What are positions limits supposed to do, exactly, and how? Here I want to focus on all month limits, not spot month limits. Spot month limits can constrain market power manipulations. I think there are better ways to do that, but at least there is a reasonable goal, and a reasonable connection between means and objectives. Can the same be said for all month limits? I am extremely skeptical, especially limits of the kind currently under consideration by the CFTC.
Here’s what the CFTC website says:
To protect futures markets from excessive speculation that can cause unreasonable or unwarranted price fluctuations, the Commodity Exchange Act (CEA) authorizes the Commission to impose limits on the size of speculative positions in futures markets.
I’ll take as given the theoretical correctness and empirical relevance of the highly dubious but positive assertion that “excessive speculation that can cause unreasonable or unwarranted price fluctuations,” and focus on the ability of position limits to address this problem. I’ll focus particularly on the kind of position limits the CFTC is currently considering, which set limits as a fraction of open interest.
It’s first important to note that all limits that have been in force and have been proposed since 1936 have or would have limited the permissible size of the position any individual entity can hold, not the size of speculative positions in aggregate in futures markets. That’s problematic when one considers several of the most widely cited examples of speculative bubbles in history. (Again, not endorsing that any or all of these are truly speculative bubbles. Just seeing if, under the assumption that they were, speculative position limits of the kind implemented and mooted in futures markets would have had the slightest impact.)
A list of such speculative bubbles might include tulips in 17th century Holland, the Mississippi Bubble in 1719, stocks in the 1920s, the internet boom, and US real estate in the 2000s. In historical recountings of these episodes, each was characterized by widespread participation by small traders, e.g., housewives speculating on RCA in the 1920s, or individuals putting their 401(k)s in internet high-fliers in the 1990s. If they were manias, they were popular–or populist–ones.
Limiting the size of the position that any individual could have held would not have constrained seriously the overall amount of speculation in these episodes, because the aggregate speculation was driven by a large number of small positions. When speculation is excessive because of an excessive number of small or medium-sized deluded traders plunge in the market, and predominately on the same side, limiting the size of any individual position will not constrain this form of excess.
In modern derivatives markets, much speculation takes place through funds or ETFs that aggregate the monies of large numbers of individual speculators into large pools. If a popular speculative mania leads to an influx of money into these pools, and if there are sufficient economies of scale in managing these pools, then limits on the size of individual positions under the control of one entity could effectively tax speculation, although in a very indirect way. If the constraint is binding on these pools, it would mean that they are not able to exploit fully their economies of scale. This would mean that they incur higher costs than in the absence of the constraint. This would raise the cost of speculation by small investors, which would tend to reduce the amount of speculation they undertake. How much would depend on (a) the difference between the cost of operating a pool at efficient scale, and the (higher) cost of operating at a scale imposed by the limit, and (b) the elasticity of speculation with respect to transactions costs.
I don’t know for certain what (a) and (b) are. (I do know that I haven’t seen anyone even point these out to be relevant factors in evaluating position limits.) My intuition is that they are small, meaning that speculative limits would not appreciably constrain the mass speculative frenzy species of excessive speculation (if it indeed exists or is empirically important).
Moreover, it is not obvious that position limits of the type currently under consideration by the CFTC would even uniformly achieve this objective. The currently contemplated limits would set maximum position size as a percentage of open interest. So permissible individual position sizes in big markets (e.g., crude oil) would therefore be larger than permissible individual position sizes in smaller markets (e.g., wheat). But scale economies relate to absolute size, not percentages of total market size. And as Adam Smith put it, economies of scale are limited by the extent of the market. This means that the difference between cost at efficient scale and cost at aggregate outputs that differ from integer multiples of the efficient firm scale is declining with market size. Thus, the kind of constraint under consideration would have a much smaller effect on the amount of speculation in big markets than in small markets, regardless of the elasticity of speculation with respect to transaction cost. (The analysis is trickier when there is considerable heterogeneity among the cost curves of different pool operators. Economies of scope–operating pools in different markets–are also relevant. Index investing raises additional complications.)
This all means that speculative position limits are not well suited to addressing the popular speculative wave species of excessive speculation, especially in large markets (like the energy markets). This is especially true of position limits based on percentages of open interest.
