F Street Capital Management LLC. Photographer: Joshua Roberts/Bloomberg
F Street Capital Management LLC. Photographer: Joshua Roberts/Bloomberg

It's hard to pick stocks that will go up, so most people can't do it consistently. It's about equally hard -- probably a bit harder -- to pick stocks that will go down, so most people can't do that consistently either. But if somehow you knew in advance what banks were going to be investigated for bank-y malfeasance, or which biotech companies were cooking their books, then you could sell their stocks and avoid losses, with consistent repeatable outperformance. Who might know that sort of thing in advance?

The decomposition of returns earned by SEC employees suggests that the abnormal returns are earned in the sell portfolio. In particular, the 12 month ahead (252 trading days) abnormal returns, using the four factor Fama-French model as the model of expected returns, of U.S. common stocks that SEC employees buy (sell) is 0.56% (-7.97%). Hence, SEC employees’ stock purchases look no different from those of uninformed individual investors ... but their sales appear to systematically dodge the revelation of bad news in the future. This fact pattern is consistent with the greater informational advantage related to potential enforcement activities that employees of a regulator are likely to enjoy over other market participants.

Oh!1 That's from this recent paper by Shivaram Rajgopal and Roger M. White, with the coolly sarcastic title "Stock Picking Skills of SEC Employees." The paper, and the related Bloomberg News article, are entertaining stuff. Statistical anomalies are, perhaps, the lowest form of evidence, but they are evidence, and this is some evidence that the Securities and Exchange Commission's employees are making use of their inside information to trade stocks.2

I guess a bigger question, though, is why are so many SEC employees trading stocks in the first place? The authors' sample, which runs from August 2009 through December 2011, consists of 29,081 transactions, of which 4,806 are trades in individual stocks (with a dollar value of $41 million) and 1,221 are in exchange traded funds ($15.7 million). Most of the rest of the trades appear to be in mutual funds, though the dollar amounts there are not clear. That's for a bit under 4,000 employees, so while it's not exactly frenzied day-trading, it's a meaningful amount of single-stock trading.

Individual stock ownership obviously creates the possibility of corruption, or the appearance of corruption, at least one of which seems to be represented by this paper. The SEC has policies requiring trade pre-approval and barring trades in stocks under investigation, but those policies are sometimes evaded, and presumably the SEC gets usable information even on companies that don't quite rise to the level of a formal investigation.

But it's also just weird that so many SEC employees fancy themselves stock-pickers. I myself don't own or trade single-name stocks,3 because I am a believer in efficient markets, and I have absolutely no reason to think that I know better than the market what stocks will go up or down.4

Obviously not everyone agrees with me about that, but there are interesting hints that the SEC does. I've written before about what I've called the SEC's increasing efficient-markets-hypothesis fundamentalism, in which, driven by an academic finance orientation that is suspicious of market "anomalies," the agency has started to look at those anomalies as evidence of malfeasance: If your fund outperforms the market, the SEC will take that as evidence that you're up to no good. What would the efficient markets fundamentalists at the SEC say about their employees' outperformance? Really, what would they say about their employees' owning individual stocks in the first place?

Most broadly of all, though, there is a real tension in the SEC's mission between protecting individual shareholders and promoting efficient and socially beneficial markets. If you think that it is important for individuals to buy and sell stocks in individual companies, confident in the knowledge that they have the same rights and protections as big institutional investors, then that will influence your views in one way. If you think it's important for capital markets to allocate capital efficiently and work without big systemic risks, that will influence your views in another way.

Those ways are, I think, different and often opposite. Individual people who own and actively trade individual stocks tend to be rich hobbyists who want to be flattered in their belief that they can compete with big institutional money managers, but who also like to think of themselves as the underdog.5 Big institutional investors, meanwhile, actually do a lot of the investing for the non-wealthy -- in mutual funds and pensions and annuities and so forth -- but they don't have the obvious underdog charm of retail investors. Making the markets work better for institutions, at the expense of retail investors, might be both more efficient and more protective of genuinely small investors. But if you view your job as protecting single-stock investors, you might resist that.

So, for instance: If you want individual investors to compete on a level playing field with the big boys, you might be suspicious of claims that activist hedge funds should be able to accumulate big stock positions in secret. "Disclose that stuff so everyone can know what you're up to," you might say. If on the other hand you care about efficient markets, you might want those activists to be able to accumulate big stakes efficiently, to provide incentives for them to work to improve companies. Big institutional investors -- not just activists, but big passive investors whose portfolios tend to benefit from activism -- seem to support this. The SEC is more suspicious of those secret accumulations.

