Christian Catalini, co-founder and chief economist of Diem Association (formerly Libra), presented at Luohan Academy's Frontier Dialogue. Darrel Duffie, a Professor at the Stanford Graduate School of Business and a distinguished fellow of the Luohan Academy, discussed Catalini's presentation. The following texts features transcripted excerpts from. It has been lightly edited for clarity and length.
Speaker presentation by Christian Catalini
It's a real pleasure to be here with all of you today. And I'd like to give a quick overview on some of the ideas that we've been working on through the Diem journey.
When we designed the MIT Digital Currency Experiment which was around the year of 2014, as can be seen on this map, the Google trends for some of the keywords was flat. It's gone through different waves. And I think in parallel with kind of rising interest in cryptocurrencies and blockchain, there's also been a rise in interest in Central Bank Digital Currencies (CBDCs). Part of this was probably accelerated by the first version of our white paperin the summer of 2020, and the reaction to what I've described as a very naive first version of the Diem vision.
But taking a step back, I think what's really interesting, and I see Joshua Gans on the call today, my coauthor on this piece, is that at its heart, blockchain technology really allows for new forms of market design. This is really driven by a reduction in two fundamental costs that the technology reduces, what we call the cost of verification, which is very intuitive, essentially the cost of verifying the digital information is accurate and up-to-date, and the cost of networking. So, for the lack of a better term, we really connect this to the ability to bootstrap [and operate] a digital ecosystem without assigning control and market power to a centralized intermediary.
So maybe a better way to talk about the cost of networking is really about interoperability. These are new types of networks that are highly interoperable, and on top of that, can be programmed to perform different functions. I think we've seen how interoperability has driven a wave of innovation in a number of different services with the internet before, but what's happening with cryptocurrencies is that to some extent, this combination of interoperability and programmability is going to a number of different verticals.
You have decentralized finance applications, which are essentially taking interoperability to financial markets and allowing a number of digital assets to trade and be exchanged in novel forms. You have a similar type of change happening in the creative industries and digital content with NFTs or non-fungible tokens. And all of these are following a very similar narrative of removing intermediaries from the picture or changing the nature of intermediation in financial services payments and of course those into digital distribution of content online.
Now, one of the challenges with the removal of intermediaries is that intermediaries play a very important role in society. Often, they're the carriers of society's values and principles. And so, in a piece with Jai Massari (the incoming SEC chairman), for example, we highlight how Defi will need to reconcile some of the capabilities of the technology with the ability to regulate markets and ensure that insiders don't exploit information that they have access to or try to manipulate the market to their own advantage, a number of protections that we've come to enjoy traditional financial systems, but haven't been ported back in the cryptocurrencies and blockchain space.
At its core, you may think that this is a technology, it's just another instance of software, but in reality, there's a crucial interplay here between code, software and complimentary institutions coming and allowing the technology to really reach its full potential.
Now there's a number of key long-term benefits of the technology. I've already stated higher interoperability. That also often comes with lower barriers to entry and the ability of really reducing switching costs and locking-in for different market participants. At its core, it's a technology that allows you to un-bundle products and financial services that are bundled together in institutions such as banks, and also there's a lot of potential and privacy. I think there's a number of new designs that when it comes to privacy allow you to embed a privacy principle from the ground up. It also comes with increased transparency and accountability. That can be a double-edged sword depending on how the technology is implemented.
At a very high level, how Stablecoin really tracks its referenced assets (for example, the US dollar) is a very interesting topic, but it's completely not correlated with other crypto-backed uncollateralized Stablecoins. It's likely to be something that looks a lot like CBDC issued by the public sector. I think there's a spectrum of how reliable these tools can be in achieving a robust payment and financial system.
Now, on the Diem front, we've incorporated some of the gold standards from the financial sector, and really trying to deliver on a payment tool and a payment network that is highly reliable and can really withstand stressful market conditions. But more important is our commitment to phase out our own coin. The key innovation of Diem is that it may support the CBDC. And in that case, essentially Diem just becomes a payment system on top of public sector rails.
Often people ask about the business model. The business model does not rely on interest on the reserve, so there's no point of running the coin [to let users keep the coin]. In the long run, and in fact, the coin will be replaced, again, with a public sector asset. It's a transactional network, and so the business model will rely on charging very small transaction fees to its participants.
Now, because of programmability, you could imagine slightly higher fees for added value transactions. Think about conditional payments, which are a big opportunity for government [as well as] citizen to make cash transfers in situations like the COVID recovery phase.
When it comes to the network, we moved away from proof of work. I mean, this has been kind of in the headlines in the last few weeks, a lot of concerns about the environmental impact of Proof-Of-Work (POW) based networks. The reason actually why we did not choose proof of work beyond the environmental impact was also because proof of work does not really have the right types of incentives for ensuring upgrades in the quality and availability of this service.
