Project Dunbar: international settlements using multi-CBDCs

Project Dunbar brings together the Reserve Bank of Australia, Central Bank of Malaysia, Monetary Authority of Singapore and South African Reserve Bank with the Bank for International Settlements Innovation Hub to test the use of central bank digital currencies (CBDCs) for international settlements. The project will also explore the development of technical prototypes using distributed ledger technologies of Corda and Quorum.

8 Nov 2021

Project Dunbar brings together the Reserve Bank of Australia, Bank Negara Malaysia, Monetary Authority of Singapore, and South African Reserve Bank with the BIS Innovation Hub to test the use of central bank digital currencies (CBDCs) for international settlements. Learn about the expected benefits of multi-CBDCs and how it could make cross-border payments cheaper, faster, and safer. See how we faced and solved critical challenges in designing a multi-CBDC shared settlement platform.

Improving cross-border payments with multi-CBDCs

Cross-border payments today can be slow and expensive. Unlike domestic payments, where banks can pay each other directly on a single national payments platform, there is no single international platform today for cross-border payments and settlements.

Instead, the correspondent banking model is used, where banks hold foreign currency accounts with each other. A single cross-border payment could pass through multiple correspondent banks, using the foreign currencies held with them. Each leg of the overall transaction takes time and effort to process, with fees levied that add up quickly and are passed on to customers, resulting in slow and costly cross-border payments.

A multi-currency common settlement platform would enable transacting parties to pay each other in different currencies directly, without the need for intermediaries such as correspondent banks.

On this common shared platform, multiple central banks issue their digital currencies (CBDCs), which can be used by participating commercial banks for payments.

Participating commercial banks can hold and transact in the CBDCs issued. This includes CBDCs in both local and foreign currencies. Banks that are connected to the domestic payment system can exchange their central bank balances for CBDC. Non-local banks (those licensed as banks in other jurisdictions, but not in the local jurisdiction) which are not connected to the domestic payment system can exchange CBDC with other banks.

In this way, all participating banks can exchange and hold different central bank balances in the form of CBDC, including those to which they usually would not have access. Banks can now pay each other directly in all the different CBDCs of different currencies

A multi-CBDC shared platform will enable international settlements on a common platform, streamlining the flow of cross-border payments and making payments faster, cheaper and safer.

Designing and developing a multi-CBDC platform

Project Dunbar brings together the collective knowledge of central banks, commercial banks and technology partners from earlier research and development work on digital currencies, to design and develop a multi-CBDC platform. The primary work took place over nine weeks, with three concurrent workstreams.

The Design Workstream is led by Accenture with support from Temasek and focuses on developing requirements and designs. Workshops with the participants alternate between open brainstorming and structured discussions to develop innovative and practical solutions. Accenture's in-house Capability Model for Financial Markets Infrastructures is used as a framework to ensure comprehensive coverage of all perspectives critical to a payment infrastructure.

There are two parallel Technical Workstreams to develop technical prototypes on two different distributed ledger technologies of Corda and Quorum. The Corda Technical Workstream is led by R3 and the Quorum Technical Workstream is led by Partior, with support from DBS, JPMorgan and Temasek. The Technical Workstreams developed prototypes to support the proposed requirements and designs from the Design Workstreams, leveraging platform functionalities already developed for digital currencies, and enhancing them to support the specific needs of a multi-CBDC platform.

Challenges in implementing a multi-CBDC platform

A multi-CBDC platform for international settlements could significantly improve cross-border payments, but there are challenges and questions regarding implementation that need to be addressed.

How do we get multiple central banks to share a common platform? Can we resolve the national security concerns that come from sharing critical infrastructures with other central banks?

Central banks sharing a platform

How do we allow non-local banks to access and make payments with CBDCs from countries where they do not have a presence? Can central banks trust banks from other jurisdictions when they do not supervise those banks directly?

Access to CBDCs by non-local banks

Payment regulations are different in each country. How do we simplify the cross-border payment flow while respecting these differences? How do we delineate the jurisdictional boundaries for different parts of a single cross-border payment?

Cross-border payment crossing multiple jurisdictions

These critical challenges may have impeded the development of shared international settlement infrastructures previously, but can now be resolved with modern technologies.

