The Sovereign Money Serenade: A Reply to David Beckworth

David Beckworth has a new post, “The Sovereign Money Blues,” in which he is critical of proposals to “open up the central bank’s balance sheet to the general public.” His critique extends to a new paper I coauthored with John Crawford and Lev Menand, “A Public Option for Bank Accounts (or Central Banking for All)” (shorter, policy-oriented version here), which proposes giving everyone (individuals, businesses, institutions) the option to open an account at the central bank.

David has had me as a guest on his podcast twice and he’s one of my favorite people in the money and banking field. He’s also a fellow Nashvillian, so I’m generally inclined to endorse whatever he says.

But I think he’s confused about this. And it’s not just David. What follows is in part a response to David, but indirectly it’s also a response to various other comments I’ve seen on this topic over the years. I see a lot of flawed assumptions and faulty reasoning in this area.

Let me say at the outset, my coauthors and I don’t deny that there would be costs to our proposal—like any other policy proposal. In the paper we have a long section on costs and objections that addresses topics like cybersecurity and fraud prevention, privacy and civil liberties, and a bunch of other issues. We also talk a lot in the paper about implications for credit allocation and monetary policy implementation, which are the topics that most concern David.

I’m going to start with some general observations before proceeding to some of David’s specific points.

1. First of all, it’s important to distinguish between Vollgeld-type initiatives, which would actually ban fractional reserve banking, and public option proposals like the one my coauthors and I have proposed, which would not ban private money creation by banks but rather give individuals and nonbank businesses the option to hold central bank money directly in “account” form. Needless to say, public options are a lot more modest. This distinction doesn’t come through in David’s piece, and I think it’s important. Now, I do think a public option could be attractive enough to “crowd out” a lot of private money—in fact this is one of its advantages—so in these comments I’m just going to assume large-scale migration from private to public money. But this would be by individual choice and not command, and that’s important.
2. The status quo bias in this area is deeply entrenched. It’s useful to step back and think for a moment about the system we have now. The Fed has (give or take) a trillion dollars of physical currency liabilities and a couple trillion dollars of “account” liabilities. The currency can be held by anyone. (David objects to “opening up the central bank’s balance sheet to the general public,” but presumably he thinks currency is ok?) But the accounts are reserved to an exclusive clientele consisting of banks and selected other big financial institutions. The account liabilities are pure base money, fully sovereign and nondefaultable no matter how large the balance (no $250,000 insurance limit here), perfectly transactable, and bearing an extremely attractive administered interest rate. This type of asset just isn’t available to individuals and nonbank businesses. Presumably we should be skeptical of special privileges in general, and especially if they accrue largely to elite financial firms. Is there an intellectually respectable basis for this? I fear people tend to backsolve to the current state of affairs and then treat it as natural.
3. One could perhaps defend the privilege on the basis of corresponding obligations—as a kind of package deal. OK, but then let’s be consistent. For example, there is widespread consensus that reserve requirements are a “tax” on banks for which they should be “compensated.” Offsetting this “tax” is explicitly why the Fed decided to pay interest on required reserves. But note the package deal conception isn’t respected here: one aspect of the institutional environment gets treated in isolation as a “tax” for which banks must be specifically “compensated” by getting government checks. Similarly, banks insist that they aren’t “public utilities” and object to any type of public service obligations. So all too often this seems to be a one way street: the privileges are said to be justified on account of corresponding obligations—but then the obligations are condemned as “taxes” or an undue impingement on “private” business.
4. I’ve understated just how convoluted and weird the current system is. Because the Fed wasn’t getting good enough passthrough on interest on excess reserves (IOER), it started “opening up its balance sheet” to nonbank financial institutions like money market funds, through the reverse repo facility. There are good reasons to think this strategy reduces passthrough to unsophisticated depositors. And we’ve now let financial market infrastructures like CME hold reserve accounts bearing IOER, despite the fact that they play no meaningful role in monetary policy transmission. By all accounts they’re loving it, which isn’t surprising. Again, what’s with the special privileges? Shouldn’t we approach this with deep skepticism?
5. All too often, these debates neglect the extent to which money-and-payments are already public. The business model at issue (bank money creation and circulation) isn’t an ordinary business, it’s entwined with and indeed underwritten by the public sector in all sorts of ways, from deposit insurance and LOLR (which mean the liabilities in question are largely public in the first place) to the Fed’s integral and dominant role in check clearing, wholesale payments (Fedwire), and even retail electronic payments where FedACH dominates private providers. Plainly this isn’t an ordinary private business model. It’s easy to lose sight of all this. I’m happy to give priority to private provisioning in most areas but this one is questionable, and the rent extraction is just immense.
6. That rent extraction comes through various channels. As noted above, there’s interest on reserves, which consists of tens of billions of dollars a year (and growing) in checks written to banks—interest that is only partially passed through to deposit rates and other market rates. Also, there are underpriced deposit insurance fees that are intermittently suspended. There are huge implicit guarantees of basically the whole money market, if we’re being realistic. And less obvious but just as important is private capture of seigniorage. The current state of affairs is, let’s just say, questionable.

