Community

Good post. The DTCC published quite a useful white paper in May in which they concluded that they couldn't afford not to move to the cloud, given the gap in security knowledge and capability. Having said that, particularly for smaller banks the complexity and impact on the current IT person/group shouldn't be underestimated. There is a good business opportunity for consultants who really understand a particular cloud platform (e.g. AWS), cybersecurity, and community banks or credit unions.

Carlo, thanks. This is a good, well-structured post, leading to at least as likely an outcome as any. There is a certain amount of arrogance about Ripple's stance that could come back to bite them. But a collaboration between SWIFT and Ripple would make sense if they can play nice together - they are radically different kinds of organization.

A few things to bear in mind:

1. SWIFT is a messaging system, not a payment settlement system. Settlement occurs through funding of nostro accounts.

2. SWIFT messaging covers more than just payments - FX contract confirmations, letters of credit, nostro statements, etc.

3. There is much more to SWIFT than a messaging infrastructure - the community, governance, standards body, etc are all needed in some form.

4. Adoption by a handful of banks (or even a few hundred) doesn't resolve the need for a truly global payments messaging infrastructure. Without a collaboration, there is a real market risk that enough volume will move from SWIFT to a blockchain-based alternative to make SWIFT unprofitable, while thousands of banks lack the resources to make the switch in the same timeframe - there could be a time during which there is no viable payment mechanism for smaller international banks. 

5. SWIFT has made a whole-scale technology transition before, when it moved to an X.25 based network. It was long and painful, but almost entirely without major problems. A move to a distributed ledger infrastructure is also eminently possible without major disruption.

My money is on SWIFT partnering with Ripple, Ethereum, or another consortium - partnership for transformation, not competition for disruption.

Some sparring is inevitable. In reality Ripple's experiments, while promising, are not valid comparisons even for end-to-end payment times (which is far from the only challenge - see my recent FinExtra post Is Blockchain Really the Answer to Global Payments?) The majority of SWIFT payments occur within 24 hours - and a good proportion within an hour or even minutes. Delays are typically caused not by the infrastructure, but by banking calendar mismatches, and incomplete or invalid payment instructions. Ripple has not yet proposed a solution to these - delivery of a smart contract doesn't ensure the beneficiary has use of the funds.

However, collaboration is essential if payments infrastructure is to be transformed. SWIFT won't be successful on its own, at least not in a timely manner. There are contributions to learning being made by many players - Ripple for sure, as well as R3, Hyperledger, and Ethereum. That Ripple are at SIBOS is good, and a little repartee increases bankers' awareness. They just need to not take it too far, and SWIFT need to avoid being too defensive. 

Well written summar of the Bangladesh incident and some of its implications. There is a potential role for distributed ledger technology, but I'm not convinced it would solve the underlying problem. The problems all lie on the periphery, the entry points to the SWIFT network. The same would still apply - somewhere or other there is going to be a transition from core banking systems and payments operations to the distributed payment delivery and settlement network. Blockchain is not a panacea, even though it will ultimately make huge contributions to improving security and efficiency of financial networks. The need for robust cyber-security policies and controls (both automated and operational) is not going away any time soon.

Well explained. Rules-based monitoring systems have to be set with very conservative thresholds, resulting in massive falst positive volumes. This creates very high operational costs, and also risk associated with "fault fatigue". As suggested in my blog How Industrial Operations AI Can Help with Banking Risk Management, banking can learn from industrial process management about anomaly detection in a machine learning environment. Several companies are working on this already including behemoths like IBM, and small startups like Amberoon.

Sanjeev and Joao, perhaps I should clarify my 10 year plan. I completely agree that a project or program with that kind of timeline simply cannot work. I would propose, though, that the technology culture of our banks has to incorporate a long-term view as well as a short-term one. Initiatives that address immediate point solutions must also be chipping away at the long-term issues. Otherwise, the long-term infrastructure challenges of banks will not be resolved until one by one they become urgent issues as the infrastructure crumbles.

Joao, I agree that large banks have a huge challenge with their organizational structure, their legacy technology infrastructure, and even their culture. That's why Ravi Kalakota recommends tactical approaches, which should be driven by the CRO but in partnership with technology, operations and creative external technology providers (e.g. Regtech). In the long term, strategic solutions to the technology infrastructure are essential, but they're not going to happen in one gigantic "let's simplify and replace everything" project. What is needed is a prioritized and phased strategy that has enterprise-wide support (from CRO, CEO, CIO and all lines of business), priority (it cannot be bumped by "urgent" projects) and funding (Board-level funding support). This is probably a 10 year program at the biggest banks! But it is essential.

In the meantime, though, there are urgent regulatory projects that require more focused MDM, development and partnering to address specific solutions. These are tactical projects. Good tactical projects are carried out in the context of the overall strategy, and are designed (even at higher cost) so that they contribute permanently to the overall strategy. This isn't easy. but it is worth the effort if the bank has sufficiently knowledgeable and skilled leadership.

Ketharaman, I respectifully disagree with your position. The most successful FinTechs in the long term will be those who understand and appropriately manage their regulatory risk, rather than hiding their heads in the sand. Managing risk may include accepting risk, as well as remediating, mitigating or transferring it. But risk acceptance (i.e. not doing anything about it) should be with full understanding of what the risk is, and determination that the company can handle future occurrence of the risked event. I think this is more like what Uber did, although perhaps not with analytic rigor. The lack of rigor in certain countries has been very painful for them (e.g. their need to retreat in Germany earlier this year).

Good post. It underlines just how much work will be necessary to significantly treduce TBML. You could also have discussed cooperation opportunities between banks, major import/export companies, customs authorities, etc., as well as overall supply chain modernization. That should keep everyone busy for the next ten years at least. But it must happen!

Supposing the use of private and shared permissioned blockchains became viable for high-volume clearing and settlement (given the scaling the challenge today), does the very immediacy of settlement create systemic risk of its own, in the speed with which irrevocable melt-down could happen?