Addressing critical challenges

Distributed ledger technologies' decentralisation capabilities and the programmable nature of smart contracts and CBDCs, tied together in a sophisticated platform design, have produced an elegant solution for central banks to share a common settlement infrastructure for cross-border payments

A shared platform implies a level of universality. Features and capabilities of the platform are universal and available to all participants. Rules and policies are applied universally and fairly across participants. To enable this universality, governance structures and decision-making powers must be designed to ensure that the diverse stakeholders are represented, and collective decisions are made fairly and equitably.

Good fences make good neighbours. A shared platform amongst equal participants is viable only if participants have a sufficient level of autonomy and control over their domain areas. Central banks demand complete sovereignty and autonomy in: (i) areas of critical functions, such as issuance of currency; (ii) the application of "local" rules and regulations at the currency- and jurisdiction-level; and (iii) data, including privacy and selective sharing of data.

The prototype platforms are designed to give central banks autonomy within the boundaries and parameters of a universal platform-level framework. Strict technical controls that even a super-user operator of the platform cannot violate are implemented to guarantee this autonomy.

In addition, the platforms are designed for resiliency, and failures at the individual level are isolated so that they do not cascade into platform-wide failures, to ensure that autonomised regions and components of a central bank on this shared platform are not infringed by failures of other central banks.

A "hybrid" CBDC access model was applied to enable non-local banks access to CBDCs even in countries in which they do not have presence. While the model was designed for non-bank corporate and retail customers in retail CBDC arrangements, the concept is relevant and can be applied similarly for non-local banks.

With this hybrid CBDC model, non-local banks hold CBDCs representing a direct claim on the issuing central banks. "Sponsoring" banks, which are subject to local regulations, perform customer due diligence processes on the non-local banks. This includes onboarding and know-your-customer (KYC) processes as well as suspicious transaction monitoring and anti-money laundering and countering the financing of terrorism (AML/CFT) processes. This allows application of control processes without the need for changes to existing regulatory policies.

While the need for correspondent banks is eliminated in the settlement process, intermediaries, in the form of "sponsoring" banks, continue to play a role in control processes such as KYC and AML/CTF measures. This limits efficiency gains of eliminating intermediaries and poses a challenge for commercial models and incentives for banks to play such sponsoring roles. Various possibilities exist, including reciprocal arrangements, obligations imposed as conditions of access and fees, and will need to be further evaluated.

Another option is "direct" CBDC access, where non-local banks can hold and transact directly without the need for sponsoring banks. This would require mutual reliance by the central banks on the supervision of non-local banks by their home regulators. Processes can be further streamlined and made significantly more efficient, but this would require changes to existing regulatory policies, and harmonisation across participating central banks.

The payments process flow was separated into settlement process and non-settlement processes. The settlement process refers to the movement of funds between the sending and receiving banks. Where transacting banks are in different jurisdictions, the settlement process occur across borders. The settlement process is governed by a common set of platform-level rules.

Non-settlement processes such as KYC and AML/CTF measures are subject to individual countries' regulatory policies. For example, when a Singaporean bank transfers AUD to a Malaysian bank, the Singaporean bank, Malaysian bank and their "sponsoring" Australian banks will be subject to Singaporean, Malaysian and Australian regulatory policies, respectively. Separating out the non-settlement processes allows them to be processed in-country and subject to their local regulations.

This separation of processes enables clear delineation of jurisdictional boundaries, adherence to regulatory policies of different jurisdictions, and streamlining of settlement processes.

Successful development of prototypes

Project Dunbar worked with R3 and Partior to successfully develop prototypes on the distributed ledger technologies of Corda and Quorum, respectively. The prototypes proved the technical feasibility of implementing a shared multi-CBDC platform. While there are still further questions to be explored and challenges to be resolved, the successful development is a first step towards the vision of a shared platform for international settlements using multi-CBDCs.

23 Dec 2021

Project Dunbar worked with R3 to successfully develop a prototype multi-CBDC shared settlement platform on Corda technology. This proved the technical feasibility of implementing a shared multi-CBDC platform. 

23 Dec 2021

Project Dunbar worked with Partior to successfully develop a prototype multi-CBDC shared settlement platform on Quorum technology. This proved the technical feasibility of implementing a shared multi-CBDC platform.

A project report, with more details on the technical design and implementation, will be published in March 2022.