What is a plumber, anyway?

Tim Harford, better known as the Undercover Economist, always has reliably entertaining thoughts. His latest piece explains Why Microsoft Office is a bigger productivity drain than Candy Crush Saga. Drawn in by that title like moths to a flame, we find the following critique:

Microsoft Office may be as much a drag on productivity as Candy Crush Saga. To see why, consider Adam Smith’s argument that economic progress was built on a foundation of the division of labour. His most celebrated example was a simple pin factory: "One man draws out the wire, another straights it, a third cuts it, a fourth points" and 10 men together made nearly 50,000 pins a day.
[…]
In a modern office there are no specialist typists; we all need to be able to pick our way around a keyboard. PowerPoint has made amateur slide designers of everyone. Once a slide would be produced by a professional, because no one else had the necessary equipment or training. Now anyone can have a go — and they do.
Well-paid middle managers with no design skills take far too long to produce ugly slides that nobody wants to look at. They also file their own expenses, book their own travel and, for that matter, do their own shopping in the supermarket. On a bill-by-the-minute basis none of this makes sense.

Superficially, this take is amusing as ever, but on reflection, I do find it a little disingenuous. Leaving the slides out of it, because actually those are part of my job, nowadays it is true that we all book our own travel and so on – but on the other hand very few of us office drones keep our own vegetable gardens or even know how to, and those that do mostly treat it as a hobby.

All that has happened is that the frontier of specialisation has moved, and what was once common knowledge for everyone is now a specialised job, while what once required specialists is now expected of everyone. Where once everybody knew how to grow their own food, now we delegate that to small groups of professionals using specific tools. Meanwhile, data processing, which used to be literally the preserve of a priestly caste, has been democratised to the point that any office job will require people to know at least the basics.

I would love to have an assistant to do my expenses and so on, and instead here I am toiling in the salt mines – but let’s face it, if your expense platform is at all decent, and you have reasonable IT skills, this should take roughly no time at all. Booking your own travel ensures that you get what you want, making your own compromises inside of the corporate travel policy.

This definitional error has some interesting consequences, as it is certainly true that most people are probably slower typists than professionals who worked in typing pools, when such things still existed. If your measurement of productivity is words banged out on keyboard per minute, it is almost certainly less efficient for professionals to do it themselves. And yet, hitting the keys yourself is always a far quicker way of getting your ideas out than dictating to even the fastest typing pool. How do you measure the productivity difference between an exec tapping out three lines on their iPhone while waiting to board a flight, versus having to wait until they get back to the office on Monday? Sure, those three lines are terse, jargon-filled, and probably stuffed with typos or interesting autocorrect-isms, but they get the point across.

All of this transformation informs Tim Harford’s predictions for 2118:

In an insightful essay from 1996, Paul Krugman predicted that there would be "no robot plumbers" in 2096. I agreed with him then. I am no longer so confident. It seems quite plausible that in 100 years’ time — and perhaps much sooner — plumbers, taxi drivers and many journalists, too, will simply have nothing serious to contribute to the labour market.

I would be seriously impressed by a robot with the combination of agility, strength, and inference reasoning required to work as a plumber. I may well be proved wrong by events (and if so, I will take refuge in probably not being around to be embarrassed by my wrongness), but I suspect it won’t quite work out that way. Instead, I suspect that the job of "plumber" is one of the safest out there, and for many of the same reasons that it was impossible to outsource: in addition to knowledge, it requires great situational awareness, problem-solving capabilities, and flexibility – all of which are weak points for automated systems.

More vulnerable are the jobs that delaminate into separate tasks, some of which will be re-bundled into new and different jobs, and others will be automated away. The job of "typist" is gone the way of the dodo because it encapsulated a single task which could either be made part of other jobs (we all do our own typing) or automated (mail merge and variables make it easy to generate and circulate even quite complex documents, without human typing).

The job market will certainly be radically different in 2118 – that prediction is fairly safe – but I expect that there will still be jobs, and people doing them - people augmented by automated capabilities.


Photo by Jouni Rajala on Unsplash