I'm so glad you said the dollar was already largely digital, I was sure that was true but constantly wondered if I was missing something.
That said, would a general federal money app be useful, even denominated in normal dollars? For example, set one up for each SSN number, then you can transfer money to it for stimulus, tax refunds, etc. If I've interpreted the news correctly – and I'm not sure I have – American stimulus cheques were actual, physical paper cheques, which seems crazy in the 21st century.
Why not just deposit funds straight into people's preferred bank accounts? But then you still have the matter of all the unbanked and underbanked, who probably need any such payments more.
India is trying this experiment, encouraging everyone to use a digital app for ordinary transactions as well as transactions with the state such as subsidies, benefits payment and so on. Right now, the issue is fraud. There are just too many holes in the system, not necessarily technical exploits, but practical ones that allow people to set up multiple identities or steal other people's identities. I'm not going to say that this kind of thing can't be fixed, but going digital is not a magic bullet.
Someone somewhere was also toying with the idea of stimulus money that would run out of time (use it or lose it). That could have been useful for the first COVID stimulus... and you could freeze stimulus spending if you noted inflation was getting out of hand...
French companies already kind-of do that with providing their employees with vacation checks and "ticket restaurant" (you can use them to buy food/pay for a restaurant if your company doesn't have a cantine... or to pay at the cantine?)
As I've said before, distributed blockchains, always and everywhere, take things that would work better if you had a trusted central authority, and then either make the system radically more expensive in energy and time (proof of work), or codify "my money works for me" (proof of stake). The "problem" that distributed blockchains are trying to address _is not a problem_. A chain of trust that ultimately reaches an entity that can be small-d democratically governed, that is answerable not just to those with wealth but to people who _don't_ already have a ton of money, is Good, Actually.
That said, postal banking also would be a good idea. And Felix Salmon has written for years about how digital payments ought to clear at par. And in that sense, a digital dollar operated by the Fed as a utility that's free to the public would be great.
Add on to this: Once you have everyone universally banked, you can start doing things like implementing UBI by way of just incrementing everybody's digital dollars account once a week. And you can do fiscal expansion and contraction by changing how much the UBI gives out.
Furthermore, you can create a Sovereign Wealth Fund that sits off to the side of the postal accounts, and implement Cory Booker's Baby Bonds concept by giving people shares in that. You also could scrap the 401(k) and all of its relatives (which have largely served to enrich the financial companies that operate the accounts for exorbitant fees) and just default everybody into an IRA that goes into a target-date tranche of the SWF.
Very nice post. At present maybe 95% of all my bills are paid digitally. I pay estimated taxes digitally. All of this is done in the absence of any kind of crypto currency regardless of who is/has established it. Once you set payments up, they are automatic based on the billing cycle.
I always ask the question, what problem is crypto trying to solve? I just cannot see the business case other than creating NFTs (collectible value) or speculation. VISA or MasterCard and perhaps big money center banks could get into crypto in a big way if they wanted to. At the end of the day, there has to be stability to digital currencies and right now there pretty much is other than inflation.
So no use really. It is disturbing, although sense eventually prevailed, to see economists, some anyway, jumping on bitcoin, blockchain, at a national level. Is it just to be novel or fashionable.
I was one of the reply guys asking you to weigh in on this, and this was exactly the kind of article I was hoping to read from you (and the conclusion I'd sort of assumed). Thanks!
I think there's one way that a central bank cryptocurrency could be very useful: improving the security of digital payments. Currently we do digital payments with card numbers - which can be trivially stolen, so we use a bunch of other identifying information too, except *that* can *also* be trivially stolen, so it doesn't really help. Twice in the past two years I've had a card number stolen after using it online - once by posting a payment to an anodyne account in the exact amount I made, presumably in an attempt to make me think it was just the other payment.
With crypto techniques you can't just repost a transaction like that with a different recipient, because the whole transaction is wrapped up in the encryption. And while it is possible to steal crypto keys, at least it's not trivial, since the relevant information is *never* broadcast, as opposed to necessarily broadcast in its entirety with absolutely every transaction you ever make.