Given that many of the conventional and political attacks on commodity speculation in the 2000s were–and continue to be–based on variations on the popular mania explanation, imposition of speculative limits will leave their supporters disappointed, for two reasons. First, because these attacks are dubious theoretically and empirically. Second, because even if such speculative waves sometimes cause unwarranted price fluctuations, limits on the positions any individual entity can hold will have little, if any effect, on the size, strength, or price impact of these waves.
This means that any defense of position limits must be predicated on a belief that a position, or a handful of large positions, that are large relative to the overall market, can cause unwarranted price fluctuations.
There are examples of this historically, of course. The Hunts come to mind immediately.
Putting aside manipulation for a moment (and it is hard to see what type of manipulation all month limits would stop), when one trader takes a disproportionately large position, it means that he has a far different view of the future course of prices than other traders. He’s taking all the action (or a disproportionately large amount of the action) on one side of the market because at the current price, his view on where prices are going is the opposite of consensus opinion (or perhaps because there’s a big imbalance among hedgers). This is contrarian speculation. If the contrarian speculator is right–he makes a lot of money. If he’s wrong–he loses a lot of money.
One justification for constraining this type of trading is a belief that such traders are more likely to be wrong than right. They are unduly confident in their own abilities to forecast prices, or unduly convinced of the superiority of their information. Their trading on these false beliefs does drive prices away from where they should be, and when reality intrudes, prices snap back. Those movements could be characterized as unreasonable or unwarranted price fluctuations.
This is a rather peculiar theory because it presumes a positive correlation between the irrationality of the investor, and the amount of money he has to invest. How did the dumb become so rich? Not that it is impossible, just that you’d expect the correlation to go in the other direction (albeit smaller than one in absolute value).
The market is, of course, a stern disciplinarian that punishes such irrational plungers; fools and their money are soon parted. The Hunts, for example, were poster children for the old joke: “Want to make a small fortune trading commodities? Start with a large fortune.” Ditto Amaranth. Given the potential costs of position limits in constraining legitimate speculation, is it really necessary for the government to try to prevent what the market ruthlessly helps to deter?
It should also be noted that there is a tension between this justification for position limits and the anti-speculative wave justification. Contrarians by definition take views at odds with the conventional wisdom. If the conventional wisdom is irrational and not based on fundamentals–as the popular mania critique of speculation believes–we want contrarians to lean against the wind. Since they are by definition at odds with the views of most others, they will take positions that are disproportionately large on one side of the market. This is the theme of The Big Short, which describes how a handful of traders bet against the real estate bubble. If such bubbles exist, we want people to lean against it.
Position limits, and those based on concentration in particular, will constrain just that kind of contrarian trading. So if you worry about bubbles driven by mass speculation, (a) position limits will do little to constrain it, and (b) will make it harder for those not suffering from the popular delusion to trade in ways that limits the price distortions resulting from it. This is a perverse, unintended consequence of limits on large trader positions.
If big speculators are right on average, or right more than they are wrong, limits that constrain their trading will not reduce deviations between actual prices and what prices “should” be, and could in fact increase the frequency and magnitude of these disparities. So this justification for position limits depends crucially on a view of the correctness of big speculators. This is something everybody has a speculative opinion on, ironically–but little or no evidence to back it up.
Another justification for limiting the size of an individual trader is that if he is wrong, the failure that results when reality intrudes could have knock-on effects. The failure of the Hunts, for instance, did have some systemic consequences (as I wrote about some months ago).
But position limits are ill-adapted to addressing this kind of problem. The real problem is the failure of a really big, leveraged player. Some players are big, but not leveraged. If they lose, they have the wherewithal to make good their losses, limiting any systemic consequences resulting from knock-on failures. If prices move against a big ETF (like USO, for instance), it’s painful for the fund holders, but their positions are effectively fully collateralized, so their losses do not create counterparty default losses that could put others into distress. Position limits constrain leveraged and unleveraged players alike. (It would be interesting to understand whether the imposition of limits could actually affect the relative importance of leveraged players in the market. If so, limits could have unintended, adverse consequences.)
Relatedly, with respect to this problem, the metric of big is not big relative to the size of the specific market in which the speculator participates. It is big relative to the broader financial system.
Consider a couple of polar cases. For a small market, let’s say Minneapolis wheat, someone could accumulate a position that is huge relative to open interest, without accumulating position that is large relative to the financial system in any way. During the peak of the price spikes in 2008, the maximum value of the open interest in July Minneapolis wheat was about $650 million. In the scheme of things, this is chicken feed, even if one person had held the entire position, let alone 10 or 12.5 percent.
