about me
game theory
... more
Subscribe Weblog

fuelingWhen you buy something you push its price higher. So when you accumulate a large position in something, you can expect it's price to rise. Further, if others note that you are doing this and you have deep pockets, you can expect others to attempt to front-run you or at the least not sell until you are thru buying. So accumulating a large position can really spike prices higher. But if you acquired a large position, pushing the price from $100 to $150, at an average price of $125, what is the expected sale price? The current price of $150 is only because of your demand. Once you are now committed to selling the exact amount you just bought, the price should plummet immediately, to $100 as people anticipate this final value: you don't make $25 per share, you lose $25 per share. It's simply not true that you can easily make money cornering goods, because while it is easy to move prices if you have a lot of money and using mark-to-market accounting make a lot of money in the short term, it is very difficult to retain these marks.

Yesterday BP traders were sued by the US Commodity Futures Trading Commission for cornering the propane market, by buying all the underlying propane in February 2004 (see WSJ article here, or pdf of the CFTC complaint here, and note all the CFTC's exhibits and audio tapes). I have no doubt BP did this, and it was coincident with a brief 50% spike in propane prices (see pic), and they bought wanting the price to go up (interestingly, most literature on market corners involve a different strategy, buying the futures and hoping to get a cascade of buying by shorts attempting to cover their positions).

Most importantly the CFTC notes that BP sought to generate a profit of $20 million...and they also note that they actually lost money because, as one of the BP internal correspondences notes:
While we called the upward price movement correctly (the Feb-Mar spread peaked at 34 cents/gal), the amount of volume we were able to move was significantly less than we predicted.
This note lays out the figures (conspicuously, the CFTC never mentions the actual profits made in this fiasco, only the expected profits). BP had 5mm BBL's, and expected to sell 2mm BBL's in Feb, and roll only 3mm BBL's, but instead they only sold 0.7mm BBL's in Feb and rolled 4.6mm BBLs, where they made 25 cents on the stuff sold in Feb, and lost 6 cents on the propane rolled (unsold). If k*(0.25*2mm-0.06*3mm)=$20mm USD, k=0.47 (where k is some gross up to account for the units they are using), then if we replace 2 with 0.7, and 3 with 4.6 , they actually lost $4.7mm (probably a little more, because they bought an additional 0.3mm barrels at inflated Feb prices)! This is the classic squeeze problem: you can push a price up, but you can't sell at the peak price, or even close to it.

This is a case for self-regulation. I'm sure BP wouldn't do this strategy again after such a failure. Now regulators are just piling on because tapes show intent. I'm sure many traders push around equities, on a much smaller scale, with the same intent, and I'm sure it's just as successful. This is not to say it's not possible, only that the nature of a successful squeeze is to use market power, and get others to think the generated pump in prices has more resources behind it than it does (just the pumper!). Otherwise, when you are fully invested, the market goes the opposite way, past your average price, and you lose. Both Niederhoffer's collapse in 1997, and LTCM's collapse a year later, were in large part because the market correctly sensed these guys had large positions, and anticipated their exit, causing them both to take much larger losses than would have been the case if the market were unaware of their precarious positions.