Is Ripple too far ahead of its time?
Ripple may have overestimated cryptocurrency's growth and underestimated industry inertia.
This piece was originally published on 11 February
- Swift's new GPI is still functionally quite limited, but is seeing much more uptake than Ripple in a shorter time.
- An immature cryptocurrency ecosystem and the regulatory climate are posing obstacles for Ripple.
- Swift and Ripple are both competing to roll out their systems to as many partners as possible. Safe, boring and easy options might have an advantage here.
As Swift and Ripple vie for the future of cross-border payments, an article published in American Banker has shed some light onto the obstacles facing both brands. At the same time, Swift has taken a step towards blockchain by linking its payments standard – GPI, or Global Payments Innovation – to R3's Corda blockchain.
It's likely to be a winner-takes-all market because getting everyone onto the same network is crucial for getting the most out of the technology.
As such, now might be a pivotal time for Ripple.
The uptake numbers are sitting in Swift GPI's favour though.
Swift says more than 450 banks, which handle over 80% of international payments, are using GPI to settle more than $300 billion in payments daily. By contrast, the Ripple network only claims some 200 members, and almost certainly only handles a fraction of a fraction of the volume of Swift GPI. And Ripple was founded in 2012, while GPI only took the stage in 2017.
If Ripple is so good, why aren't more banks using it?
What payments want
To understand where Ripple and Swift are currently butting heads, it's worth reviewing some of the basic elements of international payments.
At its heart, Swift is currently an extravagantly inefficient system in which banks bounce payments along pre-existing correspondent bank relationships until they arrive at their destination.
This is why it can take a week or more for an international payment to arrive at its destination and why they're so expensive. Payments pass through multiple banks on the way, and each will process the payment at its own rate and charge a fee for doing so.
The banks that will pass payments to each other are those with pre-existing correspondent banking relationships only. These relationships involve, among other things, opening nostro accounts with each other. These nostro accounts cover the value of the payments being sent through that pathway.
The goal for both Swift and Ripple is to make international payments faster and cheaper, but they do this in different ways.
What Swift GPI does
Swift's GPI is basically just a payment tracker. It gives more visibility into how long banks take to confirm payments, how much they skim from the top when they do and which bank a payment is currently sitting with at any given time.
This additional transparency improves the efficiency and cuts the cost of international transactions by bringing some competitiveness to the correspondent bank "marketplace." The bulk of the improved transfer speeds come from banks saying they'll try to process payments more quickly so they can look cool in front of the other banks.
Half of GPI payments are now credited within 30 minutes, Swift says.
If one of the stepping stones in a correspondent banking relationship on a major payment route starts looking like a hindrance, its erstwhile correspondents might look for a new partner in the destination to send payments more directly. Meanwhile, customer demands should also serve to pressure banks into seeking out the quickest and most cost-effective routes possible.
GPI does not directly improve efficiency. It indirectly improves it by pressuring correspondent banks to get their act together.
In time, Swift will be rolling out additional features for GPI to further bring its network into the modern world. For example, it recently added the "Stop and Recall Payment" function which allows banks to request a payment be stopped and recalled directly through the GPI interface. In the old days (i.e. prior to October 2018), a bank actually had to call up downstream correspondents – with an actual telephone – and then request that they cancel the payment and pass it back upstream.
Swift's plan is to reduce the frictions of international payments by adding these kinds of functions one by one. Eventually, you might have an international payment network worthy of the modern world.
"With GPI, we can see the charges banks have made along the way and how long they have taken to process a payment," said Manish Kohli, global head of payments and receivables at Citi. "Seeing how much each bank is deducting puts pressure on the ecosystem to reduce costs. If customers can see how long a bank is taking to move a payment, that forces banks to move toward faster processing or straight-through processing and speed up the time that it typically takes to send money across borders."
What Ripple does
Ripple is building out an interconnected network of banking relationships on top of the technical foundations of a blockchain, utilising digital currency to allow for instant near-zero fee transfers directly between any two members of the network.
This system does away with the entire idea of a correspondent banking "chain" and replaces it with a "web" that directly connects institutions. Because funds can be transferred instantly and programmatically with digital currency, there is no need for nostro accounts.
The first friction of this system is that it requires banks to convert XRP, the Ripple network's native digital currency, to and from fiat currency at either end of a trip. Ripple aims to solve this by maintaining a list of "preferred exchanges" a bank can go to for guaranteed on-the-spot conversions.
Both solutions are lacklustre in their own special way.
GPI's current limitations are quite clear. It's literally just a payment tracker, and the benefits it brings are dependent on the extent to which it can encourage banks to move faster and seek better options.
It doesn't necessarily make banks any better at processing payments after hours, it doesn't ease the need for nostro accounts, and it is basically just a neater and more informative interface for banks to process payments with.
Ahead of its time
Ripple's limitations are a bit more complicated.
Speculative chickens and eggs
First, a large part of Ripple's next-generation value comes from the use of digital currency, in this case the XRP token, whose value is driven purely by speculative market actions. The volatility isn't too problematic though, as payments can be processed and converted to fiat in the destination country almost instantly at whatever current prices are.
But this means having reliable gateways. Ripple has nominated preferred exchanges to be these gateways, such as Bittrex for US dollar exchanges. But no matter how you slice it, cryptocurrency exchanges aren't ready to start processing the world's payments, and any time it starts looking like the industry is evolving, you get something like the QuadrigaCX incident to remind banks to stay far, far away from even the most reputable cryptocurrency exchanges.
