“Best Ex” is All About PROCESS
Bloomberg’s Gary Stone makes a simple, but important point about best execution: if “best execution” reporting isn’t part of the decision-making process driving routing and trading workflows, it is useless. As Gary points out, this has not been missed by regulators either, since the FCA made this precise point in their “Thematic Review.”
From a buy side perspective, if a pre-trade cost estimate is embedded in the process for the determination of order and position sizes, firms should periodically review how their trading desk, and the brokers/strategies they use, perform versus those expectations. (And, if your firm does not use pre-trade cost estimates to determine position sizes, implement such a process!)
From a sell side perspective, this means that every firms’ monthly “best execution” committee should incorporate the same type of analysis. As a practical matter, this means both routing strategies as well as algorithms should be compared to pre-trade estimates on a statistical basis.
One important point is that pre-trade estimates are notoriously “noisy,” meaning that comparing individual trades to those estimates can create a lot of confusion. Therefore, the best way to use pre-trade estimates is as part of a contextual, statistical analysis for longer time periods. (As part of the daily process, trades that are deemed “outliers” vs various benchmarks should be reviewed, but that is separate from what is being described here.) If any readers have confusion on this topic or need help with the creation of such a process, ViableMkts is available to help or you can contact me directly.
More Mistaken Reporting about IEX
(Editor’s note – I was going to publish this yesterday, but wanted to give the author of this piece a chance to print a correction. Sadly, as of now, he has not responded to my request.)
Morningstar published a confused mash up of the Yale piece, Flash Boys and IEX, yet does get a few things right. He correctly sees that there is a difference between a dark pool, and an exchange with displayed orders, and does understand that IEX mixes half-truths into their narrative.
Unfortunately, the author also gets much wrong. IEX, contrary to his claims, doesn’t ban HFT. With regard to their displayed orders, they don’t even discourage HFT. As I have previously explained, the speed bump does not change order priority, meaning the faster traders still trade first. In addition, their venue allows HFT firms to gain a connectivity advantage within the data-center that houses their point of presence, and potentially, due to the speed-up of the SIP, provides information advantages to HFT firms co-located in Carteret.
The most important error in the article, however, comes from his dismissal of yours truly as an “interested party” equating me to IEX and their cronies. That claim is 100% false as I do not run a competing business, but rather am a consultant that could market my services to IEX or any other exchange, market center, broker dealer, or investment manager. I have spent the last few years focused on impartial analysis of execution quality, market structure, and the trading process. IEX, on the other hand, has direct self interest in all of its rhetoric and marketing pieces. As for Professor Swensen, not only is his Yale endowment invested in IEX, but his department at Yale has 2 members on IEX’s board, and this was not the first time that he has promoted IEX explicitly in the press.
The author also makes a statement that was probably fair when they were starting up, but is no longer true: “IEX, which is indeed attempting to run a clean business, in a less-than-clean industry…” Not only is IEX now a profitable, well financed company, with a famous, high-profile CEO, firms that “run clean businesses” are quite factual and limited in their marketing materials, qualify or put disclaimers on their broader conclusions, and avoid inflammatory rhetoric when making public statements, none of which describes IEX. In addition, firms attempting to run a clean business to not act to suppress dissent, which they have done on many occasions. In addition to the bullying I have previously documented against me, they have done similar things to at least 5 other people that questioned their value proposition, who have told me about this behavior on condition of anonymity. They have also told at least 2 conference organizers that they won’t participate on panels with me, have refused my challenge for a public debate, and even blocked me on Twitter. All of this behavior is a result of factual analysis showing that their marketing materials are misleading, and in their most recent case, self-contradictory.
Traders DO Understand Information Leakage, so WHY is it hard to get order data to evaluate?
ITG published a piece in Traders Magazine, picked up in TabbForum, that proves that the buy side trading community understands the importance of information leakage. Most important, they believe that schedule based algorithms such as VWAP are the largest cause of information leakage. This proves one of the points I made in my most recent feature article: “If Every Trader Gets a Trophy” is true: Despite being the most common algorithm and most common benchmark, traders know that VWAP strategies leak information.
It is interesting, however, that information leakage is cited by over 60% of the same traders as not being the most important contributor to their trading costs, while a healthy percentage do. This is likely because most buy side firms have no way to measure the opportunity cost associated with the unfilled orders routed by brokers on their behalf. (If information leakage occurs, it shows up as opportunity cost, which I am defining, in this context, as the cost of not getting filled on orders being routed or while a trading strategy chooses to wait instead of trading.) Almost all TCA is based exclusively on executed orders, so brokers that route thousands of orders to pools that provide low fill rates while markets move, or that sit on the bid or offer and chase the market, do not have such behavior measured explicitly. Thus, while information leakage is clearly part of the issue, without analyzing the order routing that causes the leakage, buy side traders are left guessing…