In addition, a central bank running the chain avoids problems like chain takeover, reversing fraudulent transactions, and probably energy use. I know Bitcoin and Ethereum couldn't handle the transaction volume; but some of the alt coins claim throughputs adequate to manage a financial system; I don't think any have been tested but, again, I'll bet a central bank can pull that off as well.
True, but there *is* an intermediary for crypto: the blockchain maintainer, as an organization. And even with the supposedly "decentralized" blockchains, there have been a few rollbacks. A central bank could potentially rollback much more easily. I'll grant it's not a trivial problem, but there are possibilities.
At least _some_ transactions are already not done with card numbers, but rather with one-time keys that involve creating a time-stamped signature. (Apple Pay is the main consumer-accessible version of this.)
We ought to encourage more migration to this design, but Visa is actively hostile to it.
Yes, we might well not need a blockchain per se for these benefits, just decent digital signing techniques. Although Apple Pay is not perfect (yet?); my husband had his card number stolen and that was enough for some fraud to go through. I assume not every system for accepting Apple Pay is using the signing system (again, yet?)
I just don't understand Visa's hostility. They spend a LOT on credit card fraud. There would be a lot less, so more profit for them, so ...?
Visa and the banks that issue cards are able to collect higher fees from merchants for the worse technology, and they're able to force the merchants to eat the cost by way of saying they can't accept cards _at all_ unless they charge the same price regardless of the payment method. The card issuers actively incentivize people through reward programs to use more fraud-prone technologies (like favoring signature debit over PIN debit).
One argument, raised by scholars Todd Zywicki and Geoffrey Manne, is that swipe fees are a way to pass on credit losses to merchants. Yet as financial blogger Felix Salmon of Reuters observes, merchants don't also benefit from the much larger credit profits derived from fees and interest payments enjoyed by banks.
Swipe fees could be increasing due to rising levels of fraud and identity theft. That, however, would represent a case for shifting away from signature debit and towards PIN credit. But signature debit is more profitable for card issuers, which is presumably why there hasn't been aggressive movement in this direction. In Salmon's view, rising swipe fees represent pure rents for the effective duopoly of Visa and MasterCard, and the Durbin amendment would help shift the balance of power towards merchants and consumers.
</blockquote>
If the Fed came in and ended this kind of parasitism of commerce, turning payment processing into a utility like water and electricity, that would be great.
FWIW, that's not how we do digital payments in Vietnam (not exactly an advanced country) so I'm not sure why you need to jump to a central bank cryptocurrency to solve that problem.
All you have to do is switch from pull to push of money.
I think one of the things about postal banking is that it can to compete with profit-driven banking to get them to behave better. Is there anything about the behavior of payment systems that really needs reining in?
This seems a little confused on a few points to me:
1. There is plenty of reason for a CDBC to use a blockchain like system. What a CBDC does not need is a *proof of work* scheme. The purpose of proof of work/stake/etc is transaction *ordering*, because sequencing is difficult in the absence of a central actor. But "block chaining" exists to maintain a transparent, immutable ledger. Transparency and immutability would (at least potentially) be desirable properties of a CDBC.
2. Our existing payments system is terrible and expensive. All in credit card processing fees are often well over 200 basis points. That's absolutely absurd, and any reasonably implemented CDBC should be able to offer its services far, far more cheaply than that.
3. The potential benefit re: banking of a CDBC is to provide competition, and force banks to actually innovate in a meaningful way. Right now banks are the only game in town for storing your funds in an accessible, safe way. Since banks are very difficult and expensive to start, that means they effectively have an oligopoly, and can offer very low yields on savings accounts, etc. Right now, people only go to the banks for storage and payment - if the government provided that to people for free, the banks would have to do more to attract capital. And anyway, there's no reason for banks to attract individual capital anyway. The federal government is already insuring most of it, so we should just end the charade, store everyone's cash at the fed, and let the fed lend to the banks, and the banks just work on their actual business: lending. Right now there is this incredibly dumb conjunction between bank's ability to attract deposits via advertising and customer service, and their net interest margin, and simplifying that whole game would be immensely valuable.
