Market Rewards Programs

Malt Liquidity 74

Middle Management

Price, by definition, is the meeting of a bid and an ask. In an ideal world, price is a function of theoretical motion up and down supply and demand curves, the intersection of which gives us an equilibrium price, which, upon arrival, gives us a seamless transaction. In practice, we call this price discovery — the moving of bids and asks in an order book (or through barter and negotiation for goods and services) that results in points of liquidity where trades occur.

In reality, of course, price discovery is facilitated by a lot more than natural orders placed. A network of middlemen in every industry serve to provide liquidity to enhance the velocity of money and facilitate price discovery. A network of payment processors and credit institutions underwrites transactions for convenience store purchases by providing immediate liquidity to the seller, temporary up-front liquidity to the buyer, and protection from fraud (because it’s the company’s money that would be stolen, not yours.) For this service, they take a cut of the transaction. In markets, we’ve talked about the role of market makers peppering the order book with quotes plenty of times:

…all market making does is facilitate transactions when there isn’t necessarily a buyer matched to a seller at the best bid/best ask price. Every transaction has to have an “other end” - if I wanted to buy Ford stock and had to wait for someone to come along who wanted to sell it precisely at the price I wanted to buy it at, transaction volume would be much lower, and prices would be much choppier, cause buyer and seller are in no way guaranteed to match. So all the market maker does is provide a bid and an ask on both sides, and when they are “hit” - one of those orders is filled - they try to seamlessly pair off an opposing trade, and the difference - the “spread” they make - is what makes this service worth providing.

The venue is far more complex and low-latency, but the idea remains the same. Liquidity — the friction between theory and reality — is what creates opportunities to profit.

Being a middleman is by no means riskless. Credit institutions are naturally exposed to counterparty risk on relatively uncollateralized debt — what good is rapidly accruing credit card interest if the borrower has no ability to repay their debt, given that it’s unsecured? — and operational risk, such as identity theft and fraud. To mitigate these risks, they can adjust the lines of credit extended to individuals based on creditworthiness (reducing liquidity offered), block certain transactions unless expressly approved (pulling liquidity entirely), and incentivize more reliable, higher value users to utilize their cards more (payment for order flow.) Similarly, market makers utilize a lot of the same tools to mitigate their risks while facilitating transactions — they may reduce the size and depth of on-screen quotes during more volatile periods (reducing liquidity offered), remove quotes entirely when prices gap around so as to not get adversely hit (pulling liquidity entirely), and incentivize non-informed, reliably profitable retail flow to get routed to them to execute (payment for order flow.) In markets, while the movement and pulling of quotes is regulated by NBBO rules and anti-spoofing measures in Dodd-Frank, payment for order flow has only recently come under close scrutiny, perhaps unjustifiably so.

In the credit card world, “payment for order flow” makes total sense — credit scores are essentially an indicator of potential flow toxicity and potential scale of usage. Through rewards points programs and premium credit cards (which have higher transaction fees), companies essentially acknowledge that a class of customers is so reliably profitable that a kickback of the “spread” they make is worth ponying up to keep them around and stabilize their revenue. Though the system has definitely been abused by churners (the MEVs of consumer finance) to the tune of costing JPM $300 million at one point for Chase Sapphire rewards, it’s a win-win — while prices may increase to offset the cost of transaction fees, merchants get increased settlement speed and ease of transaction, companies get stable revenues, and customers get emergency liquidity, fraud protection, and rewards for naturally transacting.

In markets, “payment for order flow” also makes total sense — remember what I was saying about venue and participant selection? While market makers are obligated to quote across instruments, participant selection is definitely a high priority concern — if you are trading against a party who has a better idea of what the future price movement is than you do — an informed speculator — your risk of losing money is much higher. Conversely, retail trading flow is so reliably predictable and profitable for market makers that it justifies kicking a percentage of the spreads collected to the brokers that funnels trades their way. The market maker gets stable, reliable trading profits, the brokers don’t have to build costly execution infrastructure and compete with more adept players, and retail traders get reduced price impact and price improvement.

The controversy probably arises due to how “price impact” and “price improvement” are measured, as the benefits in the market making case are not nearly as obvious as the credit card rewards program case. If I yeet a market order into a thin book, I’m at the mercy of the spread — at best, I’m crossing, and at worst, I plunge through thin air to find any resting orders. A market order routed through an MM that has purchased the flow, though, can be somewhat internalized — you won’t get worse than the spread, but there are many sources of liquidity that aren’t directly visible on the book, as we’ve discussed before:

However, in wider spreads, there may be parties willing to transact at prices better than the visible bid-ask, but who don’t want to show their willingness by placing a visible order in the book, perhaps due to not wanting to reveal a position in the market or to avoid the risk of getting picked off and ending up with undesired exposure. This is “below screen” liquidity.

