Center For The Study of Financial Regulation

Winter 2011 - Issue NO.4

HIGH-FREQUENCY TRADING

by Jonathan Brogaard, completing his Ph.D. and J.D. at Northwestern University.

Page 1

"One of the most significant market structure developments in recent years is high-frequency trading."1 -- a concept release by the Securities and Exchange Commission, January 13, 2010.

"[An] area that warrants close review is the regulatory scheme that applies to the most active and sophisticated participants in today's market structure - high-frequency trading firms."2 ~Mary Schapiro, September 7, 2010.

Speed always has been valuable in financial markets. Until recently, humans were directly responsible for the trading activities on stock markets and thus, trading decisions were bound by human processing times. Now that has changed. A large portion of market activity comes from algorithmic trading, of which the fastest subset is referred to as high-frequency trading. Those engaged in this type of trading are keen on being quick: The speed of receiving, analyzing, and reacting to information is no longer bound by human processing times, but by the laws of physics.

I have had the opportunity to study the activities of high-frequency trading firms, as well as their impact on market quality, by using a proprietary dataset from NASDAQ. The work on this type of trading is only just beginning, but the results so far show how prominent high-frequency trading has become in U.S. equity markets and how it is contributing to market quality.

Some shocking statistics have been gleaned from the data. High-frequency traders are prominent market participants, being involved in 68.5 percent of dollar volume traded. In addition, they frequently offer the inside quotes: 65 percent of the day for the overall sample, and 84 percent for large stocks. The detectable, industrywide trading strategy engaged by high-frequency traders is one based on past prices and order imbalances: After prices have risen and more shares were bought than sold, high-frequency traders are more likely to sell, and the reverse is true for their buying behavior. I estimate that in U.S. equity markets, they earn gross trading profits of approximately $2.8 billion annually.

High-frequency traders have a variety of distinguishing features. These include: (1) utilizing computing systems that quickly analyze data and subsequently generate, route, and execute orders; (2) using co-located servers, direct or sponsored access providers to interact with exchanges, and low latency, direct data feed services; (3) having extremely short holding periods; (4) submitting many orders that are cancelled shortly thereafter; (5) closing the trading day near a market neutral position; (6) frequently switching between being long and short in an asset; and (7) tightly controlling inventory.3

Back to Top

Page 2

Partly due to these common characteristics of high-frequency traders—combining high speed with sophisticated technology—new regulatory changes have been established that are meant to fortify equity markets from future rogue algorithmic trading programs, especially since the fallout of the May 6, 2010 "flash crash."4

First, circuit breakers have been placed on a number of individual stocks for a trial period.5 When a stock's price moves by 10 percent in a five-minute period, the circuit breaker will be triggered and trading of that stock will be halted on all U.S. equity markets for five minutes. Second, the Securities and Exchange Commission (SEC) now requires that registered market-makers post bid-and-offer quotes within 8 percent of the national best bid-and-offer on stocks subject to the circuit-breaker rules.6  

Third, the SEC banned "naked access" to the market, whereby a sponsored access provider allowed its clients' trades to flow through its exchange connections without being subject to pre-trade risk management by the sponsored access provider.7 These rules are meant to limit the damage that may occur as a result of rogue algorithmic trading programs. Fourth, the SEC clarified when an exchange is allowed to cancel an order that is considered erroneous.8 This fourth change directly addresses an issue that caused some high-frequency traders to withdraw from the market on May 6, 2010—a concern that trades might be cancelled after the fact.

Among other things, I examine two specific concerns about how high-frequency trading might destabilize markets, which are often cited as reasons for additional regulation. First, that while they may be liquidity providers during normal market conditions, in extreme events they abandon markets. Second, that they exacerbate price movements when they are present. I find evidence suggesting the opposite.

In my research, high-frequency traders decrease the liquidity they provide on the most volatile days to a degree, providing 10 percent less than normal, but when focusing on the most volatile periods within the day, high-frequency traders actually provide more liquidity than normal. Together, these two facts are consistent with high-frequency traders providing liquidity during large intraday price movements, but then more aggressively balancing their inventory (by using marketable orders instead of limit orders) during the rest of the day. Regarding the second concern, I find evidence that high-frequency trading may dampen intraday volatility.

Back to Top

Page 3

However, my findings suggest there is a potential cause for concern in the depth of liquidity provided by high-frequency traders. While they do provide a degree of depth, it is about one-fourth of that made available by other market participants.

My research begins to study potential concerns regarding high-frequency trading firms, but a host of additional questions remain unanswered: How are they interacting with dark pools? How do they use information across assets and across exchanges? What are the implications of the large number of cancelled orders they place? Why do most not register as market-makers? Before regulating away this new type of trader, as has been proposed, it is imperative to test whether they are improving or impairing market quality.9 It may be the case that additional regulation in specific areas should be put in place to improve market quality and stability, but the need for such regulation should first be substantiated in the data.

Reference

1. Securities and Exchange Commission. January 13, 2010. Concept Release on Equity Market Structure. Release No. 34-61358.

2. September 7, 2010. Speech to the Economic Club of New York by U.S. Securities and Exchange Commission Chairman Mary Schapiro.

3. Items (1) – (5) are discussed in the Securities and Exchange Commission January 13, 2010 Concept Release on Equity Market Structure.

4. Commodity Futures Trading Commission and SEC. September 30, 2010. Findings regarding the market events of May 6, 2010.

5. Financial Industry Regulatory Authority (FINRA) Rule 6121.

6. SEC Release No. 34-63255.

7. SEC Rule 15c3-5.

8. FINRA Rule 11892.

9. For instance, a Tobin tax has been proposed in Congress that would make high-frequency trading unprofitable.

Back to Top