Who Benefits from Securities Exchange Innovation?

The role of securities exchanges in the proper functioning of financial markets is difficult to overstate. Exchanges bring together investors, large and small, allowing them to realize gains from trade, mobilize capital, and incorporate relevant information into prices. To keep pace with the unrelenting march of technology and customer expectations, exchanges continuously innovate. Theoretical models predict that such innovation may have both positive and negative effects on market quality. Likewise, empirical studies find that while some exchange initiatives are beneficial for liquidity, others may be disadvantageous. With such varying findings, it is of interest to ask if exchange innovation is beneficial as a whole, and if all market participants capture the benefits.

We address these questions using a multi-year sample of new technological offerings by stock exchanges in the United States. These offerings range from improvements in data dissemination to enhancements in order processing by exchange engines. The results suggest that exchange innovation is generally associated with lower trading costs for investors who trade in small quantities. In the meantime, institutional investors do not appear to derive as much benefit from new exchange technologies; while their trading costs are typically unaffected, their trading volumes decline after new technologies are introduced.

Our sample of technological offerings comes from the public record of patents filed by theexchanges. In the U.S., when a company develops a new technology it files an application for a patent with the Patent and Trademark Office describing the invention and claiming an exclusive right to it. We collect a sample of all exchange filings in 2003-2021, for a total of 194 patents. The patents capture a wide range of exchange activities. For example, the NYSE patent US-9450999-B2 (https://bit.ly/4c2nsWy) describes technology for high performance data streaming, while Nasdaq patent US-11671395-B2 (https://bit.ly/3Vvajjd) claims rights to a “message tracking apparatus [that improves] the latency of a message processing system.” Like these two examples, the sample patents generally focus on enhancing the speed and efficiency of exchange infrastructure and on improving customer connectivity.

To measure trading costs, we use two data sources, the intraday Trade and Quote (TAQ) database and the Abel Noser institutional trading database. TAQ allows us to compute a set of conventional liquidity metrics such as quoted and effective spreads. The quoted spreads measure displayed liquidity, that is trading costs advertised by liquidity providers. In turn, the effective spreads capture liquidity costs that are actually incurred by market participants. Both liquidity metrics decline following introductions of new exchange technologies.

What may drive liquidity cost reductions associated with innovation? To shed light on this question, we examine two components of effective spreads, the price impact and the realized spread. The former captures adverse selection costs incurred when providing liquidity. The latter reflects liquidity provider inventory costs, fixed costs, and profits. Both components decline post-innovation, consistent with the notion that new exchange technologies help reduce market making costs and may enhance competition for liquidity provision. In addition, the data show that innovation is followed by lower price volatility and greater price efficiency.

Liquidity in the U.S. equity market gradually improves since the early 2000s (Angel, Harris, and Spatt (2015)). It is therefore important to ensure that our tests do not merely pick up this background trend. To do so, we detrend all variables and use only the detrended variables in regression tests. In addition, we verify the results in a difference-in-differences (DID) setting against a sample of Canadian securities, for which long-term liquidity trends are similar to those of U.S. equities. Our findings are robust in the cross-section and replicate for stocks of all sizes.

The TAQ-based liquidity proxies discussed so far apply mainly to market participants, who seek to trade small share quantities represented by the size of the best quotes. Meanwhile, institutional investors often trade large amounts, and for them the TAQ proxies may not be the most suitable (Eaton, Irvine, and Liu (2021)). To shed light on the effects of exchange innovation on institutions, we use Abel Noser data and estimate the execution shortfall, a metric commonly used to measure institutional trading costs (e.g., Conrad, Johnson, and Wahal (2001) and Anand, Puckett, Irvine, and Venkataraman (2013)). The results show that exchange innovation is not associated with changes in execution shortfall. Nevertheless, innovation is followed by lower institutional trading volumes. In summary, while exchange innovation may benefit the seekers of small amounts of liquidity, its effects are more ambiguous for those, who trade large amounts.

What may be behind this dichotomy? We posit that the answer may lie with tech-savvy trading firms that dominate trading in modern markets. These firms are commonly referred to as high-frequency traders (HFTs). Fierce competition for speed drives their unrelenting pursuit of innovation. Using a dataset that identifies trading by such firms, we illustrate that they are quick to react to new exchange technologies, with their share of trading volume increasing shortly after exchange patent filings.

Prior research finds that tech-savvy firms play a major role in liquidity provision, yet also points out that they are highly skilled at avoiding adverse selection and maintaining low inventories (Brogaard, Hagströmer, Nordén, and Riordan (2015), Brogaard, Carrion, Moyaert, Riordan, Shkilko, and Sokolov (2018)). Consequently, liquidity provision by such firms may be beneficial for relatively uninformed investors who trade small quantities, but not for institutional liquidity seekers, who are often informed and trade large positions. As new technologies improve the ability of these firms to avoid adverse selection and unwanted inventory accumulation, they may provide less liquidity when institutional investors demand it or provide liquidity at prices greater than what institutions are willing to pay. The probable outcome is a decrease in market participation by institutions.

In addition to liquidity provision, tech-savvy firms regularly engage in liquidity demand. Their inventory management and arbitrage strategies are highly time-sensitive and often require taking liquidity (Brogaard, Hendershott, and Riordan (2014), Chaboud, Chiquoine, Hjalmarsson, and Vega (2014), Boehmer, Li, and Saar (2018), Baron, Brogaard, Hagströmer, and Kirilenko (2019)). The ability to harness new technology may give these firms an advantage over institutions in the race to trade against outstanding quotes, potentially resulting in decreased execution probabilities for institutional flow.

