The extent of automation in FX trading systems has meant that a major market shock would lead to a slow-down in trade rather than meltdown. But it has also increased market volatility. William Essex weighs up the pros and cons.

Foreign-exchange markets are increasingly automated. Price movements between currencies can be captured as they happen and programme-trading systems have the capacity to trade without any human intervention at all. Nowadays, FX traders, whether human or automated, have recourse to automated post-trade reporting systems that deliver an updated report on their overall position much more quickly than used to be possible. And the speed of reporting is important: unless traders can track how their overall positions have been changed by a trade, they cannot continue to trade.

All of this is positive. Automation increases the speed of activity in FX markets as well as bringing costs down and increasing transparency. However, there is a potential downside. First, what are the ramifications of automation when there is an asymmetric shock to the financial system? If, for example, there is an upset in the US presidential election, or, say, Iraq blows up, it is reasonable to assume that a very large number of automated trading systems would implement sell-dollar orders. This would be likely to trigger a slide in the dollar that would in turn trigger a wave of stop-loss orders that would compound the slide.

Would this lead to a meltdown in the dollar, to the extent that its trading value reduced almost to zero? And if so, would those automated trading systems, with their capacity to operate at a faster-than-human speed, have accelerated the catastrophe? The answer to these questions is not simple. The reaction to a major asymmetric shock such as a terrorist incident would not be an immediate crash because automated trading systems all follow one core rule: when in doubt, switch off. Mark Warms, chief marketing officer at multi-bank portal FX All, says: “If there’s a big shock to the system, every electronic price will disappear, and when it comes back, it will have moved significantly.”

Putting off the inevitable

All the systems switch off, all the prices disappear, and no trading can occur. But this does not mean that all those sell-dollar orders are cancelled. They are still waiting to be executed. The switch-off rule simply defers the reaction to the shock; indeed, it allows the first part of the crash to happen without trading intervention, in that when systems switch back on and electronic prices reappear on screens, those prices are already significantly changed. What happens next? Surely all those sell orders hit the market simultaneously, so that the delayed meltdown can finally happen?

In fact, what would happen after the suspension of pricing would be a slowdown in trading. Assuming that the immediate response to the reappearance of prices was a sudden flood of orders to trade, prices would move.

Bill Moran, chief operating officer at trading technology firm EBS, says: “If there is a surge in demand as a result of a global event, we would expect prices to react accordingly.” As prices moved, so traders’ positions would change. It is unlikely that the market would lack the trading capacity to handle the flood of trades that followed a suspension of pricing.

But what matters here is not the speed at which prices would change in reaction to all that trading, but the speed at which traders’ positions would change. Nick Dyne, CEO of financial technology firm LogicScope, says: “There is a natural brake here, in that systems might not be able to provide data quickly enough to enable programme-trading systems to know their positions quickly enough to continue trading as they want to.”

This is where speed of reporting comes in. The speed of automated reporting, which will generally cover the whole range of currency exposures that a trader might have, will tend to lag the speed at which automated trading is possible.

The good news and the bad

The good news for the FX market, overall, is that automation protects against an equity-style crash. It might happen that an asymmetric shock prompted a very widespread desire to get out of US dollars or any other major currency, but what would happen is not a sudden crash in the dollar price, but a halt to trading followed by a resumption that would necessarily be slower than “crash speed” because automated reporting across a whole position doesn’t happen as fast as automated trading in a single currency.

The bad news for individual traders, and particularly those with a heavy commitment to a given currency, is not that prices will crash before they can escape from their positions after the shock, but that they will be denied any access at all to prices until those prices have changed. When they do get access, their operations might be slowed by the trading/reporting speed differential. This is bad for an individual trader because post-suspension trading will be likely to move the price towards a stable level even further away from its pre-shock level (given that all the delayed sell orders will supply pressure in one direction only), while any selling in reaction to this will be slowed.

