A recent paper by Sweden’s central bank, Riksbank, explored the potential climate impact of retail central bank digital currency (CBDC) models. While it found the energy usage of a DLT-based e-krona could be similar to that of cards, it depends on the design, particularly the degree of decentralization. It’s assumed that volumes would be on the scale of cash usage.
The authors outlined three models, including a centralized, semi-centralized and decentralized model. In all cases, the central bank is the sole issuer of the CBDC and verifies all transactions.
From an energy perspective, the paper concluded that a semi-centralized blockchain might be desirable, because a larger number of nodes for a more decentralized network would use more energy. Any blockchain adopting Bitcoin-style Proof of Work is out of the question, but Proof of Stake and any flavor of Byzantine Fault Tolerant (BFT) consensus is likely palatable.
A centralized CBDC would use similar infrastructure to today’s payments. While the authors did not rule this out as an option, the existence of this paper and previous research points to a likely adoption of DLT.
At the other end of the spectrum, a decentralized version would allow many non-central bank actors to have a high level of access to data and participation in the network. Each node would only store a subset of data related to clients of the intermediary hosting the node. This route was trialed by Riksbank in 2022 using the Corda enterprise blockchain. In this paper the authors suggest a public permissioned network would be appropriate and used Algorand as one example.
A semi-centralized CBDC model
A key advantage of the intermediate path, a semi-centralized CBDC model, is that each node would have a copy of the ledger eliminating single points of failure. There would be a relatively small number of payment service providers who have access to the network, limiting the scope for attacks. Riksbank would still operate the core system. The authors suggested a private permissioned network.
However, the specific example mentioned for the semi-centralized CBDC model was Hedera, which is best known as a public DLT network. While end users can engage in the public network in a permissionless manner, nodes with write permission are restricted to governing council members. In the Riksbank scenario, if payment providers operate the nodes, one has to assume the authors were imagining deploying a separate private network based on Hedera’s technology.
Rather than performing raw energy measures themselves, the authors relied on a 2021 report that assessed the green credentials of various blockchains. Hence, they highlighted that the study should be considered an initial exploration. Another key factor in the energy usage is whether the servers are located where it’s possible to use renewable energy.
The authors concluded, “All in all, the environmental footprint of digital technologies escalates with the involvement of multiple actors, increased usage of nodes, servers, and backup servers, and the adoption of more intricate protocols.”
