Stumbling and Mumbling

Author: chris dillow   |   Latest post: Tue, 26 May 2020, 2:30 PM


Incentivized stupidity

Author:   |    Publish date:

Two good things I've read this morning raise an under-appreciated point - that people can be incentivized to behave in ways that seem stupid.

First, there's Tom Chivers' review of Mervyn King's and John Kay's Radical Uncertainty. He says it is "completely terrifying" that they thought their book needed to be written, because economists and finance types shouldn't need to be told that "models are not 100% precise representations of reality." Boxmodels

Of course, he's right. Models are unreliable and the historic data that goes into them might not be a guide to the future - points well made by Richard Bookstaber in The End of Theory and by Nicola Gennaioli and Andrei Shleifer in A Crisis of Beliefs. But investment bankers have indeed made precisely this error, most famously when Goldman Sachs' David Viniar sais that during the crisis "we were seeing things that were 25-standard deviation moves, several days in a row".

It is trivially true that if you are regularly seeing 25 SD events, then you have horribly (pdf) mis-specified your model. So why did Mr Viniar do it?

Incentives, that's why. Goldmans wanted to take big positions in mortgage derivatives, and persuading themselves that they were safe was the way to do this. Banks need precise measures of risk to set trading limits, liquidity positions and gearing. You can't run a bank by saying "we don't know" even if this is true. And banks' bosses don't pay their underlings fortunes to say they don't know. You need at least an illusion of knowledge. And because profits are privatized whilst losses are socialized, bankers have an incentive to believe that risk is smaller than it is. Mr Viniar might have seemed stupid when he said that. But he's richer than you or me.

My next example of motivated stupidity comes from Geoff Mulgan's pessimistic assessment of whether Dominic Cummings really can improve the efficiency of government.

Geoff says:

The average tenure of many Ministerial roles is barely a year now. Since it takes a year for even a smart minister to learn how to do the job, the result is that most ministers, most of the time, are incompetent.

But Prime Ministers have incentives to do regular reshuffles. They need to remind ministers who's boss, and they maintain party discipline in part by offering backbenchers the prospect of office. If a PM promised to go five years without a reshuffle, he'd lose authority.

Geoff continues:

Without buy in from the bottom the top down changes rarely stick, even in states with authoritarian powers far beyond what UK ministers could dream of...If there is no strategy for engaging hearts and minds the programme is almost certain to fail.

But again, there's an incentives problem. People in power don't have an incentive to give underlings a voice, even if doing so would make an organization more efficient. This is why we see so much heavy-handed top-down corporate management even though we know that worker democracy is often more efficient.

And Geoff adds:

One of the big vices of Westminster and Whitehall is their valuing of words over deeds.

Again, there are strong incentives for this to be the case. It's much easier (especially for Tories) to manage the media by briefings, soundbites and press releases than it is to put in the hard graft of actually changing things.

What I'm describing here are specific examples of a general fact. Good government is (by definition) a common good, whereas maintaining power and prestige is a private one. Politics is therefore always in danger of delivering a tragedy of the commons*.

What we have in both finance and politics is therefore incentivized stupidity - or at least the appearance of stupidity**. (And not just these occupations: journalism offers us other examples of this.)

A recent paper by Christine Exley and Judd Kessler give us experimental evidence for this. They show that when people had incentives to behave selfishly - for example in splitting cash between themselves and a charity - they often behaved apparently irrationally to give themselves more. They say:

Individuals make unambiguous decision errors - acting as if they suffer from cognitive limitations or behavioural biases - in order to make more selfish choices...Behaviour that could be attributed to a behavioural bias (such as anchoring...) might instead be indicative of self-serving motives.

I say all this as a corrective to a common error, particularly among Econ101ers. This is the belief that incentives are usually a good thing - that as Adam Smith said, "it is not from the benevolence of the butcher, the brewer, or the baker that we expect our dinner, but from their regard to their own interest."

Yes, incentives can be powerful. But sometimes make us act against the public interest. Getting incentives right is not easy.

* Of course, what Ostrom said of the tragedy of the commons is also true in politics: outright tragedies are often avoided thanks to unwritten rules and conventions. But these need shoring up. They cannot be taken for grant

** Whether such behaviour actually is stupid is questionable, given that it enhances the power and wealth of those engaging in it.

Share this

More articles on Stumbling and Mumbling >>

  Be the first to like this.

I3 Messenger
Individual or Group chat with anyone on I3investor

2606  1739  974  2993 

Top 10 Active Counters
Partners & Brokers