The REAL Exchange Disruptor Strikes Again!

From its inception last decade, Bats has been a true disruptor, blending innovation, technological competence, and business acumen to deliver a platform that helped drive down the net cost of trading to its clients.  Unlike other firms that claim to be disrupting the exchange model, Bats delivered and continues to do so.  Instead of pointing fingers at “bogeyman” issues like HFT predation, order type opacity and co-location (and doing little about them), Bats, once again, has proposed an innovation that could provide a major cost reduction to the industry.  Their proposal to offer as much as an 80% discount to matched trades in the closing auction should be welcomed.

Why am I so positive about this proposal?  The answer is that it is almost a perfect “win-win” for the industry.  It should have zero impact on price discovery, yet could cut over $40 million in fees paid by brokers to exchanges annually.   In addition, in extreme situations, it positions Bats as a potential backup utility when technology issues disable one of the other primary listing venues.   I will explain how below, but want to provide some historical context first.

When Bats started, the equity markets were at the tail end of a wave of competition followed by consolidation.  There were more than a dozen electronic market centers started in the 1990s that displayed and matched orders like exchanges do today.  Called ECNs (Electronic Communication Networks) they were a form of ATS (Alternative Trading System) that brokers could use to display orders or access resting liquidity.  In the 2000s, however, merger activity threatened to suppress competition.  Instinet bought Island, forming the largest ECN, called INET.  Nasdaq followed this by purchasing BRUT, which had earlier been created out of a merger with Strike technologies.  Within a year, Nasdaq followed up their Brut purchase by buying INET, which they used as their core matching technology.  NYSE, at around the same time, purchased ARCA, which itself was formed from a merger with the Redibook ECN and a purchase of the Pacific Stock Exchange.  After this wave of consolidation, many industry participants were afraid that Nasdaq and NYSE would form a “duopoly” with the ability to raise costs and hurt broker’s profitability.  Into that breach, jumped two sets of entrepreneurs:  Dave Cummings, who formed BATS and received backing from many industry heavyweights and the duo of Bryan Harkins and Bill O’Brien who formed DirectEdge from the skeleton of the Attain ECN with the backing of Goldman, Citadel and Knight.  Both ATS’s were highly innovative, and now are part of the Bats exchange group.  While Direct Edge pioneered the first inverted (taker/maker) market, the original Bats was the most aggressive disruptor.

Dave Cummings treated Nasdaq as the enemy (mostly because it was impossible to compete with the NYSE until Reg NMS broke their stranglehold on the trading business), espousing principles of more straightforward and lower cost pricing.  They originally started operation with a pricing model of (if memory serves) 15 mil[1] rebates and 17 mil access fees for all participants.  This was both a lower “capture” rate (the difference captured on each trade) and, on an absolute basis, lower fees than any other market at the time.   They subsequently modified this to be 24 mil rebate and 26 mil access fee, after feedback from their market makers that the rebate needed to be higher or they would be less able to provide liquidity.  (I mention this for all those readers who, as part of the “maker taker” debate state that there is no evidence to prove that rebates don’t impact liquidity providers.)  Bats’s true disruption, however, was done in January of 2006, when Mr. Cummings lowered the rate to take liquidity to 20 mils and raised their rebate to 30 mils, effectively locking in a loss of 10 cents for every hundred shares traded on Bats. He promised all his clients that he would maintain the rate for the entire month or until Bats traded 5 billion shares (equating to a $5 million loss).   This bold strategy worked, as Bats signed up many new subscribers and grew its volumes to a self-sustaining level.

It is also worth reminiscing about the war of words between Nasdaq and Bats that accompanied this.  Nasdaq issued a “head trader alert” warning its users to “use caution” when trading with a market that used inverted pricing.  Dave Cummings fired back in an email to his subscribers, saying “Sounds like the Nasdaq propaganda machine is in high gear again, spreading fear, uncertainty and doubt,”.  (I could write an entire blog devoted to reminiscing about those emails, but I digress.)  The point of this story is that we have seen true disruption in the exchange space, with a market willing to operate at wafer thin margins, and for a time, at a loss, to build its business.  Bats did not charge for market data or port fees either and maintained that policy, along with their simple, low rates for many years.  Eventually, they changed to a tiered model and imposed market data and port fees, but they are still more reasonable than their two main competitors.  Those that fete IEX, should heed this story and realize that they are already the most expensive exchange to trade with on a marginal basis and have signed no contracts binding them to permanently giving away their market data or exchange access.  In short, they have talked a lot, but done relatively little to earn the moniker of disruptor…  (and the speed-bump, as has previously been described, does NOTHING to protect the displayed orders that should be the focal point of any exchange conversation)

So, what does this have to do with the Bats closing match proposal?

