The Monetary Authority of Singapore and Bank of Canada are about to release their findings in a project to join up their respective state blockchain initiatives: Project Ubin in Singapore, and Canada’s Jasper.
Ravi Menon, managing director at the Monetary Authority of Singapore, says MAS has been working with other central banks, including Bank of England, and with commercial banks to link national blockchains to enable instant settlement of cross-border payments.
“Cross-border payments and settlement of assets remain pain points,” he said in his opening remarks at this week’s MAS-backed Singapore Fintech Festival.
Operating across multiple currencies and assets, without a central trusted party, has made this cumbersome and expensive. Distributed-ledger technology holds the promise to address the problems, Menon says. Singapore’s Project Ubin has already found blockchain tech can be used to facilitate interbank payments, using a MAS-issued digital currency; and last year Ubin proved blockchain could net multi-party payments while preserving privacy.
This year, Ubin has made two advances: one is to use the technology to settle tokenized assets (which DigFin wrote about here); the other is to make cross-border payments quick and cheap.
Linking different blockchains
A group of presenters from MAS, J.P. Morgan and other participants in the study presented the basic findings of the Ubin/Jasper program during the Singapore Fintech Festival; the full report will be published this Thursday, with all the technical details. (Other central banks are also experimenting with cross-border payments, including Saudi Arabia’s.)
The teams considered multiple strategies, and presented one to the festival crowd: asset swaps using an intermediary, to facilitate atomic (small, single) corporate payments. In this case, they assumed Singapore had settled on J.P. Morgan’s Quorum as its blockchain protocol, while Canada was using R3’s Corda – the point being that countries are likely to end up with different ledger platforms, rather than one vendor’s solution conquering the world.
How could a remittance bank in Singapore effect a Singapore-dollar denominated payment to a beneficiary bank in Canada, which wants Canadian dollars? And how to make this happen in a decentralized-computing context, to preserve privacy of transaction information while ensuring against double spending?
Under the asset swaps+intermediary model, the beneficiary bank (let’s say in Canada) knows from its client to expect a payment from a remitter in Singapore, but has no correspondent banking network there.
To avoid a double-spend of the same money, both the remitter and the beneficiary banks need to transact the payment simultaneously; blockchain transactions don’t operate in sequence like a payment via SWIFT.
The beneficiary initiates the payment by encrypting details of the transaction in the equivalent of an escrow account (via a smart contract called a hash time-lock contract, or HTLC). It tells the remitter how to run a hash that can move funds into escrow.
At this point, the remitter in Singapore runs the hash to connect with an intermediary bank, using (in this case) Corda’s blockchain. The same intermediary will, off chain, digitally account for the changes in balances vis-à-vis both the remitter and the beneficiary, while simultaneously uses Quorum in Canada to reverse the flow of the payment, validating the trade back with the beneficiary bank, unlocking the escrow account and making the payment.
Part of the process is to designate the time allowed to complete the transaction, which is built into the HTLC. So in this example, payment-versus-payment must occur within T/2 (it takes each side half the allotted time to complete their side of the deal). If the funds aren’t released in time, the correspondent bank automatically returns all the books back to their original balances.
This sounds complicated but, if deployed at scale, holds a number of advantages over the status quo, says Naveen Mallela, head of digital for treasury services at J.P. Morgan, who helped present the Ubin/Jasper findings.
Using SWIFT still leaves a transaction open to error, as there is a gap between delivery and payment; and using blockchain, transactions can be fully automated and therefore run 24/7, regardless of time zones.
Corporate cross-border payments today lack transparency, involve delays, and cost $30 or more per transaction, according to MAS.
The system presented by Ubin/Jasper is similar to Ripple’s xCurrent, using XRP as a settlement token. The difference, says Mallela, is that this is backstopped by central banks, rather than by a private company, and therefore correspondent banks don’t need to pre-fund deals or hold capital in reserve to transact this way.
What the team presented is just one model under consideration: using an intermediary was the easiest way to reduce costs and mitigate credit risk.
Other models are now being studied in more detail. These include allowing parties to hold wallets with their central banks, or rely on digital cash tokens to move value across the network (akin to XRP).
Mallela says this project is not just about payments. “The bottlenecks [with the status quo] are not about where the money sits, but the enquiries” as corporations constantly want to know where the payment is, or as banks seek to run AML checks. This leads to a constant stream of manual interventions. But distributed ledger technology offers the chance to allow privacy as well as identifying flows.
“This is about distribution of information, not just distribution of payments,” Mallela said.