Banks boot up next-gen hedging bots

Automated FX hedging can save money and time, proponents argue. But corporates have qualms

  • Banks including Citi and NatWest Markets are offering corporate treasurers the ability to automatically execute FX hedges, citing greater client demand for such a service since the coronavirus pandemic.
  • The key selling points are efficiency and cost savings, banks say – with algos replacing human traders in order to reduce operational risks and maximise resources.
  • However, corporates aren’t wholly sold on the idea: many fear a loss of best execution, human oversight and governance.

A wild swing in exchange rates used to force corporate treasurers to hurriedly review their FX exposure and, if necessary, rebalance a slew of hedges.

Now, banks such as Citi and NatWest Markets are developing a new breed of automated hedging tools that, they claim, will do the job on the fly with little or no human intervention.

The sales pitch is simple: clients can plug their risk management systems into the bank’s proprietary trading platform, set the rules of engagement, and hey presto, the system identifies and executes FX hedging trades automatically.

The bank architects say it saves money for the client, and helps reduce the operational risk embedded in manual processes.

But corporate clients may take some convincing. They are concerned that a number of banks make it hard, if not impossible, to channel trades to other liquidity providers, effectively locking in the hedging flow from that client. This has raised warning flags over best execution.

And some corporates question how effective the safeguards embedded in automated processes are. Can a human intervene if the system attempts to execute an undesirable trade or fails to execute a desirable one?

Clients are also wary of releasing valuable data to banks – and to the third-party vendors that some banks use for processing the data.

As the head of FX at one large corporate says, bluntly: “Giving all of our exposure data over to a bank is something we would never consider doing.”

Systems that enable clients to automate hedging trades are not new. Barclays developed its first automated risk management platform six years ago, while Deutsche Bank’s Autobahn Maestro platform is eight years old.

What’s new is how these platforms can interface with the clients’ own systems. Previously, a client would send exposure data to a bank. Now, a bank can plug into a client’s own in-house risk system and extract data for itself (see graphic: What’s on offer?).

We see fully automated FX risk management involving the use of algorithms and machine learning a reality in a few years
Flavio Figueiredo, Citi

For example, last year Citi launched a service that enables corporate treasurers to connect their own risk management system to the bank’s FX Pulse platform. Citi can then pull data from the client’s system in order to work out what they need hedged and automatically execute trades for the corporate.

Citi’s system has limited autonomy in its trade and execution decisions. The technology still requires the client to dictate the size of trades, tenors, types of instruments, and execution venues. These rules are predefined between client and bank.

Complete automation in FX hedging is still some way off, observers believe. “At the current pace in development of bank, client and vendor technology, we see fully automated FX risk management involving the use of algorithms and machine learning a reality in a few years,” says Flavio Figueiredo, global head of FX corporate sales at Citi.

Automatic for the people

Algorithmic trading in foreign exchange markets is long-established, with an estimated 10–20% of spot FX trades now being traded by algos – contributing to around $200–400 billion in daily turnover, according to the Bank for International Settlements.

Rudi Alexis, head of foreign exchange product at Barclays, believes automated hedging is simply the next step in this trend, in much the same way as car manufacturers progressed from manual assembly to a production line of robots.

“Following 20 years of innovation, where speed and spreads have been the FX market’s focus, we have now entered an era of wider workflow automation where having a client calling a salesperson or even keying individual trades on a platform to hedge risk feels unnecessary,” he says.

The main draw for the client is cost savings, banks say. Automation means replacing (expensive) people with (cheaper) algorithms. Alexis points out that in “80% of cases” clients don’t need as big a team to manage foreign exchange execution.



Algos can also make hedging more efficient. The trade has to pass through fewer hands, which in theory, should speed up the process. This can help reduce the risk of human error, for example fat-finger mistakes or basic miscalculations. With high numbers of staff working remotely, these kinds of operational risks have increased, some argue.

“In 2020, Covid altered the day-to-day working of many treasury departments. Remote working, either from home or from split locations, not only impacted the ability of clients to access internal treasury systems, but also created operational risks around manual processes,” says Xavier Gallant, head of corporate FX and local markets sales for Europe, the Middle East and Africa at BNP Paribas.

Banks aren’t suggesting that all FX hedging can be done by bots – at least not yet. Some of the more illiquid, hard-to-trade currency pairs or instruments may still require a human touch. However, much of the most liquid trading in common spot pairs will be suitable for automation, proponents say.

The suitability of auto-hedging will also vary by client. Deutsche Bank says some corporate have automated “100%” of their FX hedging flows. Other clients use the service less, depending on their hedging needs.

