It's time for a serious conversation about central bank digital currency design

How should the next form of money be designed? In case you haven't noticed, that question has suddenly become an urgent one.

The conversation about the nature of money took a fateful turn just over a decade ago thanks to Satoshi Nakamoto and Bitcoin. But it seems unlikely that Nakamoto, whoever they may be, would have imagined how quickly traditional, centralized powers in industry and government — the very institutions Bitcoin aims to circumvent — would try to adopt digital currency technology for their own ends. One way or another, it seems inevitable that some kind of digital currency will soon be in the hands of mainstream consumers.

Discussions about how to design future corporate and government digital currencies now appear to be progressing, albeit behind closed doors. If the potential users of these platforms don't engage now, they risk ending up with systems they don't like — or, worse yet, ones that violate their human rights.

While it is difficult to predict the fate of Facebook's Libra, or really any kind of corporate currency, central bank digital currency (CBDC) will be here soon whether we want it or not. A few banks already appear to be putting the finishing touches on their own systems, and it would not be surprising to see one go live this year. Even in the U.S., conversations about a "digital dollar" have grown considerably more advanced in the context of the COVID-19 pandemic.

So what? Well, that depends on what exactly we're talking about.

China's much-anticipated digital yuan, for instance, is much more likely to disrupt the world's financial system — and perhaps even prompt a response from the U.S. — than Cambodia's digital riel, which is expected to launch this year. And given the dollar's role as a global reserve currency, recreating it in a wholly digital format will have far-reaching consequences.

The impact of CBDCs will also hinge on how these systems are designed — and on whether people find them compelling enough to use. Today, we know very little about how even the furthest-along systems, like China’s DCEP and Cambodia's Project Bakong, will address fundamental design questions, particularly those related to user identity and privacy.

The bottom line: we won't know exactly how CBDCs might change the world until they are up and running. Fortunately, in many places around the world there is still time for the potential users of these future systems to influence how they actually get designed.

Isn't the dollar already "digital"?

Central bank money is already digital, just not for "retail" customers like you and me. Commercial banks hold electronic accounts backed by central bank reserves; that's how they settle payments between each other. The digital money you transfer with a swipe of your credit card or QR-code scan is not central bank money. It's an IOU from your commercial bank or chosen financial technology company. Physical cash, on the other hand, is an IOU from the central bank.

So what would it mean to create a "digital dollar?" Right now, there's no consensus on the answer to that question.

In March, a draft economic relief bill from Democrats in the House of Representatives proposed one potential design: the bill would have compelled the Federal Reserve to make central bank money — by way of electronic accounts — available to the public, and not just commercial banks.

Then there's the Digital Dollar Project, the high-profile partnership between a nonprofit founded by former Commodity Futures Trading Commission Chair J. Christopher Giancarlo and the consulting firm Accenture. The stated goal of the project, which published its first white paper last week, is to "advance exploration" of a U.S. CBDC. It also has a different vision of the digital dollar than the House Democrats did back in March.

To coin a phrase

The difference between the two proposed designs touches on something that Raphael Auer and Rainer Böhme at the Bank of International Settlements recently included in a discussion of "key technological design considerations" for retail CBDCs: whether access to the system should be "account-based" or "token-based."

The draft House bill describes a system in which the Fed offers retail accounts. Meanwhile, Giancarlo and Accenture's David Treat argue that the digital dollar should also be available in "tokenized" form. Here's how the new white paper defines what they mean: "the act of turning an asset, good, right, or currency into a representation with properties that suffice to attest to and transfer ownership."

By this definition, a physical dollar could be thought of as a sort of token. But the traditional term for that is "bearer instrument." Giancarlo and Treat say the nation should develop a "digital bearer instrument."

During a recent interview, Treat said the difference between account-based and token-based designs boils down to a "direct transaction between two parties versus a reconciled transaction between two parties." Using distributed ledger technology (DLT) to tokenize the dollar will make it more efficient, reliable and auditable relative to a more traditional account-based approach, he said.

But don’t mistake "tokenized" for "decentralized," at least in the sense that crypto enthusiasts tend to use the word.

It's not clear, for instance, if China's system will have a tokenized yuan. Either way, it's unlikely the PBOC won't maintain access to all the transaction data. Likewise, even though users of Cambodia's system will be able to send money peer-to-peer as long as they have a Cambodian phone number, the central bank will be the system's only transaction validator.

It's all about access

According to the BIS's Auer and Böhme, CBDC design decisions can be understood in terms of a layered pyramid.

The foundational layer asks: Will the digital currency represent a direct or indirect claim on central bank money? The next layer up asks: Should the system be DLT-based or based on traditional centralized infrastructure for real-time payments? The account versus token debate begins at the third layer.

"Once the CBDC's architecture and infrastructure have been chosen, the question arises of how and to whom one should give access," Auer and Böhme wrote. Should it be token-based or account-based? The decision, they say, has implications for the way the system treats users' identities.

Traditional approaches require "strong" identities, which Auer and Böhme described as "schemes that map each individual to one and only one identifier." Alternatively, the CBDC user would need to "demonstrate knowledge of an encrypted value — an option sometimes referred to as digital tokens," they wrote.

"A token-based system would ensure universal access — as anybody can obtain a digital signature — and it would offer good privacy by default," wrote Auer and Böhme. That sounds a lot like how we understand a traditional cryptocurrency, and indeed the risks would be similar. Users could lose all their money if their private keys are lost or stolen, and regulators and financial institutions would have a harder time making sure transactions comply with anti-money-laundering rules.

Treat and Giancarlo have a slightly different perspective. "Having a token model does not equate to tokens in the wild, if you will," Treat said.

In the context of a digital dollar, he added:"The operator of the system needs to be clearly defined, regulated and supervised — with all the control mechanisms that one would envision for a systemically important financial infrastructure." In that light, the design must stem from policy decisions that must be made regarding "the endpoints for the tokens, and whether or not the wallets in which they are held are part of regulated financial infrastructure."

In other words, at least in the U.S., whether a CBDC should be account-based or token-based is only the starting point for a much more complicated public conversation about how money should be designed. Let's get that conversation started sooner rather than later.