Simple, it is another example where Bats disruptive innovation is positive for the industry.  The NYSE and other critics have charged that this proposal risks damaging price discovery, but that is complete nonsense.  As described, the Bats proposal will only accept market on close orders and not accept any limit orders.  The cutoff time for accepting these orders is set to provide more than enough time for all orders, that are not fully matched, to be routed to the primary exchange, where they will continue to become part of the price discovery process.  This means there will be zero impact on price discovery.

Price discovery is done by the auction process, which chooses the price that maximizes the amount traded, by attempting to match marginal excess demand with offsetting limit orders.   Consider the following scenarios:

  1. The primary exchange received 10 million shares of market on close orders that were fully matched (5 million shares of buy orders with 5 million shares of sell orders) and 500,000 shares of excess market on close demand to buy along with limit on close, imbalance only and regular limit orders including d-quotes.




  1. The primary exchange received ZERO shares of fully matched market on close orders (Bats matched 10 million shares) and the primary exchange received the same 500,000 shares of excess market on close demand to buy along with limit on close, imbalance only and regular limit orders including d-quotes.



If the limit order book was the same in both cases, the auction will pick the same price, regardless of the difference in the matched quantity.   The reason is that the amount of non-priced, market on close shares which pair off, are not involved in the price discovery process.  As long as the participants in the Bats matching process are not precluded from entering their MOC orders into the primary auction in time, there will be no negative effect on price discovery.  (In fact, some have suggested that, in the event there is a large imbalance of MOC orders on Bats at the earlier time, it could even help price discovery by alerting investors earlier of a potential move, thereby attracting offsetting orders to the primary listing exchange’s auction.)


The cost reduction for brokers who trade MOC orders on Nasdaq and NYSE, however, will be very real.  Nasdaq has a tiered cost structure ranging from 8.5 mils to 15 mils while NYSE charges 10 mils to all participants.  Assuming Bats offers this service at a rate of 2 mils, and the “on close” volume on NYSE and NASDAQ listed stocks stays around the current levels of over 400 million shares, that would equate to over $40 million of annual savings if Bats captures 50% of the close flow.   This is calculated by multiplying the difference of roughly 8 mils times one half of the current market on close volume times 250 trading days in the year.

In addition, there is a hidden benefit to this plan in terms of system wide resiliency.  While Bats hasn’t publicly said this, and it would take technology work for them to do so, once they are able to handle MOC orders on NYSE and NASDAQ listed stocks, that would make it much easier for them to offer a full closing auction process as a backup.  One large complexity in establishing a backup closing auction (which is an important current “single point of failure”) is that the sending systems (OMSs, EMSs, and connectivity providers) have edits or hard coded routing that limit “on close” orders to be sent to the primary listing exchange for that stock.  This proposal would require the elimination of that restriction and would likely guarantee that participants would be using Bats every day for “on close” orders.  If most participant’s systems were modified to take advantage of this functionality, it would, therefore, be a major step towards more system-wide redundancy.  Unlike the proposal for the NYSE and Nasdaq to back up each other, the broker community would have both a profit motive to do the work and incentives to ensure that the routing works daily.

The conclusion is that the SEC should approve this proposal, as an example where innovation and competition lowers costs to investors.  Arguments based on the notion that this proposal will impair price discovery are fallacies and are rooted in self-interest.   Additionally, the history lesson should be instructive to those who bandy about terms like “disruptor” when it is not appropriate and make decisions about trading based on such narratives.

[1] “mil” is trader code for cents charged per hundred shares, so 15 mils would be 15 cents on a trade of 100 shares,  $1.50 on a trade of 1000 shares and so on

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s