Citi believes automation would suit fast-growing companies that are becoming more dependent on FX currency earnings generated overseas. The treasury desks of these companies are typically limited in size, and so they don’t have treasury operations in every country they are active in.

Scrubbed and ready

But before corporates can automate FX hedging, they have to ensure their data is up to spec. The output of a corporate’s risk management system must be able to feed into a bank’s own system, which can often require an overhaul of client data.

“What we’ve found over the years is that some clients don’t have that much confidence in their data for use in an automated process so there can be work that needs to be done to review or adjust their data and data processes before their workflows go live,” says Yannick Marchal, head of Deutsche Bank’s Autobahn Maestro platform.

A reason for deficiencies in client data is that some companies are using rudimentary or ill-suited systems to track their exposures, says Fabio Madar, global head of FX sales and structuring at NatWest Markets.

“Many corporates don’t have a good handle on their exposure data as they use systems like SAP or Oracle to keep track of their exposures, which are accounting systems not risk management systems. As a result, they lose a lot of their FX information,” he says.

Often, clients will have to use specialist tech vendors such as Kyriba and Kantox to help scrub their data before they are ready to start automating their FX hedging. Clients will need to factor in the added expense of cleaning up their data in the cost/benefit reckoning.

Showing your hand

Data challenges aside, clients may be uncomfortable giving all of their hedging flow to one bank. The automated offerings from Barclays, Citi and NatWest Markets do not currently enable clients to source liquidity from other venues. Barclays says clients are free to build that technology themselves and bolt it on to the system.

Corporates are concerned that by sending all their trades to one bank, best execution is potentially undermined as they can’t encourage other banks to compete with each other on execution price.

Indeed, such concerns are why one European dealer is not looking to offer automatic electronic FX hedging.

“It’s not something we’re proposing to our client base. The FTSE 100 guys don’t want to provide one bank with all their information, they want to have everyone in competition to prove best execution,” says the global head of FX sales at the European dealer.

BNP Paribas allows clients to trade on multi-dealer platforms such as 360T and FXall. Users of Deutsche Bank’s system can trade on a matched principal basis, where the bank stands between the client and another counterparty.

What we’ve found over the years is that some clients don’t have that much confidence in their data for use in an automated process so there can be work that needs to be done to review or adjust their data and data processes before their workflows go live
Yannick Marchal, Deutsche Bank

However, offerings that give corporates a choice over which bank to execute their trades with still leave corporates feeling anxious. Even if the treasurer ends up executing their trades with other banks on a multi-dealer platform, the bank offering the automated solution would still be privy to a large amount of the treasurer’s proprietary and sensitive FX data.

Corporates may be reticent to give their proprietary information to third parties such as tech vendors, too. Many would need to feel reassured that the necessary protections are in place.

“Even when we use things like Bloomberg, it’s all iron clad and there is legal documentation and contracts that are signed before we even input our exposure numbers,” says the head of FX at the large corporate.

Meanwhile, a lack of human oversight and governance processes within automated FX hedging systems is a deterrent for some corporate treasurers. They point out that automated FX activity isn’t always preferable to manual trading and execution.

For example, technology may fail to understand the optimum time to trade Brazilian real, or that Bank X is not suitable for executing trade Y as it doesn’t have a presence in that market. They highlight that technology is, ultimately, fallible.

NatWest’s Madar agrees that automation is not a silver bullet, but he re-iterates that the automated solutions currently on offer are designed to free up corporate treasurers from the easy to execute, liquid trades. Corporates can therefore spend time providing more value on harder to execute, illiquid trades.

“There are limits to automation as by definition we’re talking about executing currencies that are very liquid. As soon as you have illiquid currencies that aren’t widely traded electronically, automation doesn’t make as much sense,” he says.

Madar points out that the 20 most widely traded currencies in FX make up about 85% of trading volumes. High liquidity in these currencies means they are suitable for automated hedging.

“So while it’s true that if you had an exposure to esoteric, emerging market currencies you’d have to do it manually, overall a much smaller part of your business is now dealt with manually,” he adds.

Editing by Alex Krohn 

  • LinkedIn  
  • Save this article
  • Print this page  

Only users who have a paid subscription or are part of a corporate subscription are able to print or copy content.

To access these options, along with all other subscription benefits, please contact customer services -, or view our subscription options here:

You are currently unable to copy this content. Please contact [email protected] to find out more.

You need to sign in to use this feature. If you don’t have a FX Markets account, please register for a trial.

Sign in
You are currently on corporate access.

To use this feature you will need an individual account. If you have one already please sign in.

Sign in.

Alternatively you can request an indvidual account here: