A recent article by Scott Patterson on WSJ.com highlighted growing concern within the financial regulatory community about so-called “Dark Pool” exchanges. Dark pools are basically markets that match buyers and sellers of equities (typically though not necessarily in large blocks) anonymously and publish the trade results only after the fact. This WSJ Live video does a nice job of illustrating how they work:
Dark pools were invented to solve a real problem, which is, as the video notes, the need for buyers and sellers to trade large blocks of stock without moving the market during the course of the trade. In other words, if I want to sell 1 million shares of Apple, and start to sell them in 10,000 share increments, today’s High Frequency Trading (HFT) algorithms will identify my strategy and act accordingly, which may not be want I want to happen. Institutional traders wanting to avoid this market signaling naturally turn to Dark Pools to avoid alerting the markets of their moves during strategy execution.
As the use of Dark Pools continues to increases – they now account for 10% of US trade volume, as the graphic below illustrates:
This tend raises many questions, both from economic and regulatory perspectives. WSJ quotes John Malitzis, executive vice president of market regulation at Finra (Financial Industries Regulatory Authority):
“We want to understand whether [dark pools] are disclosing to their customers how their orders work [and] whether customers are informed who their orders will interact with,” Mr. Malitzis said in an interview. “A big part of this is to get an understanding of practices that may or may not be problematic.”
“Problematic” in this case may be regulator-speak for possibly illegal, which is a natural concern for entities like FINRA. But Dark Pools also raise questions of economics that are interesting to consider.
The first concern that comes to mind is what effect demand-supply signals that are “invisible” until post-execution will have on financial markets? It’s natural that the HFT crowd will build new algorithms that will try to reverse-engineer what may or may not be happening in Dark Pools from ex-post facto data, but that represents a real danger to transparent market functioning. For example, let’s say one of these HFT “guessing” algorithms thought that Dark Pool “A” was helping to dump Stock “X” and so triggered a sell-off in transparent markets. That sell action triggers a sets off a route of stock X, but later it turns out that the HFT algorithm had mis-guessed. In this case, we have algorithms not just trying to outsmart a visible trading pattern — hard enough — but one that is “dark” and therefore even riskier than the typical trading scenario.
Yet another interesting issue is pricing in these Dark Pools and the relationship between Dark Pool bid-ask prices/spreads and their equivalent in normal markets. In other words, when I go to trade my 1 million Apple shares, I am not basing my offer price not on real-time Dark Pool demand signals (not possible, since they are not visible to me) but on transparent market prices, which may or may not be relevant in Dark Pool trading. The article does not discuss cross-marlet relational pricing dynamics, but I wonder if how they correlate? Are they generally identical? Do I get a better price in the DP? Worse?
Of course, in the end, this latest twist to the HFT story only adds to the conclusion that Wall Street is no place for amateur investors. If insider trading, HFT and all the other games were not enough to worry people trading at home, then the existence of black box buy-sell mechanisms, with zero relation to open markets, should be another warning that anyone who enters this world should remember the old poker saying that “if you can’t spot the sucker at the table, then it’s probably you.”
WSJ subscriber can read more here:
See also earlier post: