Streetwise Professor

Apparently “orthodox” is the sneer du jour.  A couple of weeks back I noted that IMF candidate Agustin Carstens was being denigrated as orthodox and therefore unfit to head the IMF.  And today, in a story about the Singleton oil market study I blogged about last week, Reuter’s John Kemp dismissed me, Scott Irwin, and Jeff Harris as a cohort of orthodox economists who dismiss the role of financial speculators in affecting commodity prices.

I’m only PO’d because Kemp listed Irwin first.  That wounds me to the quick.

Apparently I’m a member of an “establishment”–when do I get promoted to mastermind of an evil cabal?–that Svengali like, induced the CFTC, IOSCO, the FSA, the OECD, and presumably the IEA (though Kemp doesn’t mention it) to conclude that speculation had not distorted energy prices, or the prices of other commodities.

Who knew the vast influence I wield over official bodies world wide? I certainly didn’t.

Regarding orthodoxy, my views are certainly not considered orthodox, on say, Capitol Hill.  Or on the Champs Elysee.  Or in Brussels.

No, “orthodox” is shorthand for People Who Don’t Agree With John Kemp, whereas those who do are Brave Challengers of the Consensus.

Now to the substance of what Kemp has to say, such as it is.

First, as I made clear in my post on Singleton’s piece last week, I do not deny the possibility that speculation has “influenced prices,” as Kemp insinuates.  In fact, such a contention would be absurd, and Kemp is absurd for insinuating it.  What I have been on about is whether speculation has distorted prices.

Big difference, John.

There have been specific allegations that speculation has led to substantial distortions in commodity prices, particularly energy prices–by 10, 20, hell 50 percent.  That is what I have been responding to, and I stand by my contention that there is no evidence of such an effect.

Nor does anything in Singleton’s piece support such claims.

Variations in measures of speculation can have predictive power over returns–that is, they can influence prices–without implying that speculation has distorted them.  As I noted in my post on Singleton’s piece, limits to arbitrage-type stories predict that shocks to speculator balance sheets or hedger balance sheets can lead to associations between speculative activity and the drift of commodity prices.  This is perfectly consistent with a refinement of Keynes’s normal backwardation theory that Kemp identifies with the dreaded orthodoxy.  (Truth be told, I made an argument along these lines in a presentation at the FMAs in October of 2002.  And I’ve made it year after year in my PhD seminar on derivatives, in my discussions of incomplete markets.)

Insofar as the behavioral stories are concerned, yes,  they can also imply that variation in speculative activity have predictive power over returns.  They also can predict boom and bust cycles.  But it is notoriously difficult to determine whether these sorts of behavioral effects are what explains the predictive power of measures of speculation.

I would also note that purely rational storage models can also generate booms and busts.  Indeed, a rational storage model can match the behavior and time series properties of prices and inventories–I show that in my forthcoming book.  This is true for both pre-2004 and post-2004.  It is just flat wrong to say that the price and quantity movements observed in recent years cannot be reconciled with rational, fundamentals-driven models.  So the existence of commodity booms and busts–which way, way antedate the financialization of commodity markets–does not refute fundamentals-based explanations, or prove the existence of behaviorally driven anomalies.

Singleton points out–and Kemp emphasizes–a series of theoretical possibilities and some evidence that is consistent with theories that generate these possibilities–but which are also consistent with other explanations.  A very weak basis indeed to justify wholesale intervention in the markets.  Especially given the very vital role that speculation plays in achieving an efficient allocation of risk.

The behavioral theories are also not sufficiently well-developed to generate testable predictions about co-movements between prices and quantities and the dynamics of forward curves.  Predictions about quantities are particularly important.  As I’ve stated on numerous occasions, commodities are consumed in the here and now (in contrast to say, internet stocks), so distortions in prices should show up in distortions in quantities (notably inventories).  Do behavioral models predict this?  I haven’t seen any that have even tried.  If they do–the evidence doesn’t support it.  If they don’t–then why should we really care, as this would mean that speculation is not distorting consumption and production decisions?

And how can Kemp seriously say that Singleton’s paper “offers a richer and more realistic view of how real markets operate than the rather stylized formulations that have been popular with the old guard”?  [And who are you call old, boy?]  Has he ever looked at any of these behavioral models?  You want to see stylized?  Case in point–the Hong-Yogo model, which reverse engineers an extremely contrived–actually, artificial–behavioral setup that mimics the empirical result already produced.  Sorry, but that’s not science: model first, then test.  Moreover, as I just noted, these behavioral models don’t make any predictions about quantities, forward curves, and other crucial aspects of commodity markets.  How can you possibly consider models that don’t make predictions about such crucial variables “richer” and “more realistic”?  Especially given that these quantity predictions are what is really relevant in determining the welfare effect of speculation.

Contrast that to my work on commodities.  I take a rigorous model, derive (computationally) its implications for prices, forward curves, quantities, quantity-price comovements, volatilities, and forward price correlations, and take those predictions to the data.  I reject many of the implications of the models, and use those rejections to help understand the true richness of how prices behave.  If I’m orthodox–that’s the sense in which I am orthodox.  I’m not wedded to the models.  As I say in the introduction of the book, I look to break the models and to learn from the pieces.

And as I noted in my post on Singleton’s paper, the policy implications of the behavioral theories are ambiguous–and certainly don’t support a role for position limits on large speculators.  As I noted, if you overlook the cheesy model and believe the Hong-Yogo behavioral interpretation of their empirical evidence, behavioral effects have long been present in these markets–long before financialization.  What’s more, the behavioral models suggest that prices depend crucially on the composition of traders in the marketplace.  Policymakers can’t micromanage that composition to achieve more “rational” pricing through blunt tools like position limits.

So as I predicted in my Singleton post, people would seize on his study to advance their regulatory agenda, even though it provides only flimsy, equivocal support for that agenda.   Not surprisingly, John Kemp is leading that parade, giving further proof to  the aphorism that a little knowledge is a dangerous thing.