Will Job Polarisation Continue? I Asked Some Economists…

[I’ll carry on updating this post as I get more responses].

Over the last few decades in rich countries, the share of middle-skill jobs has been going down.

There are fewer machine operators, middle managers and so on than there were decades ago, and it looks rather like there are fewer routine middle skill jobs around to replace them.

In 1979, the most common kinds of middle skill jobs in the US – office and administrative workers, production workers and operatives – accounted for 60% of total jobs. By 2012 that figure had fallen to 46%.

The trend holds in Europe too. Between 1993 and 2010, across sixteen EU countries, the share of middle jobs went down by almost 10%.


That graph was taken from David Autor’s 2014 paper on this subject.

He also makes the striking point that in a survey in 2014, a plurality of mainstream economists agreed with the proposition that ‘information technology and automation are a central reason why median wages have been stagnant in the US over the past decade, despite rising productivity’.

If they’re right, and technology has contributed to the kind of wage stagnation that we’ve seen in the US, or even if it just means fewer opportunities for the majority to get on, it’s a serious policy challenge.

Where is this heading?

So I wondered – is this trend likely to carry on?

Are we heading toward a world where there are plenty of highly paid, high skilled jobs, plenty of low paid, low skill jobs, but not enough in between? And if so, what on earth should we do about it?

I decided to ask a handful of the economists who are closer to the data than I am.

I asked two questions:

‘With these questions, I have in mind the long term trend towards labour market polarisation in the EU and the US.

  1. Do you think labour market polarisation is likely to continue to deepen in the foreseeable future, and why? 


2. As a citizen, do you think public policy should try to mitigate labour market polarisation? If so, how?’

A few replied, their full responses are here, but this post offers a few takeaways from the responses so far.

  1. Three respondents think labour market polarisation is likely to continue, two don’t, three push back against the question.

‘Polarisation is likely to continue’ – Georg Graetz, John Van Reenen, and Craig Holmes

‘Polarisation probably won’t continue’ – David Autor,  Andrea Salvatori

‘I recommend you don’t think about it quite like that’ – Adam Corlett, Laura Gardiner and Larry Mishel. More on this in point 5 below.

2. Unsurprisingly, pretty much everyone says that there’s a role for public policy to respond to inequality in a number of familiar ways: redistribution of income or wealth, education, retraining, insurance, and so on.

It’s not obvious, though, that this inequality is a consequence of this kind of labour market polarisation.

Also, Craig Holmes has a twist. He says:

“I don’t think pay is everything, [but] in cases where the increase in productivity or reduction in costs that new technology bring about significant losses of job quality in addition to changes in pay, I would support policies that reduced these sorts of investments in the first place.”

He also points out that

“The state, as a huge employer itself and creator of employment elsewhere, can have a big impact in terms of the nature of the work it creates directly.”

3. It’s worth distinguishing between ‘the replacement of routine jobs’ and ‘labour market polarisation.’ Yes, routine jobs are the ones most likely to be automated away. And yes, those have traditionally been the jobs in the middle of the skill spectrum. But the jobs most at risk of automation might not always be in the middle of the skill spectrum. Developed countries could start creating more middle skill jobs which are not routine jobs. In other words, the replacement of routine jobs could continue without deepening the division between low and high skilled jobs.

4. Professor Graetz points out that hollowing out of employment isn’t new: previous episodes of automation were also associated with a hollowing out of employment.

5. As mentioned, some recipients push back at the question. Their core point is that it’s not obvious that polarisation is happening, and even if it is it’s not driving inequality of income.

Here’s Laura Gardiner from the Resolution Foundation on the UK:

“I think that casting this phenomenon in terms of “polarisation” is the wrong way to think about it, because there is very limited evidence that the labour market is becoming more polarised. Even if jobs that started out in the middle of the pay distribution have declined, jobs moving around this distribution, new jobs cropping up all over the place, and earnings dispersion within occupations means that the labour market doesn’t look to be a particularly different shape now than it was 20 years ago (in contrast to what’s happened in the US, to some extent).”

And on the US, here’s Larry Mishel:

“[Polarization] has not really been present in the US since 1999/2000 though the last few years show a bit of it. But it doesn’t really matter whether it does or does not since polarization has no bearing on wage trends. “Occupational employment trends do not drive wage patterns or wage inequality.” I wouldn’t be surprised if polarization resumes but we shouldn’t be obsessed about it at all. It doesn’t tell us much.”

5. Some of the most respected economists in the field are generous with their time.

Thank you to everyone who answered. And if you didn’t, you still can. I might update this post if subsequent answers change the picture or yield other insights.

The recipients’ full responses are here.