One distinction I think about is between the app the protocol. Tech apps like proprietary protocols because it locks you in. Epic EMR tried to do this before Obamacare, Twitter/fb both abandoned open standards like RSS, and Venmo/PayPal/Visa are not interoperable. Governments do a good job of enforcing interoperability, more than making consumer apps directly.
I do wonder what you think about some of our shortcomings of our current system:
ACH takes multiple business days to clear. Will cbdc solve it? Some better clearinghouse?
Merchants don't take "app" payments. I do go to a coffee shop that takes Venmo, but generally credit cards are the way to go as a US merchant. Credit cards are not as much a necessity in other countries.
Is credit card dominance in the USA a good thing, neutral, or inefficiency? Could cbdc offer a clearinghouse or protocol to replace visa network? Credit cards are such a weird thing: merchant feesore than debit, and subsidizing the high earners with rewards and burying the poor in interest payments.
Venmo, PayPal, Visa, etc are not interoperable. Are there advantages to having a common language for payments? Like ACA is doing for EMRs?
I think there is another potential benefit of a CBDC in the US. If a US CBDC is bound by the first amendment, it would be forced to facilitate a bunch of transactions that are currently being censored by the Paypals and Venmos of the world.
The most obvious example is adult content. It's a constant struggle for companies like Pornhub and Onlyfans to provide stable payment options due to the private financial censorship they often face. A US CBDC might even convince the major private payment providers to be less censorious since a regulator can't exactly make frowny faces at companies for providing financial services to adult entertainers if the government is doing the same thing.
Also of note is that the Fed has already squashed one recent attempt at narrow banking (which I think you blogged about previously?) so we already know they aren't interested in competing with banks or people having accounts directly with the government.
It's hard to see why that position would suddenly reverse just because people start talking about cryptocurrency instead of narrow banking.
"Banks exist for a reason — their riskiness allows them to do a maturity transformation, borrowing short to lend long, thus financing risky but valuable projects."
Governments do long term investment. If central banks (part of government) can outcompete private banks on the deposit taking side, then the private finance sector will be much smaller. Given its appalling performance this century, that's all to the good.
Noah, you didn't say much — if anything about the significance of the fact that the government gets a lot of seigniorage or seigniorage revenue out of it — as set out by the Bank of England's John Barrdear and Michael Kumhof here
Oh, sorry. Engert & Fung (2017) talk about the potential benefits for seigniorage revenue, and argue that CBDC would not affect this in a meaningful way! I didn't mention it because it's a bit dry and complex and the conclusion is basically that it's not important. Thanks for mentioning that!
Thanks for this Noah, but I think Engert and Fung mostly skate over the issues. Barradear and Kumhof model a substantial reduction in interest rate costs for government — because it's supplying the market with CBDC which has an intrinsic value. Engert and Fung then say that that could produce price and supply responses from banks and perhaps increase volatility. The problem with this for me is that the genuine value adding banks do in fostering trade and investment that wouldn't happen otherwise seems swamped by their collateralised lending much of the benefit of which is lost through simply bidding up (and so increasingly financialising) asset prices. Barradear and Kumhof claim very large welfare gains from lower interest rate payments and I'd need to see a lot more than Engert and Fung's dot points in Section 5.4 of their paper (which addresses itself primarily to financial stability rather than than funding costs for government) to address that claim.
I'm so glad you said the dollar was already largely digital, I was sure that was true but constantly wondered if I was missing something.
That said, would a general federal money app be useful, even denominated in normal dollars? For example, set one up for each SSN number, then you can transfer money to it for stimulus, tax refunds, etc. If I've interpreted the news correctly – and I'm not sure I have – American stimulus cheques were actual, physical paper cheques, which seems crazy in the 21st century.
It would be useful if it were near-universal. Otherwise it would lead to many of the same inequities and difficulties we have now.
