Skip to main content

An employee sits below a television screen showing stock information at the Tradeworx office in Red Bank, New Jersey November 17, 2009. New figures suggest high frequency trader activity in Canada is on the decline.MIKE SEGAR/Reuters

High-frequency-trader activity in Canada is by one measure on the decline.

The latest numbers from an Investment Industry Regulatory Organization of Canada study of HFT activity suggest there are about 60 trader user IDs in Canada that represent activity that could be HFT. Those user IDs generated 16 per cent of equity volume, 24 per cent of trading value and 35 per cent of all trades in Canada during the period from January 2012 to June 2013. That's down sharply from the 22 per cent of volume, 32 per cent of value and 43 per cent of trades that IIROC measured in a look at the months of August, September and October 2011.

About 1 percentage point of the shift is explained by a refinement of measurement criteria, but the rest of the move backs up anecdotal evidence that HFT is on decline in Canada. (ITG, a brokerage, for example put out this report in 2012 that suggested activity was on the wane).

Another interesting data point from the IIROC study is that most HFT trading appears to be "stabilizing" rather than "destabilizing," a counterpoint to the rather commonly held belief that HFTs feed volatility. On the Toronto Stock Exchange, for example, there were over 45 billion "stabilizing" trades by the user IDs in the study, and fewer than 35 billion "destabilizing" trades. IIROC defined stabilizing buys as those that were on a downtick, supporting the market, and stabilizing sells as those that were on an uptick. Destabilizing is the reverse.

All of this assumes, of course, that IIROC has captured the right traders in its study.

The first study used high order to trade ratios as a proxy for HFT. Users whose IDs showed high levels of orders for every trade were caught. This time, IIROC refined it to capture IDs with high speed changes to orders and a high number of order messages, two more hallmarks of HFT. The result was a smaller subset of user IDs, but a similar amount of activity to the larger set.

Definition of HFT is a very contentious topic, as few people hold the same one. As a panel at the conference discussed at length, there is no accepted and useful common definition of HFT.