And for perspective, Swift GPI is already processing more than double the total cryptocurrency market cap in payments every single day. Exchanges are not yet in any position to handle even a fraction of that.
To get around this, banks could keep XRP holdings of their own as a kind of nostro pool equivalent. But even though it's theoretically meant to be a deflationary currency, it's clear the markets are far too fickle for that to be a sensible choice.
So, to get around that, you might use some kind of collateralised stablecoin as the network digital currency instead. But then where does that leave Ripple? That line of thought paints developments like asset tokenisation, new banking blockchain networks and central bank digital currencies as a more likely long-term solution.
Too much transparency
The second point of friction is that the Ripple blockchain is too transparent, and anyone can see every transaction and wallet address on the network.
"Banks have indicated they do not want to communicate how much money they have on an account to everybody else," said Wim Raymaekers, head of the global banking market at Swift. "We had to keep the conversations private."
It's a strange problem to have, considering the primary function of Swift GPI is to introduce a tiny element of transparency into the system.
"With GPI, we can see the charges banks have made along the way and how long they have taken to process a payment," Kohli said. "Seeing how much each bank is deducting puts pressure on the ecosystem to reduce costs. If customers can see how long a bank is taking to move a payment, that forces banks to move toward faster processing or straight-through processing and speed up the time that it typically takes to send money across borders."
If you have your heart set on a decentralised digital currency as the medium of exchange in international payments, it would also suggest that selective privacy coins are where it's at. Solving this problem on Ripple's end would require some significant system overhauls.
Regulatory obstacles also enforce the current correspondent bank model, and a lot of risk management is built into the legalese between correspondent banks.
"If you are a bank in Oklahoma and you need to make a payment to Mexico or Vietnam, unless you have relationships with banks in those countries already, you're going to need an intermediary bank," Raymaekers said. "If the two banks don’t have the relevant KYC information, they're not allowed by their compliance departments to send or receive payments between each other because they don’t know each other."
"In this situation, the only way to get to get a payment to a bank in Vietnam is to go through a correspondent bank. Whether that's using Swift or another provider, it has nothing to do with the technology," he says. "It's to do with legal, contractual and compliance agreements between banks."
Of course, in the long run, it's almost certainly going to be a technology issue, and that technology is almost certainly going to be blockchain. As the phrase "KYC" (know your customer) suggests, compliance is about identity rather than individual transactions, and blockchain and identity management are a match made in heaven.
But that doesn't help financial institutions looking for a solution today. All banks are facing the same limitations, and to a certain extent can only move as fast as all the others. International payments are a team sport, and there are no advantages to racing too far ahead of the pack.
There's no "i" in payments.
The global payment network isn't an amorphous blob. Rather, it's a collection of over 10,000 banks and financial institutions, and whichever standard they collectively elect to use is the one that will take over.
Without a strong degree of support from that collection of banks, the benefits of switching to a new standard are limited. This is because even if Ripple was the best thing in the world, banks still have to keep paying the cost of maintaining their old Swift network until there's enough support for an alternative that they don't have to use Swift at all. This is why Western Union found its Ripple tests to be lacking.
No matter how good it is, being the first to switch from Swift to Ripple would be like tearing out your landline as soon as mobile phones were invented. You're definitely on the right track, but it's going to be years before you can talk to anyone except technophiles.
"It [Ripple] needs 11,000 banks to be speaking the same language and have that technology embedded for a bank to successfully send payments to another bank," Kohli said. "That’s a very long haul for all banks to start adopting such alternatives," Kohli said.
Compared to the more grandiose long-term promises of blockchain and digital currency, GPI is almost hilariously unimpressive. "The biggest thing to happen to correspondent banking in 30 years," to use Swift's words, is just a payment tracker.
But "unimpressive" also tends to mean, simple, reliable, easy to implement and low risk, all of which probably suit banks just fine. It's difficult to get 11,000 banks to turn on a dime, so keeping things easy and uncomplicated is key to achieving that crucial uptake.
And as Kohli described, the industry initially thought GPI would only bring a little bit of transparency. But now it's generally accepted that this marginal improvement in this one area can also open the doors for improvements in cost, speed, convenience, security and cost.
And as Raymaekers concedes, international payments can still be improved considerably through more conventional avenues. Blockchain systems in general are still only as good as the information that goes into them, and if banks only update records once per day, the improvement a blockchain brings are marginal.
Blockchain and digital currency allows for an incredible degree of payment and payment data entry automation, but once again that's still off in the future.
"What banks need to do is to improve the speed by which they communicate the status of the account to you," Raymaekers said. "They realize that they have to change their back-office processes in order to provide more frequent updates."
Ahead of its time
The theme running through all of the main obstacles facing Ripple are that it came in a little ahead of its time.
Reliable XRP settlement is challenged by immature crypto exchange infrastructure, its level of transparency is too much too soon and the regulatory environment around payments still encourages the use of "safe" existing correspondent banking relationships.
There's a lot of inertia in the global payments system, and Ripple might have put its back out trying to move that enormous weight.
It may have still pushed hard enough to move Swift though. Plus, being ahead of your time isn't too bad if you live long enough.
Disclosure: At the time of writing, the author holds ETH.