour is due to the plummeting cost of capital goods, particularly those associated with computing and information technology.

By one reckoning the price of cloud-computing power available through Amazon’s web services has fallen by about 50% every three years since 2006. Google officials have said that the price of the hardware used to build the cloud is falling even faster, with some of the cost savings going to cloud providers’ bottom lines rather than to consumers—for now, at any rate. The falling cost of computing power does not translate directly into substitution of capital for labour, but as the ICT industry has developed software capable of harnessing these technologies, the automation of routine tasks is becoming irresistible.

From the end of the second world war to the mid-1970s productivity in America, measured by output per person, and inflation-adjusted average pay rose more or less in tandem, each roughly doubling over the period. Since then, and despite a slowdown in productivity growth, pay has lagged badly behind productivity growth. From 2000 to 2011, according to America’s Bureau of Labour Statistics, real output per person rose by nearly 2.5% a year, whereas real pay increased by less than 1% per year.

The counterpart to this eclipse of labour is the rise and rise of capital. In a landmark book that became an unlikely bestseller, Thomas Piketty, an economist at the Paris School of Economics and an authority on inequality, argues that economics should once again focus on distribution, as it did in the 19th and early 20th centuries. In those days the level of wealth in rich economies often approached seven times annual national income, so income earned from wealth played an enormous part in the economy and caused social strains that sometimes threatened the capitalist system. In the decades following the first world war old fortunes were wiped out by taxation, inflation and economic collapse, so by 1950 wealth in rich economies had typically fallen to just two or three times the level of annual national income. But since then it has begun to creep up again.

Mr Piketty acknowledges that inequality today is different from what it was 100 years ago. Today’s great fortunes are largely in the hands of the working rich—entrepreneurs who earned billions by coming up with products and services people wanted—rather than the idle gentry of the early industrial era. Yet even if the source of the new wealth is less offensive than that of the old, the eclipse of labour could still become a disruptive social force. Wealth is generally distributed less equally than capital; many of those getting an income from work own little or no wealth. And Mr Piketty reckons that as wealth plays a bigger part in an economy, it will tend to become more concentrated.

The decline in the role of wealth in the early part of the 20th century, Mr Piketty observes, coincided with a levelling out of the wealth distribution, as for the first time in modern economic history a broad, property-owning middle class emerged. That middle class has been a stabilising force in politics and society over the past 70 years, he reckons. If it were to disappear, politics could become more contentious again.

Labour in America would have lost out to capital even more dismally except for soaring pay among a small group of high earners, according to a study in 2013 by Michael Elsby, of the University of Edinburgh, Bart Hobijn, of the Federal Reserve Bank of San Francisco, and Aysegul Sahin, of the Federal Reserve Bank of New York. The typical worker has fallen behind even more than a straightforward look at the respective shares of labour and capital suggests.

One explanation for that is the changing nature of many jobs. In recent years economists such as David Autor and Daron Acemoglu of the Massachusetts Institute of Technology have pioneered a new way of looking at work: analysing occupations in terms of the tasks they involve. These can be manual or cognitive, routine or complex. The task content determines how skilled a worker must be to qualify for work in a particular occupation. Mr Autor argues that rapid improvement in ICT has enabled firms to reduce the number of workers engaged in routine tasks, both cognitive and manual, which are comparatively easy to programme and automate.

A manufacturing worker whose job consists of a clear set of steps—say, joining two sheets of metal with a series of welds—is highly vulnerable to being displaced by robots who can do the job faster, more precisely and at lower cost. So, too, is a book-keeper who enters standard data sets and performs simple calculations. Such routine work used to be done by people with mid-level skills for mid-range pay. Over the past generation, however, technology has destroyed large swathes of work in the middle of the skill and wage distribution, in a process economists call labour-force polarisation.

The hole in the middle

As recently as the 1980s demand from employers in rich countries was most buoyant for workers with a college education, less so for those with fewer qualifications and least so for those who had at best attended high school. But from the early 1990s that pattern changed. Demand still grew fastest for skilled workers and more slowly for less-skilled workers, but the share of employment in the middle actually shrank. In the 2000s the change became more pronounced: employment among the least-skilled workers soared whereas the share of jobs held by middle- and high-skill workers declined. Work involving complex but manual tasks, like cleaning offices or driving trucks, became more plentiful. Both in America and in Europe, since 2000 low-skill, low-productivity and low-wage service occupations have gained ground.

Highly skilled work, on the other hand, has become increasingly concentrated in jobs requiring complex cognitive or interpersonal tasks: managing a business, developing a new product or advising patients. As non-routine work has become more prized, supply and demand in the labour market have become increasingly unbalanced. Many cognitively complex jobs are beyond the abilities even of people with reasonable qualifications. The wage premium for college graduates has held steady in recent decades, but that is mainly because of the rising premium earned by holders of advanced degrees. The resulting competition for lower-level work has depressed wage growth, leading to stagnant pay for typical workers.

Technology has created a growing reservoir of less-skilled labour while simultaneously expanding the range of tasks that can be automated. Most workers are therefore being forced into competition both against each other and against machines. No wonder their share of the economic pie has got smaller, in developing economies as well as in the rich world.