The CFTC is supporting Congressional efforts led by Sen. Levin of MI to pass legislation that would allow it to impose position limits and obtain position information on energy derivatives traded on ICE (and presumably other electronic OTC trading platforms) in order to prevent energy price manipulation:
The CFTC is seeking mandatory daily reporting to the agency of large ICE trades, limits on the number of ICE energy contracts a single trader could control and exchange rules to force a trader to unwind large contract positions.
To recycle a line I’ve used before, they want to reduce manipulation in the worst way–and they’ve found it. Position limits and position reporting are inefficient ways to reduce the frequency of manipulation, as compared to readily available alternatives. There are so many defects in these remedies, it’s hard to know where to begin.
For one thing, position limits must be set based using very crude information about the market conditions that will prevail when the position limits are in effect. Sometimes the limits will be too large to constrain much manipulative behavior. Other times, the limits will be smaller than necessary to combat manipulation, and will be so small that they constrain the legitimate activities of market participants.
This last effect is very important. It is economically efficient for some traders to accumulate large speculative positions. These speculators absorb risk from hedgers. Position limits constrain speculators’ ability to take on risk, meaning that hedgers will find it costlier to transfer risk to speculators. They will hedge less as a result. That’s a real cost.
I have just completed a paper that shows that even if position limits constrain manipulation, they reduce overall welfare because of they interfere with the efficient allocation of risk. In essence, position limits throw out the baby with the bath. Sure, they constrain the rent seeking, manipulative activities of speculators, but they also constrain speculators’ efficiency enhancing risk bearing function too. In the standard mean-variance/Anderson-Danthine hedging/speculation equilibrium model of my paper, the latter effect outweighs the former, and position limits reduce welfare. Thus, even if they are successful in reducing manipulation, that doesn’t mean that they’re good public policy.
This possibility has been totally ignored in the debate, such as it is, over position limits in energy. This reflects a general ignorance of the socially beneficial effects of speculation, and indeed, the hostility to speculation that has dominated the Congressional and regulatory intellect for decades (stretching back into the 19th century.) Risk transfer is a primary function of derivatives markets, and position limits interfere with this primary function. This is a real cost, but one that is too often ignored.
Position reporting wins no prizes for efficiency either. Presumably the theory is that the ability to monitor positions gives regulators the ability to identify incipient manipulations, and intervene before things get out of hand. Now let’s deal with the reality.
The first reality is that how big is big–i.e., how big a position must be before it poses a manipulative threat–depends on variable market conditions, and how the trader acts. Just seeing that somebody holds a big position tells you absolutely bupkus. As a result, regulators will start to scrutinize big positions only when price distortions become manifest in the marketplace. Well, by the time this happens, a lot of the damage is already done. Maybe the regulators can keep things from getting worse, but the idea of preventing manipulation before it affects prices is a pipe dream. Even when armed with position information, regulators only become engaged when price distortions become apparent.
And even when they intervene, there is no guarantee that they will succeed. They can jawbone, but sometimes regulators can be fooled, persuaded–or intimidated–into inaction by manipulators. Think that doesn’t happen? Want to buy a bridge?
Regulators can also be wrong, and it is important to remember that when attempting to intervene during an alleged manipulation regulators necessarily have worse information than is available after things play out. The manipulative “end game” (as I call it in my ’93 JOB piece on manipulation) often provides the most powerful information on whether a manipulation in fact occurred. For one thing, the price movements post-manipulation–the “burying the corpse” effect–often provide the strongest evidence that prices were distorted. (After a corner ends, the manipulated price collapses relative to other prices as the excess deliveries are dumped on the market. These deliveries are the “corpse” of the corner, and disposing of them is referred to in the biz as “burying the corpse.”) (And to reiterate, Bill, I didn’t make up that phrase. P. D. Armour did. Sorry for the inside humor, folks, but I couldn’t resist–although I did restrain myself from making a really cutting remark.) Also, the alleged manipulator’s behavior–whether he takes deliveries at prices that appear uneconomic–provides the strongest evidence on his motives and intent.
Errors are important. Regulators can intervene and force liquidations of large positions when there is really no manipulation, or they can fail to intervene when a manipulation is actually in progress. Both kinds of errors are costly. Given the lack of crucial information before things play out, for a given probability of one type of error, the other type of error is higher when regulators try to intervene before the fact than afterwards. As I’ll discuss momentarily, civil (or even criminal) sanctions imposed after the fact can utilize better information, reduce the frequency of one or both kinds of error, and reduce the frequency and severity more efficiently than early intervention.
Reporting is wasteful for other reasons too. It is costly to collect this information, most of which will never be needed to detect or prevent a manipulation.
Moreover, traders can avoid position limits and reporting by moving their activities into the non-electronic OTC market, using voice brokers or just bilateral deals with other traders. To the extent that it is more efficient to execute transactions electronically on a platform like ICE, this displacement of trading activity is inefficient (though it is a boon to voice brokers.)
