Hold on… this doesn’t make sense - Frequent Batch Auctions, Part I

These comments form an informal response to the recent publications FCA Insight “Big bucks from small change,” and Occasional Paper No. 50 Quantifying the High Frequency Trading ‘Arms Race’: A new methodology and estimates (Aquilina, O’Neill and Budish, Jan 2020). They will also reference “The High-Frequency Trading Arms Race: Frequent Batch Auctions as a Market Design Response * (Budish, Cramton, Shim, 2015), and “A Theory of Stock Exchange Competition and Innovation: Will the Market Fix the Market?” (Budish, Lee and Shim, Dec 2019) as the studies share some degree of authorship, theme and motivation. Collectively, I’ll refer to them as the “FBA Papers.”

In these papers, it is alleged that Central Limit Order Books (“CLOBs”) lead to a $5b “latency arbitrage tax,” increased costs to investors, destroyed social welfare, among other things. These papers also suggest that matching buyers and sellers of financial instruments solely via "lighthouse-style", discrete-time, Frequent Batch Auctions (“FBAs”) would be a viable substitute to the status quo. 

These suggestions are dangerously flawed, and over a few posts, I will explain why. The core of the arguments against existing market structure is summarized by the FBA Papers’ authors below, quoted from their January 2020 FCA Occasional Paper. It draws sweeping conclusions related to the costs imposed by the continuous matching market structure of CLOBs:

We find that the “latency arbitrage tax”, defined as the ratio of daily race profits to daily trading volume, is 0.42 basis points (0.0042%). If this still sounds like small beer, then consider it in real terms across the whole market. The annual sums at stake in latency arbitrage races in the UK equity market are about ￡60 million a year. Extrapolating globally, we find that the annual sums at stake in these races across world equity markets are about $5 billion per year.

-Aquilina, O’Neill and Budish, 2020

?? Doh!! Where to begin?! Moe, please pass me a beer. A large beer. 

(Note: ESMA’s definition of a Frequent Batch Auction differs from that referred to in the FBA Papers. ESMA’s definition is more conventionally referred to as an unscheduled “Periodic Auction” by practitioners. In this series, I’ll use the FBA Papers’ definition, which does not integrate Periodic Auctions alongside Continuous matching, and proposes it as a standalone market structure that specifically excludes continuous matching).

WHY IT’S IMPORTANT TO PUSH BACK ON THIS

At the moment, I reside thousands of miles from the nearest market center. I have no “axe” in this argument, but I do care passionately about market structure, fairness and working together constructively to improve markets. I care about informed debate.

For those reasons, I am compelled to urge caution in jumping to conclusions from the aforementioned articles due to their shortcomings in:

1. assumptions,
2. methodology, and
3. conclusions 

In doing so, I wish to raise awareness of conclusions that should not be drawn from these works, particularly as Europe resets on both sides of the Channel post-Brexit with a potential MiFID 2.1 or 3.  Attention to those drafting new rules!  

Apologies for some “pour” beer puns. But they started it!

PRACTITIONERS POINT OF VIEW

Like the FBA Papers’ authors, I believe the terms of today’s speed-based arms race is wasteful in many ways, and risks homogenizing liquidity provision. It motivates for-profit exchanges to discriminate service offerings with ever atomizing speed granularity, which isn’t the core function of what the exchanges are there to do. I say this as a person who has both worked for an exchange, and signed the contract to join the original Spread Networks “Flash Boys” fiber optic line in the USA while working for a high-speed trading firm.  

In 2010, reluctantly, five or so firms joined Spread on Day 1, and enjoyed a 100x increase in service costs. In return we would not increase our profitability, but we could remain in the game. We went live and we knew the service would soon be obsolesced by microwave networks.  

A few years later, the decline of the London-based operations of the firm where I worked was directly attributable to not deploying a pan-European microwave network. As we lost our ability to find the signals in the noise, I lobbied exchanges to work together to provide more democratized microwave access between key data centers and trading venues. I argued, at the time, that by ensuring affordable pricing and availability, they would be able to retain a greater number of diverse trading participants. 

