The CFTC Has a Very Hard Row to Hoe

The CFTC made all the front pages by filing manipulation charges against Parnon Energy, Arcadia Petroleum LTD and Arcadia Energy (Suisse), all affiliates of the trading firm Arcadia.

Let me say at the outset that I am in favor of addressing manipulation by ex post actions like those filed by the CFTC, and have long made that argument.  Let me also say that in 2002, Arcadia was accused of manipulation of Brent crude, and price patterns at the time were symptomatic of a corner, and per public reports, Arcadia’s trading was consistent with an attempt to corner.  Thus, I have no objection to the type of action the CFTC has brought, and don’t put it beyond Arcadia or a company like Arcadia to engage in manipulative conduct.  Indeed, I wish that the CFTC would rely more on ex post actions and less on things like position limits or ex ante jawboning to reduce the frequency and severity of market manipulation.

That said, based on the complaint, I conclude that the CFTC case is a tenuous one.  It is tenuous because of the novel nature of the claim, and the weak support for that claim provided by the price and trading data cited in the complaint.

The CFTC does not allege a corner.  Instead, it claims that Arcadia engaged in deceptive conduct.  It bought cash WTI forward contracts it did not need, then sold those contracts in the last three trading days of the WTI “cash window.”  The CFTC alleges that the buying deceived the market by suggesting buying interest that was not there, creating a false perception of market tightness.  This perception of increased tightness caused backwardation to increase.

Arcadia’s subsequent sales allegedly disabused the market of its (allegedly false) belief, causing backwardation to fall.  Arcadia allegedly lost money on these cash WTI purchases and sales, but made up for it through first buying and selling nearby spreads, and once the front month contract stopped trading on NYMEX, shorting the (new) nearby spread.  The creation of the perception and subsequent destruction of the perception of tightness, the CFTC alleges, generated profits for those spread positions.

Thus, the CFTC alleges a trade impact manipulation, rather than a corner.  These are always a greater challenge to prove, and the proof the CFTC adduces is hardly convincing.

In particular, the price data do not provide strong support for the allegation.  Consider the first alleged episode, in January, 2008.  Per the complaint, Arcadia bought 4.1 million barrels of February WTI cash crude.  During this period, it was also buying the February-March spread.  But in the 3 January-8 January period, when this trading–notably the cash trading, per the complaint–was allegedly creating an artificial perception of tightness in the market, (a) front month cash WTI fell by about $3.30/bbl, (b) the February-March (GH8) backwardation actually fell by 4 cents, and (c) front month WTI fell relative to front month WTS and LLS.  The spread did jump up 25 cents between the 8th and the 10th, and the WTI-LLS front month spread jumped by over a dollar, but this is after Arcadia’s purchase of the cash WTI contracts, and the purchase of the bulk of the GH8 spreads.  So to believe the CFTC, one has to believe that the market only became aware of the impending tightness with a lag–which is hard to reconcile with its assertion that it became aware that the demand was chimerical immediately on the sale of Arcadia’s cash position.

Moreover, by 15 January the GH8 spread had fallen to a level below its value prior to the commencement of Arcadia’s buying.  Then, when Arcadia began to sell the GH8 spread, the spread actually began to widen, going from 17 cents to 64 cents during the period that Arcadia was selling it. February WTI actually rose relative to February WTS during these days of selling February WTI as well.

This is all very hard to square with any assertion that the market was responding to Arcadia’s trading activity.  After all, NYMEX February futures and WTI cash forwards are effectively perfect substitutes, both allowing the buyer to take delivery of 1000 barrels of WTI in Cushing during February, 2008.  The market was showing greater signs of tightness at the very time Arcadia was reducing its demand for February WTI.  It will be devilish hard for CFTC to argue that purchases of near perfect substitutes have different price impacts, and for it to argue that the market was very sensitive to Arcadia’s trading of February WTI when its sales are associated with rising backwardation and its purchases are associated with falling backwardation.

