Saturday, October 31, 2015

How digital disruption affects jobs and wages

A lot of people worry about the bad economic consequences of the digital revolution. Among the worriers is Dr Andrew Leigh, the shadow assistant treasurer and a former economics professor at the Australian National University.

Leigh made his concern clear in the "distinguished public policy lecture" he delivered this week at Northwestern University in Chicago.

But whereas most people worry that the digital revolution will lead to mass unemployment, Leigh's concern is that it will make our incomes a lot more unequal.

It's not surprising that people observe all the workers whose jobs are taken by computers and worry about widespread joblessness. As Leigh observes, this concern has been around at least since 1811, when disgruntled Nottingham textile workers wrote to factory owners under the pen-name of Ned Ludd, threatening to smash machines if they continued to be used.

But economists soon learnt not to worry. Why not? Speaking to an audience of economists, Leigh regarded it as too obvious to need explaining.

But let me fill you in. New technology leads to increased productivity – more goods and services produced per worker.

This constitutes an increase in the community's real income. When that increased income is spent, more jobs are created.

So whereas non-economists see only all the jobs that have been lost as industries X and Y digitise, economists understand this is just the most visible part of a more complex process in which jobs aren't so much destroyed as "displaced" – taken from some industries and moved to others.

This is why, after 200 years of labour-saving technological advance, we're still only up to having 6 per cent of the labour force unemployed (or about twice that if you add in underemployment).

Of course, this is the economy-wide outcome. The new jobs being created elsewhere in the economy may be very different to the jobs being lost. So this still leaves a problem for those individuals whose skills fitted the old jobs but not the new ones.

This is where Leigh comes in with his concerns about the effects of a newer idea – "skill-biased technological change" – on the unequal distribution of income between workers and, hence, families.

This is the idea that digitally driven technological change tends to disadvantage workers with less skill, and advantage those with more skill. It tends to lower wages for those with less education and raise wages for those with more education.

But the story's a bit trickier than that sounds. Research by David Autor, of the Massachusetts Institute of Technology, suggests jobs can be divided into three categories: manual, routine and abstract.

Abstract jobs – which typically involve problem-solving, creativity and teamwork – tend to be paid a lot more than manual jobs, with routine jobs – occupations such as bookkeeping, administrative support and repetitive manufacturing tasks – in between.

Autor has found that, over the past 30 years in America and the past 20 in Europe, it's routine jobs that have shrunk most. Why? Because they're the jobs that can be done most easily by a computer.

It's turned out that manual jobs – such as cooking, cleaning, being a security guard or providing personal care – are much harder for computers to do. For instance, the problem of shape recognition means that, a best, it takes a robot 90 seconds to fold a towel.

Robot hairdressers do a job similar to what you'd do if you drank a bottle of tequila and tried cut your own hair without a mirror, Leigh says.

He says the job characteristics that are hardest for computers to mimic include those involving communicating clearly with co-workers, showing empathy to clients and adapting to new situations. A lot of manual jobs require these skills.

Many studies – including some Australian ones – show that recent decades have seen a polarising or "hollowing out" of employment. There are a lot more abstract jobs (particularly managers and professionals) and modest growth in the number of manual jobs, but many fewer routine jobs in the middle.

But Leigh says this loss of mid-skill jobs doesn't mean the pain has been greatest for mid-skill workers and middle-income families.

Why not? Because what happens to wages is a product not just of the (declining) demand for mid-skill workers, but also of the supply of workers willing to do low-skilled manual jobs. And as job opportunities have declined for mid-skill workers, more of them have become willing to do manual work rather than be jobless.

So it's been wages at the bottom that have grown most slowly, not wages in the middle. (Because our wage-fixing system is more regulated, this is probably truer in the US than it is in Oz.)

At the top, Leigh says, it's altogether a different tale, with technology actually adding to the skills of the most skilful, making them more productive and so adding to their pay. A top surgeon, for instance, can use technology to do a better job and do more operations per day, thus adding to the demand for his (rarely her) services.

This may partly explain why chief executives' pay is rising, according to Leigh. The biggest firms have got bigger in recent years, and this is partly explained by better technology making it easier to manage larger and more far-flung businesses. As companies get bigger, the boss's pay gets bigger.

This is skill-biased technological change. Technology also helps explain the rise of "winner-takes-all" job markets for such people as actors, pop stars and top sportspeople.

People want to see the very best, much more than the almost-as-good, they'll pay more to do so and technology makes it possible.

At a time when technology is working to make the rich a lot richer and the poor only a little less poor, should we be "reforming" the tax system in ways that add to this income inequality or reduce it?