A few years later, I joined a major European exchange as the Head of Equities. That Exchange boldly chose to provide exactly those microwave services I’d been asking for (before I joined, and sadly after the firm I previously ran had retreated from Europe). As a client of the Exchange, I recognized that they worked tirelessly to maintain a vibrant ecosystem for Liquidity Providers (LPs), sell-side market access providers, stat arb hedge funds, and most impressively global and local wealth management firms of all sizes. Now on the other side of the table, I enjoyed working more directly with the broad trading community. Together with our clients, we developed several innovative changes to our order books, including new forms of price-size-time matching, hybrid and segregated order books, all targeted to facilitate the needs of less technically equipped local investors and sophisticated global liquidity providers. We specifically contemplated “stale-quote sniping” risks discussed within the FBA Papers, and constructed market structure to resolve the issue.

It must be said, however, never were FBAs proposed by our clients as a substitute or supplement to the CLOB or our other order books.

Encouragingly, from my new perch at the exchange, I was able to see that the market was moving on from speed. Many emergent liquidity providers famously eschew high speed networks (look around, you’ll find them). Presumably, in lieu of speed-sensitivity to read and react to where the market is, they use incredible computing power to accurately anticipate where fair market value will migrate across diverse instruments before it happens. Contrary to the FBA Papers, it is clear to me that speed alone does not declare the winners and losers in liquidity provision. Not anymore. But speed remains a very important tool for some.

This evolution towards short-term predictive market making has accelerated subsequent to the recent FCA Insight papers’ nine-week data sample from late 2015. So too has market behavior evolved after the new minimum tick increments mandated via MiFID 2 RTS 1. We should be cautious about drawing conclusions about the UK market, much less for the entire world from these narrowly-scoped and outdated data sets. More anon.

IF THERE ARE PROBLEMS WITH CLOBS, THEY CAN BE SOLVED IN OTHER WAYS

From a practitioners point of view, it's not clear that the alleged problems related to latency arbitrage warrant the radical proposals that might destroy previously accumulated market efficiency gains, as well as jeopardize the promise of future innovation from competitive market models. 

Market-Makers/LPs are diversifying their business models away from speed

As mentioned above, many LPs are moving beyond speed to assist in their market making and mitigate risks related to market structure and competitive pressures. The use of short-term predictive market making strategies, versus strictly reactive mechanical strategies, is thwarting the impact of sniping (not in all cases, but to a material proportion). 

LPs offer bi-lateral trading as Systematic Internalisers (SI’s”) in Europe. The bi-lateral nature of SI’s provide a layer of protection for all participants, including the end investor and the liquidity provider. Should a counterparty “snipe” at stale quotes, or only decide to trade at times that might be disadvantageous to the LP, the LP is able to adjust their pricing to that counterparty. Similarly, liquidity seeking counterparties (buy-side or their agents) can test to determine the performance of their specific bi-lateral counterparties, provide direct feedback or take their business elsewhere.

Finally, LPs also now deploy an increased breadth of trading and business strategies, helping diversify the impact of market making losses. These include wholesale diversification (e.g. becoming agency brokerages), licensing algorithms, deploying liquidity taking strategies based upon statistical arbitrage techniques, and migration to expanded asset classes. Many LPs provide two-way markets without much profit, but due to their liquidity-providing volume receive tier-discounts at exchanges and clearing partners that benefit these other strategies.

Execution platforms offer different forms of market microstructure

Globally, including Europe, market centers are collaborating with their members to establish creative execution destinations and techniques to meet the varied needs of participants, some of which are specifically architected to protect participants from latency arbitrage. These market structures include periodic auctions, speed bumps, dark trading with mid-point matching. In Europe, these also include Cboe’s Equity BlockTrading, several venues’ various flavours of periodic auctions, swim lanes bifurcating LPs and natural liquidity at Aquis and SIX Swiss Exchange’s volume-priority EBBO, RFQ’s on Instinet Blockmatch, and Equiduct's volume-weighted VBBO market structure. As mentioned above, SIs are a valuable source of liquidity that fall under the MiFID II regulatory regime.

De-escalating the arms race via regulation

While liquidity provision does become more costly as a result of the arms race, these costs are not passed along to the end investor. Liquidity providers who rely upon speed must pay high fixed costs, mostly in the form of low latency infrastructure, in order to maintain competitiveness and mitigate risks associated with being top-of-book. These are the costs of doing business as an LP. As Budish, Lee and Shim address in their December 2019 paper, exchanges are in potential conflict with their member-customers on this issue (they are not the sole extractors of technology rent, who also include third party telecoms providers). There are many potential ways to address this conflict issue, but a wholesale move to FBAs is not likely the answer, and most certainly not the only one.

