Robots and Jobs: Not Time to Panic

L: This guy is not quite ready to take ALL the jobs. I know that we keep hearing the scary predictions of tech folks (like Elon Musk, for heaven's sakes) who are warning that the jobless future is imminent. A 2013 study by researchers at the University of Oxford (yep, that's the British one) reported that nearly half of all jobs in the US were at risk of being fully automated over the next 20 years. These forecasts are darn-right hair-raising!

So, to make you feel a bit better, I shall provide a public service and tell you about an article that ran in the September issue of Wired. (The author is James Surowiecki (@jamessurowiecki) who is also the author of The Wisdom of Crowds and a senior story producer at VICE News Tonight.) It has a couple of titles but they all make the same case: Robots Won’t Take All Our Jobs (Robopocalypse Not). The cool thing about this article is that it makes a case that any economist would love.

If job losses due to automation were increasing, you would find changes in two economic indicators; the first is productivity and the second is unemployment. We're not seeing either.

Since productivity is defined as the output per hour of HUMAN labor, as machines take over human jobs (like the Japanese company SoftBank that staffs their stores totally with talking bots named Pepper), productivity ought to soar since output is happening with a lot less human labor. But, at least at the moment, productivity is absolutely not soaring. In fact, economists keep writing a steady stream of papers trying to explain why productivity is so weak. (BTW, GDP would also go up if/when productivity goes up and that's not happening either.)

If machines are taking all the jobs, then humans don't have them, right? But unemployment is below 5% right now. Here's a quote from the article:

If automation were truly remaking the job market, you'd also expect to see a lot of what economists call job churn.... But the amount of churn since 2000 has been just 38 percent of the level of churn between 1950 and 2000.

Weirdly enough, the median time that employees stick with an employer today is similar to what it was in the 1950s.

So why the panic? Well, job losses definitely occur due to automation. For example, Goldman Sachs reported that autonomous cars/trucks could put 300,000 humans out of work--but that's over a 25-year time horizon, which does provide time for the economy to adapt. In fact, much of the argument here is that automation has historically moved much slower than predicted (MUCH slower) and much of the current manufacturing lob losses were due to the rise of China rather than the rise of automation.

The summary is, "yes, job losses occur because of automation, but no, it's not happening now in overwhelming numbers nor is it expected to happen dramatically any time soon." So, be aware but do not panic!

This is a really good article and if you're at all interested in the topic, I'd suggest reading the whole thing. My favorite example is that there are more than 400,000 ATMs in the US, they've been around for 45 years(!), and the actual number of bank tellers rose between 2000 and 2010. The Department of Labor has forecast an 8 percent drop in the number of tellers over the next decade. (Not 100%, notice that?)