Since I was last with you at these conferences, Rich, I’ve been honored to hold a number of roles. Most recently I spent three and a half years at MIT, where my research and teaching centered on the intersection of finance and technology.

One thing that I’ve come to believe is that technology and finance have coexisted in a symbiotic relationship since antiquity, bringing greater access, innovation, and competition to our markets and spurring economic growth.

Our central question is this, though: As new technologies come along and change the face of finance, how do we continue to achieve our core public policy goals?

Today, I’d like to focus on one area where finance and technology are intersecting in new ways: equity markets.

Today, technology has changed how market makers interact, how trading platforms compete, how investors access those markets, and the economic incentives amongst these various market participants. Further, retail investors can trade over commission-free brokerage apps. That wasn’t the case even a few years ago.

Underpinning many of these changes is also an insatiable demand for customer and transaction data by companies.

Today, though, we are often relying on rules written in an earlier period. Rules mostly adopted 16 years ago do not fully reflect today’s technology. I believe it’s appropriate to look at ways to freshen up the SEC’s rules to ensure that our equity markets reflect our mission: to maintain fair, orderly, and efficient markets, while ensuring we protect investors and facilitate capital formation.

The question is whether our equity markets are as efficient as they could be, in light of the technological changes and recent developments.

I’d like to talk about two observations in those markets: segmentation and concentration. Then, I’m going to turn to two developments: the rise of payment for order flow and gamification. Lastly, I want to talk about the requirements for best execution in the context of the national best bid and offer (NBBO).

Today, our markets essentially have three different segments. While the public generally thinks of lit markets, like Nasdaq and the New York Stock Exchange, those big public markets only accounted for about 53 percent of trading volume in January.[1]

So where’s the other 47 percent — trading interest that’s not displayed on the lit markets?

About 9 percent of January’s volume was executed on alternative trading systems, also known as dark pools.

That leaves about 38 percent, most of which was executed by off-exchange wholesalers. Just seven wholesalers accounted for the vast majority of this group.

If Rich or I decide to go onto a brokerage platform to buy five shares of a company, if it’s a market order, more likely than not, it won’t be routed to Nasdaq or the New York Stock Exchange. A lot of people are surprised to learn that the vast majority of such orders go to these wholesalers.

Further, wholesalers have many advantages when it comes to pricing compared to exchange market makers. The two types of market makers are operating under very different rules.

Exchange market makers, for example, must compete with each other on an order-by-order basis to offer the best price. Wholesalers are able to price their segmented order flow simply by referencing the NBBO, which is a much less competitive benchmark.

Now, let me add a second observation. Within the off-exchange market maker space, we are seeing concentration. One firm has publicly stated that it executes nearly half of all retail volume.[2] There are many reasons behind this market concentration — from payment for order flow to the growing impact of data, both of which I’ll discuss.

Market concentration can deter healthy competition and limit innovation. It also can increase potential system-wide risks, should any single incumbent with significant size or market share fail.

These two observations raise several questions in my mind. Does this segmentation and concentration best promote fair, orderly, and efficient markets? Does it help the price discovery function or capital formation for issuers? Does it best protect investors?

Next, let me turn to the first of the two developments: the rise in the use of payment for order flow.

While January’s markets events brought much attention and debate to payment for order flow for retail trades, I want to note a second kind of payment for order flow: payments from exchanges to market makers and to brokers, also known as rebates.

Both types of payment for order flow raise questions about whether investors are getting best execution.

Let’s discuss payments from wholesalers to brokers.

Brokers profit when investors trade. For those brokers who have these arrangements —and not all do — higher trading volume generates more payment for order flow.

What makes the current zero-commission brokerage environment different is that investors do not see their costs as they’re executing trades, so they may perceive them as free.

Further, as a significant and growing share of retail orders are routed to a small, concentrated group of wholesalers, certain market makers have more data than others.

As we’ve seen in many other parts of our economy, a few central actors gain advantage through the increasing use of data — whether in search, e-commerce, or in this case, the data from transaction flows.

Payment for order flow raises a number of important questions. Do broker-dealers have inherent conflicts of interest? If so, are customers getting best execution in the context of that conflict? Are broker-dealers incentivized to encourage customers to trade more frequently than is in those customers’ best interest?

Some of these issues were highlighted in the SEC’s settled enforcement action against Robinhood last December.[3] As described in the Commission’s order, certain principal trading firms seeking to attract Robinhood’s order flow told them that there was a tradeoff between payment for order flow and price improvement for customers.

Robinhood explicitly offered to accept less price improvement for its customers in exchange for receiving higher payment for order flow for itself. As a result, many Robinhood customers shouldered the costs of inferior executions; these costs might have exceeded any savings they might have thought they’d gotten from zero commission trading.

The second development I wanted to mention is gamification and other behavioral prompts, which we’re seeing across mobile brokerage apps. Many of these features encourage investors to trade more, leading to more payment for order flow for the brokers.

Some academic studies suggest, however, that more active trading or even day trading results in lower returns for the average trader. Thus, I have asked staff to prepare a request for public input for consideration on these issues.

Lastly, let me turn to the requirements for best execution in the context of the NBBO.

Broker-dealers are obligated to seek the best execution for their customers’ orders — not just better execution.

It’s interesting to note that the United Kingdom,[4] Canada,[5] and Australia[6] don’t allow broker-dealers to route retail orders to wholesalers in return for payments. The European Securities and Markets Authority has also raised concerns about these potential conflicts of interest in payment for order flow and best execution.[7]

Again, it’s best execution — not just better execution. But it’s best execution in comparison to what?

That brings me to the NBBO. In fulfilling the requirement of best execution, brokers must consider, among other factors, prices currently being quoted. The NBBO is designed to aggregate information across different exchanges. I believe there are signs, however, that the NBBO is not a complete enough representation of the market.

First, as evidenced in January, nearly half of the trading interest in the equity market either is in dark pools or is internalized by wholesalers. Dark pools and wholesalers are not reflected in the NBBO. Moreover, the NBBO is also only as good as the market itself. Thus, under the segmentation of the current market, nearly half of trading along with a significant portion of retail market orders happens away from the lit markets. I believe this may affect the width of the bid-ask spread.

Further, as it relates to the lit markets, while the definition of odd lots is planned to change under the SEC’s 2020 Infrastructure rulemaking, the NBBO still doesn’t include many of the exchange prices, such as odd lots and non-displayed orders. Additionally, by SEC rule, the NBBO must be priced in penny increments.

Wholesalers are able to transact at sub-penny increments. As a result, wholesalers may operate on an unequal playing field when competing for order flow.

I’ve asked staff to make recommendations for the Commission’s consideration on best execution, Regulation NMS, payment for order flow (both on-exchange and off-exchange), minimum pricing increments, and the NBBO, with the aim of continuing to make our markets as efficient as possible.

Before I close, I briefly want to mention one other aspect of equity structure — that’s central clearing.

I’m reminded of an old saying in the markets: “Time equals risk.” These events have prompted questions about whether we can lower risk by shortening the time of our settlement cycles from T+2.

The good news is, though it will take a lot of work, we now have the technology to further shorten the settlement cycles, not only to T+1, but even to same-day settlement — T+0 or “T+evening.”

I believe shortening the standard settlement cycle could reduce costs and risks in our markets.

I’ve also asked staff to consider whether there are other gaps in central clearing that the SEC can address — such as in money movements, netting, or the posting of collateral.

Rich, I look forward to your questions. I covered a lot on our equity markets, but there’s certainly more to discuss about treasury markets and non-treasury fixed income markets — and how technology is giving us the opportunities to bring greater transparency, competition, and resiliency to those markets.

ENDNOTES

[1] See Cboe Global Markets, available at https://www.cboe.com/us/equities/market_statistics/historical_market_volume/ (providing downloads of historical reported volume in NMS stocks by all self-regulatory organizations) (“SRO Volume Data”); see FINRA OTC (Non-ATS) Transparency Data, available at https://otctransparency.finra.org/otctransparency/OtcDownload (providing downloads of historical reported volume in NMS stocks by FINRA members) (“FINRA Member Volume Data”).