Measuring Block Trading Performance
Dr Darren Toulson, LiquidMetrix
This talk, given by Dr. Darren Toulson, focussed on how to monitor, measure and quantify the benefits of trading on block venues, and ways in which benefits and performance of trading blocks can be quantified versus other ‘algorithmic’ ways of trading.
Assessing the performance of block trading from 2017 until 2018, the discussion showed some interesting statistics, and highlighted areas for both buy- and sell-side to be aware of in the year ahead.
“A theme of MiFID 2 was that off-book trading should ideally be reserved for trading larger (block) sizes. The most concrete example of this ‘nudge’ was limiting of non-LIS volumes that could be traded on Dark Pools, the so called “double volume Caps”
Even prior to MiFID II, many Buys Side Firms were favourably inclined towards trading Block Sizes. Perceived advantages include:
More likelihood of the counterparty being another Buy Side and less likelihood that the counterparty will be a HFT/active market making firm;
Less likelihood of information leakage as fills are large, possibly reported with a delay and (usually) crossed at market mid-price so it is not clear who the aggressor was;
The possibility of getting very large order sizes done quickly with minimal market impact.
The buy side were predisposed to trading LIS and in blocks. The new mechanisms of trading and mechanisms in MiFID II, is balancing the good and bad.
“The Achilles Heel of block trading is finding the other side of the trade, but fill rates continue to disappoint. You can expect some, but not all, of your trades to execute.”
“So, with the pressure of new MiFID II rules and the general preference of many buy-side to trade blocks, the environment seems set for an increase in block trading.”
The data set
“One difference is that ideally analysis should include orders that are unfilled. This is important if we want to analyse fill rates and the opportunity cost adjusted performance measures.”
Results of order characteristics:
Over 2017 there was a gradual rise in order sizes;
In addition to this trend the 2 key events, MiFID ii and DC resulted in further step increases in block volume order sizes;
By the middle of 2018 average order values were about double to what they were prior to MiFID ii (for each of our 3 size measures. (2018).
“Historical spreads show that the majority of orders are in blue chip stocks (<10bp spread). The proportions of value trade in spread bands has remained fairly constant before and after MiFID II was introduced.”
“For two unnormalized measures of fill success, there was a significant increase in performance from middle of 2017. After MiFID I, the headline rate of % orders filled / partially filled actually decreased somewhat”
“The percentage of volume filled did continue to increase slightly post MiFID II. However, as we will see, these statistics are somewhat misleading as order sizes and trade volumes rose significantly after MiFID II.”
“Using normalised measures of fill rates, we can see that the improvement in fill rates post MiFID ii is significantly clearer. In terms of average value filled per hour and % ADV filled per hour, rates post DVC leapt upwards to nearly treble their values.”
Opportunity costs for resting orders.
“We have noted earlier that there is a significant market movement in the direction of restoring Plato orders, especially for unfilled orders. This movement has in fact increased in recent months such that it averages around 12 bps.”
“For unfilled orders there is a cost of about 12bps due to market prices being significantly worse by the time of the order cancellation. The key point here, however, is that these costs only appear if the block orders are not resting conditionally.”
“Order sizes have significantly increased in the lead up to MiFID II and, these larger sizes have been maintained. Fill sizes increased both in the lead up to MiFID ii and sharply after the DVC caps were introduced in March 2018.”
“Fill rates expressed as simple counts of filled orders and also normalised counts such as %ADV traded per hour have more than doubled over the last 12 months.”
“There is significant movement in the underlying stocks in the direction of Plato orders, especially in the case of unfilled orders. This movement appears to be specific to the stock, general market movements are not aligned with the side of resting orders.”