On June 5, Mary Jo White, Chair of the U.S. Securities and Exchange Commission, sketched some proposed changes to regulate high-frequency trading (HFT). Her full speech is available from the SEC website. Some analysis can be read in the New York Times and Bloomberg News.
Synopsis of White’s comments
White surprised many observers by stating that investors are doing better in the algorithmic trading regime today than they did in the “old manual markets.” She noted that for institutional investors, the cost of executing a large order is roughly 10% lower than in 2006, and the spreads between bid and ask prices are “as narrow as they have ever been,” thus benefiting retail investors too. White concluded that “all of these market quality metrics show that the current market structure is not fundamentally broken, let alone rigged.”
While the trading volume of HFT is now very high, this is largely accounted for by frequency of trades and not leverage, since the HFT traders typically hold stock very briefly and make their money on transaction fees. White may be right that the marginally higher prices retail investors thus experience are more than balanced by greater liquidity and lower spreads.
White then mentioned some changes that are already in the works. First of all, the SEC has already proposed stricter requirements for the technology used by exchanges and trading venues. Final regulations are being prepared. It is also implementing rules to ensure the robustness and resilience of markets, including a 10-minute recovery standard.
She then addressed questions of HFT and “fairness”. She started by saying that “the SEC should not roll back the technology clock or prohibit algorithmic trading,” but that the SEC is assessing whether some aspects of computer trading are working against investors rather than for them.
One of the proposed rules is to clarify the status of unregistered active proprietary traders and subject them to rules similar to those required of dealers, and to require Financial Industry Regulatory Authority (FINRA) membership for dealers that trade exclusively in off-exchange venues. They will also require exchanges and FINRA to include a “time stamp” in the consolidated data feeds that indicates when a trading venue processed the display of an order or the execution of a trade.
Another potentially controversial proposal is to include “batch auctions” or other solutions that minimize the inherent advantage of HFT. Another is to include “affirmative or negative trading obligations” for HFT firms employing the most sophisticated trading tools.
In a larger sense, White is concerned about the fragmentation of U.S. equity markets. There are now 11 exchanges and more than 40 “alternative trading systems,” for instance. The interconnections between these venues raise the potential for a malfunction in one system to disrupt others. White is also concerned about the increasing percentage trades that are handled by “dark trading venues” — rising from 25% in 2009 to 35% today. Indeed, the dark trading venues raise serious concerns about transparency, since “investors know very little about many trading venues that handle their orders.”
One final concern is managing “broker conflicts,” namely potential conflicts of interests when fees and payments are not clearly passed through from brokers to customers.
Impact of “Flash Boys”
It is widely believed that these concerns have their roots, at least in part, with the recent publication of Michael Lewis’ book Flash Boys. Brad Katsuyama, the Canadian market trader who was featured in the book, was particularly concerned with sophisticated order types that gave other trading operations an advantage, especially when executed in “dark pool” venues. He thus argued that the market was “rigged” (which may be the source of White’s comment above). See our earlier Mathematical Investor blog for details on this and Scot Patterson’s informative earlier book Dark Pools.
Also, the SEC may still be weighing the impact of the May 6, 2010 flash crash, when the Dow Jones industrial plummeted more than 1000 points (9%) in just a few minutes, only to recover most of these losses a few minutes later. This and two or three other less severe incidents in the past few years have led some to worry that modern algorithmic trading has introduced a potentially serious instability into financial markets.
Questions
It is worth pointing out the care with which the SEC is approaching the issue. As noted above, White clearly rejects the claims of Katsuyama and others that the market is fundamentally and irretrievably “rigged” against ordinary traders.
But White also implicitly acknowledges that previous changes implemented by the SEC have resulted in their own set of problems. For instance, previous regulatory changes forcing the buyer to buy the cheapest offered price enabled high-frequency traders to get early warning from a small cheap purchase that a much larger purchase is coming, thus enabling them to stake a position profiting from the upcoming transaction.
White’s comment on “frequent batch auctions” is intriguing. While there are no details as yet, she may refer to a proposal requiring market venues to collect all bids in a certain time slice (say a few milliseconds), then execute them in random order, thus negating some of the inherent advantage that high-speed traders currently enjoy.
Still, it is clear that the SEC has its homework to do: Questions that arise include:
- How can these regulations truly be crafted to ensure fairness?
- How, exactly, can one define and/or quantify “fairness”?
- How can the needs of individual investors be balanced with the needs of large institutional investors?
- How can all of this be enforced?
We await details!