10 May 2012

Don’t Trust Mainstream Economics

It was almost exactly 20 years that statistics prepared by Paul Krugman about the extent of rising income inequality began to show up in the New York Times on page 1. Back then we--or at least I--thought that this would quickly reverse itself: I remember one dinner at which Claudia Goldin challenged Paul, asking him to agree that rising inequality had made going to college such a no-brainer that we would soon see a flood of increasing investment in education that would bring income distribution back into balance. She was wrong. I was wrong. The American income distribution righted itself, at least for white guys, after the Gilded Age. It did so with mass education and the rise of the social democratic welfare state. What is different this time?
This paragraph belies some astonishing ignorance, held by one of the better-known mainstream economists.
In the first place, DeLong assumes that the laws of supply and demand do not apply to college graduates.  How else to explain his assumption that continuing to increase the supply of college-educated labor would either a) drive up the price of marginal college-educated labor or b) drive down the price of fundamental college-educated labor?  That there was such a high degree of price inequality in the college-educated labor market should have suggested that market was near saturation (side note: this is perfectly exemplified in the current labor market for lawyers, in that wages are generally low, but there is a high degree of inequality among income-earners).

In the second place, DeLong assumes that there is a natural balance of income distribution.  There is not, because humans are not equal to one another in terms of drive, ability, temperament, sociability, intelligence, and other factors that are tied together in determining one’s income.  As such, it is unreasonable to assume a trend towards income equality in general.  However, the presence of extreme inequality should at least be checked out, if for no other reason than to ensure that politicians aren’t enriching themselves or their cronies by defrauding the populace.  But it shouldn’t be assumed that high degrees of inequality are themselves bad.  Basically, inequality should be viewed as a consequence, not a cause.

Thus, DeLong is unable to understand why the policies that appeared to work the first time around don’t work the second time around.  Here’s a hint:  the market for educated workers has obviously reached its saturation point.  Also, people aren’t equal.  Expecting more equality and more education to fix the problem is therefore stupid, and DeLong is stupid for not seeing why.  The sad thing is that he is not alone; there are many experts asking the same question, unable to provide the answer.  They are not to be trusted; in fact, they should be ignored.


  1. Very good analysis Simon.

    Economics is a hard, hard profession. Economic activity operates within a social context and thus is influenced by it. Therefore the good economist must not only have a grasp of basic economic laws, but also of philosophy, history, psychology and current events.

    The multivariate contingency of economic outcomes means that what works in one set of circumstances does not work in another. The problem is that this multivariate analytic capacity is very, very rare and only acquired with considerable effort. Secondly, the specialist nature of modern education tends to treat subjects outside a discipline's core trivially. Economics is a generalist's profession.

    Most economists are one trick ponies. Their solution to any downturn is to press on the liquidity pedal. Their mental capacity and education is simply not up to task. Nobel Laureates included.

  2. @SP- Thanks. As always, good to have you around these parts. (Also, I enjoyed your post on charity.)

    One thing that I've noticed complicates economics is the relationship among multiple variables. Simply put, it's hard to come up with highly specific policies that will be successful because the relationship between variables is too complex. General rules work better, but they make for rather simple policy. And you can't bill as many hours for coming up with simple policies.