Why not just deposit funds straight into people's preferred bank accounts? But then you still have the matter of all the unbanked and underbanked, who probably need any such payments more.
If you had bank account information associated with your most recent tax return, the stimulus payments were deposited directly into that account.
India is trying this experiment, encouraging everyone to use a digital app for ordinary transactions as well as transactions with the state such as subsidies, benefits payment and so on. Right now, the issue is fraud. There are just too many holes in the system, not necessarily technical exploits, but practical ones that allow people to set up multiple identities or steal other people's identities. I'm not going to say that this kind of thing can't be fixed, but going digital is not a magic bullet.
Someone somewhere was also toying with the idea of stimulus money that would run out of time (use it or lose it). That could have been useful for the first COVID stimulus... and you could freeze stimulus spending if you noted inflation was getting out of hand...
French companies already kind-of do that with providing their employees with vacation checks and "ticket restaurant" (you can use them to buy food/pay for a restaurant if your company doesn't have a cantine... or to pay at the cantine?)
As I've said before, distributed blockchains, always and everywhere, take things that would work better if you had a trusted central authority, and then either make the system radically more expensive in energy and time (proof of work), or codify "my money works for me" (proof of stake). The "problem" that distributed blockchains are trying to address _is not a problem_. A chain of trust that ultimately reaches an entity that can be small-d democratically governed, that is answerable not just to those with wealth but to people who _don't_ already have a ton of money, is Good, Actually.
That said, postal banking also would be a good idea. And Felix Salmon has written for years about how digital payments ought to clear at par. And in that sense, a digital dollar operated by the Fed as a utility that's free to the public would be great.
Add on to this: Once you have everyone universally banked, you can start doing things like implementing UBI by way of just incrementing everybody's digital dollars account once a week. And you can do fiscal expansion and contraction by changing how much the UBI gives out.
Furthermore, you can create a Sovereign Wealth Fund that sits off to the side of the postal accounts, and implement Cory Booker's Baby Bonds concept by giving people shares in that. You also could scrap the 401(k) and all of its relatives (which have largely served to enrich the financial companies that operate the accounts for exorbitant fees) and just default everybody into an IRA that goes into a target-date tranche of the SWF.
Very nice post. At present maybe 95% of all my bills are paid digitally. I pay estimated taxes digitally. All of this is done in the absence of any kind of crypto currency regardless of who is/has established it. Once you set payments up, they are automatic based on the billing cycle.
I always ask the question, what problem is crypto trying to solve? I just cannot see the business case other than creating NFTs (collectible value) or speculation. VISA or MasterCard and perhaps big money center banks could get into crypto in a big way if they wanted to. At the end of the day, there has to be stability to digital currencies and right now there pretty much is other than inflation.
So no use really. It is disturbing, although sense eventually prevailed, to see economists, some anyway, jumping on bitcoin, blockchain, at a national level. Is it just to be novel or fashionable.
I was one of the reply guys asking you to weigh in on this, and this was exactly the kind of article I was hoping to read from you (and the conclusion I'd sort of assumed). Thanks!
👍👍👍😊😊😊
Thanks Noah, perfect explanation
Thanks!!
I think there's one way that a central bank cryptocurrency could be very useful: improving the security of digital payments. Currently we do digital payments with card numbers - which can be trivially stolen, so we use a bunch of other identifying information too, except *that* can *also* be trivially stolen, so it doesn't really help. Twice in the past two years I've had a card number stolen after using it online - once by posting a payment to an anodyne account in the exact amount I made, presumably in an attempt to make me think it was just the other payment.
With crypto techniques you can't just repost a transaction like that with a different recipient, because the whole transaction is wrapped up in the encryption. And while it is possible to steal crypto keys, at least it's not trivial, since the relevant information is *never* broadcast, as opposed to necessarily broadcast in its entirety with absolutely every transaction you ever make.
In addition, a central bank running the chain avoids problems like chain takeover, reversing fraudulent transactions, and probably energy use. I know Bitcoin and Ethereum couldn't handle the transaction volume; but some of the alt coins claim throughputs adequate to manage a financial system; I don't think any have been tested but, again, I'll bet a central bank can pull that off as well.
