By Kelvin To, Founder and President of Data Boiler Technologies
Conscientious faithful words grate upon the ear. In an effort not to be lost in the ocean of comment letters to the SEC, we at Data Boiler want our views to be heard. In this article we use analogies to promote communication and understanding of the complex subject of market structure reform. Please do not block us from shedding light on taboo topics. Constructive debate helps avoid the phenomena of “blind people and an elephant”.
Modernize Rule 605 Disclosure about Order Execution
How do retail consumers buy their cell phones or computers? Geekbench and Antutu are examples of benchmark services to gauge hardware or system specs. Yet, most consumers are not aware of these benchmark tests. Which product has the fastest performing microchip, the most RAM, storage space, are often NOT the most critical factor in consumers’ purchase decisions. No-one likes to pay extra or receive an inferior price. However, the “most favorable price” does not represent a need to call for a heightening of consumer or investor protection law.
Policy makers should NOT prescribe or endorse certain statistical benchmarks. Some benchmarks use the average rather than the median (tail risk). Other shopping comparisons emphasize different aspects, such as customer service. One group may prefer one set of benchmarks that make their brands look better, while the other wants to use a different benchmark for promotion. People hold implicit biases that favor their own in-group. Not all broker-dealer execution quality is equal. An analogy of it is – not all ETFs are created equal. The SEC should preserve its independence. If any constituent including the regulators want to have comprehensive metrics produced, let’s have the vendors compete for their business.
A large High Frequency Trading firm was among key stakeholders in petitioning for this modernization of Rule 605. We are concerned that the more detailed order execution information in the proposal will enhance certain constituents’ ability to model the market. Then they could use sophisticated algorithms to craft out order flow for the elites within a certain echo chamber rather than promoting competition.
The summary report and the more detailed order execution information create a false sense of comfort about order execution practices and quality. It contradicts the standard investor disclaimer – “Past performance is NOT indicative of future results”. BestEx reports may demonstrate ‘where’ an order should be routed, but not ‘when’. Giving away vast amounts of information to free riders increases vulnerabilities. Instead of focusing on the 80% knowns, learn from Universa’s Black Swan strategy – “tail hedge acts like insurance with asymmetric and explosive downside protection”. Discovery of the 1% unknowns with timely warning of irregularities would make the market safer.
Reg. NMS – Tick Size/ Minimum Pricing Increments
The industry seems to have consensus on half a penny instead of the SEC’s proposed tenth of a penny. The SEC is forcing almost all stock to trade at least 4-8 ticks wide. An attempt to finesse the way queue priority is operated and/or prohibit “gaming” or “step in front of” of the queues. Artificially altering the queue (equal waiting line at all checkout counters, except leaving much room for the Exchanges to selectively use tier rebates and other perks to divide the cake with the elites in hurting the other “content” creators) may affect the “apparent”, NOT the real supply and demand for securities.
Liquidity vanishes when pricing increments are too small, that hurts the institutional market. Do not dismiss the voice of issuers about minimum tick size. Although the Commission views the wholesalers (Market Makers internalizers) as the middleman, intermediary costs in reality encompass: connections with more trade venues (including auction markets), increase market data subscription fees, demonstrating BestEx, the inelastic demand of Transaction Cost Analyzer, the reliance on liquidity sourcing, execution services and other ‘tools’, for example, the use of split, cancel, derivatives and other means to adjust to the more convoluted market environment.
Going beyond the 1,337 NMS stocks that are tick-constrained would substantially increase quote volume but would not necessarily translate into matched trades. The quote fading problem would exacerbate, it has ever since IEX goes all-in on Data Revenues, Quote Fade and (Virtual) Rebates. Technology advancements do NOT reduce “flickering” concerns because Tech in itself is neutral. Average investors would have a harder time navigating the market. Asset maximizers (‘farmers’: fund industry, retirement, insurance sectors) would have challenging time achieving economy of scale without losing out to the ‘hunters’ amid the NMS tectonic shift. All lead to a further consolidation of asset managers.
Do not use “price improvement” interchangeably with the “willing seller willing buyer standard” per 75 FR 3597. Over emphasizing “most favorable price” undermines the other aspects of the 4Vs (Volume, Variety, Velocity, and Veracity). Rather than attempt to price control or inadvertently calibrate the wrong prescriptions, the SEC should consider principle-based rules, like the music industry’s “Four-Part Test”. In addition, protected quote and last sale should be based on a constant dollar amount.