And so, Diem is sort of a hybrid network where, to some extent, we're relying on the off-chain reputation of the initial members to run and secure the validator notes. And this is kind of where the association members come into the picture. They're there to drive utility and also secure the validators, but over time, this is really not designed to be a walled garden. It's modeled after the open technology standards of the internet. It forces interoperability between different wallets. So, users from a small wallet will benefit from access to the user base on a larger wallet with much stronger network effects. The same is happening on the merchant side. So, a merchant could to switch between different payment service providers and kind of retain all the functionality of a pay with Diem flow. The idea is really to drive competition and to use that competitive force to lower prices, new products and services, and really, improvements in quality.
Now, relative to existing Stablecoins, this is really a network that is optimized for payments, and we've made a number of changes from the early version of the white paper to really account for all of that. I won't go into too much detail, but I want to highlight a few key questions.
So why build a Stablecoin at all? Well, you should think of this as a retrofitting exercise. We have slow, expensive payment systems, just like the card networks which have a number of legacy costs including interchange. By retrofitting them with much faster, more efficient interoperable digital rails, we can really accelerate progress on that dimension and reduce really fragmentation in payments, and then in the medium term, just programmability. And as a result, there's a lot more you can do when you add conditional payments, expert services and other functional IT on top of the network.
What are we innovating on is higher standards for consumer protection, the design of the Reserve, and figuring out a design that does not interfere with macroeconomic policy and financial stability. We don't think that that's kind of a sustainable one, and in fact, we want to work to get a number of local governments in trying to find a way for the network to be accessible without threatening current substitution or other kind of key risks.
The other big area of innovation is robust financial crime and compliance standards. This is going to be the first network that not only would implement sanctions at the protocol level, but would combine things like the travel rule and other measures to combat financial crimes.
Now, we think of Diem very much as a compliment to the public sector journey to our CBDCs. Some people have compared it to the collaboration that you've seen in space between NASA and SpaceX. But if you look at some of the documents coming out from the Bank of England, in their discussion paper on CBDCs, they really stressed that a CBDC without private sector involvement is unlikely to really meet their design principles. And this is something that we also believe in as a driver of re-complementarities between the public sector really building the core infrastructure and then the private sector driving innovation and use cases on top of it.
So very high level, what are some of the comparative advantages? Of course, as to the public sector and monetary policy, they are providing stability and preserving value. There's no point in having a Stablecoin If you do have a CBDC token that's clearly superior and provides better consumer protection.
If you think of the public sector providing the core settlement there, then on top of that you can have all sort of additional protocols that take advantage of that and really customize that network for different private vehicles.
A really important role for the public sector is also setting rules across those dimensions like consumer protection, financial stability, competition, market integrity, and also availability of the infrastructure under stress conditions, forcing and really encouraging interoperability between different networks. So, if you have a CBDC token, can it be used across many different networks that are competing with each other? And on a global perspective, harmonizing legal frameworks. For example, the work by the FSB (Financial Stability Board) on global Stablecoins has been extremely helpful in terms of driving some degree of convergence on some of these topics.
On the public sector side, this is where really being able to develop the consumer business facing products and experiences, it's one where the private sector can move much faster. Experimenting with programmability is also something that probably a CBDC token may have a hard time given the risk involved, and also accelerating time to market.
So again, going back to the picture of the early internet, this is an era where public and private sectors are actually compliments. And I think building on that history, countries that are really interested in accelerated their journey should explore these combinations given that, on the other hand, you do have more permissionless cryptocurrency networks that do not provide some of the same protections and are still growing at a very fast rate.
Discussant presentation by Darrell Duffie
It's a great pleasure to be here, especially in front of this group, and especially to have the chance to discuss Christian's work, very important work at Diem. He and I have had a lot of conversations about this over the past few years.
I want to sharply focus my remarks on one theme, because I see a lot of influencers in the audience and I want to influence them with one question, “Why can't banks do this?” I'm going to try to make the case banks are perfectly positioned to do this, but they won't. It's not that they can't. And let me get into this.
So, here's your prototypical payment situation. Alice needs to pay Bob, the baker, $8, how is she going to do it? Well, for centuries, Alice sent a message to her bank saying, "Send the $8 out of my account, put it into Bob's account." Now, today, from Alice's point of view, this is working fine. From Bob's point of view, it's not working at all well. He's paying huge fee and he's getting the money slow. The system is not working well. And there are similar problems on the B2B side, on the cross-border side, and in many other places.
There's nothing wrong with this diagram. This wiring diagram works perfectly. Alice, as customer one of bank one, can get the money to Bob, who's customer seven of bank three, through the current payment rails in principle, at least according to this diagram, without difficulties and inexpensively. The technology exists today to do that, but it's not happening.
