All Yous Guys Pay Up! (Response to Larry Tabb Open Letter to Senator Levin)


Larry Tabb has just published an open letter to Senator Carl Levin in which he supports the practice of payment for order flow (PFOF), and urges Sen. Levin to reconsider his proposal to ban the practice.  While we at Themis are ardent readers and consumers of so much of the high quality product that Tabb Group produces in general, and what Larry produces in particular, this letter and stance disappointed us. Needless to say we disagree with Tabb, and wholeheartedly agree with Senator Levin.

Larry’s argument runs as follows:

1)      Banning payment for order flow will just create new challenges and create new conflicts of interest.

2)      “I know that PFOF sounds bad, but the SEC says it’s legal.”

3)      Retail order flow is valuable, but banning PFOF will not make it less desirable or sought after by market participants who are seeking to trade very short term.

Larry says that retail orders are valuable and easy pickings for internalizing market makers (like Citadel, UBS, Citigroup, and Knight-Getco) to make money off of because they are small in size and… errr… dumb (negative alpha), as opposed to the orders from institutions that result from Wall Street research, are large, and tend to be the same direction (herded as Larry puts it).

Larry says that at least under the PFOF mechanisms, executions get some tiny amount of price improvement compared to what they would get if they were sent to a stock exchange. Banning PFOF would mean that online brokers would have to have trading desks, and have expensive technology. They might even be tempted to keep the dumb retail orders in house, and have their own prop trading arm to trade against it (like all the big banks do currently J ).

And Larry’s pro-PFOF stance makes some questionable assumptions in our opinion, and without data to back it up. This morning we want to reiterate our anti PFOF stance, and contrast it with Larry Tabb’s.

We at Themis have long been against the practice of payment for order flow, including maker taker on the exchanges as well as PFOF in general.  We have made that opinion known in numerous articles and media interviews, in our book Broken Markets, and as well to our regulators. Earlier this year, we addressed the issue again with this post on our blog, What Happens if Payment For Order Flow Dies:

“Conflicts of interest would be drastically reduced. There is no revenue to be gained by not solely acting as agent for the order initiator. TD Ameritrade doesn’t get to make some $200 million per year selling the free option to trade against their own customers’ orders. Agency brokers don’t get to make money by selling institutional child orders to ELPs for rebates either. Buyside traders and agency brokers don’t send orders to the highest rebate giving exchanges for limit orders, or to inverted exchanges that give rebates for marketable limit or market orders. Orders go to destinations based on their market pricing at the time. If NYSE is offered at $20, a limit buy order at $20 goes to NYSE. If BATS is offered at $20, a limit buy order at $20 goes to BATS.”

Larry is concerned about the profitability of big bank market makers and high frequency market makers. We are not.

Larry thinks the dumb retail order would get filled at the far side of the spread all the time on the exchanges, and not receive the amazing 1/100th of a penny price improvement it gets from Citadel. We disagree. In an order driven market, some retail orders will be limit orders, and some market orders, and some marketable limit orders. They will interact on the exchanges sometimes at the bid, sometimes at the offer, and sometimes at the midpoint (the stock exchanges do allow for midpoint orders!) We believe that the retail order will be better off on the public lit exchanges!

Larry thinks eliminating PFOF will create other conflicts of interests elsewhere in order routing, still. We disagree. We believe if PFOF is eliminated everywhere, including the stock exchanges, there will be no order routing conflicts, period. Might spreads widen a penny? Maybe, maybe less than that. The high frequency market makers and big banks will adjust, however. They will quote and prop trade on the public exchanges, just perhaps at a slightly wider spread. And those exchanges will allow for glorious interaction of all kinds of order flow: retail, institutional, HFT, quant, stat arb, value, etc.. Those order books will be efficient – in good times and bad.

Larry ignores the concept of price discovery in his letter and pro PFOF stance. We believe that payment for order flow harms the price discovery process by magnifying the segmentation of order flow in the market place, and limiting the ability for all that order flow to cross-interact. When retail brokers sell their customers’ orders to the highest bidder, they are actually blocking market forces since they are restricting the number of participants that can interact with that flow.  We believe that price discovery is enhanced when diverse pools of liquidity (institutional, retail, quant, etc.) are allowed to interact with each other.  True prices could then be set by all market participants, not just a subset of participants who are looking to pick off uninformed flow.

Larry ignores what happens to markets dominated by payment for order flow when under stress. Perhaps Larry does not remember the May 6th Flash Crash. Perhaps he does not remember that the SEC-CFTC Flash Crash report laid much of the blame for that afternoon’s events squarely at the feet of the high frequency cherry-picking internalizers. Those internalizers cherry picked which retail order to interact with – for profit, and exhausted the rest out to a stressed marketplace with no regard for price – even at stub quotes of a penny. We think that despite the steady and complacent march of the stock market upwards in recent years, it is a mistake to ignore that fairly recent memory. The payments that market makers send out to the online brokers essentially give them the right to receive a sneak peek at market orders, and a free option. In times of calm, it is highly profitable to take the other side of the slow, dumb, negative alpha order against a stale old quote; we get that! In times of stress, however… well we all have seen that movie.

Larry Tabb is an innovator, a great thinker, and a treasure to our industry. He deserves all the credibility he instantly receives on any market structure issue he opines on. But today he is wrong, and the stakes for the marketplace are too large to not let his opinion, shared by all the big banks, go unchallenged. This debate is much larger than the business models of Themis Trading, Tabb Group, or any of the large banks and internalizers; it needs to be had sans concern for anything except what is best for the marketplace as a whole – in good times and in bad. We have data already on how PFOF can make stressed markets horrific (May 6th). Perhaps the time is right for a pilot program which removes all PFOF, so that we can then have the needed data for our regulators to make a truly informed decision on what’s best for the marketplace as a whole. We believe Senator Levin, other lawmakers, and our regulators are quite wise for taking a fresh look at what payment for order flow has done to our marketplace.