Reflections on High Frequency Trading
Here are a few thoughts on high frequency trading [HFT] – not a thorough analysis of the problems and solutions, but rather a brief outline to encourage further discussion.
What is HFT?
The modern version of HFT, as described by Malkiel and Leavitt in an April 11 Wall Street Journal op-ed, “involves the placement of high-speed computers in close proximity to stock-market servers to give some participants the ability to buy and sell stocks faster than the blink of an eye.” The favored participant purchases early access to information from either a public exchange or a so-called dark pool, which is essentially a private exchange established by investment banks. With privileged access, the HF trader can learn of a pending order before it is executed, and then earn a small profit by buying or selling ahead of the order. For example, a sell order exists at $100.00; an HF trader learns of a pending buy order at $100.02, which allows the HF trader to earn $.02 by first buying at $100.00 then selling at $100.02.
What are the pluses and minuses of HFT?
Benefits of HFT can include market efficiency, increased liquidity, and lower transactions costs. To illustrate: Assume orders have been placed to sell 5,000 shares of XYZ at $15.02 and 5,000 shares at $14.94. Assume further, a pending buy order for a minimum of 10,000 shares at $15.00. The HF trader, acting as market-maker, might buy 10,000 shares from the two sellers at an average price of $14.98 and re-sell the shares to the buyer for $15.00. The result would be a narrowed bid-asked spread, reduced trading costs, increased volume, and enhanced liquidity.
On the other hand, as Malkiel and Leavitt point out, HFT is a form of insider trading known as front-running whereby “optimally positioned traders can see trade orders from other investors before they are executed. They can execute a purchase just ahead of those orders and run the price up just a bit, pocketing the difference.” In addition, HFT has been blamed for increased market volatility.
Has HFT caused increased volatility and “flash crashes”?
Early information is new information; and new information, by its very nature, can trigger market volatility. Every major announcement is potentially destabilizing, and the persons who first acquire the information have an advantage – usually to the detriment of retail investors. Yet government intervention to suppress new information is rarely beneficial. Increased volatility, taken alone, does not justify regulation. At a minimum, there should be a showing that HFT either generates large economic losses or involves fraud, breach of contract, or breach of fiduciary duty.
Because price often responds to increased volume, one way to dampen volatility would be to slow down trading. For instance, HF traders might be charged a transaction fee or required to hold a security for a specified period of time. Such regulations would also mitigate front-running, but they could diminish market efficiency and raise trading costs to the detriment of all investors. Drawing again from Malkiel and Leavitt: “In Europe, when trading taxes were implemented, trading volume and liquidity fell, and bid-asked spreads increased. Similarly, in Canada, when fees were increased on high-speed traders, spreads increased and liquidity decreased.”
With respect to flash crashes, an SEC-CFTC joint report determined that HFT actually improves conditions by absorbing some of the sell pressure. Moreover, the Dow’s 600-point decline in a few minutes during May 2010 – which was reversed within a half-hour – was attributed to a single $4+ billion futures sale by a large money manager. In fact, flash crashes were common long before HFT. The exchanges have already implemented “circuit breakers” to minimize their impact.
Does HFT entail front-running?
Ordinarily, front-running allegations are based on breach of fiduciary duty. The front-runner is under a legal obligation not to profit from inside knowledge of his client’s pending order. By contrast, the HF trader is transacting for his own account; he has no client, and so could not have breached a fiduciary duty. His trades are based on information that he has purchased from an exchange.
If, however, the exchange did not have the right to sell the information, then the HF trader’s early access might have been acquired illegally. Typically, investors have written agreements with their brokers, but not with exchanges; so the sale of access by the exchange is not a direct contractual breach. Nonetheless, that sale might still be impermissible if contracts are deemed to contain an implicit term that provides for ownership of the information by the investor who initially placed the order. In that case, it’s the exchange, not the HF trader, that has acted unlawfully by selling information it doesn’t own.
What framework should control the regulatory environment?
Accordingly, ownership of the underlying property rights is one criterion for government regulation of HFT. If the investor owns the information on his pending orders, his rights have been violated by the exchange; and any sale of the information must be subject to his consent. Effectively, that might shut down HFT because the investor is the party whose trades are being front-run. But if the exchange owns the information, then its sale does not violate anyone’s rights and traders should be allowed to compete for faster access.
Under those circumstances, the only other basis for government regulation would be significant welfare losses – e.g., if the costs of front-running and higher volatility outweigh possible improvements in market liquidity, pricing efficiency, and trading costs – in which case investors unable or unwilling to compete against HF traders might have to be protected.
Who has property rights to information on pending orders?
The Coasean answer to the property rights question is that the initial assignment of ownership doesn’t matter: If transactions costs are low, bargaining by the various parties will direct resources toward their highest valued use. But here, transactions costs among affected investors, brokers, exchanges, and HF traders are likely to be prohibitive. Moreover, libertarians are concerned with distributive shares – i.e., who benefits and who bears the costs – not just aggregate resource allocation.
A Lockean rule would assign ownership to the originator of the information. The economically efficient rule is that ownership should vest in the party least able to avoid harm if the right were to vest elsewhere. And the libertarian rule is that the victim of any harm has the right to claim compensation. Application of those rules is unclear in this instance, and perhaps even conflicting. When that happens – i.e., when rights theory doesn't provide clear guidance – a utilitarian analysis can inform the initial assignment of a property right. In other words, an economic cost-benefit assessment and a rights-based assessment will merge. (That also happens when we evaluate, say, speed limits or any other safety regulations.)
In brief: Preventing exchanges from selling access to early information is equivalent to establishing a property right in the originating investor. Perhaps that's desirable, but it requires careful scrutiny of costs and benefits that, so far, hasn’t been evident in the literature.