So what is better? Dropping the hammer on manipulators after the fact, that’s what. Law and Economics scholars (notably Steve Shavell) have identified conditions under which deterrence through the imposition of sanctions ex post is more efficient than ex ante prevention. Most importantly, deterrence is efficient (relative to prevention) when the offense is detected with high probability, and malfeasors are not judgment proof.
Both conditions characterize manipulation in derivatives markets. As Judge Frank Easterbrook once wrote, an undetected manipulation is an unsuccessful manipulation. A successful manipulation distorts prices–and everybody in the market can observe it. Moreover, most manipulators are well heeled, and can pay legal judgments levied against them.
By waiting to act until after a putative manipulation runs its course, regulatory or judicial authorities–or private plaintiffs harmed by the manipulation–can rely on the best information available on the price impact of the manipulation–including the “burying the corpse” effect–and on the alleged manipulator’s actions. Thus, there should be fewer errors (incorrect convictions, or incorrect acquittals) after the fact than when regulators attempt to intervene during a suspected manipulation.
Litigation is expensive, to be sure. But the expenditures can be targeted, and only incurred when the harm occurs. Position information need only be collected via discovery when the evidence of a manipulation is manifest, and from the likely culprits (who can be identified pretty easily, as there is a lot of information about who is doing what is floating around the markets, and some manipulative behavior is hard to miss, e.g., who ends up owning all the deliverable supply.) This is cheaper than collecting information from everybody (virtually all of whom are completely innocent of anything) all the time (when usually nothing amiss is happening).
So, in a nutshell, deterrence rules, and position limits/reporting are for fools. And guess which way Congress and the CFTC are moving? Draw your own inferences, people; do I have to draw a map?
Ironically, the news about the Levin bill and the CFTC advocacy thereof comes almost simultaneously with the announcement that BP has agreed to pay $300 million(!) (subscription required) to settle civil charges that it manipulated the propane market in 2004, and that four BP traders are facing criminal charges (ditto) arising out of the same incident. I have mixed feelings about the criminal aspect of this case, but the imposition on a hefty fine on BP is the way to go as a way to deter this sort of conduct going forward. Three hundred large large is some major change–especially considering that the alleged (and I guess now admitted) BP corner attempt was apparently unprofitable because the company ended up with a much bigger corpse to bury than it had anticipated. Surely any trader contemplating cornering, squeezing, or hugging a market in the future will think once, twice, and a few more times, before squeezing the next time. It should also be noted that all this activity took place in the OTC market, and that it did not take place on ICE, or to my knowledge, any electronic trading platform. Nonetheless, the manipulation was identified, the culprit fingered, and the hammer brought down. No need for watching everybody all the time, most of said watching effort being completely unnecessary because much of the time manipulation isn’t a serious worry.
For those of you with way too much time on your hands, I have a much expanded analysis of the deterrence vs. prevention debate in my book The Economics, Law, and Public Policy of Market Power Manipulation. (Shameless plug! Now up to #1,302,119 on Amazon! Only 3 copies left! Order soon!) I also presented some rigorous numerical analysis of the power of detection ex post vs. ex ante in the working paper version of what eventually turned into my 2004 ALER article on the Ferruzzi squeeze. I think that version is still posted on my UH web page.
One last comment. I find the obsession with energy market manipulation to be somewhat curious. The US government has manipulation problems much closer to home–the US Treasury securities market. Manipulation problems in this market became so chronic in the early-2000s that two high ranking Treasury officials raised the subject in public speeches, and the FRBNY called bond market heavyweights to a Come-to-Jesus meeting, where the assembled congregation was sternly warned to go away and sin (that is, manipulate) no more. Despite the importance of the Treasury markets, and their size, Congressional interest in this problem seems de minimus as compared to the constant hyperventilating over energy market manipulation. Presumably this has to do with the fact that energy prices affect pretty much everybody, whereas the Treasury markets (especially the repo market) are pretty much inside baseball; Congress has to be seen Doing Something about energy prices even though there is little if any connection between persistently high energy prices and the kinds of manipulation that are allegedly the targets of legislation, and even though (as discussed above) the legislation is an inefficient way to reduce those kinds of manipulation.
I was thinking of inserting an apposite Mark Twain quote here (you probably know the one I’m thinking of), but won’t gild the lily. Like Mom says, if you can’t say anything nice . . . I know I’ve broken that rule a lot already in this post, but I’ll follow it just this once.
Correction. I wrote that to my knowledge, none of BP’s propane transactions were consummated on an electronic system. I took another look at the CFTC’s civil complaint against BP, and noticed that market participants utilized an electronic trading platform called Chalkboard to trade propane. I don’t know the breakdown of volume between voice brokers and Chalkboard.