Or: If you view your mission as protecting individual shareholders in individual companies, you will be very concerned about insider trading, and prioritize it over other sorts of malfeasance, because insider trading "steals" from unsophisticated investors who trade the other way.6 And not just real insider trading, either: You would question any sort of informational advantage that big investors get. Favored access to research analysts for big clients, for instance, would trouble you: Why should big institutional clients get any better information than retail stockholders? Seems so unfair. If you viewed your mission as promoting capital markets efficiency, you would be blithely untroubled to hear that professional investors purchase valuable services as part of their business of professionally investing.

Or, in market design, you might think it really important that every investor who wants to buy stock have access to the best price available in any market at the time she places her order. So you might implement something like Regulation NMS, which requires brokers to route orders to the exchange showing the best price. This, it turns out, has led to a lot of creepy high-frequency trading dynamics, with some possibly horrific consequences for market stability. But it feels so good to protect the little guy!

The SEC's mission really is to protect investors, so you can understand why they might have an -- occasionally counterproductive -- bias towards protecting individual investors. But the SEC staff's active trading of individual stocks suggests another possible source of that bias: The SEC is full of individual hobbyist traders itself. Of course they want the markets to work well for people like them.

1 Later:

SEC employees differ from corporate insiders in the pattern of their trading returns, however, and appear unable to capture gains in their buy portfolios. Rather, buy portfolio returns are statistically indistinguishable from zero in all versions of abnormal returns (the t-statistics for buy side abnormal returns are 0.29, 0.47 and -1.1). However, sell portfolios of SEC employees earn strong negative abnormal returns, ranging from -4.14% when the Fama-French three factor model is considered (t-statistic = -3.31) to -9.7% (t-statistic = -4.8) when the CAPM is the model used to compute “normal” returns. If SEC employees are trading on privileged information, it appears to offer insight on downside risk rather than upside potential, which would be expected of an agency tasked with investigating potential malfeasance in corporate governance and financial reporting. In that sense, the SEC employees seem no different from naïve individual investors in terms of the securities they pick to buy.

Also fascinating is the sector breakdown:

Buy and sell transactions are not equally distributed in these sectors. Panel C shows that sells heavily outweigh buys in (i) banking ($0.3 million of buys v/s $2.4 million of sales); (ii) financial services ($0.2 million of buys v/s $1.7 million of sales); (iii) insurance ($0.2 million of buys v/s $0.73 million of sales); (iv) pharmaceuticals ($1.7 million of buys v/s $3.1 million of sales); and (v) machinery ($0.8 million of buys v/s 1.3 million of sales). In contrast buys dominate sells in the following sectors: (i) computers ($3.8 million of buys v/s $2.1 million of sales); (ii) chemicals ($0.51 million of buys v/s $0.29 million of sales); and (iii) paper ($0.27 million of buys v/s $0.06 million of sales).

The data is from August 2009 through December 2011.

2 In particular, the authors look at the 56 SEC enforcement actions during their time period, and finds that SEC employees traded in advance of six of them -- with significantly more sells than buys.

3 Except that I own some Goldman stock, which is not my fault, and I sell it whenever I can.

4 "If I knew what the markets were going to do, I wouldn't be working here!" is something that you'll hear a lot from investment bankers, accountants, corporate lawyers, certainly bloggers, really everyone in financial markets outside of professional investors. But maybe not the SEC? They know what markets are going to do, apparently.

5 As I've written before, about big institutional investors getting advantages over retail investors:

“Joe Sixpack” is a terrible name for a market-timing retail investor. The model here is not [the] median U.S. household – who has zero dollars of equities! – getting screwed by Steve Cohen investing in his personal account or whatever. The model here is big professional investors who run your pension and 401(k) ... having an advantage over rich hobbyists who think they can time the market. Joe CaseOfLafite, really.

6 Similarly, you'd spend a lot of time prosecuting penny stock scams, which hurt rich boobs but don't really have many systemic consequences. I can't really criticize this, though, since penny stock scams do tend to be funny.

To contact the writer of this article: Matt Levine at mlevine51@bloomberg.net.

To contact the editor responsible for this article: Tobin Harshaw at tharshaw@bloomberg.net.