Beth and Marc Goldberg Professor of Law
Cornell Law School
Dear Professor Omarova,
As you probably know, I’m among the critics of the plan for “democratizing money” that you outline in your forthcoming Vanderbilt Law Review article. I disagree with it enough to be unwilling as yet to join many of our mutual friends in supporting President Biden’s own plan to make you our next Comptroller of the Currency. I’m writing to explain my position to you, and also to them. I want to make it clear that my stance is entirely a reaction to your plan, and not a reflection of any doubts I have concerning either your general qualifications or the quality of your scholarship.
According to your own description of it, your plan would ideally see public FedAccounts “fully replace—rather than uneasily coexist with—private bank deposits.” Consequently it would “likely cause a massive contraction in bank lending” to businesses and individuals. Most if not all of the lending now done by banks would instead be done by the Fed, either directly or through Fed purchases of securities issued by a National Investment Authority. Some of the Fed’s lending would consist of “New Discount Window” (NDW) loans to “qualifying lending institutions” (QLIs). But in order to “maximize the allocative impact” of this arrangement, the Fed would be free, not only to determine which banks qualify for loans and what collateral they must post, but to explicitly prefer banks that fund particular firms or activities, while refusing funding to others. In short, despite allowing some banks to continue to make loans, the plan would have bureaucrats take the place of the thousands of bankers who now decide where bank credit goes. There is, indeed, more to your proposal than this; but it is these aspects of it that concern me most.
In my opinion, the changes you advocate, were they to come about, would have harmful, if not disastrous, consequences. By saying so, I don’t at all mean to suggest that bankers today allocate credit flawlessly: far from it. I know that they sometimes fail to get credit to certain credit-worthy applicants, while lending recklessly to less worthy ones; and I understand that the government, besides attempting to correct such errors through regulation, must sometimes compensate for them through its own credit programs aimed at supporting certain groups or industries. I also understand that, owing either to market failures or to the effects of government guarantees and other kinds of regulation, bankers’ profit motive doesn’t necessarily cause them to steer credit to where it does the most good.
But it hardly follows from such truths that a comprehensive Fed takeover of deposit banking would boost “productive economic enterprise, in the long-term interests of the American people.” To claim that it would is, I think, to underestimate both the extent to which the profit motive encourages competing commercial bankers to invest the public’s scarce savings productively, and the difficulties bureaucrats are bound to encounter in their efforts to do the same.
“The public’s scarce savings”? Yes, I hold that commercial banks are credit intermediaries and not just credit creators—a view that has lately come under fire, and one you yourself consider a “widespread misconception.” Still I beg to differ. It’s true that, whenever a commercial bank makes a loan, it “creates” deposits by crediting the borrower’s account, and that in this sense (to quote Zoltan Jakab and Michael Kumhoff) “[t]he only ‘resource’ that banks require to make a loan is a keyboard or, in earlier times, a pen.” It’s also true that banks don’t just lend funds previously deposited with them, and that the banking system and the Fed together determine the nominal extent of bank lending. But the equilibrium real (that is, inflation-adjusted) extent of bank lending is nevertheless a function of people’s real demand for bank deposits; and every individual bank’s contribution to total lending is ultimately dependent on its real share of the market for deposits. That’s why commercial banks typically compete aggressively for “core” deposits, and why those that fail to attract such deposits, or that try to succeed just by creating credit ex nihilo, tend to go out of business—sometimes spectacularly—despite being equipped with pens or keyboards. Fiat-money issuing central banks are an exception. If one thinks in terms of real equilibrium deposits, determined by the demand side of the market, instead of nominal deposits, the credit intermediary view of banks comes to the fore, where it appears to be more subtle and more correct after all than the claim that banks aren’t really intermediaries at all.
I raise this point because it causes me to disagree profoundly with your understanding of your plan as one for “democratizing” money. If every commercial bank were indeed ultimately a credit “creator” rather than a credit intermediator, I might see merit in your suggestion that giving the Fed a monopoly of deposits, and corresponding control of the flow of credit, would be equivalent to giving “the people” more control of that flow. I might even understand your claim that your plan would “finally render the orthodox notion of ‘financial intermediation’ a reality.” But if individual commercial banks’ command of credit already ultimately depends on people’s own willingness to direct credit to them, by electing to hold certain real average deposit balances with them, this ceases to be so. Instead, allowing people to choose among thousands of bankers in disposing of their savings, or that part of their (real) savings they wish to keep in the form of deposit balances, appears more “democratic” than compelling them to place them all at a single, central bank.
