Streetwise Professor

Almost exactly a year ago, I testified before the House Ag Committee and wrote an oped for the WSJ on the effect–or lack thereof–of speculation on oil prices.  The issue was high on the political agenda at the time, as oil prices hit $147/bbl, and gas prices were above $4/gal.  The collapse in prices during the financial crisis put the issue on the backburner, but it is back with a vengeance.  To loud Congressional hosannahs, the CFTC has announced its intent to impose a far reaching regime of position limits on energy and other commodities “in finite supply“–pray tell, what traded commodities are in infinite supply?

So, I guess I have to roll that rock up the hill, yet again.  (I have posts on the subject of “excess speculation” and position limits from 2006, 2007, 2008, and 2009.)  

At the risk of becoming even more of a repetitive bore than usual, I’ll restate my objections to the intellectual basis to these policy initiatives, and the policy initiatives themselves.  Before doing so, it is worthwhile to define “excessive speculation,” something those that hurl around the phrase with abandon almost never do.  I define “excessive” speculation as that which drives prices away from the competitive price consistent with available information.  That is, excessive speculation distorts prices.  Now, my objections:

• The mechanism by which speculators allegedly distort prices is almost never specified, and when it is, it makes no sense.  Advocates of the “excessive speculation” hypothesis routinely say that speculators inject artificial demand into the market.  Given that most speculators almost never actually take delivery, and thus offset their positions as they become prompt, they are not influencing either supply or demand for the physical commodity.  This is a point that has been recognized since at least 1902 in a report authored by the US Industrial Commission, which included several Congressmen and Senators.  Which I guess goes to show that Congressional IQ exhibits an inverse Flynn Effect.  
• This is especially the case for commodity index traders, who almost all trade cash settled instruments, and hence can in no way influence supply or demand.
• That is, none of the boatloads of money allegedly flowing into commodity funds actually goes into, you know, actual commodities.  Some of the money in commodity-linked investments is purely notional (e.g., in commodity swaps, or index swaps).  Some goes into T-bills (in the form of margin, or in fully collateralized commodity trades).  
• The evidence of a distorting speculative effect is laughable.  Distortions in prices should lead to distortions in quantities, e.g., in commodity stocks.  No reliable evidence of such an effect exists.  Indeed, I have looked at one commodity–copper–in close detail as part of a chapter of my upcoming book on commodity price modeling.  People were screaming about a speculative bubble in copper last year at the same time as the (louder) screams about a speculative bubble in oil.  But copper stocks (and the stocks of other industrial metals) were drawn down sharply at the same time prices were spiking: this is exactly what one would expect to observe with a positive demand shock.  Indeed, I am working on calibrating a rational expectations theory of storage model to copper data, and this model can explain the twists and turns of prices and stocks during the extreme events of the last several years.  That is, speculative irrationality is not required to explain the movements in copper prices AND inventories over the 1997-2009 period.  (I’ll post more on this when I finish a couple more parts of the analysis next week, hopefully.)
• Some of the “evidence,” logically ungrounded as it is, doesn’t even support the assertion of a speculative effect.  For instance, a common piece of “evidence” is that prices rose right along with measures of long speculative interest in the oil markets.  (This was a Michael Masters favorite).  Set aside the correlation-causation post hoc/propter hoc problems, and the potential for spurious regression problems whenever one examines the relation between two highly autocorrelated series.  Just look at the graph in this WSJ story:

Sure, prices and long non-commercial positions were both trending up in 2005-2008 (not that that means a damn thing).  But look at what happens in late-08 and early-09.  Prices and long spec positions are mirror images: prices were plunging when long interest was rising, and then, just as long spec interest peaked and began to fall, prices ceased their fall and moved strongly upwards.  How can you possibly explain this if the market dances to the speculators’ tune?  The speculative Pied Piper story says that these measures should move up and down together, not in opposite directions.  Now, this Masters style analysis is B.S., but those who used it to rationalize their assertions of a speculative effect are hoist on their own petard: they vouched for the validity of non-commercial long open interest as a measure of the “excess” demand for oil caused by speculation, but the most recent experience is decidedly inconsistent with that story.  Not that that they have even paused for thought, but continue to flog the same story.  

