On Oct. 21, the SEC voted unanimously to propose regulations that would increase the transparency of dark pools. The proposed rules have not been released yet by the SEC, but public announcements and congressional testimony by SEC staffers have provided a thumbnail sketch.
At the outset, and in the interest of transparency, I hereby disclose that I have no love for dark pools. Nonetheless, from what we can discern so far, the proposed regulation takes aim at the culprit and shoots an innocent bystander.
Dark pools are electronic communications networks–or ECNs–that enable participants to quote and trade equities without displaying their quotes to the public markets. The first dark pool that I can recall was the ill-fated Optimark system that emerged in the late-1990s. Optimark was well-financed and enthusiastically received. It also was a commercial failure.
All that changed with Regulation NMS, a rule change that was intended to foster efficient price discovery by increasing transparency in the listed markets. Once Reg NMS went into effect, dark pools emerged like bats swarming from their caves at twilight. And they were instantly profitable. There are now 40 or so dark pools, and in contrast to other parts of the trading industry in these tough times, they are generally doing well. On a related note, Reg NMS and dark pools also provided abundant profitable opportunities for high-frequency algorithmic trading, including flash orders, but that is a topic for another day.
Reg NMS made dark pools profitable because of its trade-through rule, which requires the execution of better-priced orders before inferior-priced orders can be executed. This part of the rule was intended to encourage institutions to submit large orders into the public markets by preventing other market participants from "trading around" them. Instead, it has had the opposite effect.
The rule raised the transaction costs of executing large orders in the public markets, because it requires smaller, better-priced orders to be executed before a large order at an inferior price can be executed as a block. However, the trade-through rule only applies to orders that are part of the public quote stream. Dark pools could execute transactions among their participants without interacting with the public markets. Reg NMS therefore transformed an unsuccessful business model into an instant success.
The "astonishing" success of dark pools is explained by a very simple principle: Institutional investors do not like to show their trading interest to the public. They dislike revealing their trading interest so much that they are willing to pay extra money to avoid the public markets. This may seem odd, since institutions need to get their trades done, just like all other market participants.
The best explanation for the antipathy of institutional investors to the public markets was expressed by Bernie Madoff, back in the day when he was known for founding and managing a successful and legitimate trading operation, rather than for his criminal expertise running Ponzi schemes. Bernie used to say that he made a career from handling small orders of less than 5,000 shares. Small orders, according to Bernie, contained no information. Large orders, on the other hand, were not simple trades. Instead, they conveyed information, and Bernie’s business originally was not set up to deal with that information. Later, he relied on this concept to market something called "order slicing," which was intended to shred larger orders into small, anonymous pieces.
Large orders contain two valuable pieces of information. First, large orders may cause a temporary liquidity shortage, which will move the market. So, if other traders know a large order is coming, they will purchase or sell in advance–a manipulative practice known as "front-running"–to take advantage of this shortage. Second, large orders are generated by professional investors, whose actions are based on more knowledge than your average "retail" participant. Other investors can exploit this knowledge by making similar trades at the same time. This type of "free ride" on the knowledge of others, without doing any of the work necessary to obtain this knowledge, is called "rent extraction" by economists. Institutional investors attempt to disguise their trading interest to avoid manipulative front-running and to prevent others from exploiting their research and thereby diminishing its value.
Participation in the public markets, therefore, imposes a cost on institutional investors. Simple economics tells us that they will be willing to pay an amount up to the value of that cost if the same trades can be accomplished away from the public markets.
The SEC is aware of this behavioral calculus. Its new proposal, therefore, purportedly will allow quotations for values greater than $200,000 to remain in the dark.
I don’t think that exception will work, because institutional investors are not willing to place quotes of that size into any electronic trading system. Instead, they use electronic trading systems to accomplish something similar to "order slicing"–they shred large orders into smaller pieces. The problem they will have is that in a fully disclosed marketplace, other market participants can quickly discern that someone is placing a large number of similar orders into an ECN and detect the presence of the large, shredded order. And then the trade-through rule will continue to raise the costs imposed on institutional investors from interacting with the public markets.
In the absence of dark pools, institutions are more likely to return to the embrace of the large investment banks, which already execute most large institutional orders through over-the-counter transactions with other big investment banks–the "upstairs" markets, as they are fondly known.
The fact is that the large investment banks already offer lots of services to institutional investors that are not available to the general run of retail investors. First, their corporate finance operations offer institutional investors access to public offerings in debt and equity securities that retail investors rarely get to sniff. Second, the big investment banks make their research analysts available to large institutions, and not just by providing the sort of "street research" tossed out as crumbs to the herd. Instead, the banks’ analysts prepare and share spreadsheets that contain elaborate predictions of the effects of various factors on an issuer’s earnings. This enables an institution’s analysts to obtain valuable confirmation of their own analyses and additional information that would otherwise be difficult to obtain. Investment banks also invite institutional investors to attend industry conferences where officers of public companies explain their businesses in detail that is difficult to glean from cold public disclosure.
Of course, all of this comes at a cost. Investment banks pay for all of this with commissions received from institutional investors on upstairs trades placed with the banks’ trading desks. And institutions will continue to send most of their trades through the trading desks of large investment banks, despite the additional commissions costs this practice entails. Dark pools have provided an alternative execution venue for institutional investors who believe they have paid enough commissions to receive the "good stuff" from the investment banks, and would like to pay less to execute other trades as a way to moderate their trading costs.
Naturally enough, the big investment banks would prefer that their customers did not have such a lower-cost alternative and are eager to see this source of competition limited in any way possible. The proposed rule plays right into their hands. Having nowhere else to go to trade privately, institutions will tend to send more of their order flow to the large investment banks. That result can be expected to further improve the profitability of the banks’ trading desks (supporting ever-more-breathtaking bonus compensation).
I doubt that this is what the SEC intends. It seems likely they are hoping that institutions will continue to shred their orders and send orders up to $200,000 to ECNs, so that the investing public will be able to interact with these orders, allowing the markets to achieve more efficient price discovery.
We have witnessed a decade or so of market structure reforms, most of which were intended to cause institutions to mingle their orders with the public markets. All of them have failed because institutional order flow comes laden with valuable information. Until a rule change comes along that confronts this reality, regulations intended to eliminate the upstairs market will be exercises in regulatory vanity.
Stephen J. Nelson is a principal of The Nelson Law Firm in White Plains, N.Y. Nelson is a weekly contributor and columnist to Traders Magazine’s online edition. He can be reached at sjnelson@nelsonlf.com
