Technology and the Automation of Jobs

Ivan Vanchev

Tuesday, May 27, 2014

A recent study was quoted in an Economist article a few months ago regarding the rise of job automation and the social impact it could have in the near future. For the less acquainted with the topic, innovation - considered widely the major source of long-term economic growth - always brings along a certain level of job displacement and transition, as industries upgrade their equipment and processes. This short-term hit in employment is exactly that - short and generally inconsequential; labor moves on to more productive jobs and helps bring the economy forward. However, the study by Oxford University researchers Carl Frey and Michael Osborne reaches a unsettlingly high estimate for the amount of jobs that could get automated in the next two decades - 47%. We had a quick talk inspired by a Huffington Post article on the topic. Several good points were raised, and once we realized that a proper discussion of the topic required the equivalent of publishing a paper on the matter, the chat quickly unwound to talk of brighter things in life.

Plamen Ivanov: Hmm, I'm reading some reddit title...

Ivan Vanchev: aha?

Plamen Ivanov: ... about how 47% ot all jobs will be automated by 2034. So I was thinking of the approaching problem of organizing our (well, Western mainly) society. This automation may at some point create real problems.

Ivan Vanchev: Well, yes. If it isn't handled well.

Plamen Ivanov: Mhm.

Ivan Vanchev: Or it can allow us to transition into a post-scarcity economy.

Ivan Vanchev: :D

Plamen Ivanov: Although on the other hand there is plenty of time before they automate a lot of things.

Ivan Vanchev: Indeed...

Plamen Ivanov: We are plenty automated already - tractors, trains, cars, ships, etc...

Ivan Vanchev: ... Plus, there's no way the number quoted is a global value.

Plamen Ivanov: ... and every time, someone loses her job.

Ivan Vanchev: Well, these things are usually short-term.

Plamen Ivanov: I guess now things are accelerating.

Ivan Vanchev: I mean economically there's nothing to worry about; people transition to other jobs.

Plamen Ivanov: Yes, the question is whether employment shrinks too quickly - then, we'd have a long period of high unemployment.

Ivan Vanchev: That's why we have social safety nets.

Plamen Ivanov: Well, sure, but the problem is that if it's a too long of a period, safety nets would need to support a disproportionate amount of people. The ones working would need to support more and more unemployed. At least that's how I understand the issue.

Ivan Vanchev: I can't imagine why there'll be that big of a chunk of unemployed people all of a sudden. It's not like there's a massive push for automation in many industries. We need to look at the numbers.

Plamen Ivanov: I am just saying what the worst case scenario is.

Plamen Ivanov: I think the problem seen by some is that if you cut out a major part of the work force, you will lose too much of the tax (depending on how the upper bracket is taxed).

Ivan Vanchev: Even if all those jobs are automated that means that someone would get the profits from the labor savings...

Plamen Ivanov: Otherwise, as you said, if the rate is low enough, it's nothing new.

Ivan Vanchev: you'll get more tax revenue from them.

Ivan Vanchev: Well, I don't like talking with pulling stuff out of my ass, but...

Plamen Ivanov: Haha, that's the only way I can about this topic.

Ivan Vanchev: ... it seems reasonable to me that the biggest chunk of endangered jobs won't be well-payed, well-educated people...

Plamen Ivanov: Plus, this is mostly looking at extremes.

Ivan Vanchev: ... but people at the bottom end - service industry, etc. - who already don't really pay much tax.

Plamen Ivanov: Yes.

Plamen Ivanov: Well, there are also dangerous jobs which likely pay plenty (e.g. miners) but are dangerous.

Ivan Vanchev: Good point...

Plamen Ivanov: Such jobs may be more significant than simple service industry.

Ivan Vanchev: However, they're

  1. already automated to a high extent
  2. not a big employing industry nowadays
In developed economies the biggest chunk of people work in the service industry. I guess the two points are connected :) But the point remains.

Ivan Vanchev: Yes... according to the Bureau of Labor Statistics, 0.6% of US employment is in the mining industry.

Plamen Ivanov: So, I guess, without the following data we can't really effectively hypothesise:

  1. what things can get automated in the next 20-30 years
  2. how much people are working those jobs and won't be able to work them
  3. how will the above two affect the tax revenue
  4. optional: how many other jobs can be opened for people in 2)
They are also talking about some office jobs here.

Ivan Vanchev: Well, the BLS has projections into 2022, and I can't see any drastic changes in employment across industries, i.e. anything out of historic trend.

Plamen Ivanov: Ha, sorry beit, here I am getting into waters I don't know how to thread. Plus, if the study is to be believed, it is claiming the governments are not ready (so by extention they also wouldn't have predicted it correctly).

Ivan Vanchev: Indeed!

Ivan Vanchev: The wording of the huffington post article quote is could be up to 47% which sounds a little speculative to me.