FX trading focus : HFT : Dan Barnes

FX: THE PHANTOM LIQUIDITY MENACE.

 

The sterling flash crash is indicative broker liquidity has gone; the buyside must look for alternative counterparts in FX. Dan Barnes reports.

The flash crash in sterling /dollar on 7 October, 2016 saw the pound fall to $1.1840 at 12.09am BST from $1.2609 at 12.06am BST, then bounce back to finish the day at $1.24, according to Bloomberg data. These events have occurred in several key markets before with US equities in 2010 and US treasuries in 2014 being the most notable. The latest occurrence is under investigation by the Bank of International Settlements (BIS) following a request from the Bank of England, but if it follows the pattern of previous investigations it will find a common pattern as the cause.

be35_web-image07-42-48

There will be several contributing factors cited including thin liquidity, a high proportion of automated trading as well as liquidity provision by high-frequency trading (HFT) firms and phantom liquidity – the placing and cancellation of a high volume of orders by HFT firms. The reduction of traditional broker dealer liquidity will also be on the list. A source close to the BoE reports that initial indications point to electronic trading issues rather than a ‘fat finger’ error.

It has been documented that the FX markets faced some of these circumstances on 7 October. In a note issued on 5 October 2016, around 40 hours before the sterling flash crash, Bank of America Merrill Lynch FX strategists warned that the market was looking increasing fragile “as phantom liquidity creates the illusion of stability.” The firm reported that its volume-based analysis showed market liquidity had materially worsened with market impact of a given volume some 60% greater than in 2014, while volatility events had increased in frequency and amplitude. The sterling crash was conducted in Asian trading hours when liquidity tends to be at its thinnest, exacerbating these factors.

be35_web-image07-42-56

“The HFT firms, and e-Commerce algos and hedgers at the banks, have been very successful in establishing a 24/7 view of a currency in terms of transparency and reference point,” says Darryl Hooker, co-head EBS Brokertec Markets. “They are not necessarily equally as additive in terms of putting liquidity to those prices. So in Asia you can see a great price in sterling at 7pm New York time, but then you try and hit that price for ten pounds and you would probably only get one done.”

Take your partner

For the buyside FX trader, thinner sellside liquidity, the consequent potential for greater market impact and volatility, increase the importance of finding a wide circle of liquidity providers. The growth of electronic trading is making that choice easier, says David Scilly, head of fixed income & currency dealing at First State Investments. “The electronic environment allows buyside traders flexibility to make their own choices around how they execute. There is enough flexibility in most e-trading systems for us to be able to choose whether or not we do an RFQ, whether we take advantage of available liquidity at the time or whether we work an order into the system over specific parameters.”

Where broker-dealers are falling back, electronic liquidity providers (ELPs) are moving in. These firms are part of the wider HFT demographic, which was estimated to be approximately 40% of volume in spot FX across 2014-15, having fallen slightly from about 42% in 2012-13, according to analysis by Aite Group.

Representing such a considerable chunk of actionable liquidity, makes HFT firms a potentially valuable counterpart for investment managers. The heads of trading on buyside desks have matured their ability to interact with HFT liquidity, often due to experience beyond the institutional investor market. The result is that relationships between buyside firms and HFT firms are getting stronger.

be35_web-image07-43-05

“In the public view HFT was often seen as ripping off buyside firms, generating decent profits because of little knowledge at buyside firms about HFT activities,” says Christoph Hock, head of multi-asset trading, Union Investment. “This is clearly not our view. From our perspective HFT is not bad for markets per se. But it is important to differentiate [between types].”

He draws a distinction between electronic liquidity providers (ELPs) which offer automated market making and provide additional liquidity in markets, and firms that are active in markets with pure proprietary strategies like latency arbitrage and short-term momentum trading.

“In this field we implemented measures in the past to give us a high level of protection against these activities,” Hock says. “For seven years I worked myself in the hedge fund space and looked at these kinds of market making and prop strategies in detail, so I am quite familiar with the dos and don’ts.”

For traders seeking to avoid HFT altogether, some platforms offer mechanisms that limit the types of trading strategy that can be used, in order to remove the risk of any predatory practices.

Dan Marcus, CEO of trading platform ParFX says, “FX trading counterparties rely on a range of trading venues to act as neutral facilitators of liquidity and the onus is on these venues to put mechanisms in place to protect their customers against any forms of disruptive trading behaviour. However, this was not occurring in many instances.”

be35_web-image07-43-18

ParFX was developed in partnership with several large currency trading institutions and incorporates a number of preventative measures and mechanisms that discourage ‘disruptive’ trading, including free market data distributed in parallel, to prevent any counterparty gaining an advantage, and the introduction of a randomised pause of 10-30 milliseconds.

“The pause is meaningful enough to nullify disruptive traders whose strategies rely solely on speed,” says Marcus. “Essentially, what this does is deliberately randomise trading instructions so the fastest, latency-dependent traders have no guarantee of being the fastest. However, it is meaningless to those that come with a genuine trading need, who seek firm, executable liquidity and who compete on strategy.”

The liquidity picture

The challenge firms have is not only the risk posed by counterparties, but how to determine real market liquidity, where a large volume of orders might be cancelled, creating a phantom liquidity picture and potentially increasing market impact for the unwitting.

Overall spot market trading activity has been in decline, falling by 19% to $1.7 tn per day in April 2016 from April 2013 according to the BIS Triennial Survey. It noted that the decline was the first picked up in a report since 2001. While liquidity does not correspond directly with volume, a shallower market makes bigger orders easier to detect.

Working with non-bank sources of liquidity will help, but it will not be a salve. Daily spot turnover with hedge funds and proprietary trading firms as a counterparty declined 29% in the three years to April 2016 to $200 bn. Forwards and FX swaps also saw a decline in trading with those counterparties, falling by 29% and 37% over the same period.

Given these circumstances, finding liquidity will be down to a combination of the right tools and platforms. Scilly says, “You don’t necessarily have full transparency into liquidity as a user, but increasingly you are being directed through EMSs and through other smart trading platforms to execute in a certain way, using the systems’ internal workings to take advantage of the liquidity that is available.”

There is a more radical option, notes Hooker. A regulatory intervention that forces internalised liquidity back out in the market, just as ‘dark trading’ in the equity markets has faced volume caps under impending MiFID II rules.

“If regulators or central banks said banks were not allowed to internalise more than 50 per cent of flow, the amount of liquidity that would pour back into the market would be tremendous in terms of helping to ease these stress events and show that there was real liquidity out there,” he says.

©BestExecution 2017

[divider_to_top]


Related Articles

Latest Articles