Here is evidence of why it's not happening. This is a bit complicated. It's from the McKinsey Global Payments Report, the latest one. It's a series of very useful reports. What I want you to take a look at is first the headline. The cost of payments in the US as far as the users go, that's payment revenues, is 2.4% of GDP. That is not a small number, and that's in North America. In Europe, Middle East and Africa, 1.3%. What accounts for that difference? It's not that north America has much, much better payments than Europe. It's simply that you're dealing with incentive issues with respect to margins and two-sided markets.
For example, speaking of two-sided markets, let's look at the credit card segment in the US. I'll just spend a little moment on that. 33% of the payment stack in North America is on consumer credit cards, 10% in mostly Europe. Why? Well, because in North America, there's an enormous markup on the use of credit cards. Interchange fees are well above 2%. There's something that's not working right in the US system.
Now I'm going to finish my remarks with my diagnosis of the problem. The main problem is that you have a situation in which the banks are capable technically of doing this, but they don't have the proper incentive to do it and there are coordination failures. You have walled gardens, which Christian mentioned, and you have a Rochet and Tirole’s two-sided market situation in which the payment service providers face consumers on one side and merchants on the other, and they have the opportunity to stick it to the merchants in order to profit from them and give the consumers a relatively painless and low-cost experience.
Why are central banks in most countries today is deeply against letting disruption occurring in their payment systems, whether they should let in fintech payment service providers like Alipay and Tencent - pay, let them have central bank accounts, or encourage private Stablecoins like Diem, or go to the “nuclear option” of introducing CBDCs? Most central banks do not want the responsibility for making these decisions. They do not want the responsibility for protecting data privacy, for doing anti-money laundering, KYC (Know Your Customer, a technical term in the anti-money-laundering business), and they definitely don't want to disrupt banks. They would love to have the banks just take care of this problem.
Central banks keep saying when they say if they design a CBDC, they will do it without disrupting banks. I think that's a misunderstanding. The whole point is to disrupt banks. We need to disrupt banks in order to get proper payment services at reasonable prices.
One of the reasons that central banks say they don't want to do this is that if they disrupt banks, then banks will not have a cheap deposit source of funding, and therefore credit provision will go down. Let me repeat that. We're all economists. Do you really think that it makes sense that if banks are subsidized on the deposit side, that they're going to make more loans? If you were at Chase and you had the opportunity to make a losing loan, would you do it because somebody on the deposit side said, "Go ahead. We're going to make profits over here in payments and deposits. You can go ahead and lose money on the loan side." No, it doesn't make any sense at all.
Another reason that central banks say they don't want to do it is that they're worried about creating Bank-runs (i.e., people cashing out of bank out of fearing the safety of their deposit). I think that problem is not as serious a problem as it's made out to be.
So, we're in this situation now where central banks have to come to grips with this very serious problem. If you readGovernor Brainard (Lael Brainard, member of the Federal Reserve Board of Governors) the speech this week, you'll see that eventually the United States is coming around to the view that it's going to have to disrupt banks and introduce a CBDC if the banks don't do it themselves.
The banks have very profitable franchises, and if they don't move quickly, they're going to lose some of those. And I think you will see some movement, but I don't know how to handicap this race.
Let me finish by just a couple of remarks on Christian's comments about moving to faster programmable smart-contract based payment systems. The banks can do this as well. There's no reason that bank deposits can't be tokenized and made into smart money. Let's be cautious about bifurcating the payment system into several chunks, which is not efficient. There are lots of enormous network externalities, positive network externalities in payments, and having everything run one set of payment rails, or as nearly as possible, is actually a much more efficient way to do it.
Let's also be cautious about what we do when we make programmable money on smart contract settlement. There's a really good paper by Rob Townsend, Michael Junho Lee, and Antwan Martin that draws some cautions, and I'm actually working on this problem as well. There are unintended consequences of balkanizing payments and settlements by using programmable money and smart contracts. Not that it can't work well, but it needs to be done thoughtfully and carefully. [I recommend that you] read Rob's paper.
It's also the case that when you split payment networks between private and central bank, private Stablecoins, central bank digital currencies, and bank rail payments, you are increasing the need for pre-funding each of those payments systems with liquidity, and the total amount of high-quality liquid assets that need to be held in the form of a payment medium goes up a lot.
So, let's be cautious about these potential unintended consequences of moving away from a bank rail payment system. Much better would be somehow get the banks to do this, whether through regulation or incentives, but I'm not optimistic that's going to happen. And regulators in central banks need to get ready to come in with their own CBDCs, if necessary, to get movement in this area.
 Financial Stability Board, 2020, Regulation, Supervision and Oversight of Global Stablecoin Arrangements
 Garratt, R., Lee, M., Martin, A., & Townsend, R. M. (2019). Who Sees the Trades? The Effect of Information on Liquidity in Inter-dealer Markets. The Effect of Information on Liquidity in Inter-Dealer Markets. FRB of New York Staff Report, (892).
For more information, please visit Luohan Academy's youtube channel: Luohan Academy