Streetwise Professor

I seldom comment on anything having to do with precious metals because, well, the subject tends to bring out, how to say it?, er, enthusiasts.  Monomaniacal enthusiasts.  But I feel somewhat compelled to do so, because Adrian Douglas, a board member of the Gold Anti-Trust Action Committee (GATA) structured his comments to the CFTC about metals position limits around one of my older writings on manipulation. (Douglas’s comments are all over the web–that’s just one link.)  (Douglas delivered some impromptu testimony at the CFTC hearings on the subject.)

Douglas starts out all nice and stuff:

I recently read some of the work of Craig Pirrong, a recognized expert on commodity markets and market manipulation.

But then says:

Despite Pirrong’s alleged [thanks!] expertise in commodity markets, he considers that manipulation is mainly instigated by the “long” market participants in what is colloquially called a “corner.” He writes:

“Other speculative activities sometimes called manipulative are far more ephemeral than corners, and are of dubious practical relevance. For example, farm interests and farm state legislators frequently assert that large short sales of futures contracts by speculators are manipulative, and cause prices to fall below their ‘true’ value.

“Such ‘bear raids’ are profitable for the raiders only under very restrictive conditions. In order to realize a profit, it is necessary to sell high and buy low — that is, the short seller must eventually buy back his positions at a price which is lower than the price at which he initially sold. Since the number of contracts sold is equal to the number of contracts subsequently bought, this can happen if, and only if, the futures price responds asymmetrically to the speculator’s purchases and sales. That is, the price decline caused by the speculator’s sales must exceed the price rise caused by his subsequent purchases.

“There is no credible evidence that such an asymmetry exists or has existed in futures markets. Moreover, it is even difficult to construct a theoretical model that exhibits this property. As a result, it is highly unlikely that short manipulations of the type that is criticized so vigorously by the opponents of futures markets are a practical concern. Indeed, futures industry experts have been nonplused by the allegations of widespread ‘downward’ manipulation as far back as 1921, when there was no regulation; most recognized the real danger of squeezes and corners, but were deeply skeptical of the possibility of short manipulations.

“Nonetheless, the primary impetus behind the regulation of futures markets in the early 1920s was the collapse in agricultural prices after the end of World War I. Despite the skepticism of the industry witnesses, the promoters of the legislation regulating futures markets, such as Senator Capper and Representative Tincher, both of Kansas, were convinced that short-selling speculators were largely responsible for this collapse. As a result, Congress was intent upon preventing manipulative short selling. However, since it could not distinguish legitimate short selling for hedging purposes, for instance, from illegitimate short selling, Congress simply proscribed ‘manipulation’ and passed the buck to exchanges by requiring them to prevent what Congress could not define — or face the closure of their markets.”

I find it astonishing that an alleged expert could make such a claim. In the futures market there has to always be a buyer and a seller for every contract; that is, a “long” for every “short.” This means that there are as many longs as shorts. Whatever model can be proposed for long-side manipulation must, by symmetry, be possible for the short side. If one can drive prices up by buying a large amount of contracts, one can also drive the market down by selling a lot of contracts.

The prejudice that manipulation cannot be effectively carried out on the short side is a fundamental barrier to any intelligent and productive discussion of manipulation of the precious metals markets.

What the right hand giveth, the left hand taketh away.  Although Douglas then goes on to use other things I wrote in that article to support his arguments for a crackdown on manipulation in the gold market, which suggests that I’m not all bad:)  (Some friendly advice, Adrian: citing me is probably not the best way to persuade Gary Gensler and Bart Chilton.  Just sayin’.)

Mr. Douglas needs to read a little further, and a little more carefully.  I was addressing one type of short manipulation that was commonly alleged in the 1870s-1930s; the “bear raid,” in which a manipulator drives down prices by selling large quantities of futures and then profits by  . . .

Well, that’s the problem.  It’s hard to see how this could be profitable, because the manipulator has to buy back the contracts.  If selling contracts drives down the price, why wouldn’t buying them back at the end of the manipulation drive up prices?  And given the frictions (bid-ask spread, etc.) wouldn’t transactions costs make this unprofitable?  That’s why I said that there has to be some (unexplained) source of asymmetry in price impact to make such a strategy profitable.

I would agree that if such a strategy is profitable for a short, then the mirror image strategy would be profitable too.  The problem is, I don’t see either strategy being profitable, as I don’t see reliable evidence of the necessary asymmetry.

Manipulation can work by driving down prices in one market can enhance the profitability of other contracts with prices tied to that price.  (Again, the same thing can work in reverse.)  But that’s different than what I describe in what Douglas quotes.

Moreover, both large longs and shorts can sometimes manipulate by exercising market power in the “delivery end game.”  A long manipulation of that type–a corner–is what I focus on in the article Douglas quotes, and what is undoubtedly the most empirically important kind of manipulation.