A public bank account option might be reckoned still more democratic. But your ideal plan wouldn’t make central bank accounts optional: your article instead “advocates full migration of demand deposits onto the Fed’s balance sheet” (my emphasis). In my opinion, such a plan, far from really acknowledging “the people’s” sovereignty, would be a reversion to the medieval treatment of money issuance as a government “prerogative,” meaning something governments alone are entitled to do, even if their citizens would rather have private enterprises do it.
But my main objection to your plan isn’t that it’s actually less “democratic” than the status quo. It’s that, instead of encouraging “productive economic enterprise, in the long-term interests of the American people,” it would almost certainly do the opposite. For although commercial bankers often err in deciding who to lend to, I believe they are far less likely to do so than a body of government-appointed bureaucrats.
There are at least three reasons for this. First, if they invest credit unwisely, commercial bankers stand to lose their shirts—though admittedly this is less likely for those whose banks are considered “too big to fail.” Government bureaucrats, on the other hand, can usually make bad loans with impunity, or at least without facing any pecuniary losses; and central banks never fail.
Second, by dividing responsibility for lending among thousands of independent bankers, according to their success in attracting deposits, instead of placing it entirely in the hands of one set of officials, the present system reduces the risk of systemic errors: one bad commercial bank needn’t spoil the whole bunch, whereas one bad central bank is equivalent to a spoiled bunch of commercial banks.
Lastly, central bankers’ lending and investment decisions, instead of depending solely on the judgment of technocrats, are likely to be heavily influenced by political considerations—considerations that could prevent even the most savvy and well-intentioned bureaucrats from investing people’s savings in ways far removed from those in accord with their long-run best interests. Of course, one can imagine having planners who get everything right. But in practice what matters is how flawed real-world planners stack up against flawed bankers.
That, despite many notorious setbacks, commercial bank lending has contributed immensely to worldwide economic growth and development seems indisputable. The evidence here starts with Adam Smith’s famous account, in Book II, chapter 2 of The Wealth of Nations, of the advantages of bank-supplied money, which, as I’ve noted elsewhere, he credited for Scotland’s extremely rapid growth during the first three-quarters of the 18th century. (In England, in contrast, banks contributed less to development, while banking crises were more common, thanks to a concentration of privileges in the Bank of England, and the corresponding legal emasculation of other, “country” banks.) The Scottish and English cases are two among many others reported on by Rondo Cameron and his coauthors in a pair of volumes entitled Banking in the Early Stages of Industrialization (1967) and Banking and Economic Development (1972). The main lesson of these volumes is that, where they have been allowed to do so, commercial banks have been instrumental in fostering economic development.
Later econometric research by Robert King and Ross Levine confirms that lesson. Of particular relevance here is their finding of a very strong relationship between a country’s economic growth and a variable representing what Levine elsewhere describes as “the degree to which the central bank versus commercial banks are allocating credit.” By way of illustration, King and Levine observe that
if in 1970 Zaire had increased the share of domestic credit allocated by banks as opposed to the central bank from 26 percent to the mean value for developing countries in 1970 (about 57 percent), then Zaire would have grown 0.9 percent faster each year in the 1970s, and by 1980 real per capita GDP would have been about 9 percent larger than it was.
King and Levine also report a more-or-less equally strong relationship between economic growth and the proportion of credit allocated to private rather than state-owned enterprises.
Even heavily regulated commercial banks are capable of promoting growth. Take early U.S. national banks. They were required to back every $90 of their outstanding notes with government bonds, and prohibited from having branches that would have made it far easier for them to diversify their loan portfolios. Yet, according to Scott Fulford, their presence in a community sufficed to increase per capita production there by 10 percent! (See as well this more recent study by Chenzi Xu and He Yang.) Tragically, the post-bellum South proved the opposite point: because northern states got the lion’s share of early national bank charters, the total number of which was strictly limited, the South long found itself bereft of banks, and suffered corresponding underdevelopment.
Of course the tremendous growth of securities markets since the start of the 20th century has dramatically reduced banks’ relative importance as a source of business funding. But this hasn’t made bank lending unimportant, or less productive. A 2015 study takes advantage of the “natural experiment” consisting of the partial removal, in 1994, of barriers to interstate branch banking, to study the effect of improved access to bank credit on firm-level productivity. It finds that the change led to a significant increase in the productivity of firms located in states that allowed out-of-state banks to cross their borders, with smaller firms experiencing the greatest gains. Bank lending also continues to enhance business productivity in other industrial economies. For example, a recent IMF study using Italian data finds “sizable, persistent, and robust” effects of increased bank credit growth on Italian firms’ productivity.
