2011-11-17

Zero Marginal Product of Labor?

Bryan Caplan and Arnold Kling have been going back-and-forth on the concept of "zero marginal product of labor." Today Caplan summarizes his main objections to the idea.

My observations:
  1. Caplan's arguments seem to hinge on the idea that once someone's MP(L) is zero, it's zero forever, which basically misses the whole point of Kling's PSST theory. (It's sort of like saying, "Assume PSST is false. Ergo, PSST is false!" But of course I have blogged about this kind of assumption before.)
  2. Caplan seems unwilling to approach the problem from the standpoint of microeconomic models, which I think resolve the whole issue.
My Defense of PSST
Here's my disclaimer: I know very little about PSST, other than what I have read on EconLog. So take this "defense" with a grain of salt. But if I understand the theory correctly, then the idea is pretty simple and frankly inarguable.

Consider the following story:

Once upon a time there were a group of economic consultants and a Chief Statistician who produced statistical models for a niche market. One day, company management hit upon the idea that they could make more money by offering one, comprehensive online statistical model to all of their clients for a subscription fee, rather than paying junior economists to develop labor-intensive spreadsheets every time their clients had a question.

So the Chief Statistician developed the model and translated it into an online, Javascript-based platform. The company fired all its economic consultants, retained the Chief Statistician to maintain the model, and revolutionized the consulting industry by offering online models developed by a great statistician, rather than expensive custom models developed by consultants.

Meanwhile, the fired consultants found that every other company had moved to online models, too. It took them some time, but eventually they brushed up on their Javascript coding skills, and found jobs with firms who were willing to hire them to program the models or conduct business analysis (or both).

What we see in the above story is the following:
  • A sectoral shift toward more efficient, less labor-intensive technology
  • A group of employees whose MP(L) goes from quite high to effectively zero
  • Approximately no change in real output
  • A group of employees who must acquire new skills to remain competitive in the job market and to engage in a sustainable pattern of production
Now, the above story is a real-world example (I'll let you guess who one such "economic consultant" was.)

Caplan seems to believe that the economic consultants in my story "didn't have ZMP because they found other jobs." But of course their MP(L) was zero at their old employer, and only became positive again when they realigned to new labor market conditions. The process took some time. This is fully consistent with Kling's theory.

As Kling noted a few days ago, Caplan may score some debate points by maintaining that the consultants' MP(L) never really dropped to zero, subject to Caplan's assumptions. But doesn't that miss the whole point of Kling's theory?

I feel like Caplan is just prodding Kling a little on a technicality. But this is very much an absurd technicality. MP(L) can be zero if profit is zero, regardless of quantity-produced-per-hour. MP(L) can also be zero during periods of time in which the employee produces nothing (like vacation, or sabbatical, or unemployment, or...).

My point is that everyone knows that there is only real and permanent ZMP for invalids and corpses. Kling's PSST model doesn't claim that an employee's marginal productivity is zero now and forever, and they are screwed. Kling's theory suggests that when the macroeconomy experiences a bubble, a whole lot of people are engaged in activities that aren't very productive or profitable; and when the bubble bursts, these folks will have to find more sustainable employment elsewhere.

So what can Caplan really object to?