Citadel: Tick size pilot risks trading secrets

Regulators have failed to calm market maker Citadel Securities’ concerns that a program to test the impact of wider stock-pricing steps will expose confidential aspects of broker-dealers’ trading strategies. 

The Financial Industry Regulatory Authority (Finra) and several other self-regulatory organizations (SROs) on October 3, 2016 launched a long-awaited two-year pilot program to widen the minimum quoting and trading increment – or tick size – for some small-capitalization stocks.

Under the pilot, the tick size is widened from a penny ($0.01) to a nickel ($0.05) for specified securities listed on national securities exchanges. For some pilot securities, only quoting will need to occur in $0.05 increments, while for others, the nickel increment will apply to both quoting and trading. The data collected from the pilot will be used by the SEC, national securities exchanges and Finra to assess whether wider tick sizes enhance the market quality of these stocks – by providing less volatility and greater liquidity, for example – for the benefit of issuers and investors.

The information will be made publicly available (see Data, below). Finra originally planned to do so on a monthly basis, but industry members complained that firms were at risk of having their trading strategies uncovered by competitors. In response, Finra has proposed pushing back publication of the data to within 120 calendar days after month-end.

In a recent comment letter, Citadel Securities expresses its appreciation for Finra having acknowledged the importance of maintaining confidentiality when publishing the tick size pilot program data. But Adam Cooper, the firm’s chief legal officer, says the change is insufficient to address the confidentiality concerns raised by market participants. Specifically, Citadel remains worried that, even if the data is aged by up to 120 calendar days before being published on Finra’s website, market participants will still be able to easily determine the identity of broker-dealer trading centers.

This would be in violation of the tick size pilot plan, which explicitly provides that the ‘data made publicly available will not identify the trading center that generated the data,’ Cooper says. ‘Based on conversations with Finra and other market participants, we understand that Finra is intending to publish the data in a disaggregated format using only anonymized ‘dummy’ trading center identifications to mask the identity of a particular trading center. We further understand that the same dummy ID would be used for each trading center throughout the duration of the tick size pilot. In our view, publishing the data in a disaggregated format with a static dummy ID will allow for the identification of broker-dealer trading centers with relative ease.’

For example, he writes, a market participant could direct an order in a particular tick size pilot symbol to a specific trading center, and then use the execution data to find that specific order in the published data based on time and execution price.

‘Furthermore, there does not appear to be a public interest justification for the use of static dummy IDs as opposed to other publication formats that would better protect anonymity,’ Cooper continues. ‘It is incumbent upon [Finra] to ensure trading centers will in fact remain anonymous in the published data, consistent with the tick size pilot plan requirements.’

If the identity of a specific broker-dealer trading center can be determined from the published data, a significant amount of insight may be gained regarding the proprietary trading strategies of that trading center, according to Citadel. Cooper says any market participant, including competitors, could use the data to obtain information regarding the trading center’s:

• Relative aggressiveness or passiveness in specific symbols
• Use of hidden orders
• Fill rates
• Routing practices.

Citadel recommends that, to ensure the published data doesn’t reveal the identity of a trading center, it will be published only in aggregated form. This could be done by separately publishing aggregate data relating to exchange trading centers and aggregate data relating to off-exchange trading centers, Cooper writes.

‘Aggregation in this manner would continue to provide a useful dataset to the public in order to evaluate the effect of wider trading increments, while substantially reducing concerns around the ability to reverse-engineer the trading strategies of a broker-dealer trading center,’ he adds. ‘In order to facilitate robust academic study of the tick size pilot, it may be possible to provide academics with access to disaggregated data, with an anonymized identification for each trading center, provided [Finra] takes steps to categorically ensure the data is used only for such limited purposes.’


DATA
Under the tick size pilot, broker-dealers must, where relevant, supply certain data, including:

• Daily market-quality statistics categorized by security, order type, original order size, hidden status and coverage
• Market and marketable-limit order data such as exchange code or trading center identifier, ticker symbol, date, time of order receipt and order type
• Daily market-maker registration statistics categorized by security, including ticker symbol, SRO and number of registered market-makers
• Daily market-maker profitability statistics categorized by security, such as total number of shares of orders executed by the market-maker and raw market-maker realized trading profits
• Total daily market-maker profitability statistics categorized by security, including total raw market-maker realized trading profits.