Wednesday, May 6, 2015

Incentives and behavior: psychology vs. economics


 - by New Deal democrat

This morning the NY Times reviews the book "Misbehaving" by Richard Thaler, the current President of the American Economic Association, in which he pillories the dominant Chicago school style of economics:

After so-called behavioral economics began to go mainstream, Professor Thaler turned his attention to helping solve a variety of business and, increasingly, public policy issues. As these tools have been applied to practical problems, Professor Thaler has noted that there has been “very little actual economics involved.” Instead, the resulting insights have “come primarily from psychology and the other social sciences.
To the extent that economists fought the integration of behavioral insights into economic analyses, it seems that their fears were founded. Rather than making the resulting work less rigorous, however, it simply made its economic underpinnings less relevant. Professor Thaler argues that it is actually “a slur on those other social sciences if people insist on calling any policy-related research some kind of economics.”
Professor Thaler's narrative ultimately demonstrates that by trying to set itself as somehow above other social sciences, the "rationalist" school of economics, actually ended up contributing far less than it could have.
Ouch!  That's gotta sting.

Contrast this with a statement by Brad Delong just a few days ago:
Arthur Burns was right: you [economists] are better-positioned than any other group to help us make the right choices, at the level of the world and of the country as a whole, but also at the level of the state, the city, the business, the school district, the NGO seeking to figure out how to spend its limited resources–whatever.''''[Why?] I would say:....
  • Economists know about systems, and how systems work:
  • Supply and demand
  • Opportunity cost
  • General equilibrium
  • Incentives and behavior
(my emphasis)

Excuse me??? Didn't Thaler just hand you your head on point number 4?

Here's a nice, simple question.  How do you get more of a behavior?

  • Economists say: increase the incentive
  • Psychologists say: variable reinforcement
These are two apparently contradictory answers.

Now, I suspect that if a psychologist put a rat in a cage with two bars dispensing food pellets, and one dispensed a pellet on a variable schedule, and the other always dispensed two pellets, the rat is going to press the lever that gives him more food.  Score one for Econ.

But suppose, after the rat learns to press that lever, I divide the cage in such a way that, once the rat chooses one bar, it is very difficult to get over to the unchosen bar.  The rat makes its choice ("two pellets all the time, please"), and now I start cutting back on the pellets, and only periodically and variably dispensing a pellet.  Now I am going to get a ton of bar-pressing for diminishing rewards. Score one for Psych.

Now think about annual compensation schemes by large corporations.  The metrics constantly vary, and the weighting of each metric varies.  The employee-rat having made its choice (by accepting employment), the employer-experimenter tries to maximize the desired behaviors by the employees by variably dispensing ever-slightly-decreasing and sporadic rewards. Since the employee doesn't know which job aspect will be the most rewarded in the next year end compensation scheme, the employee tries to do the most of all of them.  Variable reinforcement, thank you very much.

And to bring this back to the big economic issue, one thing we know about humans is that we try - asymmetrically - to avoid actually realizing losses.  In 1929, in order to forestall repossession of goods bought on installment payment plans, consumers stopped spending money in far larger proportion to the initial downturn.  Similarly, in September 2008, in order to prepare for the sudden emergency that Bush et al told them might only be days away, consumers en masse simply stopped spending.
To refresh your memory, on September 19, 2008, the NY Times reported that, "As the Fed chairman, Ben S. Bernanke, laid out the potentially devastating ramifications of the financial crisis before congressional leaders on Thursday night, there was a stunned silence at first. Senator Christopher J. Dodd [said] the congressional leaders were told “that we’re literally maybe days away from a complete meltdown of our financial system, with all the implications here at home and globally."  And on September 24, it reported, "President George W. Bush on Wednesday warned Americans and legislators reluctant to pass a historic financial rescue plan that failing to act fast risks wiping out retirement savings, rising foreclosures, lost jobs, closed business and “a long and painful recession.”

Professor Delong's statement that economists best know about incentives and behavior stands in stark contrast with his concession a few days earlier that a new macroeconimc paradigm was needed, and he honestly didn't know what that should be:
This entire structure is now in ruins. Our social-democratic income distribution has been upending in an appalling manner by the Second Gilded Age. Our government, here in the U.S. at least, has been starved of proper funding for infrastructure of all kinds since the election of Ronald Reagan. Our confidence in our institutions' ability to manage aggregate demand properly is in shreds--and for the good reason of demonstrated incompetence and large-scale failure. Our political system now has a bias toward austerity and idle potential workers rather than toward expansion and inflation. Our political system now has a bias away from desirable borrow-and-invest. And the equity return premium is back to immediate post-Great Depression levels--and we also have an enormous and costly hypertrophy of the financial sector that is, as best as we can tell, delivering no social value in exchange for its extra size. 
We badly need a new framework for thinking about policy-relevant macroeconomics given that our new normal is as different from the late-1970s as that era's normal was different from the 1920s, and as that era's normal was different from the 1870s.
So long as economists think that they are the ones who know the best about human incentives and behavior, they are going to continue to stumble along in the dark and their public policy prescriptions are likely to inflict much unnecessary harm on millions of people.