• Not to mention that numerous analyses, relying primarily on Granger causality, have failed to find any reliable connection between changes in speculative positions and price changes.  Or the analysis (by the CFTC!) showing that commercial traders tend to be market leaders and speculators are followers.

The calls for government regulation have gone international, and gone beyond calls for mere position limits.  Gordon Brown and Nicholas Sarkozy penned a WSJ editorial advocating widespread governmental “supervision” (Euphemism alert!) of the energy markets (including cooperation with OPEC):

For two years the price of oil has been dangerously volatile, seemingly defying the accepted rules of economics. First it rose by more than $80 a barrel, then fell rapidly by more than $100 before doubling to its current level of around $70. In that time, however, there has been no serious interruption of supply.

The oil market is complex, but such erratic price movement is cause for alarm. The surge in prices last year gravely damaged the global economy and contributed to the downturn. The risk now is that a new period of instability could undermine confidence just as we are pushing for recovery.

Where does one even begin the ridicule (beyond noting the point that whenever French and British politicians agree, it must be a really bad idea)?  Regarding volatility.  Uhm, the past two years have been among the most volatile in the memory of most living people–you have to go back to the 1930s to find anything remotely similar.  In 2007-FH2008, Chinese (and Asian growth generally) greatly spurred demand for commodities.  Note that in addition to commodity prices, shipping charter prices skyrocketed, even though the price of something perishable like transportation on a ship can hardly be distorted by speculative buying, because it has to be consumed and hence is impossible to hoard.  That was followed by a financial collapse and economic recession of severities unseen since aforementioned 1930s.  Look at every measure of uncertainty, notably such things as the VIX volatility index.  These things have been at dizzying heights since last August (and had begun their rise even earlier, in 2007).  What would be weird is if oil prices (and commodity prices generally) HADN’T been volatile during this period.  

Regarding “serious interruption of supply.”  Uhm, Gordo, Nicky–there’s this concept called “demand.”  Believe it or not, it’s one of the “accepted rules of economics.”  Don’t take my word for it.  It’s in all the textbooks.  Price evolution in the recent period is clearly demand driven, not supply driven.  Prices spiked when demand spiked.  Prices plummeted when demand cratered.  

And there have been supply shocks.  Nigeria is an ongoing source of supply shocks, and during the price spike last summer losses of sweet Nigerian crudes in the face of very high demand for distillates (driven in part by European regulations regarding clean diesel) led to big spreads between benchmark sweet crudes like WTI and Brent and sourer crudes.  What’s more, OPEC output cuts (especially by Saudi Arabia) played a key role in the firming of prices following the price collapse in October-January.

I should also note that recent research by Lutz Kilian (nicely summarized in his JEL piece) shows that movements in oil prices hew more closely to movements in proxies for demand than measures of supply.  James Hamilton disputes Kilian’s contention that supply shocks are unimportant, but Kilian does demonstrate that demand is clearly an important price driver.  (I also have some reservations about some of Kilian’s analysis, most importantly his conclusion that changes in demand for precautionary inventories is also an important driver of prices.  I don’t disagree with the logic, having written several posts and an academic paper–which will also be another chapter in the book–showing rigorously how that can work in practice.  My problem is that Kilian does not measure factors that could affect precautionary inventories directly, but instead attributes oil price movements not explained by supply or demand measures to an “oil market specific” factor which he asserts is related to the precautionary inventory effect.  I am currently working on developing some more direct measures that would influence precautionary inventory holdings.  I have a good handle on measuring demand side risks that could influence such holdings, but measuring supply side risks is far more difficult.)

