On January 3, 2018, the revised Markets in Financial Instruments Directive, or MiFID II, became effective across EU member states. This comprehensive and far-reaching regulation will shape European capital markets in years to come. Among other things, MiFID II puts several restrictions on dark pools in European equity markets: (i) Broker Crossing Networks are essentially banned; (ii) dark pools that rely on “reference prices” on exchanges can only execute trades at the midpoint of exchange best bid and offer; and (iii) dark pools are subject to volume caps of 4% for a single venue and 8% across all dark pools (colloquially referred to as the double volume caps). On the other hand, MiFID II keeps the “Large in Scale” (LIS) waiver, so sufficiently large transactions can still go through without being counted toward, or affected by, the double volume caps.

Jargon and technical details aside, these MiFID II rules essentially push dark trading to return to basics: the matching of large institutional orders to reduce price impact (for both sides). Price impact—the very act of buying or selling moves prices adversely—can be quite costly for institutional investors, especially in today’s market where alphas are hard to generate and high-frequency traders watch every market movement at the microsecond level. By reducing the price impact of trades, investors enhance returns.

Figure 1 below illustrates the basic economics of block-sized dark trading between investors, adopted from Duffie and Zhu (2017), “Size Discovery”. The two thin lines illustrate typical inventory paths of a large buyer (red solid) and a large seller (blue dashed) if they can only trade on a public exchange. Their concern of adverse price impact causes them to slow down and trade piecemeal, using algorithms to break up the large orders into small pieces and execute them gradually. Investors’ inventories eventually converge to the targets, but they still incur some price impact costs and face the risk of sudden price movement while execution is ongoing.

Figure I. Illustration of Block Crossing

Block trading, or “size discovery”, gives investors another option. Imagine that the large buyer and the large seller “meet”—on a secure electronical platform under anonymity—at the beginning of the trading day. They are offered the opportunity to trade any quantity at a fixed price, say yesterday’s closing price. Fixing the price overcomes the concern of price impact, and the two large investors can trade a chunky amount immediately, upon mutual agreement, as shown by the thick blue line and the thick red line. Clearly, adding this one-shot size discovery mechanism reduces inventory imbalances, which reduces investors’ costs of price impact and the risk of sudden price movements.

Even before its implementation, MiFID II pushed the industry to more seriously consider block trading mechanisms. As a result of concerns that banning Broker Crossing Networks and the implementation of the double caps would constrain execution quality, new block venues and block order types emerged in the lead up to the implementation of MiFID II. Three new LIS venues were established: Turquoise Plato Block Discovery (TPBD) in 2014, CBOE LIS in 2016 and Euronext Block in 2017. These venues use conditional order types, which allow investors to rest large undisplayed orders while simultaneously working these orders via algorithms. If a block match is found for the large conditional order, investors are asked to “firm-up” their conditional interest. Investors can set a minimum execution size allowing investors to be rewarded for providing large size, and minimize the risk of other investors stepping-ahead of, or anticipating their order flow. These new venues supplement existing venues aimed at facilitating block trading, namely Liquidnet Negotiation, Liquidnet H20, and SIX Swiss Dark. Each of these venues exhibits average trade sizes substantially above the LIS-waiver thresholds. Although there are some variations in the way these venues set the trade price, the common feature among them is that they seek to provide the “size discovery” described by Duffie and Zhu. 

Figure 2 shows that there is a growing appetite for these block executions. Over the period January 2015 to November 2017, block activity has grown from approximately 2.8% of dark trading activity to 11.4% of dark trading activity. Although not all of this activity meets the new MiFID II LIS thresholds, the level of activity that does meet these thresholds is expected to continue to grow after the implementation of MiFID II.  

Figure 2. Block activity in Europe as a fraction of total dark volume

In addition to innovations relating to LIS trading, concerns about the MiFID II dark trading restrictions have also contributed to the launch of high frequency batch auction venues and a proliferation of new Systematic Internalisers. The impact of these developments on execution quality is at this stage unclear.

Although the big picture remains that more transparency is overall desirable, these market responses to MiFID II highlight the need for regulation to acknowledge large investors’ concerns about disclosing their intent too early. Fine-tuning dark trading mechanisms is no trivial task, but the basic function of dark trading remains as relevant in today’s markets as it ever has. Post MiFID II, dark pools have every reason to return to basics.

Haoxiang Zhu is an Associate Professor of Finance at the MIT Sloan School of Management.

Carole Comerton-Forde is a Professor of Finance at the UNSW Business School.