(Non) performativity in economics

Aditya Chakrabortty writes:

Economists didn't just fail to spot the financial crisis – they helped create it. They provided the intellectual framework and drew up the policies that helped caused the boom – and the bust.

He gives the example of how (some) academics made the case for big bosses’ pay.

What he’s getting at here is the concept of performativity. Academic economics does not merely describe the world, but also helps create it. Markets in options, and index tracker funds, for example, are academic creations; see the work of Donald Mackenzie, such as his book, An Engine Not A Camera.

However, economics is not always performative.

Let’s take Aditya’s example - that economists have described “how markets work best when they are left alone.”

There is a large, old and powerful body of theory which shows that, under some (restrictive) conditions, this is true. Back in the 1950s, Gerald Debreu (pdf) and Kenneth Arrow showed that a set of complete state-contingent markets can be Pareto optimal. The converse is also true. As Greenwald and Stiglitz showed (pdf), an economy with incomplete markets will be sub-optimal.

If economics were always per formative, you would therefore expect there to be at least widespread markets in major contingencies.

And there are not. There’s a market for my labour (I hope), but not a market for my labour, contingent upon there being a great depression. Although I can insure my house losing value because of fire, I cannot insure against it losing value because of a fall in demand for houses in Rutland. I can insure against inflation, thanks to index-linked gilts, but not against recession or inequality. As Robert Shiller pointed out in his wonderful books, Macro Markets and The New Financial Order, markets for coping with major economic risks are lamentably under-developed.

This poses the question. How come economics is performative in some regards (option pricing, bosses’ pay) but not in others, such as contingent markets?

It’s certainly not because Arrow-Debreu theory is new or marginal. It is much older, and far more widely taught in universities than the work of Jensen and Murphy cited by Aditya.

Nor is it because state-contingent markets would be an obviously bad idea. Yes, we know from the theory of the second best that it is not necessarily efficient to remove a single market distortion, so it’s theoretically possible that the introduction of a few such markets would be sub-optimal. But whether this would be the case in practice takes some proving. And anyway, plenty worse ideas than these have been turned into economic reality.

There is, of course, a simple answer to this question. Economics is performative when it serves private interests, but not (necessarily) when it serves public ones. Traders immediately saw the usefulness of the Black-Scholes option pricing formula, and bosses quickly saw the use of Jensen and Murphy’s work. But the benefits of better state-contingent markets accrue to millions and cannot so easily be captured privately. They are an example of a Nordhausian innovation - one with high social benefits and low private benefits and which do not therefore exist. Paradoxically, markets are incomplete because of a market failure - that there’s a positive externality to creating complete markets.

I could, of course, put this more crudely. Economics is performative when it serves the interest of the powerful, and not performative when it doesn’t. In this sense, the problem is not with economics, but with a class structure that causes the “real world” to be a corrupted and perverted form of a market economy.

Comments

[Luis Enrique]

yep, as I understand it Jensen and Murphy recommended lots of things that didn't happen (bosses having to hold shares for long durations, getting paid zero if performance is poor) and all that didn't get "performed" in reality, just the pay bosses lots bit did.

[PaulB]

Ingenious financial models for pricing mortgage CDO tranches persuaded major banks to take on the risk of correlated mortgage defaults in the US market. And that hasn't proved to have high social benefits.

Insurance depends on diversification of risk. Option-writing depends on hedging of risk. Neither applies to the Rutland housing market. A seller of the insurance you want would just be another AIG waiting to be bailed out. The best way for this risk to be carried is by individual house-owners. They have a hedge of a sort in that they are short the house they're going to move into when they move out of this one and sell it.

[Tom P]

Thing is some of the concepts have become unchallengeable in my corner of the world. Try arguing with a corp gov person at an asset manager - agency theory is all they have, so they struggle to see much wrong (apart from the odd Fred Goodwin) wrong with pay in PLCs.

[George Hallam]

"In this sense, the problem is not with economics, but with a class structure that causes the “real world” to be a corrupted and perverted form of a market economy."

This sounds like a cop-out.

Why not say that the problem is that economics as a discipline is structured to serve the class interests of the rich.

[Alex Marsh]

Mackenzie's work (which is excellent btw) also indicates that you have to take account of the private interests of the economists. Economists were paid handsomely to advocate, on the basis that the economy would be more efficient if such markets existed, for changes in the law so that settling futures contracts in cash not goods was legal (previously it had been defined as gambling). Similarly, economists didn't only develop the theory around CAPM but also set up companies to provide information to the market based upon the model.

I don't disagree with your position, but you could also argue that these are ongoing processes and some of the products/markets you mention just haven't arrived yet. Shiller of course is associated with the Case-Shiller house price index, and for the last five years at least there have been attempts to develop derivatives in house prices that hang off the C-S and other such indices. Case has been an advocate of developing retail products off the back of such new financial instruments which will allow individual homeowners to hedge. It may happen eventually.

[NAME REDACTED]

So you are saying that the problem with this state sponsored enterprise (economics is mostly state funded) is that it is too influenced by power? Isn't this the libertarian argument against state funding?

[Ted Rogers 321]

In fairness to Donald he does also describe moments of 'counterperformativity': situations where the application of financial models produce real world results which are the opposite of the model's propositions. The back half of his and Milo's Black Scholes paper discuss this at length (the enduring volatility skew, post crash). His work on the 0.3 default correlation assumption in the structuring of CDOs also suggests a kind of counter-performativity at work.

[digital options]

Very interesting post. the Gerald Debreu and Kenneth Arrow theory is very unique and interesting. And I also liked what you said about the problem being not with economics, but with the class structure.