Canadian Alternatives: Buy-side Case Study


Michael Thom, Equities Trader, Genus Capital Management offers a look into the Canadian equities world, including perspectives on dark pools as well as algo implementation and usage.

Inverted pricing models
We have just seen the introduction of more innovative pricing models in Canada, essentially since the launch of TMX Select. For most buy-side participants like me, we do not see our tick fees as rebates because they are bundled into the commissions we pay to our brokers. This is an exciting development for participants that thrive on different market structures, but I would not say that we particularly benefit from this market model. From an intellectual perspective, it is interesting to wonder what will happen as a result of these developments, but I would not say it has any immediate net benefit to us or our clients.
Trends for Dark Pools in Canada
Canadian regulators have taken the right approach. There are lessons to be learned from other jurisdictions where dark liquidity was left to develop and regulators then had to play catch up. I applaud the Canadian regulators for giving their approach to dark liquidity critical thought before it gets to the point of significantly damaging market quality. Regulators in Canada are at a point now where if they change the regulations significantly, venues and firms would be able to adjust. The debate over the trade-at rule in the US shows that whole business models are built around sub-penny pricing and trading not at the touch. I do not think that is where we want to go in Canada.
I am a little cautious around some of the regulators’ specific proposals on minimum size. I am more in favor of the minimum increment being set at a half penny. The minimum size is the more difficult concept because anything that functions around a single pivot size, either in value or number of shares, can disseminate information through trading around that pivot point.
Although to my knowledge very few participants choose to structure their orders in such a way, it should be up to market participants to build into their orders the minimum execution quantities for dark pools as they see fit. I do not think a lot of buy-side participants are currently building their orders or customizing their third party algorithms to that level of detail. From where I sit, it is not a perfect solution, but this compromise might be the best of the difficult alternatives.
It is important to point out that they are not putting in a minimum size right away. The architecture is built to allow the regulator to, on very short notice or if they start to see some compelling data points, put limits in place without going through the full comment and review process, which is all very prudent. They are giving themselves the tools to deal with all possible market outcomes. Flexibility does not come easily to regulators. Typically, they adopt very specific proposals and if those proposals fail, it is back to square one, whereas here they have given themselves a degree of latitude which is commendable.
Simplifying Algo Implementation The algo and DMA providers who are winning our business are those who can give us transparency right down to how they are interacting with each individual venue, what order types they are using and how they are implementing venue specific idiosyncrasies. If a venue has very unique order types, our providers should say how they are using those and why they made the decision to use the order types they did. Providing a transparent, empirical basis for decisions regarding algo structure, architecture, order types and routing is really important. Many decisions go into building quality algorithms and routing, and those who will share the data behind it are my providers of choice. Algo providers seem to now be more willing to tailor and be empirical about constantly improving the product to fit a firm’s or a trader’s trading styles. That is where algorithmic trading is headed, as it relates to buy-side, and we are just starting to see the leading edge of that in Canada.
The level of attention to the customization process differentiates a very small number from the bulk of providers. Canada has had a very interesting development in terms of its algorithm trading scene. You had the first pioneers who are the incumbents and now you are seeing the second and third generation players come in and really taking it to the next level. There has been an increase in third party algo usage or white labeling in Canada, but I think those who have made the investment to get a decent product off the ground are being rewarded for it. 
The average Canadian buy-side trader’s ability to differentiate between a good algo provider and an average one is still developing. While there are a lot of commoditized algo products, it is when you start to tailor it to your execution and liquidity gathering philosophy that you begin to see the differentiation. For the most part, any algorithm will get your trade done, it is just a question of which ones and which providers will consistently perform better over time. 
Pre-trade TCA
All our order flow does get pre-trade TCA snaps. Everything that goes through our desk gets a pre-trade assessment on it. We do not however rely on pre-trade TCA alone as most implementation cost estimate models we use tend to break down somewhat with larger orders or in less ordinary environments. Instead, we will look at the unique characteristics of the order with respect to the liquidity environment, as opposed to how it is profiled based on historical data. TCA is one input into how our trading decisions are made, but it is not the only one. In certain cases, we look at the trader’s knowledge of where and how to source liquidity much more than whether or not they are within a certain range of their benchmark or implementation cost estimate.
‘Dark’ Future for Canada?
Before we start to see more dark venues in Canada, we will need to see some validation of the business models that are already running. I think regulatory uncertainty is going to have to resolve somewhat before we will see more ATS and dark growth. Historically, Canadian participants have been slow to adjust, in terms of turning on venues and exploring new order types. From where I sit, I see the growth in number of venues will be slow, which is not to say we will not see any new venues open up or new order types emerge. But there are simply too many uncertainties for prospective entrants, with the pending Maple-TMX merger and forthcoming regulatory changes clouding the outlook. I do not think you will see a lot of bold moves into this space until some of that resolves itself.
To VWAP or not to VWAP
On the buy-side, I think algo usage will continue to increase moderately. Algo usage responds to so many different factors in terms of what the market is doing, volatility, the upstairs block market (which has traditionally been very dominant in Canada) and other factors. It also responds to how much PMs are comfortable taking a more passive strategy as opposed to seeking out liquidity aggressively by other means. Buy-side algo usage definitely tends to be more passive because VWAP and scheduling strategies are still predominant. We will continue to see growth, but it is going to be moderate, relative to the last three or four years, when it became a significant part of the market. I rarely use VWAP algos except in very limited situations, but I see a lot of fellow buy-side traders being held accountable to that benchmark, especially when that is the benchmark their PMs and management have been sold on from a policy standpoint.
Click to contact the author: