High Frequency Trading and Market Efficiency
“Overnight” Change, 40 Years in the Making
Technological change can appear to the general public as an overnight revolution but in actuality stem from years of intense, behind-the-scenes development. For instance, widespread awareness and adoption of the Internet began in the 1990s, though its roots lay in the early 1960s and the creation of ARPANET, which itself built on years of earlier progress. In similar retrospect, calls for electronic exchanges can be traced back at least to the early 1970s: “[A] stock exchange can be embodied in a network of computers, and the costs of trading can be sharply reduced, without introducing any additional instability in stock prices, and without being unfair either to small investors or large investors” (Black, 1971).
Financial theory has developed simultaneously with changes in markets. Theoretical work on nonsynchronous trading (e.g., Lo & MacKinlay, 1990; Miller, Muthuswamy, & Whaley, 1994; Scholes & Williams, 1977) serves as the intellectual foundation for HFT. More work is needed on formal theories, since traditional ideas such as Modern Portfolio Theory (Markowitz, 1952) are horizon specific. On the practical side, program trading was another step in the transition to computer-based trading, along with the first ECNs in the late 1990s and later the decimalization of prices on the major exchanges. Hardware improvements, algorithmic breakthroughs, and creative competition all factored into the recent boom in HFT. Viewed in the proper context, HFT is an extension of decades of theoretical advances and market evolution.
Effects on Markets
The term efficiency carries a significant semantic load in finance research, but its primary meaning here is the swift incorporation of information into prices. In this sense, there is strong evidence that HFT improves efficiency. Statistically, prices exhibit less mean reversion (Castura, et al., 2010) and more closely resemble a random walk (O’Hara & Ye, 2011), both implying greater market efficiency. Parallels exist in other disciplines: voice recognition, robotics, data storage, drug design, and GPS all have become incredibly more efficient at informational tasks, thanks to development in computing power. HFT is the financial analogue of these advances.
Another metric of a market’s efficiency is the cost of transactions. The inverse relationship between trading frequency and price spreads is well established (beginning with Demsetz, 1968), in part because the higher liquidity lowers the inventory risk for a market maker. Empirically, this appears to be true, given that HFT has to a great extent supplanted the NYSE specialist. Quoted spreads have also narrowed in recent years. However, there is an important caveat that effective spreads have apparently widened in some cases, which may be due to high frequency traders managing to front-run orders through interpositioning algorithms (Hendershott, Jones, & Menkveld, 2011). Similarly, some long-term institutional investors have complained that HFT has increased transaction costs (Arnuk & Saluzzi, 2008, 2009).
Proponents of HFT tout its improvements in liquidity, and there is unequivocal evidence that trading volume has grown tremendously in recent years. However, volume remains only one aspect of the multi-faceted construct of liquidity — which is also a function of market breadth, depth, and resilience. Furthermore, there is concern that HFT faces no affirmative obligation to trade and could withdraw liquidity in turbulent markets. HFT obviously improves some measures of liquidity in some periods, the ongoing question (still in the relatively early stages of understanding) is the overall impact on market liquidity.
Similarly, the effect of HFT on market volatility remains controversial. Zhang (2010) argues that HFT increases volatility, while other studies suggest that it may dampen volatility (e.g., Chaboud, Chiquoine, Hjalmarsson, & Vega, 2009; Hendershott & Riordan, 2011). The diversity of HFT strategies suggests that some algorithms might dampen while others exacerbate volatility. This is an important area for further research.
Order flow can be affected dramatically by algorithmic strategizing in computer systems. Predictably, the price competition rules of Reg NMS have allowed a proliferation of trading venues and competition, so that as much as 30% of all equity volume now takes place away from the traditional exchange (O’Hara & Ye, 2011). The question for the field of finance is whether market fragmentation due to HFT damages markets. In their analysis of trade reporting facility data, O’Hara and Ye (2011) conclude that fragmentation reduces transaction costs, increases execution speed, and generally improves market efficiency. The informational linkages through monitoring and arbitrage appear to be sufficient to overcome any feared problems due to the fragmentation of order flow. Computer algorithms are capable of monitoring, searching, and implementing trades across a range of venues that far exceeds the capability of a human trader.
Thus, further development in computer science can only improve algorithms, power, speed, and market efficiency. HFT algorithms attempt to rebundle and route orders for optimal execution, which is an issue akin to the NP-complete problem of bin-packing, a cutting-edge question and momentous challenge for computer scientists. True market efficiency will thus become possible only if it can be demonstrated that P = NP (Maymin, 2011) so that such complex algorithms can be conducted in polynomial time. Financial theorists should monitor and participate in this vital question at the intersection of mathematics and computer science.