Reg. NMS – Access Fees and Rebates Incentives
The focus of “Hu-Murphy Paper” is off. Access fee rebate, Payment for Order Flow (PFOF), and market data/ market structure issues are all intertwined. Schwab’s empirical evidence proved that “Order routing revenue and price improvement are NOT zero-sum”. The noumenon of rebate incentives serves as royalty payments for the use of others’ copyrighted material. The prevailing market problem is “who owns the data”. The SEC failed to grasp the correct context, which causes bias with the regression analysis on how the treatment group is chosen or omitted. Policy makers should consider Market Makers (MMs), Alternative Trading Systems (ATSs), Single Dealer Platforms, and Exchanges as different “streaming platforms” (Viacom CBS, Disney+, Showtime, Cable-TVs) in order to have the right focus.
We agree that access fee rebate and PFOF incentives ought to be recalibrated. However, the recalibration objective should NOT be about trimming billions of incentives to go around in the market. Where can broker-dealers squeeze their margin to cover substantial income loss, plus other increased costs? The basic forces of supply and demand suggest that income reduction would cause a decrease in demand. Recalibration should be about the fairness of “who gets what”.
Incentives affect quote quality (see Table 5 of NASDAQ article about widening spread and reducing depth). Block trading activities would be reduced when there are insufficient incentives to go around or if the NBBO is narrowly defined. The proposal failed to account for the effects of cross-subsidization in market making or implications for thinly traded securities. The proposed fees and rebates determinable at the Time of Execution may push trading venues to be more defensive. The proposal does not curb price discrimination or the use of other perks to skew the market. Hence, the SEC suggested 80.3% reduction in access fee rebates, plus the expected cut in PFOF, do NOT address the “who gets what” fairness issue.
Variety helps reach a wider audience and avert the continuous down trend in the number of FINRA registered firms. Making all “streamers” the same undermines the different players’ roles and their frienemy dynamics in fabricating the fragmented market under Reg. NMS. Healthy markets need both farmers (assets maximizers) and hunters (performance optimizers). A consistent Copyright Licensing Mechanism would align and addresses the economic viability of a constituent to exploit its economy of scope and/or economy of scale, and harmonize different market centers.
Reg. NMS – Accelerate Odd-lots/ Round-lot, Indicator of Better Priced Orders
603(b) of Market Data Infrastructure Rule (MDIR) is problematic. Ingesting a lot of data raises the concern about bandwidth connectivity. What you see may not be what you get. Lacking depth-of-book data undermines the usefulness of Odd Lot data. The SEC letting NYSE launch 100G colocation service in April 2020 when MDIR is based on 10G connectivity, is example of how the Commission exacerbates the gap between the ‘haves’ and ‘have-not’.
The SEC should NOT be taking the stand to protect the exchanges’ “agency market business model”. We argue that forprofit exchanges are indeed operating a “Jukebox model” to extract rent, hurting all, especially the smaller players. Page 193 of the proposal acknowledged that “the proposal could increase the demand to purchase depth of book data … result in more market participants purchasing data from exchange depth of book proprietary data feeds than do currently.”
Best available price should include access fees and other incentive components. We agree with Prof. James Angel that “There is no reason that brokers should not be allowed to take fees into account and route to the best available all-in price.” Proposed rule 600(b) requires identifying of the best odd-lot orders. An odd-lot NBBO creates ambiguity. The SEC would get lambasted, while free riding speculators ask for evermore “indicators” that they amplify herd behaviors by using the echo chamber to disrupt the US market. Pass through rebates and/or absorbing market centers’ fees are a matter of standard that we again advocate for a Copyright Licensing Mechanism.
The industry is craving for evermore data and the race to have the fastest speed have caused infrastructure costs to increase exponentially for both data suppliers and buyers. We doubt these data vault, cloud initiatives, or upgrade from 40G to 100G, 400G, 800G, or even 1.6T speed would truly deliver any economic benefit. Our trading community has become subservient to these infrastructure providers over time. We believe using time-lock encryption for a secured and synchronized starting line in data distribution would promote fair, reasonable and non-discriminative (FRAND) principles and save the industry from unnecessary wastage of infrastructure investment.
