Are we nearing the tipping point for a runaway labor market collapse?
“Runaway” processes, in which vicious-circle feedback mechanisms move economic systems and societies to extremes, have been getting more attention recently. The economic crisis, and especially the months-long panic triggered by the run on Lehman Bros. in late 2008, have reminded us that such processes are very real and can wreak havoc without much warning.
There’s one potential economic runaway that I’ve been thinking about a lot lately. It’s the kind that would eventually limit itself (see image) but would still quickly reach disastrous extremes. The idea here is that technological trends, exacerbated by the recent recession, are leading the economy to consume its own consumers. In this view, the sharp drop in US employment starting last year may not be something we’ll snap back from, but instead may mark the beginning of a cataclysm from which there will never be a natural, market-driven recovery.
Technology will soon be displacing human labor faster than human labor can adapt. In the past, advances in technology eventually increased employment. This happened because an advance in technology led to increases in profits and investment, which ended up creating more jobs than were lost due to the technological advance. But these days technology is evolving so quickly that humans are having trouble retraining in time to find new jobs before those jobs too are displaced by new technology. Already grocery checkout clerks, telephone receptionists and toll booth operators are pretty much obsolete. And before the labor market recovers from the current recession, robots and other AI systems will have begun moving into other skill areas. Soon commercial systems should be able to use tools and control heavy equipment (including vehicles), carry objects in all kinds of terrain, and monitor their environments with better-than-human sensory acuity.
Government policies and the inefficient health care market have hastened this process by making human labor ever more expensive. Think of rising health insurance and all the other overhead costs that employers must pay, not to mention the threat of litigation if an employee is fired.
The companies out in front of this wave are those that use technology to make a lot of money while employing relatively few people. The tech industry has the advantage here. Google has only about 20,000 employees and a market capitalization of $150B+ — thus, about $7.5M/worker. Facebook is said to have a market valuation of about $10B and fewer than 1,000 employees ($10M/worker). My favorite example is plentyoffish.com, run by one guy, Markus Frind, from an apartment in Vancouver, and with annual revenue of about $10 million — thus, a value per worker that must be close to $50M, i.e., up in the realm of film and sports stars. Compare those enterprises to an old-tech dinosaur like Delta Airlines, which has 60,000 employees and a market cap of only $7B (~$100K/worker). But as time goes on, even the airlines will use more machines and fewer people for every dollar of revenue, starting with back office and baggage handling staff and ending with maintenance techs and flight crews.
In some technology-intensive market sectors, the economies of scale needed to compete will soon be out of reach of all but a few players. Consider Google’s gigantic, ever-growing server/storage infrastructure. Or the creeping dominance of Amazon for almost any online purchase. In these sectors, the dominance of one or a few players forces industry consolidations (and layoffs). New growth here will be reflected chiefly in additions/upgrades to technological infrastructure, i.e., more machines, not more people.
Soon AI systems will become sophisticated enough to perform even creative intellectual tasks. You wanted to be a writer? Or a scientist? Or a hedge fund trader? Too bad. Soon even human entrepreneurs will start to become obsolete, which will mean that in almost every market, AI systems will be able to supply even tiny demand-niches (think: the “long tail”) more cheaply than human business-owners can.
This process will generate economic growth for a while. But that growth will be concentrated among the tech-driven winners, and in time it will start to reduce human employment, thus eating away at the consumer buying power on which its own growth depends. Ultimately it must choke on itself. Thus the “ouroboros” metaphor.
There may be no stable “equilibrium” in this process. Will governments be forced to use taxation to compensate humans for their obsolescence? How will we compete against high-tech, low-population societies that won’t have such a need to tax their industries? I fail to see an obvious sustainable solution here—assuming that the majority of humans still have some influence over their elites—other than a retreat into autarky and perhaps even a devolution into small, no-growth societies along the lines once suggested by E.F. Schumacher.
The key event here is the tipping point at which innovation in the broad economy stops creating jobs and starts killing jobs instead. We won’t see that tipping point except in hindsight. But it seems to me that as long as our society and markets are freely evolving we have to reach that tipping point someday, because machines clearly are evolving faster than humans (however much we drive ourselves to keep up), and already have inherent advantages as economic producers (i.e., don’t have human-like needs). And once we do reach that tipping point, everything else in this scenario follows. So I think many people now sense that while technology has boosted prosperity in the past, this time things are going to be different—this time there’s going to be a lot of upheaval without any assurance of a happy ending. And thanks to the 2008-09 recession, the upheaval may have started early.
Originally published September 27, 2009