Regulatory Challenge and Public Perception
A major challenge of HFT is regulatory concern for the maintenance of fair and orderly markets. Order entry errors and other potential glitches in automatic execution raise fears of potential market problems such as the “Flash Crash” of May 6, 2010. There are suspicions of an unfair advantage for some market players, such as flashing orders, a practice that allows some participants to trade on order book information before it is publicly disseminated. Other questionable activities include spoofing and co-locating, representing methods that traders can utilize to game the system or gain an unfair advantage. The trading that arises from these behaviors has earned the moniker toxic order flow. The goal of responsible regulation must be to minimize the use of manipulative games and unfair informational advantages, without squelching the benefits of HFT. But regulating and monitoring HFT will be difficult in light of the sheer volume and pace of trading as well as the proprietary nature of trading algorithms.
As a final consideration, the financial community must take into account the public perception of HFT. It is human nature to fear that which one does not understand. Just as areas labeled terra incognita on Medieval maps were often illustrated with fantastical beasts, HFT lays beyond the ken of the average investor and is thus sometimes imagined to be confusing and dangerous, even a rigged game. This perception might lead investors to withhold or withdraw funds, thereby damaging markets. An important parallel exists in theories of insider trading, which have demonstrated both that legalized insider trading can aid informational efficiency (Leland, 1992) and that lack of public confidence can lead to market breakdown (Bhattacharya & Spiegel, 1991).
Public perception should be of vital concern to financial academics and practitioners because a populist outcry can lead government and regulators to “do something” rather than take thoughtful and reasoned measures. Furthermore, education will be necessary, though likely not sufficient, to maintain fair and orderly markets. Thorough research, appropriate regulation, and clear communication can simultaneously calm fears and improve markets.
This is a time of great change in global financial markets. Markets are simultaneously more fragmented and closely tied together, as high-frequency algorithmic trading allows competition for order flow but also creates more linkages through global arbitrage. These changes are irreversible. Indeed, they represent an ongoing evolutionary process of market development. Properly handled, HFT should improve market efficiency and provide a range of other benefits for market participants by encouraging immense creativity and technology adoption, attenuating risk exposure (thanks to short holding periods), and providing significant profits. Computers are tools that simply enact perfectly logical principles, according to their programming. The future will show how market participants are able to use such technology to improve market functioning and efficiency.
1) Arnuk, S., & Saluzzi, J. (2008). Toxic equity trading order flow on Wall Street: The real force behind the explosion in volume and volatility. White Paper, Themis Trading, LLC.
2) Arnuk, S., & Saluzzi, J. (2009). Why institutional investors should be concerned about high frequency traders. White Paper, Themis Trading, LLC.
3) Bhattacharya, U., & Spiegel, M. (1991). Insiders, outsiders, and market breakdowns. Review of Financial Studies, 4, 255 – 282.
Black, F. (1971). Toward a fully automated exchange. Financial Analysts Journal, 27, 29 – 35 & 86 – 87.
4) Castura, J., Litzenberger, R., Gorelick, R., & Dwivedi, Y. (2010). Market efficiency and microstructure evolution in U.S. equity markets: A high-frequency perspective. Working paper (RGM Advisors).
5) Chaboud, A., Chiquoine, B., Hjalmarsson, E., & Vega, C. (2009). Rise of the machines: Algorithmic trading in the foreign exchange market. Technical report, FRB International Finance Discussion Paper No. 980.
6) Demsetz, H. (1968). The cost of transacting. Quarterly Journal of Economics, 82, 33 – 53.
Hendershott, T., Jones, C., & Menkveld, A. (2011). Does algorithmic trading improve liquidity? Journal of Finance, 66, 1 – 33.
7)Hendershott, T., & Riordan, R. (2011). Algorithmic trading and information. Working paper.
8) Leland, H. (1992). Insider trading: Should it be prohibited? Journal of Political Economy, 100, 859 – 887.
9) Lo, A., & MacKinlay, A. (1990). An econometric analysis of nonsynchronous trading. Journal of Econometrics, 45, 181 – 211.
Markowitz, H. (1952). Portfolio selection. Journal of Finance, 7, 77 – 91.
10) Maymin, P. (2011). Markets are efficient if and only if P = NP. Algorithmic Finance, 1, 1 – 11.
11) Miller, M., Muthuswamy, J., & Whaley, R. (1994). Mean reversion of Standard & Poor’s 500 Index basis changes: Arbitrage-induced or statistical behavior. Journal of Finance, 49, 479 – 513.
12) O’Hara, M., & Ye, M. (2011). Is market fragmentation harming market quality? Journal of Financial Economics, 100, 459 – 474.
13) Schapiro, M. (2010). Testimony concerning the severe market disruption on May 6, 2010. Congressional testimony.
14) Scholes, M., & Williams, J. (1977). Estimating betas from nonsynchronous data. Journal of Financial Economics, 5, 309 – 327.