My brilliant former student Eli Dourado wonders why I’m so hostile to the Zero Marginal Productivity (ZMP) theory of high unemployment:

It is hard to think of another idea that is more Caplanian. This is after all the man who pointed out that “the lower deciles don’t contribute that much to the economy, anyway.” …

The ultimate revenge of the nerds is developing tools that make jocks literally useless

ZMP coheres well with my elitism,  but coheres poorly with my rejection of high-IQ misanthropy.  The lower deciles contribute little to the economy in percentage terms, but the absolute value of their contribution is still hundreds of billions of dollars. 

In any case, when I judge specific hypotheses, I try not to put too much weight on coherence with my overall worldview.  The best reasons to reject ZMP are:

1. Non-economists have been falsely predicting the technological obsolescence of human labor for centuries.  Economists have been correcting them for about as long.  Clearly ZMP’s psychological appeal far exceeds its true relevance.  This creates a very high burden of proof for anyone who claims that the ZMP nightmare is finally a reality.

2. Low-skilled Americans clearly didn’t have ZMP a few years ago.  Indeed, their marginal product remained enormous by world and historic standards.  How could their productivity have changed so much so quickly?* 

In his post, Eli points to precisely one factor that did change quickly in recent years: the “cyclical fall in aggregate demand.”  But if that’s the key shock, ZMP is just a misleading redescription of Keynesianism.  Isn’t the whole point of ZMP supposed to be that boosting aggregate demand won’t reduce unemployment nearly as much as Keynesians think?

3. Eli also mentions “nonconvexities,” especially the “fixed costs of employee management.”  But low-skilled American workers clearly remain much more productive than median workers in Third World countries.  And the median worker in Third World countries is profitably employed despite the fixed costs of employee management.  Note further that in countries with low labor costs, employers happily hire low-skilled workers to do all sorts of tasks that First World employers mechanize or do without.

4. Most economists aren’t entrepreneurs.  The fact that we can’t think of a productive job for a low-skilled worker is weak evidence for ZMP.  But if even ivory tower economists can think of productive jobs for low-skilled workers, that is strong evidence against ZMP.  And thinking of such jobs is easy.  My first stab: How about as personal servants for high-skilled workers?  In Third World countries, the middle class routinely hires live-in housekeepers, drivers, and so on.  In the worst-case scenario, we can learn from them.

5. If nominal wages for low-skilled workers had fallen massively and high
unemployment persisted, I can understand why someone might start to
wonder (not believe, but wonder) if such workers had
ZMP.  But nominal wages haven’t fallen massively.  They haven’t fallen
at all!  ZMP is literally a resolution in search of a paradox.

* Note the tension between stagnationism and ZMP.  If technological progress is stagnating, how can low-skilled workers be getting obsolete so rapidly?