Darkness. It evokes shadows, evil and foreboding. It was the ninth of the ten plagues of Egypt. If something is done in the dark, it is unknown, hidden and may be nefarious. King Lear is said to be Shakespeare’s darkest play, so dark that for many years it was not performed. And as the king says – “There’s hell, there’s darkness”.
When exactly the term “dark pool” emerged is unclear. In 1986, Instinet launched its after-hours equities crossing network and the following year ITG launched Posit. Both were early dark pools. Attempts to use terms such as non-displayed liquidity were too long and clumsy; somewhere in that timeframe, “dark pool” gained traction. The term may not have helped, and probably will not help, approvals in some jurisdictions particularly in Europe.
In reality, we accept that secrecy and confidentiality are an essential part of business. If every exploratory business conversation was published for all to see, much business would never happen. We rely on outside auditors and compliance officers to ensure such conversations are legal, not public records. So it is in financial markets – immediate transparency can damage liquidity and widen spreads for commercial transactions. The introduction of TRACE, which required prices to be published in the corporate bond market, is a well-known example of how liquidity can be impacted.
At the end of the presentation that I gave to BIS Markets Committee in January, 2014 ( (which you can view here), I forecast three trends: increased algorithmic trading, growth of DMA and growth of dark pools. Trading through dark pools already represents greater than 30% of US equity volumes, so why not in FX?
On the sell-side, dark pools are growing. Two major interbank dark pools are operated by BGC and FXall. BGC reported a 50% year-on-year increase in fully-electronic spot revenues in Q4 2013, in a period when EBS volume was down 20%, Thomson Reuters Matching down 12% and CME FX futures were down 8% from a year previously. I can only attribute BGC’s growth to their mid-point matching dark pool which stands in contrast to the other three platforms which are lit markets.
The value proposition of interbank dark pools increases proportionately to the concentration of volume among the top few banks. If major banks need to hedge risk, they have essentially three alternatives aside from dark pools: the first is to skew in the favor of a customer attracting the customer to provide an offsetting order. That is the essence of internalization and provides the best option – spread capture, no brokerage and low risk of a footprint in the market. But such customers may not always be available. The second is to use an ECN or an exchange, such as EBS, Reuters and CME, in which case information is transmitted to a wide audience very quickly. The third is to call another market maker, usually a competing bank. If they opt for this third option, information is leaked to a competitor. The more the market becomes concentrated, the higher the proportion of the whole market that a single competitor represents and the more attractive dark pools become.
Indications are that the market continues to become more concentrated at the top end, with Greenwich Associates estimating that the top four banks accounted for 46.7% of FX volume in 2013. Dark pools, such as that operated by BGC, provide a means for large banks to trade with each other and limit information leakage.
The question then arises: will there be FX dark pools on the buy-side? Here the jury is out. Certainly, the buy-side was the staple of equities dark pools. Liquidnet admitted no sell-side participants for several years. In some ways, however, buy-side dark pools already exist in FX. They are the large banks. When a large bank crosses two customer orders, what is functionally different to a dark pool? The only information leakage is to the bank providing the service, and that bank has an incentive not to move the market against its customer if it wishes to retain the customer’s business.
An additional reason we have not seen non-bank buy-side dark pools is that many customers have a dependency on the major banks for other services. These customers wish to provide FX customer business to the major banks as it enables the banks to keep prices low for other services – particularly those services which must be charged explicitly as a fee rather than be included in a spread.
There is an argument that non-bank exchanges are less conflicted than banks that offer exchange services. The fact that dark pools are not transparent means it is important that non-conflicted organizations offer them. Two organizations are uniquely placed to launch buy-side dark pools. Both Thomson Reuters and Bloomberg have wide distribution to real money managers and intuitions that are the traditional participants in equity dark pools. If they launch it will be interesting to see how a non-bank dark pool would work in FX.
As Edgar says in Lear, the prince of darkness may yet be a gentleman.
‘Thoughts on Basel’ Series – Part 3 of 3
Catch up on the whole Series:
Part 2: Traders in ever greater numbers , FX and the second law of thermodynamics – read here
Part 1: It no longer pays to be the reference rate – read here