And contrary to what Douglas claims, I did show in my 1993 J. of Business piece (and in a related chapter in my manipulation book) there IS asymmetry in market power manipulation.  The conditions that make long market power manipulation profitable (inelastic supply and elastic demand of the deliverable commodity) tend to make short manipulation unprofitable (and vice versa).  Therefore, one kind of manipulation should predominate in a particular market.  Short market power manipulation is usually found in perishable commodities, such as potatoes and onions.  Perishability makes the demand for these commodities very inelastic in the short run, which makes driving down the price relatively easy.  (Historical aside: frequent short manipulations led to a federal ban of trading futures on onions, the only commodity so honored.  There’s the answer to your question, Renee:)

Since gold (and, to a lesser degree silver) are held as stores of value (investments), their demand should be relatively elastic, making short market power manipulation very difficult and unprofitable.

Ironically, in the testimony before the CFTC, the main scare story the gold bugs told related to long manipulation.  Noting that the open interest in gold derivatives exceeds the physical gold available for delivery, they said that there would be extreme price disruptions if there was demand for delivery on a large number of contracts.  But this would be a long market power manipulation that drives prices UP, not a short manipulation that drives prices down, and which GATA claims has happened for years.

And that’s another problem.  GATA alleges that central banks, in cahoots with bullion dealers, have manipulated gold prices down for years.  But most manipulations are very short term affairs that have relatively short-lived effects on prices.

The GATA story of how manipulation works also doesn’t make much sense.  GATA was most exercised in the 1990s and early-2000s, when gold producer hedging was quite common.  Producers would short gold forward to bullion dealers.  The dealers would cover this long position by borrowing gold from central banks, and then selling it.  GATA interpreted the gold sales as an additional flow of gold onto the market that depressed prices.

This is mistaken on many levels.  The stock of gold remained unchanged by these transactions, and the price of gold is determined by the size of the stock, and the demand for the stock.*  Moreover, just mechanically, a gold loan is a simultaneous sale and future repurchase, not an outright sale; the sale and subsequent repurchase are essentially a wash.

If anything, the gold loan mechanism, and the resulting possibility that a single firm can accumulate a long position that exceeds readily deliverable supplies, can make long market power manipulations possible.  There is some evidence in gold lease rate data that such long manipulations have occurred in recent years.  (This is analogous to the repo squeezes that occur in the Treasury market, and which were rife in the mid-2000s.)  But these have the opposite effect on prices than GATA alleges, and these effects tend to be temporary.

If gold hedging affects prices, it does so indirectly, and in a non-manipulative way.  By permitting producers to manage risks more effectively, hedging presumably allows them to increase output (e.g., it permits them to obtain financing on better terms to expand mines.)  This would tend to reduce prices, by increasing the rate of growth in gold stocks.  But this is not manipulative.  It represents a socially beneficial reduction in the costs associated with financial frictions.  This cost reduction is welfare enhancing, whereas true manipulation is welfare reducing.

So, the GATA story doesn’t wash with me.  It is just a repeat of the same allegations that have been directed at commodity derivatives markets for years, and suffers from the same fundamental misunderstandings of the ways these markets work.  The GATA allegations are made more lurid by the involvement of central banks, and the secrecy with which these institutions operate.  But at the end of the day, there’s nothing to see here folks, so just move along.  Not that the gold bugs ever will.  But I hope that having written this, I can.

* GATA and other gold bugs have chronic difficulties distinguishing between stocks and flows.  They point to the fact that the volume of trades in gold exceeds actual gold stocks.  The aptly-named (since it is high variance) ZeroHedge breathlessly repeats such things, claiming that this is evidence of a Ponzi scheme.  Did these guys just fall of the turnip truck?:  this is true for virtually every major commodity market.  For a variety of reasons, the number of transactions, and frequently the open interest, is a multiple of the underlying deliverable supply.  Douglas states that this has developed slowly over many years:

The futures markets were conceived to allow commodity producers to reduce risk by being able to contract to sell their yet-to-be-produced commodity at some time in the future. By the same token, it allowed commodity users to reduce risk by being able to contract to buy a yet-to-be-produced commodity in the future. The futures markets also served to perform the function of price discovery by matching future demand with future supply at a market-clearing price.

However, gradualism has resulted in the futures markets morphing into giant casinos where less than 5 percent of contracts get settled with a physical commodity delivery. The futures markets are now totally divorced from their intended function. The futures markets have become such a farce that the delivery of physical commodities has become an inconvenience that hinders speculation and market manipulation.

Uhm, not exactly.  The facts that the volume of trades, and open interest, exceed deliverable supply, and that very few contracts are settled by delivery, have characterized commodity markets since the birth of futures markets in the late-1860s.  They have been the subject of comment and frequent criticism since that time.  This is not, contrary to what Douglas says, a new thing: the “morphing” took place about, oh, 1867.