I’ve been suggesting here and elsewhere for some time now that we’re likely to see significant unemployment as a result of advancing automation technology. Anytime this argument is made, economists are likely to bring up what’s known as the “Lump of Labor Fallacy.”
The idea here is that people like me are falling into the trap of assuming that there is some fixed amount of work that needs to be done in the economy. Those of us who are economic rubes believe that if we automate or offshore some of those jobs (or allow immigrants to do them), then that means we’ll have unemployment—since, after all, there’s only so much work that really needs to be done. This limited and fixed requirement for work is referred to as a “lump” of labor. (The Lump of Labor Fallacy seems to be closely related to the “Luddite Fallacy” which I discuss at some length in my book, The Lights in the Tunnel — get the free PDF).
Anyone who suggests that automation may present a problem in the future is nearly certain to be accused of falling for the Lump of Labor Fallacy. This is true even if the person doing the suggesting is a trained economist. In August of 2009, Gregory Clark, who is on the economics faculty at U.C. Davis, wrote an article for the Washington Post in which he suggested that job automation would create a massive underclass that would need to be supported via taxation and redistribution. (Actually, Clark’s argument was a milder case of what I have been talking about because he seems to believe that the impact from automation will be limited to those without advanced education and skills. I think that’s wrong. Automation is coming for nearly everyone: we’re going to see knowledge workers with college degrees get hit hard within the next decade.)
As soon as Clark’s article came out, the reflexive accusations of “Lump of Labor Fallacy” quickly appeared: Tim Worstall, Will Wilkinson, EconoSpeak. Is it really true that anyone who worries about the impact of technology on employment is “committing the fallacy?”
As Tim Worstall points out in his comments on a post I wrote for Angry Bear, human beings have unlimited needs and desires while resources (including human labor) are limited . This implies that if you automate the jobs that are involved in fulfilling our current needs and desires, then we’ll quickly decide that we want something else—and that, of course, will mean that labor will shift into producing whatever becomes the next flavor of the moment.
I basically agree with what Tim is saying. Therefore, I am not committing the Lump of Labor Fallacy. I don’t belive that the amount of “work” that needs to be done is in any way limited. It may well be infinite. I just think that machines will be able to do the work. Or I think many of our desires will be delivered digitally—and therefore autonomously. Human labor may well be a limited resource, but what if it becomes a largely superfluous resource? The amount of sand in the world is limited too, you know.
Think for a moment about our evolving needs and desires. A great many people (for reasons that continue to elude me) seem to “need” to spend a great deal of time on FaceBook. We need to receive Tweets. We really need video that streams directly to our laptop because it sucks to have to stand in line at the Blockbuster store. While there are certainly some jobs that are created by these new desires, let’s face it: the vast majority of the “work” gets done automatically by giant server farms and by fiber optics.
If we project that digital fulfillment trend all the way to its possible conclusion we end up with some form of advanced virtual reality (VR) technology. This could potentially mean that a computer might be able to interface directly with your brain and create simulated experiences that were basically indistinguishable from reality.
If truly advanced VR ever arrives, it will introduce all kinds of interesting (and disturbing) economic questions: If a virtual experience is just as good as the real thing, will there still be demand for tangible goods? If you can live like a billionaire in the VR world, will there still be a strong incentive to seek wealth? Why not live in a rat hole but “live” in the Playboy Mansion? Will we even have a real-world economy? Perhaps everyone will just stay plugged in…until the lights go out.
It may be a good thing that VR is a long way off. In the meantime, there can be little doubt that machines, and computers are going to play an increasingly important role in producing tangible goods and services. And they will get nearly all the work when it comes to our evolving digital desires. After all, if you lose your job at the widget factory, you’re unlikely to find work delivering Tweets.
There’s no lump of labor. There’s plenty of work to be done. And machines will do it.