A conflict of public interest
Analogies to movie titles are purely coincidental… The Markets in Financial Instruments Directive (MiFID I) was introduced in 20071 to harmonise financial markets, strengthen investor protection and increase competition throughout Europe.
One of the consequences of the implementation had been a grown demand for dark pools, trading venues which circumvent equity pre-trade transparency requirements. The pre-trade transparency regime, i.e. the obligation to make public current bid and offer prices and the depth of trading interests at those prices, had raised concerns among market participants. Predominantly large investment managers expressed fear that pre-trade information around (large) orders can be abusively used by competitors – which is why MiFID had introduced the large-in-scale waiver (LIS). Dark pools have been allowing them (and others) to generally avoid pre-trade transparency by making use of another waiver: the so-called reference price waiver (RPW). The RPW, one of four2 waivers from pre-trade transparency under MiFID, allowed venues to match orders at prices that are references to prices determined by other markets.
More than a decade lies between the implementation of MiFID I and its revision on 3 January 2018, MiFID II – and one of the most severe financial crises in capital market history. Including all delegated and other implementation acts, the revised regulatory framework fills roughly twenty thousand sheets of paper. Industry concern that a withdrawal of the waivers could be stipulated in MiFID II, did however prove unfounded.
In that context, it is worth looking into how dark pools have been influencing market structure and market functioning. Buy side firms have been arguing, waivers are supportive of liquidity in that they bring participants to the market that would not have been there otherwise. Second, the maintenance of waivers would mean lower costs, preventing the broker from incurring spread costs in association with accessing a lit order book that otherwise always puts the buyer at a disadvantage. Third, the absence of waivers would expose the investor to a higher risk of the market moving against his interest (due to the pre-trade information then being shared with the market). The words ‘participants’, ‘buyer’ and ‘investor’ in the above asset management industry arguments have been set as italics intentionally – because they refer to the retail investor in each case.
Assessing the extent to which the above are valid arguments requires examining dark pools’ implications on liquidity and price formation from a market microstructural perspective. In theory, dark pools are associated with uninformed order flow while on lit exchanges informed traders are looking for fast, reliable execution. This notion assumes that trading without pre-trade transparency is of particular benefit for traders looking to benefit from price improvements. Following this model assumption, retail order flow is benefiting from dark pools since it avoids the risk of adverse selection and higher costs of trading on lit venues. This theory has demonstrably proved outdated in several respects. Technology, to name but the most significant factor, has led to an increasing number of market participants that focus on gathering all available pre-trade information in the market and to ‘front-run’ on it based on super low latency trading mechanisms.
In order to limit the amount of trading under certain waivers (the RPW and NTW3) and thereby to prevent harming the price formation process, MiFID II had introduced the double volume cap (DVC) mechanism. This mechanism caps the transactions that are allowed to be executed under the RPW and NTW at 4% at a venue level and at 8% for all venues throughout the European Union. If those thresholds are reached, competent authorities must make sure that dark pools must suspend trading for a period of six months.
It will not surprise you that market participants are looking for ways to circumvent this rule, too. As ESMA4 has noted back in 2018 already:
“Trading flow previously executed under one of the two waivers covered by the double volume cap, is in particular flowing to systematic internalisers and periodic auction trading systems.”
ESMA added that, since the first suspension of dark trading in March 2018, trading volumes on periodic auction trading systems have tripled. It may be assumed, however, that ESMA, if deemed necessary, will adopt further measures to intervene.
As mentioned above, the regulatory framework has left enough space for large block trades to stay exempt from pre-trade transparency such as the large-in-scale waiver (LIS). This is noteworthy because the majority of orders in dark pools does not qualify for the LIS, i.e., they do not justify an exemption from pre-trade transparency. Moreover, prices in dark pools are not resulting from demand and supply within the dark pool, they are derived from lit market references. I.e., orders in dark pools inhibit price formation to the extent they would otherwise (when executed on-venue) contribute to the lit market order book depth as they then would represent a degree of information available before the price is being discovered.
Dark pool prices, conversely, are a sole function of the execution at a certain level – inevitably increasing volatility and having an adverse effect on retail trading costs through spreads and liquidity. Another, recurring phenomenon that has become manifest again very recently should discourage retail investors from engaging in dark pools: if market volatility adds to the less liquid execution infrastructure of dark pools and if the market is in selling mode, the order will not be executed – sometimes the platforms cannot be kept up at all.
To end on good news: we can disabuse the market of the misconception that the emergence of new lit markets, such as Spectrum, would have an adverse effect on liquidity. Fragmentation in visible order books – as a study of the Centre for Economic Policy Research (CEPR)5 has found – improves global liquidity, whereas dark trading has a detrimental effect. But this is a topic for our next issue…
To be continued...DOWNLOAD THE PDF