The Impact of Dark Pools on Access to Desirable Liquidity

Share

Emma Quinn, AllianceBernstein’s Head of Asia Pacific Trading discusses accessing liquidity through dark pools, aggregation and asset allocation.
Trading Volumes, Liquidity and Asset Allocation
I think that you’ll see trading volumes rise when you get an asset allocation back into equities, and people have more conviction in the markets. The reason that there’s just no liquidity in the markets is not because people are worried about exchange mechanisms or aspects like that, it’s about the macroeconomic environment and the allocation into equity.
I don’t think that we’re going to see volumes in other asset classes recover faster than allocation into equities as we’ve already seen that allocation change. People are either bullish or bearish, and are set for what they think is going to happen. And so we are in a position that people will just trade around their positions without making any significant move either way until we get some clarity on the macroeconomic environment.
The Rapid Expansion of Dark Pools and Access to Desirable Liquidity
We use dark pools to access liquidity for orders we would not normally place in the central limit order book. I think dark pools aid price discovery. There has to be post-trade transparency but once that happens you’ve actually got more transparency on a market than you normally would. In this sort of environment, because you’re not putting out so much into a central limit order book, what used to be 10% of average daily volume is now 30% of average daily volume, you’re obviously leaving more in a dark pool if your order size hasn’t changed. I do think dark pool liquidity aids you, as your expected cost is going to be lower and thanks to post trade transparency in dark pools the market sees a block trade that it would not have seen.
With regard to desirable liquidity, I think the onus is on the buyside to actually put in parameters that can minimize risk. Obviously, you don’t want to go into a dark pool blindly. The same thing could be said of going on to the central limit order book. The same thing can happen to you on a central order book as can in a dark pool, if you’re not smart about the way you trade in a fragmented environment you leave yourself open to be gamed.
Best Execution
We have an unbundled commission policy and as such our traders are not limited to paying based on a research vote. We have the discretion to use the broker that will give us the best execution outcome. This discretion is important and enables us to focus purely on the best execution outcomes for our clients.
Impact of Direct and Indirect Costs Imposed on Buy-side Traders by Liquidity Fragmentation
We spend a lot of time on quantitative trading strategies and both post and pre-trade cost analysis – we have pre-trade expected costs in our trader management system and we also look at post trade – both daily and weekly as it is not enough just to look at one trade in isolation as so many factors can contribute to whether you have got a trade right or wrong.
Some of the cost of fragmentation has already been borne by the buy-side and sell-side, such as having to have smarter systems and employ quantitative trading. Brokers are now wearing additional costs with some regulators looking to recoup the costs that come with the increase in surveillance costs for a fragmented market. The brokers may have made savings due to the fact that we now have multiple markets and with that came a compression on exchange fees, but they could well and truly be paying that out now to regulators.

 
Aggregated Liquidity
One major consideration when using dark pools is whether you go in under your own name or under that of an aggregator and what are the benefits or costs of each approach. When entering a dark pool under an aggregator’s name, then, obviously, no one at the broker knows who you are regardless of whether they are sitting in the dark pool. However, you need to consider if you go into a dark pool in your own right, if that will afford you any differing treatment, and the cost benefit of that needs to be considered.
In terms of developing systems, we have done a lot globally. In New York, we have worked to build aggregators in a very fragmented market. We use dark pools now but we need to consider how we go about aggregating them effectively.
Other costs that need to be taken into account include the tiering system, how many pools we access and the cost of accessing the pools via an aggregator as opposed to directly. In addition, we need to look at how an aggregator’s smart order router works. Does it just sit there passively or is it actively moving orders around, cutting from one pool and placing them in another? How is it prioritizing orders?
Market Structure Changes and Access to Liquidity in India and Australia
Many people have pulled out of India due to the uncertain tax regime and that issue has taken the spotlight ahead of market structure changes.
In Australia, I think market structure changes just mean that people have had to be smarter about the way that they trade. You can’t just passively place an order. You have to make active trading decisions. You’re going to be considering the block trade, the dark pool, the algorithm, before you pass it off and put it into a central limit order book.
Algorithms
We have made some changes to the way that traders access algorithms and the way we choose algorithms we participate in. I do believe that there is a natural level of how many orders fit the criteria to make them suitable for that type of trading.

