January 29, 2016
It continues to amaze me that we have people simultaneously running around terrified that the robots will take all the jobs and at the same time that we will not have enough workers to support a growing population of retirees. (In some cases, it is the same people.) In the first case, productivity would have to go through the roof for a job shortage to be a problem. In the second case, it would have to go through the floor, since even modest rates of productivity growth swamp the impact of the aging of the population.
Thomas Edsall struggled with this issue earlier this week when a cyber-attack was preventing me from getting in my two cents. Now that I’m back, let me firmly throw my hat in with the middle position.
First, the robots taking all the jobs story is almost absurd on its face. How fast do we think productivity will grow that demand and reduced hours cannot keep pace? Productivity grew at a 3.0 percent annual rate from 1947 to 1973. We saw rapid growth in pay and living standards and very low rates of unemployment. Do we think the story would have looked worse if annual productivity growth was 4.0 percent?
It is almost impossible to imagine a story where productivity growth suddenly jumps from its current rate of less than 1.0 percent annually to a pace so rapid that we are losing jobs left and right due to improvements in technology. It is possible to tell a story where the Fed raises interest rates to slow the economy and job creation even as technology is displacing more and more workers.
That is a plausible story given that we have had several members of the Fed’s Open Market Committee that sets interest rates who have been worried about hyper-inflation. But the problem in that case is crazy-bad Fed policy, not robots taking jobs. And, we do the country a horrible disservice if we imply that the problem is somehow technology rather than the people running the Fed.
On the other side, the techno-pessimists essentially want us to believe that we have few or no opportunities to reduce our need for labor, while still maintaining our living standards. That seems to contradict what I see almost everywhere I go.
To take my favorite easy targets, combined employment in restaurants and retail is just over 27 million out of total private sector employment of 121 million. This comes to a bit over 22 percent. Imagine this was cut in half.
Does that sound far-fetched? This would mean many fewer restaurants and longer waits at the restaurants that still existed. On the one hand, this would imply a decline in living standards since we would have to go further to get to a restaurant and wait longer for our food once we got there. However, the ratio of restaurant workers to total employment increased by more than 25 percent between 2000 and 2015. This presumably meant that we had to travel less to get to a restaurant and spend less time waiting when we got there. That gain was not picked up in the GDP accounts. (We did pick up the value of the service, but not the quality gain from having it closer and being able to eat more quickly.) If we are being consistent with our accounts, the reversal of this increase also would not appear in the GDP accounts.
It is possible to envision a similar story with retail, although there was no comparable expansion in the sector in the last 15 years. Suppose there were fewer convenience stores and the ones that did survive didn’t stay open 24 hours a day? Department stores and Walmarts could have fewer people chasing us around asking if they can help us. And all those stores that now fill airports, in which almost no one ever seems to shop, could go out of business.
Again, this would amount to some decline in living standards. After all, someone goes to convenience stores at 2:00 in the morning and buys shoes at the airport store. But most of us probably would not be hugely bothered by losing this option.
Okay, where am I going? Suppose that we could snap our fingers and cut employment in these two sectors by 50 percent. Given their share of the economy, this would amount to an 11 percent increase in measured productivity. That would be a big deal, even if it occurred over the span of a decade, effectively adding more than a percentage point to annual productivity growth. And, this doesn’t require robots doing amazing things or computers with sophisticated artificial intelligence. It just means better using resources (workers) that we have today.
Furthermore, I feel pretty confident that if we could raise wages enough at the bottom end, we would see something like this drop in employment fairly quickly. After all, if the convenience store clerk is getting paid $20 an hour, no one will hire them to work the midnight shift to service the three people an hour who come in for milk and bread. So this gives us a pretty good start on our productivity quest, even before we talk about self-driving cars and trucks and all the other good things that technology is likely to bring us over the next two decades.
If it sounds like I’m being cavalier about jobs, let me get back to the Fed and the problem of demand in the economy. We do know how to create demand: it’s called spending money. The government can spend pretty much what it wants, as long as we don’t overheat the economy with too much demand. At the moment, folks suffering from deficit fetishes are preventing the government from spending more money, which would create more demand and employment. This is a really huge problem. But again, it’s not the robots, it the policy.
So let’s keep our eyes on the ball and not get distracted by folks making wrongheaded claims about us being doomed as a result of surging productivity growth or condemned to poverty as a result of collapsing productivity growth. The problem is simply bad fiscal and monetary policy. That is simple to fix even if the politics may be very hard. But the first step is looking at the problem with clear eyes.
Comments