The thing about this is, when you lose intermediation you lose contestability. You can't get any of your money back after an identity theft.
True, but there *is* an intermediary for crypto: the blockchain maintainer, as an organization. And even with the supposedly "decentralized" blockchains, there have been a few rollbacks. A central bank could potentially rollback much more easily. I'll grant it's not a trivial problem, but there are possibilities.
At least _some_ transactions are already not done with card numbers, but rather with one-time keys that involve creating a time-stamped signature. (Apple Pay is the main consumer-accessible version of this.)
We ought to encourage more migration to this design, but Visa is actively hostile to it.
Yes, we might well not need a blockchain per se for these benefits, just decent digital signing techniques. Although Apple Pay is not perfect (yet?); my husband had his card number stolen and that was enough for some fraud to go through. I assume not every system for accepting Apple Pay is using the signing system (again, yet?)
I just don't understand Visa's hostility. They spend a LOT on credit card fraud. There would be a lot less, so more profit for them, so ...?
Visa and the banks that issue cards are able to collect higher fees from merchants for the worse technology, and they're able to force the merchants to eat the cost by way of saying they can't accept cards _at all_ unless they charge the same price regardless of the payment method. The card issuers actively incentivize people through reward programs to use more fraud-prone technologies (like favoring signature debit over PIN debit).
Sadly a lot of Felix Salmon's best writing on this was on his Reuters blog, and they've taken it all down. I found one of his pieces on SeekingAlpha, here: https://seekingalpha.com/article/210643-interchange-and-free-checking
https://www.forbes.com/2010/06/10/visa-mastercard-bank-fees-opinions-columnists-reihan-salam.html
<blockquote>
One argument, raised by scholars Todd Zywicki and Geoffrey Manne, is that swipe fees are a way to pass on credit losses to merchants. Yet as financial blogger Felix Salmon of Reuters observes, merchants don't also benefit from the much larger credit profits derived from fees and interest payments enjoyed by banks.
Swipe fees could be increasing due to rising levels of fraud and identity theft. That, however, would represent a case for shifting away from signature debit and towards PIN credit. But signature debit is more profitable for card issuers, which is presumably why there hasn't been aggressive movement in this direction. In Salmon's view, rising swipe fees represent pure rents for the effective duopoly of Visa and MasterCard, and the Durbin amendment would help shift the balance of power towards merchants and consumers.
</blockquote>
If the Fed came in and ended this kind of parasitism of commerce, turning payment processing into a utility like water and electricity, that would be great.
FWIW, that's not how we do digital payments in Vietnam (not exactly an advanced country) so I'm not sure why you need to jump to a central bank cryptocurrency to solve that problem.
All you have to do is switch from pull to push of money.
I think one of the things about postal banking is that it can to compete with profit-driven banking to get them to behave better. Is there anything about the behavior of payment systems that really needs reining in?
Not really.
I kind of thought as much, I was just hoping there might somewhere somehow be something more to the CBDCs.
This seems a little confused on a few points to me:
1. There is plenty of reason for a CDBC to use a blockchain like system. What a CBDC does not need is a *proof of work* scheme. The purpose of proof of work/stake/etc is transaction *ordering*, because sequencing is difficult in the absence of a central actor. But "block chaining" exists to maintain a transparent, immutable ledger. Transparency and immutability would (at least potentially) be desirable properties of a CDBC.
2. Our existing payments system is terrible and expensive. All in credit card processing fees are often well over 200 basis points. That's absolutely absurd, and any reasonably implemented CDBC should be able to offer its services far, far more cheaply than that.
