The Post-Work Myth

This post is part of our Work Symposium. An introduction to the symposium can be found here; all of the posts written for the symposium can be found here.

The idea that Progress is going to do away with the need for our labor is not new.

The idea might be older than economics itself.

The actual history of the world has been, however, that people while there may be an overall downward trend in the average number of hours worked per person, people as a group have continued to work.

Efficiency improvements in farming led to fewer people working in farming, people’s growing consumption demands enabled a bunch of folks to get manufacturing jobs. And when manufacturing became more mechanized and efficient, excess demand manifested in the catch-all of “services”.

And somewhat bizarrely, efficiency improvements don’t always reduce employment in a sector. People once lamented jobs lost as bank tellers were replaced by ATMs. These efficiency improvements along with other innovations actually greatly increased demand for the financial industry’s services and overall employment in the sector.

More generally, when you figure out how to make something with fewer people, you can decrease costs (or improve the product). This can increase demand, sometimes dramatically. Depending on the steepness of the demand curve, people might increase their consumption enough that you actually have to hire *more* workers as a result of per worker productivity improvements.

Forgive the made-up example, but imagine that you run a clothing producer before the invention of the sewing machine. Your workers have to stitch by hand. It is entirely possible that people will only buy one set of clothes and use that set more or less forever. Demand is such that you don’t have to hire that many workers.

After the invention of the sewing machine, you are able to reduce your labor costs significantly on a per-item basis and thus charge significantly less. People might then change their purchasing behaviors and want multiple sets of clothing’s in different colors or for different occasions. Demand shoots up even faster than the productivity of your workers shot up (assuming demand is relatively elastic).

If this new buying behavior becomes a culturally enforced habit, you might have to hire even more workers than you had before you got the sewing machines.

That’s a version of what happened with banks and ATMs. That a machine will do your job doesn’t mean you will be out of a job. It means you will be out of that particular job.

This is not a theoretical phenomenon. For most of America’s manufacturing history, companies have invested more capital into making their workers more productive, and employment increased anyway because it turns out people had a near insatiable demand for manufactured goods. It is only in the last handful of decades that productivity improvements have been coupled with job losses in manufacturing, and I’m not even sure those losses hold up if you look at global manufacturing employment.

Of course, that doesn’t mean that this time that the robots won’t all put us out of work. But it’s worth noting that the post-work thesis is very old and it hasn’t happened yet. Like “peak oil”, the logic is sound, and it might come true someday, but the timing has been so lacking that it might not even matter if the idea is correct.

People are horrible at predicting what technology will do. A lot of 90s pundits thought the internet would fix everything because everything would be physically connected. Everything else is just software, right? Now, interoperability remains the biggest single issue in information technology today. (One CIO told me that his position should stand for “Chief Interoperability Officer”.) The people who don’t get this are the same people who don’t understand why gun registries don’t automatically check for outstanding warrants and the no-fly list.

Luckily, being wrong about every single generation of technology you encounter isn’t a job that is likely to be automated away.

Someday perhaps more people will value a marginal hour of leisure than what they can earn with a marginal hour of work. Notice that that is a far less dire-sounding claim that most post-work-ers claim. The post-work mythologists claim that the value of a marginal hour of work will fall to zero as we 3D print our driverless cars from within our robot-constructed houses.

Why though couldn’t it be simply that we’d rather spend an extra hour on Occlus Rift glasses than log another lucrative hour of work. Sure, pay is stagnating for some, but entertainment options multiply far faster. If people stop working, we might need to “blame” the high value of leisure rather than low pay.

But my money is on continued work for most, at least for our lifetimes. Then again, James Jobs-are-going-to-disappear Hanley says much the same thing:

I don’t think it will come in my lifetime, but I think in my lifetime we have been and will continue to be on a trajectory toward it.

It is a wise policy to make bets that you won’t live to pay out on, but such hedging implies there might not be anything for this generation to worry about.