While I’m not an advocate of prescriptive regulation, there are merits to considering the degree to which exchanges should profit from “featurization” of speed-related service offerings. As ESMA considers how to ensure the price of market data is offered on “a reasonable commercial basis”, so too should they consider applying this principle to other speed-related to service offerings such as co-location, order entry capacity and non-display data usage. This can be done under market regulatory or anti-competitive grounds, if required.

If escalating non-discretionary fixed costs stemming from this conflict of interest were to be considered material enough to require regulatory deterrents, a healthy by-product would be trading venues’ and regulators’ re-focus on bonafide market structure and micro-structure innovation. As long as Europe retains its historical preference for predominantly principles-based regulation (notwithstanding much of MiFID II’s relatively prescriptive approach), this would be welcomed by most within the industry.

FBAS’ IMPACT ON MARKET STABILITY: “AN IMPORTANT FUTURE RESEARCH TOPIC”

The FBA Papers’ authors themselves recognized the “important” need for further examination of their proposal’s impact on market stability. Yet, in the five and a half years since the FBA proposal first blurred the market structure conversation, the most recent FBA Papers do little to further inform the debate; however, they do seriously risk distracting it with inflammatory headlines derived from:

1. Flawed hypotheses, assumptions and methodology
2. Erred, incomplete and academically negligent conclusions

I intend to touch on these points. After doing so, we can reframe what’s “small beer”, and where an “untapped keg” sits lying in wait.

There’s a lot to unpack here, so it’ll probably take a few posts. This isn’t an academic paper, and I’m not an academic. Nor am I a blogger. My comments are rooted in experience leading market structure, commercial and regulatory affairs at two global electronic liquidity providers, a major European bank and a major European exchange. I’ve seen more sides of the trade than most. As a practitioner, I haven’t had the luxury to make oversimplified assumptions, nor recommendations for single point solutions in isolation of wider context, particularly for problems of elusive empirical materiality. I’ve lived in the real world, and I have innovated alongside truly objective peers, counterparties, competitors and partners across the industry. We haven’t always agreed over every detail; however, we all recognize we need it to work for all sides of the trade in order to have healthy, enduring markets.

In sum, it is my belief that intraday auctions, including FBAs, could form part of a portfolio of potential market microstructure offerings. That is for market participants to determine, with oversight of objective regulation. However, it is equally clear that the authors, in the five and a half years of promoting the hypothetical FBA idea, have yet to attempt to prove (and certainly not actually prove), how it could be “the solution” in the world we live in, with its highly competitive, globally distributed and interconnected markets.  

That’s a good stopping point, keeping this less filling and still tasting great. More later on the issues related with the FBA Paper’s assumptions and methodology, as well as the potential effects of an FBA/anti-CLOB future.

Cheers! Let’s raise our glasses to People and Technology!

Read the entire series:

• Hold on… this doesn’t make sense - Frequent Batch Auctions, Part I
• The $5 billion heist that doesn’t exist - Frequent Batch Auctions, Part II
• Trade offs and potential impact on market stability - Frequent Batch Auctions, Part III
• Enough to matter: UK workplace pension-savers overcharged, let’s put FBAs into perspective, Part IV

About the author:

Eric Dorre is a management consultant based near Seattle, WA. His career has focused on pragmatic, data-driven problem solving and operations. Since beginning his career as a consultant at Oliver Wyman and Diamond Technology Partners (now PWC), Eric has worked alongside some of the world's most accomplished data scientists, inventors and developers in research and development (Applied Minds, Inc) and quantitative trading firms (Madison Tyler now Virtu Financial, RGM Advisors, now DRW), for whom he led global equities trading from Los Angeles and London. He led Deutsche Bank Global Markets Fixed Income MiFID II Market Structure and Regulatory Change, and served as Head of Equities for the Swiss Stock Exchange. Eric earned his undergraduate degree in Geography from Dartmouth College, and MBA from the Anderson School of Management at UCLA with a focus on Decision Sciences.

Contact:

[email protected]

www.highburyassociates.com