3. The potential benefit re: banking of a CDBC is to provide competition, and force banks to actually innovate in a meaningful way. Right now banks are the only game in town for storing your funds in an accessible, safe way. Since banks are very difficult and expensive to start, that means they effectively have an oligopoly, and can offer very low yields on savings accounts, etc. Right now, people only go to the banks for storage and payment - if the government provided that to people for free, the banks would have to do more to attract capital. And anyway, there's no reason for banks to attract individual capital anyway. The federal government is already insuring most of it, so we should just end the charade, store everyone's cash at the fed, and let the fed lend to the banks, and the banks just work on their actual business: lending. Right now there is this incredibly dumb conjunction between bank's ability to attract deposits via advertising and customer service, and their net interest margin, and simplifying that whole game would be immensely valuable.
Great read!
One distinction I think about is between the app the protocol. Tech apps like proprietary protocols because it locks you in. Epic EMR tried to do this before Obamacare, Twitter/fb both abandoned open standards like RSS, and Venmo/PayPal/Visa are not interoperable. Governments do a good job of enforcing interoperability, more than making consumer apps directly.
I do wonder what you think about some of our shortcomings of our current system:
ACH takes multiple business days to clear. Will cbdc solve it? Some better clearinghouse?
Merchants don't take "app" payments. I do go to a coffee shop that takes Venmo, but generally credit cards are the way to go as a US merchant. Credit cards are not as much a necessity in other countries.
Is credit card dominance in the USA a good thing, neutral, or inefficiency? Could cbdc offer a clearinghouse or protocol to replace visa network? Credit cards are such a weird thing: merchant feesore than debit, and subsidizing the high earners with rewards and burying the poor in interest payments.
Venmo, PayPal, Visa, etc are not interoperable. Are there advantages to having a common language for payments? Like ACA is doing for EMRs?
I think there is another potential benefit of a CBDC in the US. If a US CBDC is bound by the first amendment, it would be forced to facilitate a bunch of transactions that are currently being censored by the Paypals and Venmos of the world.
The most obvious example is adult content. It's a constant struggle for companies like Pornhub and Onlyfans to provide stable payment options due to the private financial censorship they often face. A US CBDC might even convince the major private payment providers to be less censorious since a regulator can't exactly make frowny faces at companies for providing financial services to adult entertainers if the government is doing the same thing.
Also of note is that the Fed has already squashed one recent attempt at narrow banking (which I think you blogged about previously?) so we already know they aren't interested in competing with banks or people having accounts directly with the government.
It's hard to see why that position would suddenly reverse just because people start talking about cryptocurrency instead of narrow banking.
"Banks exist for a reason — their riskiness allows them to do a maturity transformation, borrowing short to lend long, thus financing risky but valuable projects."
Governments do long term investment. If central banks (part of government) can outcompete private banks on the deposit taking side, then the private finance sector will be much smaller. Given its appalling performance this century, that's all to the good.
I don't find Venmo, PayPal, Zelle, etc easy to use or very trustworthy (strong cyber/data protections?), but do I trust a new Fed option more?
Noah, you didn't say much — if anything about the significance of the fact that the government gets a lot of seigniorage or seigniorage revenue out of it — as set out by the Bank of England's John Barrdear and Michael Kumhof here
https://www.bankofengland.co.uk/working-paper/2016/the-macroeconomics-of-central-bank-issued-digital-currencies
Are they hyperventilating?
Oh, sorry. Engert & Fung (2017) talk about the potential benefits for seigniorage revenue, and argue that CBDC would not affect this in a meaningful way! I didn't mention it because it's a bit dry and complex and the conclusion is basically that it's not important. Thanks for mentioning that!
Thanks for this Noah, but I think Engert and Fung mostly skate over the issues. Barradear and Kumhof model a substantial reduction in interest rate costs for government — because it's supplying the market with CBDC which has an intrinsic value. Engert and Fung then say that that could produce price and supply responses from banks and perhaps increase volatility. The problem with this for me is that the genuine value adding banks do in fostering trade and investment that wouldn't happen otherwise seems swamped by their collateralised lending much of the benefit of which is lost through simply bidding up (and so increasingly financialising) asset prices. Barradear and Kumhof claim very large welfare gains from lower interest rate payments and I'd need to see a lot more than Engert and Fung's dot points in Section 5.4 of their paper (which addresses itself primarily to financial stability rather than than funding costs for government) to address that claim.