Enough backdrop, now to respond a bit more directly to David:

1. David sees a Hayekian “knowledge problem” with sovereign money, relating to “centralized economic planning.”  I’m surprised to hear this from David in particular. Last I checked David wanted nominal GDP level targeting! Why is the “knowledge problem” any different here? In principle, monetary policy works no differently in the event of a shift toward sovereign and away from private money. The monetary authority would look at macro variables and adjust, just like it does now—buy or sell assets, adjust administered rates, communicate intentions, and so forth.
2. David then has a couple of paragraphs that I think confuse key issues. Here’s the first one: “Money, currently, is the byproduct of many decentralized decisions among a large number of borrowers and lenders in very different circumstances. Specifically, banks create money through loans to businesses and households in various locations based on local, idiosyncratic economic conditions. To believe this complex process could be easily replaced by a central bank committee deciding how much money should be created is incredulous.” I have a bunch of issues here. The paragraph assumes that a shift toward more sovereign money would necessarily involve the central bank in individualized credit allocation decisions. But let’s think carefully about this. For the sake of argument, suppose there are enough Treasury securities to accommodate the entire desired money supply (“desired money supply” just meaning the money supply produced by implementing your preferred monetary policy mandate, whatever it may be). Also suppose the central bank is the sole money issuer (again, our proposal would merely give people the option to hold central bank accounts, but set this aside). So the central bank balance sheet grows to $14 trillion or whatever, and its assets are exclusively Treasuries. How are “loans to businesses and households” allocated in this system? Are they divorced from “local, idiosyncratic economic conditions”? Obviously not. They’re just made by lenders that don’t have “money”-type liabilities.
3. It’s crucial not to lose sight of the fact that lending markets are competitive. Deposit banks or “money issuers” have no monopoly on extending credit; they coexist with myriad other financial institutions and investors that make loans and buy bonds. If profitable lending opportunities exist, the market should be expected to ferret them out. In a sovereign money scenario, if the Fed buys more (say) Treasuries in exchange for newly issued money, there will be more funds outstanding available for investment. To the extent that financial markets are efficient, these funds will make their way (directly or indirectly) to lending markets if that is their optimal use. Loans don’t have to be, and very often aren’t, financed by “money” creation. The credit markets are way bigger than the banking system proper; the U.S. bond market alone is worth $40 trillion give or take, dwarfing any money measure you care to cite, and that doesn’t even count lending markets. So we’d still have decentralized, market-based credit allocation in this sovereign money world.
4. Plainly we wouldn’t want the central bank involved in direct lending to individuals and nonbank businesses. But to the extent the high-quality liquid bond markets aren’t big enough to accommodate the desired money supply, the central bank can outsource portfolio management. Indeed, as I’ve argued in the past, one way of thinking about a bank charter is as a means of outsourcing the credit allocation that arises as a byproduct of the decision to issue money in exchange for credit assets. (I think the whole history and structure of U.S. bank regulation are best understood through this lens, but this is a topic for another day.) As we explain in our paper, this problem wouldn’t be materially different under a sovereign money system from how it is today.