Instinet’s Glenn Lesko looks at the rise in popularity of dark pools in Asia and how dark aggregators should be used to most effectively access them.
The rapid expansion of dark pools has had a significant, positive effect on desirable liquidity. From our experience, dark pool liquidity in the Asia-Pacific region is largely high quality, and the participants that we have seen in dark pools are primarily large-scale global investors.
In Australia and Japan, the use of dark pools is largely driven by the same factors as in the rest of the world – faster exchange technology with typically tighter spreads and lower volumes have resulted in less depth in the market and reduced visible liquidity. Exchange competition has fragmented displayed liquidity in both of these markets as well, which was a factor that drove dark pool development in the US and Europe. Finally, in Australia, the block tradition of the market is also a factor, as it can be difficult to find sufficient liquidity on the exchange outside the top 10 or 12 stocks.
In Hong Kong, dark pools are adding value for very different reasons. Trading spreads are quite wide, so investors are saving spread as most of the dark pool trades are done at the midpoint. They also help investors avoid the transparency issues that result from the exchange’s display of broker names.
In terms of liquidity, we don’t see dark pools as increasing fragmentation as much as fulfilling genuine market needs. Brokers have always had their own liquidity; that liquidity was fragmented whether it sat on their block desk as was the case historically or is placed in their dark pool as it is today. The fact is, trading via displayed exchanges has never been the best way to access all liquidity in the market.
Tools and Systems Available to Aggregate Liquidity
Today there are only a few dark pool aggregators available in Asia, which is similar to the situation in the US four or five years ago. This may change in the future, but currently the big brokers – which only relatively recently deployed their dark pools in Asia – are focused on getting returns from that investment rather than aggregating other firms’ dark liquidity for clients.
The dark pool aggregators that are available offer access to multiple venues where dark liquidity resides, be it in a bulge bracket-broker dark pool, agency-broker dark pool or on exchange (as ‘hidden’ or ‘iceberg’ liquidity). The venues accessed by these aggregators are generally fully customizable, and are therefore leveraged by a wide range of traders – from block traders looking to get the largest possible block closest to the price at which the order came in, to those benchmarked to volume-weighted average price (VWAP) or participation weighted price (PWP). We measure the performance of our dark aggregator every month to see how it is performing versus implementation shortfall, PWP and VWAP.
Dark aggregators typically prove to be to most effective way to access as much dark liquidity as possible. Dark pool penetration is somewhere between 5-10% in Japan, about 3% in Hong Kong and 3% (and growing fast) in Australia. However, in Q1 of this year, clients using our dark aggregator executed 18% of their flow in dark pools in Hong Kong, 14% in Australia and in excess of 11% in Japan (rising to 22% if PTSs are included). While many point to the relatively low dark liquidity usage figures in Asia- Pacific, the simple fact is that the results can be significantly higher for those using the right tools.

Risk Mitigation Strategies
The list of blue-chip institutional clients using our dark aggregator in Asia demonstrates the quality of flow in dark pools in the region. However, like in any type of venue, there may be undesirable flow found as well – such as participants engaged in gaming, information seeking and other predatory behaviours – so it’s not advisable to interact indiscriminately with any and all dark liquidity. Furthermore, there is no obligation on behalf of pool operators to disclose the identities of their participants or even a foolproof method of determining which pool participants present risks to other clients. So rather than attempting to dictate the terms around the type of flow they interact with, we advise clients to use a multifaceted technological solution to increase the benefits and mitigate the risks.
First and foremost, clients should know where their broker is routing and why. This does not just apply to dark pools, but all venues. Best execution policies, where they exist in Asia, don’t specifically address venue access, so all routing decisions are made at the discretion of the broker. This means client interests can be compromised by a desire to internalize, substandard routing technology and venue pricing.
Also, some pools onward route to others, taking control away from the broker with whom the client has entrusted their order (for that reason, we have policies in place to ensure we do not access pools that onward route).
Next, we recommend using an aggregator that includes intelligent anti-gaming capabilities to manage your interaction with dark venues. Functionalities such as support for minimum fill sizes, randomisation, and analytics that evaluate short-term pricing models, volatility and spread changes before sending order slices are generally quite valuable.
It is also advisable to monitor results in real time. The FIX Protocol provides tags on both a live and post-trade basis that allow clients to view the dark pools in which their orders are being filled. So even if using a product that is aggregating over 20 dark pools, for example, clients are able to accurately evaluate the liquidity with which they are interacting on a real-time basis.
Finally, we urge clients to supplement this with post-trade analytics, which allow the risk versus benefit trade-off to be effectively weighed. A venue’s benefits can typically be accessed by analyzing standard trading performance metrics including fill size, price and frequency, while its risk can be measured by evaluating any changes in price, quote, volatility or a reversion to other prices during and after interactions.
Benefits and Costs
In our view, the benefits of dark pool trading far outweigh any additional costs. The real cost is that institutional traders must now be familiar with more sophisticated tools to access fragmented markets, but the benefits are that clients can access more liquidity, often with price improvement. For example, in a recent month clients in Asia using our dark aggregator executed an average of seven times more than the primary market’s consideration during their order’s trading interval. They also enjoyed an average spread capture of 49%, or about 10 bps versus what they would have paid to cross the spread. Slippage across market caps was -11 bps versus the arrival price and performance versus Interval VWAP and participation-weighted average price (PWAP) (20%) was +10 bps and + 3 bps respectively.1
[1] Statistics from Instinet’s Insight trade analytics suite.