Q: How does the correspondent banking model have to change and modernise in light of new technologies – such as blockchain – and the digitalisation of trade?
A: Customers expect more today from their banking experience, and correspondent banking is no exception. Correspondent banking has to respond to changing customer expectations – for instance, by banks offering a faster, more predictable and more transparent means of making cross-border payments. We are working hard on this front with gpi, our global payments innovation service, which is predictable in terms of timing and which allows customers to track their payments.
We have already seen more than 1 million gpi payments in total, and it works. Customers using it know how long a payment will take and how much it costs, and even though it’s still very early days, the vast majority of gpi payments are reaching the ultimate beneficiary (not just their bank) within one day. Most are even reaching end beneficiaries within a few hours.
Over the next few years, we expect gpi to become the norm for correspondent banking, at least for payments made between major economies. Beyond predictability, speed and transparency, customers also want information, accessibility and flexibility – and here, banks’ and fintechs’ ability to exploit the advantages of open application program interfaces [APIs] and overlay services is going to be key.
Blockchain is exciting – but most blockchain work, like our own, is at proof-of-concept stage – so whether blockchain has a big role to play at the core of correspondent banking remains to be seen. If it makes sense, we’ll do it.
Q: What role do you see fintechs playing in trade finance? Will they help more small and medium-sized enterprises [SMEs] to export for the first time?
A: I would expect fintechs to play an enabling role, collaborating with banks to develop solutions for SMEs. We are used to seeing fintechs being portrayed as disruptive competitors to existing players, but underneath the hype there is a very collaborative relationship between fintechs and the banking sector. And ultimately it makes more sense for this to continue, particularly in areas such as trade finance where financing is required. If it’s more efficient, safer and cheaper for SMEs to export, they will likely be more inclined to do so – but trade finance isn’t the only consideration; SMEs are likely to be key beneficiaries of services such as gpi.
Q: Tell us about the moves to facilitate instant payments in the US and other markets and how Swift is involved.
A: We are very much engaged in the US move to instant payments, both on the standards front, and in developing solutions that will enable our member banks to re-use their local Swift infrastructures. What is notable in the move to instant payments – which is really picking up speed at the moment – is that quite divergent approaches are being taken in different markets.
The chosen models vary enormously. They range from bilaterally cleared models running over dedicated networks with instant settlement through the central banks (as is happening with NPP Australia, where we are providing the network, the orchestration mechanism and the addressing database) to centrally cleared solutions with deferred netted settlement that enable connectivity through multiple networks.
In the euro area we are seeing another two models: the Eurosystem’s TIPS system, which clears centrally and settles gross and instantly, and the European Banking Authority's RT1, which is prefunded and settles on a net deferred basis. Another option might entail bilateral clearing and net settlement. At the end of the day, it is up to each market to determine the best format. We stand ready to support any model that our communities around the world want to implement.
What’s also exciting about the move to real-time in the US and elsewhere, is how this will help speed up cross-border transfers. Domestic payment market infrastructures of course have a critical role to play in facilitating cross-border payments. As soon as an international payment reaches a domestic or regional market, payment market infrastructures typically come into the picture for local clearing and settlement. As more real-time domestic systems emerge and implement local market practice for gpi payments, banks will effectively be able to tie real-time domestic systems together, thereby creating a near-real-time experience for their customers.
Q: Cybersecurity has been a critical issue for the industry and for Swift. Are banks doing what they need to on the security front and how will Swift’s payment controls service assist?
A: For the most part, banks are ahead of other industries in relation to cybersecurity and in understanding cyber threats, but there is always more that can be done. Just over a year ago Swift launched a customer security programme to help banks protect themselves against cyber threats and increase cyber security co-operation within the financial industry. More recently we introduced a set of security controls, some of which are mandatory and will set a security baseline for banks.
We are asking customers to attest to their level of compliance with these mandatory controls by the end of 2017. It is an important undertaking for the industry and the resulting transparency will benefit everyone in terms of cyber risk management.
A further measure, payment controls, will be available in 2018. Payment controls will bring additional safeguards to ensure payment instructions are in line with business expectations and don’t represent a significant or unacceptable business risk. Users of the service will benefit directly from the reduced fraud risk that it provides, and their correspondents will see benefits through improved certainty associated with the payments that they receive.
Q: What are the opportunities and risks in applying artificial intelligence [AI] to the payments business?
A: AI can provide the payments business with opportunities for process improvement, task optimisation, data classification and data understanding. It can also provide business insights that can be used to drive innovation, improve the effectiveness and efficiency of processes, and mitigate compliance and fraud risks, including cyber fraud. We see applications in a range of areas – for instance, natural language being used in payment processing, to strengthen the effectiveness of sanctions filters, or automate alert investigations for compliance or cybersecurity systems.
AI can also enable new services or solutions. For example, gpi collects data about the processing of payments, which provides opportunities for AI-based data discovery to optimise payments in new ways; for instance, enabling banks to offer new services that consider delivery timing or certainty, message fees or other factors.
I think we will also see AI play an increasing role in compliance and fraud detection – for instance, in our new payment controls service – where it is important to learn patterns of behaviour over time. The advent of open APIs is also creating huge opportunities – not only for fintechs, but most particularly for those already expert in handling big data (such as Facebook, Amazon, etc), who will be able to mine banking data using AI at little to no cost.
But we shouldn’t overstate its capabilities. First, AI requires data that is relevant to the problem and is complete and accurate; and it works best in areas that are well understood by people. Even when AI is used for data discovery, a human still needs to interpret the findings.
And there are risks. Systems that learn from big data do not always infer correct or accurate decision-making rules, resulting in the wrong decisions or systematic bias. This needs to be considered as part of any design and implementation, and is especially important in areas relating to regulatory compliance where model validation – the need to understand exactly how a decision model operates – is of critical importance.