Price improvement, on the other hand, is based on your own resting orders. When you set a limit price, you cannot execute at worse than that price — of course, if the price gaps through, you may not transact at all, but if the price trades through and clears your order at the level then you will get filled at that price. However, if the internal spread is tighter than the visible bid ask (which might be a penny wide), with internalized flow, you might be able to execute better than your limit price by a few fractions of a penny. It’s a win-win — brokers get a cut of the spread cost and don’t have to invest in infrastructure themselves, market makers get reliable flow, and the retail customer does receive better execution. Which is why it’s surprising to me that the SEC has proposed new rules on how to manage retail orders:

The centerpiece of the SEC’s plans is a proposal for brokers to send many small-investor stock orders into auctions. This would enable a mix of high-speed traders and institutional investors such as hedge funds or pension funds to compete to fill the orders, with the idea that investors would get better prices as a result—higher prices if they are selling shares, or lower prices if they are buying.

The auctions would apply to so-called marketable orders—in which investors buy or sell stocks at the currently available price—less than $200,000 in size and placed by investors who average fewer than 40 trades a day. They would be required to last between one-tenth and three-tenths of a second, roughly the duration of a blink of an eye, and would likely be run by exchanges. 

I’m not sure why allowing institutional investors and hedge funds direct access to retail orders helps anyone. Big investors already route their execution through institutional trading desks — is this supposed to give them more access to retail book depth provision? Would this even be meaningful in terms of saving retail spread cost?

The main argument seems to be that everyone should have access to indiscriminate uninformed flow, rather than a small group of traders like Virtu and Citadel who dominate retail flow (with Virtu holding about ~25% and Citadel about ~40%). But is this really such a big deal? As I’ve written before,

High frequency traders are not the industry titans that investment banks, mutual funds, hedge fund managers, or even college endowments are - they’re service providers to make electronic markets more seamless. Virtu, one of the largest and most well-known market makers due to their propensity for never losing money during a trading day, probably has a revenue of around 4.5-5 billion in perhaps the most historically market-maker friendly year ever. Compare this to Harvard’s endowment alone, which is 42 billion, or Bill Ackman’s Pershing Square, which handles about 12 billion. The fees on these AUMs alone would translate to well into the 9 figures, if not well into the billions after performance incentives, which is pretty close to how much Virtu probably makes on that 5 billion in revenue.

Virtu employs about a thousand people. Pershing Square employs 30. Are these the people who need unfettered access to retail flow in the name of “making the game fairer”?

The PFOF model allows brokerages to compensate retail traders through zero- and lowered- commission trading. Furthermore, PFOF allows reliable trading for retail investors regardless of the liquidity of the instrument — on a stock like AAPL, sure, batch auctions probably allow spreads to come inside pennywide a bit more, especially given that the SEC is also trying tick size requirements

A second proposal, advanced with a 5-0 vote, seeks to level the playing field between exchanges and wholesalers, according to the SEC. It reflects longstanding complaints by executives from the New York Stock Exchange and other exchanges that they are handicapped in their ability to attract individuals’ orders. 

The proposal centers on the price increments in which stocks are quoted and traded, called “tick sizes.” Currently, for most stocks, exchanges can only quote prices in increments of one penny. The proposal would allow exchanges to quote prices in tighter increments, ranging from one-tenth of a penny to a whole penny, with the most actively traded stocks generally getting the tightest increments. 

However, would this remain true for a very thin stock or option? Off the top of my head, this auction approach would likely result in the necessity of a larger player to participate to tighten the spread. If they don’t show any interest or volume, is the retail order simply at the mercy of the spread? What about real-time access to below-screen liquidity?

Under the proposal, auctions could be skipped if wholesalers can beat a key price metric for filling an order. Wholesalers could also scoop up orders for which auctions fail to produce good bids. 

This key price metric appears to be midpoint fills, which, admittedly, are kind of an art form for the retail trader to receive right now. But in the case of the failed auction, the market maker would be the trader facilitating your fill anyway. I don’t see what this is supposed to improve, other than adding a very slight delay.

To me, the SEC efforts seem misguided. In the interest of “leveling the playing field”, their rules seem to propose pitting the actual sharks — the Ackmans, Cohens, and Einhorns of the world (two of which were implicated in massive insider trading cases) — against retail trades rather than allowing them to be facilitated smoothly with a slight kickback to everyone involved in making the flow happen. How is further subsidizing buy-side price impact better for retail? It’s like banning credit card rewards points to increase competition between credit institutions, while ignoring the monopolistic practices of payment processors — who is this really benefiting? Turns out, in the fight to pick up the pennies, the SEC was the steamroller all along.