The GH8 cash WTI spread did indeed go from backwardation to contango during the cash window of 25-28 January, when Arcadia was doing its selling of Feb cash WTI.  But February WTI actually rose during this period, and indeed rose on 28 January when Arcadia did most of its selling.  Again, this is not symptomatic of a market that was suddenly convinced that demand tightness had declined.

It must also be noted that the collapse in the February-March cash spread was equally large for LLS ($.985), and more than twice as large for WTS ($2.335).  Thus, the spread collapse was not limited to the product that Arcadia was trading.

Perhaps CFTC will argue that Arcadia’s deceptive buying had lifted and then crushed LLS and WTS spreads as well, but any manipulative buying should have a larger impact on the commodity actually bought, and the 25 January-28 January spread moves are not consistent with that.

What’s more, making this argument would present the CFTC with an acute problem.  It defines “market dominance” based on the size of Arcadia’s position relative to the WTI market.  If LLS and WTS are such close substitutes for WTI that buying one impacts the prices of the others because they are effectively one market, then for the purpose of determining market dominance Arcadia’s position must be measured relative to that market–WTI, WTS, and LLS, plus any other close substitutes.  It would be much harder to prove market dominance under such an expanded market definition.

In brief, the price and trading patterns for the January episode are not consistent with the CFTC allegations.  Prices fell throughout, even after Arcadia allegedly fooled the market into believing that conditions were tight.  Prices and backwardation did not rise when Arcadia was buying.  Backwardation was rising when Arcadia was actually selling the spread.  The spread collapse during the cash window that the CFTC apparently believes is its “Gotcha” piece of evidence was mirrored by collapses as large or larger in other types of oil.

I also note that the CFTC has provide no evidence that any of the price and spread moves were statistically significant.

The spread and trading co-movements are more equivocal during the March, 2008 episode.  The April-May (JK8) spread did rise when Arcadia bought it, and during the period it commenced buying April cash WTI.  But the biggest move in the spread, from $1.16 to $1.94 occurred when Arcadia sold its JK8 NYMEX spreads.  It was buying 3.5 mm bbl of April cash during this 14-19 March period,  but it was selling 14.4 mm bbl of JK8 spreads–meaning that it sold 10.9 million more barrels of April WTI than it bought during this period–so how could this contribute to an explosion in the perceived tightness of WTI–as indicated by the 78 cent increase in the backwardation? Note too that the WTI April cash flat price was falling $5.73 during the period it was buying the 3.5 mm bbl of April cash crude–another thing that is hard to square with a perception of increased tightness.

Furthermore, the April-May LLS spread fell by more during the cash window ($2.29) than did the WTI spread ($2.015).  This is very hard to square with any coherent manipulative theory, and as noted before, any story in which sales of WTI cause declines in the price of LLS would require a market definition under which it would likely be very hard for CFTC to prove market dominance.

I should also mention that even ignoring this problem, the CFTC’s market definition is suspect.  It compares Arcadia’s position to the company’s estimate of flows of WTI in Cushing during the delivery month.  But this ignores stocks of oil in Cushing, which were about 17 mm bbl during this period.  Now some of these barrels were not WTI, but some if not most of these barrels should be included in any calculation of the size of Arcadia’s purchases relative to the size of the market.

Corner manipulation cases are hard: the CFTC has never won one.  Trade impact manipulation cases in which it is alleged that buying or selling created false perceptions of demand are even harder to analyze and prove.  Thus, just based on the nature of the allegation alone, the CFTC has filed a very challenging case.  When one looks at the evidence the CFTC presents in its complaint, the odds become even higher.  For the January episode in particular, the most straightforward interpretation of the evidence cuts squarely against the allegations.  This will be a very difficult case for the agency to win.

There’s another lesson here that has been lost in all of the hue and cry over the filing of the complaint.  CFTC has been examining the oil market with a fine tooth comb going back to 2005 if memory serves.  If this is the best case they can find after all that, the oil market must be pretty damn clean.

• Become A Funded Futures Trader