In recent months, the oil market and world equity markets have exhibited unprecedented positive correlation (something I called attention to in October-November).  Oil and equities swooned together; they rallied together March-June; and have been selling off together in the last couple of weeks.  This reflects the dominant role of demand in commodity prices at present.  Both equities and commodities are responding to rapidly evolving views in a highly uncertain situation about the likelihood and strength of economic recovery.  They are also responding to government policies, notably aggressive quantitative easing strategies that raise the specter of inflation.  

Now, you could argue that this is speculative.  And indeed it is.  Both equity market and commodity market participants are speculating on the future course of the world economy.  But is it destablizing speculation?  

What’s more, to the extent that energy prices (and commodity prices) are reflecting information/perceptions/sentiment about the future course of economic recovery, and are reflecting the same information/perceptions/sentiment as equity markets and bond markets, why would anyone think that position limits on commodity derivatives would break this linkage?  Once embedded in equity prices, information/perceptions/sentiment about future economic growth become public information that will be embedded in prices even if speculators are forced to trade less.  

The arguments against position limits as a policy are well known.  Speculators absorb risk from hedgers, and supply liquidity.  They therefore impair ability of the markets to serve their essential risk shifting function.  They therefore increase the costs of managing risk, and impose harm on hedgers.

This makes a mockery of the distinction that politicians and regulators draw between the “good hedgers” with price exposures inherent in their businesses that need to be managed, and “bad speculators” who have no underlying exposure to manage, but are taking on risk like gamblers in a casino.  But there is a symbiotic relationship between hedging and speculation.  To whom do hedgers transfer the risk if not to speculators?  And if there are no speculators, how is it possible to hedge risks?; after all, hedging doesn’t make risk magically disappear, it just transfers it to somebody willing to hold it–a speculator.  Thus, politicians and regulators who invariably say that they just want to protect “legitimate” commercial participants from the predations of evil speculators are more full of it than usual.

That said, there is an apparent puzzle: many of the most ardent critics of speculation are physical market participants.  But it’s not necessarily a puzzle, for the same reason that nobody should be amazed that Walmart supports Obama’s health care plan: many firms benefit from regulations that make markets less efficient.

In the commodities case, here are a couple of examples.

It is well known that some commercial market participants are informed traders who make money on the basis of their information.  A Cargill, for instance, has an extensive information network that allows it to trade profitably.  (I sat at the trading desk of Continental Grain some years back, and observed that the traders had access to information around the world about weather conditions and grain flows due to their physical presence in numerous markets.  This gave them an information edge.)  Such firms don’t hedge mechanically.  They trade futures and other derivatives strategically to optimize a risk-return trade-off.  Their information advantage allows them to time the placement and unwinding of their hedges so as to make money.  

Competition from informed speculators tends to erode the competitive/information advantage that informed commercial firms have.  Hamstringing such competition could provide a greater benefit to such firms than would result from the loss of liquidity from speculators.  

As another example, a reduction in market liquidity resulting from speculative position limits almost certainly has disparate impacts on commercial firms.  Some firms, due to capital structure, operational leverage, real options, or other reasons, get more benefit out of hedging than other firms in the same industry.  A regulation that reduces market liquidity raises everybody’s costs, but the first type of firm’s costs go up more than the second type of firm’s.  Thus, the second type of firm may actually support regulations that make hedging costlier or less effective because it raises their rivals costs more than its own costs go up.  

So, the support of some commercial firms for restrictions on speculation is not necessarily a puzzle.  In my view, it more likely reflects an opportunistic use of the regulatory process to distort competition in a way that profits some firms.  This is a ubiquitous phenomenon, observed in many industries and many regulations: why should derivatives trading be any different?  

Well, that’s all I have to say for now, though I doubt this will be last word on the subject, hardy perennial that it is.  Like Sisyphus, I am resigned to the unending task of rolling the rock of enlightened reason up the slope of benighted ignorance, only to see it slip back down again:)