The market for trading stock in the United States is remarkably complex. As of July 2023, there were 16 registered public exchanges like the New York Stock Exchange, 70 alternative trading systems (ATS) that act like private exchanges, and numerous other entities where equities securities can be bought and sold. Large institutional investors use their own trading desks and the trading desks of investment banks to source liquidity—that is, to exchange stock for money or vice-versa—across this array of choices. Individual investors (i.e., retail traders) typically rely on their brokers to route their orders, matching buyers and sellers with nearly similar bid and ask prices (with the trading platform getting a small payment for its services).

There are economies of scale in the technology associated with this order routing and execution, meaning that the larger a trading entity is, the more cheaply and readily it can match buyers and sellers. Retail brokers have come to rely on a set of electronic marketplaces known as “wholesalers” to execute their customers’ liquidity-demanding orders promptly at, or better than, the quoted price. The largest of these wholesalers are Citadel, G1X Susquehanna, Jane Street, Two Sigma Securities, UBS, and Virtu Financial. The Securities and Exchange Commission estimates that over 90 percent of retail orders were routed through those six wholesalers in the first quarter of 2022.

In both historical and absolute terms, retail traders currently enjoy excellent execution quality both in terms of transaction speed and the price at which orders fill. Five retail brokers in the United States stopped charging commissions for equity trades in October 2019. Most other discount brokers did likewise, largely because many of them now charge the wholesalers for the opportunity to facilitate all those trades.

There is abundant empirical evidence that wholesalers offer high-quality executions. SEC Rule 605 requires that execution venues report the execution quality of orders that are routed to them. In 2022, the largest wholesalers’ Rule 605 reports showed a $3 billion savings for retail customers from quick execution.

However, because the reports cover only a portion of the orders handled by wholesalers and ignore some important additional services offered by wholesalers, we have estimated that the actual retail customer savings might be five times the officially reported number. The fact that wholesalers offer customer savings that are not officially recognized suggests that the wholesaler market is quite competitive. A 2022 paper by Anne Dyhrberg, Andriy Shkilko, and Ingrid Werner also provides evidence consistent with strong competition among wholesalers.

SEC’s proposed rule / Nevertheless, the SEC has expressed concern about concentration in the wholesaler market. In late 2022 it proposed the Order Competition Rule. This rule would all but prohibit wholesalers from immediately filling most orders “internally,” meaning from their own inventory rather than submitting the orders to a qualified auction operated by an “open competition” trading center such as a registered exchange. The SEC posits that retail traders will receive better prices if the order is exposed to competition from additional liquidity providers and estimates a current market structure annual “competitive shortfall” of between $1.12 billion and $2.3 billion, yielding a 0.15¢–0.47¢ per-share benefit to the proposed rule.

The SEC has received many comment letters claiming those benefit estimates are heavily overstated. We have a different concern about the SEC analysis: we think it significantly understates the proposed rule’s cost.

The SEC used real-world data to infer (with a considerable degree of uncertainty) which public-reported trades result from a retail investor. After identifying alleged retail trades using this algorithm, the SEC estimated how frequently the quoted price would move against the trader during the auction period—that is, the final transaction price would shift (usually just slightly, what is known as “quote fade”) from the initial buy or sell orders. In these cases, customers are harmed should the auction fail because they now trade at worse prices than they would have if the auction had not been imposed. The SEC then assumed the cost to the customer is 1¢ per share when an auction fails instead of actually measuring the cost. The expected cost of auction failure is the frequency with which quotes move against the trader multiplied by the cost when the quotes move against the trader—assumed to be a penny. The SEC concludes that the expected potential cost of failed auctions is 0.046¢ per share, an order of magnitude smaller than the upper bound of their benefit estimate.

Testing the SEC’s analysis / We tested the SEC’s analysis by using a sample of actual retail orders in May 2022. We found that the potential expected cost per share of failed auctions is considerably greater than the commission’s estimate of 0.046¢ per share. It was frequently as large as the commission’s lower bound estimate of the proposed rule’s benefit of 0.15¢ per share, and potentially greater than the commission’s upper bound on benefit of 0.47¢ per share. We estimated the annual cost of the proposed rule to be $1.12–$1.97 billion under the range of proposed auction lengths.

Why did we and the SEC reach such different conclusions regarding the potential cost of failed auctions? The possible reasons are numerous but begin with the fact that we assessed auction failure costs relative to the order-receipt-time quote and not (as the SEC did) the trade-time quote. Although wholesaler trades generally happen quickly after order arrival, quotes also move quickly, which means delaying the measurement of quote fade to trade time diminishes the cost estimate. At the longest proposed auction period (three-tenths of a second), we found that adverse quote changes affect over 20 percent of our actual retail shares compared to less than 5 percent of the SEC’s inferred retail trades.

We also measured the actual amount of quote fade (per share) and found that it frequently exceeded the 1¢ assumed by the SEC. Using the Commission’s preferred mid-quote benchmark price, the average per-share cost of failed auctions was 31 percent higher than the Commission assumed. If we consider the actual price the retail trader received in the current trading environment, the average per-share cost of the auction is 1.98¢, nearly double the SEC’s assumed cost.

Taking the likelihood of an auction failure and multiplying by the associated average cost per share produced an expected cost per share of 0.35¢. In comparison, the SEC’s estimated expected cost was 0.046¢ per share, 7.6 times larger. If all the order flow that wholesalers currently execute is forced into the proposed auctions, that would impose a $1.968 billion annual cost on retail investors.

Conclusion / Given our estimates, we are concerned the SEC’s proposed rule would fail a cost–benefit test, even if the Commission’s benefit estimates are accurate. If the SEC is concerned about order quality in wholesalers, we suggest instead that it proceed with another proposal, the Disclosure of Order Execution Information Rule, which would expand execution quality statistics reporting. That would allow market participants to better judge these venues’ quality themselves.

Readings

  • “On the Potential Cost of Mandating Qualified Auctions for Marketable Retail Orders,” by Robert Battalio and Robert Jennings. Journal of Investing, forthcoming.
  • “Retail Order Execution Quality under Zero Commission,” by Samuel Adams and Connor Kasten. SSRN Working Paper no. 3779474, 2021.
  • “The Retail Execution Quality Landscape,” by Anne Dyhrberg, Andriy Shkilko, and Ingrid Werner. Working paper, 2022.