Reducing the price gap

On the plus side, there are two factors that will reduce the distance between the pre-shock price and the stable level it reaches after suspension and post-suspension trading. The first is “non-correlated interest” – or interest in trading that is not affected by the shock. This might be, for example, trading interest from corporates managing the currency requirements that arise from their trading activity. Such interest will not be correlated to any shock, in that, for example, corporates will go on needing dollars whatever happens in Iraq.

Automation will allow such corporates to place orders (for example “buy dollars when possible at the prevailing market price”) for execution at the earliest opportunity. This would have been less efficient pre-automation. Mr Warms says: “When prices were made verbally, everything could get pulled verbally and there would be no source of liquidity whatsoever. With technology, interest that’s non-correlated to an asymmetric shock finds its way faster to the market.”

It finds its way faster to the market in the sense that such interest can be programmed into an automated trading system and thus acted upon automatically. It does not have to wait for a trader to contact a willing counterparty over the phone. Corporates will go on wanting to buy dollars for their own business reasons even when everybody else is trying to sell them for market or price reasons, and the same is true for other major currencies.

Distinguishing features

Given the size of the FX markets (overall daily volumes are estimated in trillions), non-correlated interest can be assumed on both the buy and sell side of any currency. Simon Wilson-Taylor, global head of State Street’s Global Link, says: “The FX market is distinguished from other markets by the fact that there seem to be more participants with different viewpoints operating in the FX market than anywhere else.”

Those different viewpoints give rise to different trading orders and different reactions, if any, to any shock. Also, behind the non-correlated interest, the second positive factor here is the power of the central banks. In a worst-case slide, where a currency’s trading value does change significantly, central banks have the power to step in and, by buying in a selling market, restore a given price level.

However, even without an asymmetric shock to destabilise FX markets, there are some signs that automation is increasing their tendency towards volatility. Automation has enabled a wider range of asset managers to enter the FX market. Among these are hedge-fund managers and a number of other participants with relatively short time horizons and relatively

little commitment to long-term currency management. Thanos Papasavvas, head of currency at Credit Suisse Asset Management (CSAM), says: “There is more leveraged money in the currency markets, which is not coming in necessarily with a strategic three-month or six-month view; they may have a one-week or even one-day time horizon.”

1890.photo.jpg

Thanos Papasavvas, head of currency at Credit Suisse Asset Management

Such short-term participants will tend to commit to a single position rather than a diversified spread, and will tend to react more quickly to short-term price movements – even where such movements occur within a long-term price trend. Their “straight in, straight out” activities thus increase FX volatility.

Liquidity management

Another automation-related issue is liquidity management. There has been a proliferation in the number of front-end electronic trading platforms provided to market participants by the major FX institutions.

There are a lot of these platforms. The snag is that they are just portals back to a much smaller number of liquidity providers, in the sense that a portal, or e-trading screen, can be placed on a trader’s desk much more cheaply and quickly if it is linked back to existing liquidity than if some new source of liquidity needs to be found to back it.

Paul Chappell, managing director of FX specialist C-View, says: “The proliferation of electronic platforms is disconcerting, in that they’re often piggy-backing off the same liquidity. One bank might be providing live streaming prices. These get dispersed through a variety of electronic means, which can suggest that there’s greater depth than there actually is.”

In a volatile market, if every e-platform takes in its limit of sell orders, they could all be pitching for the same buy-side liquidity. If that happens, only a fraction of orders would be filled. And that’s a pretty big shock that’s just waiting to happen.

PLEASE ENTER YOUR DETAILS TO WATCH THIS VIDEO

All fields are mandatory

The Banker is a service from the Financial Times. The Financial Times Ltd takes your privacy seriously.

Choose how you want us to contact you.

Invites and Offers from The Banker

Receive exclusive personalised event invitations, carefully curated offers and promotions from The Banker



For more information about how we use your data, please refer to our privacy and cookie policies.

Terms and conditions

Join our community

The Banker on Twitter