Transaction Costs and Sound Bites


Modern Markets, the HFT lobbying group, is sending out emails and blogs again; they want you to listen to what William Baxter, Fidelity’s Head of Program Trading, said back in August at SIFMA’s 2014 Market Structure Conference. Mr. Baxter contends that markets today are more efficient, cheaper to trade, faster, and that the buyside has much more choice (algos and venues). He then proceeds to quote some ITG transaction cost data in large cap stocks from his program trading peer group:


We agree with Mr. Baxter. Current large-cap market structure is favorable for:


a) Retail trading large cap stocks.

b) Program trading desks that tend to rebalance baskets composed of smaller orders in large cap stocks.


What Mr. Baxter states is fairly obvious, and acknowledged for years. Large-cap bid-ask spreads have declined dramatically from the 1990s till present – with nearly the entire decline in the bid-ask spread having taken place prior to the implementation of Reg NMS.

We would like to add some clarifications we feel are important.

a) The decline in large cap bid-ask spreads, and large cap impact costs for smaller orders, and efficiency of buyside trading desks indeed have improved. This is because of automation generally, decimalization, Reg ATS, the order handling rules, the end of Rule 390, and then enabled by all that competition introduced by ECNs, ATSs, and HFTs.

b) The vast majority of those improvements have occurred between 2004 and 2006 – a period where, according to Tabb Group, HFT represented about 20% of the volume.


It is nearly 9 years since the lion’s share of these improvements have occurred. We would like you to:

a) Think of all the extra risks, order type proliferation, fragmentation, quote traffic, order routing abuse, and bad HFT-acting that has come post 2006.

b) Weigh in your minds if the marginal costs, including the costs you now incur policing your order flow, and the costs of increased attention to market structure, offset the de minimus improvements in transaction costs since 2006.

We would like to close this morning note by pointing out that Professors Angel, Harris, and Spratt have updated their Equity Trading in the 21st Century study. Glance through the report. You will see that net effective spreads have increased since 2006. You will see that quote to trade ratios have increased dramatically. You will see that substantially more orders trade in the dark. Finally, you will see how they document that trading costs of large orders – 1,000,000 shares of a $30 stock – have not decreased since 2006! This data includes trading across all stocks – not just MSFT and BAC.




So, what’s out point? Frankly the observance of a single point in time spread in the very largest capitalization stocks experienced by folks trading small orders is not representative of the very large segment of the institutional community who invest and trade in large sizes. Those benefits have not increased since 2006. And let’s not even get started on trading costs in mid and small caps…

Let’s all keep the debate honest and be wary of sound-bites.