5. OK, now for the next paragraph, which I also have issues with: “But it gets worse. Under a sovereign money regime,  estimating the amount of money needed would actually become more complicated. This is because financial firms wanting to originate loans would need 100% reserve backing. Money, in other words, would be needed for both transaction purposes and for loan-backing purposes. So not only would the central bank need to know where and how much transaction money demand there was, it would also have to know where and how much loan-driven demand for money there was.  It would be very easy for central banks to get monetary policy wrong. Central bankers, in short, would need near omniscience to do their job well in a sovereign money regime.” The phrase “financial firms wanting to originate loans would need 100% reserve backing” is an odd construction. Reserves of course are an asset not a liability of the private sector, so there is no sense in which they “back” loans. What I think he means is just that financial intermediaries would need to procure funding by some mechanism other than issuing “money.” Yes, but so what? Most businesses don’t fund themselves by issuing money. They usually fund with equity and longer-term debt of some sort. Credit portfolios can be and often are funded this way. And I honestly don’t understand how David’s distinction between “money needed for transaction purposes” and “money needed for loan-backing purposes” has any bearing on monetary policy. Presumably monetary authorities should be concerned with macro aggregates (inflation, unemployment, NGDP, or whatever) in pursuit of their mandate. Again, this wouldn’t change at all with a large shift toward more sovereign money.
6. So I don’t think David’s “Knowledge Problem” holds up. What about his “Public Choice Problem”? Here I can be briefer. David’s argument here really just amounts to an argument against having a central bank in the first place. Yes, the Fed is very powerful and has a large balance sheet; under sovereign money its balance sheet would be considerably larger. We should try to design systems and commitment devices to mitigate these problems—like a reasonable amount of administrative independence to insulate the Fed (somewhat) from politics, and statutory restrictions on its ability to lend directly to individuals and nonbank businesses. We already have these things. They’re not perfect but neither are they ineffectual. And keep in mind that the federal government is already explicitly or implicitly backing virtually the entire “private” money supply. $7 trillion in deposits are insured explicitly, and the rest (as well as nonbank money-claims) have implicit government backing. That these are contingent liabilities rather than “on balance sheet” is immaterial; the public choice problem is already with us and has to be managed.
7. David says under sovereign money the central bank would have “the sole power to determine who gets money.” This is a strange thing to say. Does the Fed have the sole power to determine who gets base money today, or who gets a private loan, or who issues securities into the marketplace? None of this changes with large-scale migration to central bank accounts. (I don’t think David is making a privacy or civil liberties point here; if he is, we address this in detail in our paper.)
8. Finally, David’s “Tradeoff Problem.” He’s worried about the central bank becoming “the mother of all too big to fail institutions” and the possibility of a “central bank run.” What he means isn’t redemption risk—the issuer of base money can’t default in any ordinary economic sense—but inflation risk. This is reminiscent of the fearmongering over QE and currency debasement from a few years ago. David, mind you, was on the right side of that debate! But he seems to be falling into the trap here.

To conclude: One’s view of sovereign money proposals depends to a large degree on how satisfied one is with the status quo in money and banking. I think the status quo in the U.S. has huge problems. The instability of private “money” has been the preeminent cause of severe, acute recessions in U.S. history. Rent extraction in the current system is enormous and accrues to privileged financial interests. We have a large unbanked population that is excluded from the mainstream money-and-payments system. Cash management is hard for large institutions because there is no pure asset called “money” they can realistically hold in nondefaultable form. Payments are slow and expensive in the U.S. “Opening up the central bank’s balance sheet” could substantially mitigate these problems.

If we only confined ourselves to narrow, incrementalist, technocratic tweaks to existing arrangements, we would never have had a central bank to begin with. More muscular approaches to public policy problems are sometimes in order.