Order Competition (Retail Auction)
The Commission has acknowledged that there are additional complexity and connectivity costs to market participants arising from the introduction of qualified auctions. The proposal contradicts with 79 FR 5592 footnote 711 (Volcker Rule), it said “The Agencies are not adopting a ‘transaction-by-transaction’ approach because the Agencies are concerned that such an approach would be unduly burdensome or impractical and inconsistent …”
Auctions may not have a price guarantee, and price improvement on auction may only occur briefly for few orders being filled. The frustration is analogy to deeply discounted items on ‘Black Friday Sales’. The market is constantly picking up signals. A failed auction is an analogy to the collapse of SVB after its failure to raise funds in the lit market. The cost and adverse consequences of mandating qualified auctions for marketable retail orders is quantified in this empirical research.
Without sufficient diversity or variety of participants in the market, price discovery or veracity leaving up to the auction process may indeed intensify “gamification”. The auction market is usually crowded with Specialists and Professional players. Is retail the “Dim Sum” for many predators?
The concept of using a discrete-time (periodic batch) auction to replace continuous–time limit order books (CoB) to deemphasize speed as a key to trading success is NOT a novel idea. According to this 2014 CFA post, it points out that “pre-trade transparent constantly updating price in real-time, providing true price discovery, which complicate the auction process … Budish agreeing that latency arbitrage could continue on parallel continuous market exchanges. In this case, the proposal is similar to IEX’s speed bump, which is designed to protect traders only on the IEX platform.” No control over interaction with derivative trading that is under the CFTC’s jurisdiction is the fatal flaw.
Auction only works in an “all or nothing” mode, i.e. mutually exclusive with CoB. We did try to consider if the SEC and CFTC may consider joining forces to set, for example halt or close the market at 11:X am Eastern Time to host an auction at 11:X+. The halt or close of the CoB trading will add to a higher chance to fill meaningful order volume. The auction at 11:X+ is like the opening auction, 9:30am, the CoB continues and allowing the Exchange to not reinvent the wheel. However, unfilled orders by auctions would be inexplicable regardless of the type and number of auctions. Policy favoring auction markets is unfair in the eyes of those who wanted immediate liquidity from CoB.
Best Execution Policies and Procedures + Annual Reports
It is not economically viable for every broker-dealer to have dynamic price shopping capabilities like Booking.com or Ticketmaster. Yet, trade strategies and risk models all need to be upgraded if these proposals are adopted. In addition, all players would need significant changes to their trade reports. The SEC underestimated the costs to comply by a big margin.
Under proposed Rule 1100, “the term ‘market’… encompass the wide range of mechanisms … this description of ‘market’ is expansive …” We believe the Commission may have the Proposed Investor Protections in Communication Protocol Systems and ATSs and Amendments to Exchange Act Rule 3b-16 Regarding the Definition of “Exchange” in mind (see our 2022 comment letters to the SEC, ESMA, and FCA). MiFID-II calls for ‘sufficient steps’ to ensure the favorable execution of client orders, while the US requires ‘reasonable steps’ which is more bureaucratic and subjective. The ebb and flow would cause the US market to lose a competitive edge.
The Commission did acknowledge on pages 190 and 335 that “increased costs could be passed through to customers in the form of higher commissions or reduced services [and] could … result in higher barriers to entry and potential exit of small broker-dealers.” Costs and benefits are not justified. The SocGen $7.2 Billion loss in 2008 should serve as a reminder for rubberstamping on annual reports and the uselessness of these policies and procedures.
It is humanly impossible for the SEC examiners to objectively decipher if there may or may not be BestEx violation or conflicted transactions (even if given the CAT system). Timestamp tolerance at 50± millisecond, lack assessment of MMs risk profile, not consider RENTD, no automated checks of “flipping-a-switch” between clients versus counterparties, etc. would cause the best talents to not able to determine “where” and “when” in indicating the “most favorable price”. Officials craving for evermore power may lead to unethical behaviors in exploiting the governed.
Other Remarks and Conclusions
We believe “What Gets Paid and Who Gets What” must be based on clear delineation of rights and obligations. Which type of trading venue should have what capabilities in order to maximize overall reach and efficiency of global capital markets, it depends on three things:
(1) How much an interactive or non-interactive streaming platform is willing to carry, i.e. a broader or narrower “catalog”;
(2) Nimbleness to maneuver around in crafting niches by rewarding “content” creators to reduce unknowns; and
(3) Ability to grow the number diversified market participants and growing the overall pie.
The SEC’s 1,656 pages of complicated proposals if adopted would create a vicious cycle when echo chamber generates more bifurcated behaviors. Our counter suggestions would achieve Pareto improvement (someone better off without anybody worse off or win-win) for market efficiency gain.
