Center For The Study of Financial Regulation

SPRING 2012 - Issue NO.8

High-Frequency Trading and Price Discovery

by Terrance Hendershott, Associate Professor, University of California, Berkeley.


Adobe Acrobat (SM)  Download the article

 

Page 1

Historically, financial markets require intermediaries to provide immediacy to outside investors. These intermediaries often were given special status and located on the trading floor of exchanges. The automation of stock exchanges increased markets’ trading capacity and enabled intermediaries to expand their use of technology. This reduced the role of traditional human market-makers and led to the rise of a new class of intermediary, typically referred to as the high frequency trader (HFT). Like traditional intermediaries, HFTs are central to the trading process, have short holding periods and trade frequently. Unlike traditional intermediaries, however, HFTs are not granted privileged access to the market that is not available to others.

The substantial, largely negative media coverage of HFTs and the so-called “flash crash” on May 6, 2010, raise significant interest and concerns about the role that HFTs play in the stability and price efficiency of financial markets. In a new paper, Ryan Riordan and I examine the role of HFTs in the price-discovery process using data from NASDAQ that identifies the participation of a large group of HFTs in each transaction.1

Overall, HFT profits from predicting future price changes over relative horizons of between 10-30 seconds. This is the standard positive role in price efficiency played by speculators. This is driven by the marketable (liquidity demanding) HFT orders. Passive HFT limit-orders negatively predict future prices changes. HFT marketable orders’ informational advantage is sufficient to overcome the bid-ask spread and trading fees to generate positive trading revenues.2 For non-marketable limit orders, the costs associated with negatively predicting future price changes are smaller than the bid-ask spread and liquidity rebates, so these HFT orders also generate positive trading revenues.

The role of traders in determining prices is typically discussed in terms of finding the efficient price and in causing prices to diverge from the efficient price. The Securities and Exchange Commission’s (SEC 2010) concept release on equity market structure primarily expresses concern regarding short-term volatility, particularly “excessive” short-term volatility. Excess volatility is best thought of as pricing errors that are defined as the difference between the efficient price and the observed price. Such pricing errors could result from long-term institutional investors’ need to adjust their portfolios, often referred to as price pressure. If HFT trades against these pricing errors, it can be viewed as supplying liquidity and reducing long-term investors trading costs. If HFT trades in the direction of pricing errors, it can be viewed as increasing the costs to those investors. HFT trading in the direction of pricing errors also could arise from attempts to manipulate prices.

We find that overall HFT trades in the opposite direction to pricing errors, as HFT buys when market prices are below the efficient price and sells when market prices are above the efficient price. HFT trades in the direction of changes in efficient prices by buying when the efficient price is rising and selling when the efficient price is falling. These imply that HFT is associated with improvements in price efficiency, both in terms of helping to get new information into prices and reducing pricing errors. Both of these relations are driven by HFT marketable orders.

 

Page 2

More informative stock prices can lead to better resource allocation in the economy. However, HFT’s information is short-lived at less than 30 seconds. If this information would become public without HFT, then the potential welfare gains may be small. However, the fact that HFT predicts price movements at such short horizons does not demonstrate that the information would inevitably become public. It also could be the case that HFTs compete intensely with each other to quickly get information that is not obviously public into prices. If HFTs were absent, it is unclear how such information would get into prices, unless some other market participant played a similar role. Joe Stiglitz refers to this in his Nobel Prize acceptance speech as, “if markets were fully informationally efficient—that is, if information disseminated instantaneously and perfectly throughout the economy—then no one would have any incentive to gather information, so long as there was any cost of doing so. That is why markets cannot be fully informationally efficient.”

Reducing pricing errors improves the efficiency of prices. Just as with the short-term nature of HFT’s informational advantage, it is unclear whether or not intraday reductions in pricing errors facilitate better financing decisions and resource allocations by firms and investors. One important positive role of smaller pricing errors is if these corresponded to lower implicit transaction costs by long-term investors. Examining non-public data from long-term investors’ trading intentions would help answer this.

One important concern about HFT is its role in market stability. Our results provide no evidence that HFT contributes to market instability in prices. To the contrary, HFT overall trades in the direction of reducing transitory pricing errors, both on average days and on the most volatile days during a period of relative market turbulence (2008-2009).

While our paper’s results illustrate interesting relations between HFT and stocks prices, care is needed in interpretation. Without HFT, other investors almost certainly would change their behavior. Would the resulting prices without HFT be less efficient? The results suggest so, but we cannot prove it as we do not observe the market without HFT. Algorithmic trading, which includes HFT, has been shown to improve market efficiency,3 but “natural” experiments to establish this for HFT have not yet been found.

HFTs are a type of intermediary. When thinking about the role HFT plays in markets, it is natural to try to compare the new market structure to the previous market structure. Some primary differences are that there is free entry into HFT, HFTs do not have a designated role with special privileges, and HFTs do not have special obligations. Without such privileges, there is no clear basis for imposing the traditional obligations of market makers on HFT. When considering the optimal industrial organization of the intermediation sector, HFT more resembles a highly competitive environment than traditional market structures. A central question is whether there were benefits from the old, more highly regulated intermediation sector that outweigh reduced innovation and higher entry costs typically associated with regulation.

When thinking about regulation of HFT, the answers to many important questions are not yet known. If HFT becomes competitive (zero profits), will HFT then resell their technology as brokers? Could this lead to efficiency without imposing costs on other investors? Do dark pools and batch auctions limit part of the costs of technology investment for slower and less non-technologically savvy investors? Significant volumes are already traded in these ways; e.g., the opening and closing auctions. Finally, if HFT should be regulated, can regulations specifically target behavior viewed as negatively affecting markets without impacting the efficiency brought by other aspects of HFT; e.g., the price efficiency associated with marketable HFT orders?

 

Page 3

References

1)  Hendershott, Terrence, and Ryan Riordan, 2012, “High Frequency Trading and Price Discovery,” working paper, University of California, Berkeley. Nasdaq makes this data set is available to academic researchers and it has been used in a number of studies. The data represent a limited set of HFT, e.g., excluding some small finance firms and large integrated firms, in a single market. Unless there are reasons to believe that HFT trade differently on other markets or the not identified HFT follow different strategies, the dataset is representative of HFT.

2) Whether HFT profits should be viewed as a tax on other investors is an interesting question. If one views trading as a zero sum endeavor then HFT profits represent a loss to other investors. If HFT are viewed as intermediaries providing immediacy to other investors, then HFT profits can be seen as the value of the service they provide. Both of these sources of revenues can be associated with marketable and non-marketable orders.

3) Hendershott, Terrence, and Ryan Riordan, See, for example, Hendershott, Terrence, Charles Jones, and Albert Menkveld, 2011, “Does Algorithmic Trading Improve Liquidity?” Journal of Finance 66, 1-33.