Persuade Me

This post started out in life as an attempt at persuading those who believe that our primary problem in producing enough jobs is technological unemployment that they were mistaken. After I’d worked on it for a while, I realized that I was setting an impossible task for myself. So I decided to reverse the purpose of the post and outline what it would take to persuade me that the primary reason that more people are unemployed now than was the case in 2007 is that their jobs have become obsolete due to technology.

First, even though the nominal unemployment rate has now fallen below 7%, it is simply incontrovertible that fewer people are working than was the case in 2007. That’s true whether you conside it as a percentage of the whole or in absolute numbers and I do not believe that demographic arguments are sufficient to explain the difference. See this graph.

Therefore for the primary problem to be technological unemployment, you’d need to show some technological development since 2007 that has resulted in what we’ve seen. I do not believe there is such a development.

Note that my position is not that technological unemployment is irrelevant to our present predicament but that it is fractional.

Here’s Paul Krugman’s position on the subject:

What I’d say about America now is that we have big problems, very much including too much talent going into financial fiddling, too few people who actually make stuff — actually, I worry as much or more about machinists as I do about scientists and engineers. But that observation has virtually no bearing on high unemployment right now. So I’d hope we can walk and chew gum at the same time, appreciating the structural problems but not letting that understanding get in the way of fighting the immediate jobs crisis.

I’m largely in agreement with that but I think that Dr. Krugman is too determined to view all of our societal problems through the prism of partisan politics. I don’t think that our “structural problems” go back only to 2007 or even to 2001. I think they go back to the early to mid 1990s with the most important problem caused by China’s pegging the yuan to the dollar. That simultaneously had the effect of tremendously increasing the rewards to “financial fiddling” and reducing the marginal returns to labor, something also effected by the massive immigration from Mexico we’ve seen over the last several decades.

So, the second thing you’d need to do to persuade me our primary problem is technological unemployment would be to show that over the period of the last twenty years technology has been a larger factor in pushing employment down than Chinese mercantilism, immigration, too many dollars chasing too few secure investments, or malinvestment. That’s an empirical question that can only be demonstrated with numbers.

Finally, I could produce reams of material showing that precisely the same claims have been made in just about every major economic downturn over the period of the last century and a half. See this quote from a paper in the Journal of the American Statistical Association in 1935. They thought the problem in the 1930s was technological unemployment. I think that whatever we believe the actual reasons for the end of the Great Depression might have been it wasn’t a sudden increase in skills on the part of the American labor force.

Therefore, the third thing you’d need to do to convince me that our main problem is technological unemployment is to demonstrate that this time really is different.

Now, there’s one area in which I’m in complete agreement that technological unemployment is a major issue: retail. Over the last decade and a half there have been several revolutionary changes in retail which enable vendors to sell enormously more stuff without adding employees. That just demonstrates that fiscal stimulus targeted at increasing retail sales won’t have much effect on employment. We need to be growing, pumping, mining, and making more stuff and buying more of the stuff we grow, pump, mine, and make.