Do institutions fail to benefit from exchange innovation because they do not pay attention to new technological developments? To shed light on this question, we examine patent citations with the assumption that market participants affected by a patent will cite it. We find that while most patents are cited primarily by the tech-savvy trading firms and firms that manufacture technologies used for trading, a smaller group of patents is cited mainly by institutional investors. Examining these two patent groups separately, we show that the patents cited by trading and technology firms tend to conform to our main findings. They primarily benefit those who trade small quantities and are associated with lower institutional trading volume. Meanwhile, patents cited by institutions are not linked to any adverse effects for small traders, yet are associated with greater institutional trading volumes. This result suggests that institutions are not entirely disinterested in new technology and are capable of harnessing it. However, the relatively small number of patents cited by institutions indicates that most technologies may be challenging to appropriate for institutional benefit.

Finally, we seek to understand the driving forces behind exchange innovation. Innovation may occur as a routine practice, with exchanges continuously developing new product offerings, or as a strategic response to various challenges such as declining volume or market share and innovation by the rival exchanges. Our results are consistent with both routine practice and strategic response explanations. Nevertheless, the strategic response effect is rather weak. Furthermore, the innovating exchanges do not seem to directly benefit from their new technology; their own trading volumes and market shares do not increase. Overall, the relentless advancement of technology and the need to meet the expectations of tech-savvy customers responsible to liquidity provision seem to be the primary drivers of exchange innovation efforts.

Our findings contribute to an ongoing discussion of modern market structure and the securities exchange industry. Even though exchange innovation by the incumbent exchanges likely does not have nefarious intentions towards any particular group of market participants, it may inadvertently benefit those, who are best equipped for technological change. In this respect, establishing markets that cater to the needs of the institutional community may be socially optimal. This logic echoes Biais, Foucault, and Moinas (2015), who model interactions between investors that operate at different trading speeds and examine the benefits of segmenting markets into those that cater to fast and slow traders.

Related literature

Several theoretical models consider the effects of advancements in exchange technology and predict that such advancements may be both beneficial and detrimental. Menkveld and Zoican (2017) show that increasing exchange engine processing speed may have both positive and negative effects on liquidity. In their model, the direction of the effect may vary over time and in the cross-section and depends on the mix of liquidity traders and news events. Pagnotta and Philippon (2018) show that increasing exchange speed may lead to lower trading fees and greater investor participation. They however caution that purely technological improvements to trading technology may lead to limited welfare gains.

Perhaps the closest to our study, Cespa and Vives (2022) model an environment, in which exchanges supply technology services to various market participants. The fast adopters of new technology use it to improve liquidity, generating a positive welfare effect. In the meantime, traders who are unable to adopt new technology face lower payoffs. Our results echo the implications of this model. While exchange innovation benefits small liquidity consumers by allowing liquidity providers to reduce adverse selection and inventory costs, it also results in lower market participation by institutional investors, who seem to often struggle with adopting new technology.

The empirical literature examines a number of technological enhancements to the exchange trading process. The outcomes of such enhancements vary. Hendershott, Jones, and Menkveld (2011) show that exchange-driven automation of liquidity provision results in liquidity improvements. Conversely, Hendershott and Moulton (2011) and Foucault, Kozhan, and Tham (2017) find that exchange technology enhancements that benefit liquidity-demanding strategies have adverse liquidity effects. Brogaard, Hagströmer, Nordén, and Riordan (2015) and Conrad, Wahal, and Xiang (2015) study co-location, a practice that allows trading firms to place trading servers near exchange matching engines. They report that the practice is mainly used by liquidity suppliers and therefore benefits liquidity. Overall, empirical findings vary across technologies and practices, making it difficult to draw conclusions about the net effects of technological advancement. Our analysis of a multi-year sample of exchange technological offerings therefore brings the literature closer to understanding the general effects of exchange innovation.

Our study is related to the literature that examines liquidity costs and to a sub-set of this literature that focuses on institutional investors. Anand, Puckett, Irvine, and Venkataraman (2013) suggest that despite the overall trend to lower liquidity costs in the 21st century, institutional investor trading costs did not decline. This result echoes the earlier findings by Goldstein and Kavajecz (2000) and Jones and Lipson (2001), who show that regulatory actions that reduce spreads may not have similar liquidity-boosting effects for large traders. Eaton, Irvine, and Liu (2021) generalize this finding to caution that changes in institutional trading costs do not always align with changes in conventional TAQ-based liquidity metrics.

Our arguments are also rooted in the literature on sophisticated trading firms often referred to as HFTs. Such firms fulfill several important functions, among which liquidity provision and ar- bitrage stand out as dominant (Brogaard, Hendershott, and Riordan (2014), Chaboud, Chiquoine, Hjalmarsson, and Vega (2014), Brogaard, Hagströmer, Nordén, and Riordan (2015), Boehmer, Li, and Saar (2018)). HFTs are highly skilled at avoiding adverse selection and maintaining low inventories (Korajczyk and Murphy (2019), Van Kervel and Menkveld (2019)), and therefore their liquidity supply is likely more beneficial to uninformed investors seeking small amounts of liquidity than to institutional investors.

Read more