Now consider a big market, like crude oil. At the peak in July, 2008, the value of NYMEX open interest was on the order of $200 billion–nearly 3 orders of magnitude bigger than for MW. Somebody with 12.5 percent of this market (the upper bound under previous CFTC proposals) would have had a notional position value of about $25 billion–that’s serious money, which, if the position holder is leveraged, could lead to some knock-on problems if it went south. (Of course, $25 billion is an upper bound on the loss a long could suffer. For both longs and shorts, the uncollateralized credit exposure is far smaller. Perhaps sufficiently small that even such a large position is not systemically risky even if held by a leveraged player.)
Note that under the new proposals, where OTC markets are included, the disparity would become even more extreme. The OTC positions tied to CL are large, absolutely and relative to the NYMEX open interest, as compared to OTC positions tied to MW absolutely and relative to MGE open interest. (How much larger, nobody really knows at present. That’s one reason why some commissioners are reluctant to act until they know more about the size of the OTC market.)
So under this theory of the systemic risk associated from the failure of a big speculator, you might conclude that a limit would be sensible for crude but not for Minneapolis wheat. But the commodity position limits will be imposed on markets of all sizes.
In this regard, it is interesting to note that press reports state that the limits the CFTC is considering would hit far more agricultural traders than energy and metals traders. Given that the latter markets are a lot bigger than the former, and hence would pose a far greater systemic risk, this means that the proposed limits are very poorly adapted to addressing any putative systemic risk posed by “large” speculators.
One last thing. In October, I had a discussion (disclosed on the CFTC website) with the CFTC position limits team. I basically argued that in designing such limits, it was imperative that the limits be designed based on a diagnosis of the problem to be solved. This led to a discussion of potential problems associated with large positions. A lot of the scenarios that were discussed revolved around how traders with big positions might exploit microstructural frictions to make money.
Yeah, maybe. But I am skeptical that somebody is going to accumulate a huge position, with all of the capital required and risk entailed, to play liquidity games for ticks. Moreover, many of the big players that some critics of speculation are obsessed with, e.g., the “massive passives” (ETFs and index funds) that keep Bart Chilton up nights, most certainly do not play these games. They pretty much trade in a mechanical fashion, rolling on a schedule, and investing and liquidating based on customer money inflows and outflows. So again, there seems to be a huge disconnect between the paranoia about massive passives, and the types of behavior that some at the CFTC look to prevent using position limits.
There’s a common pattern to all this. Come up with a scenario in which speculation distorts prices. Then evaluate how position limits of the type currently under consideration would operate under this scenario. And then conclude that these position limits would either not address the purported problem, or could actually be counterproductive and make the purported problem worse.
Given that speculative limits also interfere with the legitimate uses of derivatives markets, the very real possibility that they have virtually no offsetting benefit makes this a very bad policy choice.
Supporters of a specific position limit proposal should be required to provide a reasonable justification for that specific proposal. This justification should first spell out in detail exactly a scenario or scenarios under which speculation distorts prices. It should then show how–exactly–their proposed limit will reduce the likelihood of these scenarios.
I’ve gone through several alternatives above, and argued that there is a complete mismatch between the diagnosis and the prescription; position limits appear to me to be the financial equivalent of leeches or faith healing. But maybe my imagination is unduly limited. Maybe there are other stories/scenarios that I haven’t considered. I’d be glad to entertain them–but I would also have to be convinced how a specific position limit proposal would improve market performance under these scenarios.
Of course, I also believe that those advancing these scenarios should provide rigorous theoretical justifications and empirical evidence to demonstrate that their scenarios are real possibilities. But I’m pretty confident that supporters of limits will be unable to meet even the weaker burden of showing how position limits cure what they identify as ailing the market, no matter how implausible, not to say crack-brained, that diagnosis is. And until someone proves otherwise, position limits will remain a solution in search of a problem.
* My daughter Renee gave a paper on the political economy of regulation of speculation, focusing on position limits, at the Instituto Bruno Leoni Mises Seminar 2010 in Sestre Levanti, Italy last October. Her paper compared the drive for position limits in the 1930s to that in the late-2000s. The audio of her presentation is available online. Her paper has been accepted for publication in the Journal of Economic Affairs.