Here in the United States, loans and lines of credit, mostly secured from banks, supplied more than half of small business’s external funding needs in 2019. And commercial banks were, according to a 2017 survey of small business credit by the Fed’s Board of Governors, “by a wide margin, the most common source of virtually every credit product included in the survey.” The same survey notes that, besides being especially risky, lending to small businesses is complicated by their “informational opacity,” meaning the lack of public information about their financial well-being that can assist lenders in making underwriting decisions. For this reason, as many studies have shown, lenders are generally more likely to extend credit to small businesses they’re well-acquainted with, and especially those that have long kept deposits with them. Any wholesale migration of deposits to the Fed is bound to end such relationship lending temporarily. What’s more, given the agency problems that make relationship lending easier for smaller banks, the fact that under the proposed plan only QLIs could secure funds from the Fed, and could do so only “against qualifying high-quality collateral” (which would presumably exclude loans of any sort), the Fed’s deposit takeover is likely to reduce relationship lending for good.
Even large businesses that can borrow in money markets still rely on banks as providers of “liquidity insurance,” in the shape of substantial lines of credit they can draw on during times of economic turmoil. Immediately following the outbreak of the COVID-19 crisis, for example, the use of such lines of credit led to a $700 billion, or 30%, jump in banks’ C&I loans. (The Fed, in contrast, took months to get its own Main Street Lending Program going, and ultimately extended only $17.5 billion in credit through it.) It’s worth noting here that banks’ ability to supply such liquidity insurance appears to go hand-in-hand with their ability to deal in demand deposits, for reasons Anil Kashyap, Raghuram Rajan, and Jeremy Stein compellingly explain. According to their article’s abstract,
since banks often lend via commitments, their lending and deposit-taking may be two manifestations of one primitive function: the provision of liquidity on demand. There will be synergies between the two activities to the extent that both require banks to hold large balances of liquid assets: If deposit withdrawals and commitment takedowns are imperfectly correlated, the two activities can share the costs of the liquid-asset stockpile.
It’s true that many people, often drawing on Douglas Diamond and Philip Dybvig’s famous 1983 article, regard banks’ traditional combination of bank lending and deposit taking as inherently dangerous because it makes them vulnerable to panic-based runs that might even overwhelm a nation’s entire banking system. But while Diamond and Dybvig did set out to prove this, as I explained not long ago in a series of essays here, they never quite managed to do so. Nor has anyone since. In fact, bank runs resulting from sheer panic of the sort Diamond and Dybvig try to explain have seldom led to bank failures, let alone financial crises. Not that runs themselves have been rare: in the United States they’ve been all too common, especially before the advent of deposit insurance. But most have been runs on banks that were already in trouble, or that had close ties to other, troubled banks; and too often the source of their troubles was one or more misguided government regulations. Even the notorious runs of the early 1930s generally fit this pattern. That bank failures, or impending failures, typically cause runs, instead of the other way around, and that some relatively lightly-regulated banking systems of the past have been remarkably panic-free, is the main reason why, unlike proponents of narrow banking, I don’t consider commercial banks’ practice of funding loans with deposits to be fundamentally unsound.
Indeed, banks’ “core” deposits (meaning those that come to them, rather than those they create when making loans), including demand deposits, far from being an unstable source of funding for bank loans, are actually one of the most dependable sources of such funding. Other sources, and wholesale funding in particular, are far less stable and reliable. For this reason banks that are forced to rely more heavily on other funding sources tend to be less profitable, and more failure-prone, than others. I therefore doubt very much that, instead of having to give up any large part of their lending business, banks operating under your plan could simply “fund themselves with different loans.” I include here loans from the Fed, which though extended for longer periods would not be automatically granted even to qualifying banks with the requisite collateral. Instead, the availability of such loans would be subject to all sorts of temporary and permanent “targeted adjustments.” At the same time, I worry that some—perhaps many—of the banks that do secure extended Fed credit at the NDW’s “preferential rates” will turn out to be insolvent, for that’s what happened when the Fed last allowed banks to secure extended credit from it at non-penalty rates.
So much for commercial bank lending. What, then, does experience tell us about government bureaucrats’ ability to allocate credit in a way that promotes productive enterprise? Because central banks generally aren’t in the business of lending to businesses, one must instead consider the success of other official banks, including so-called “development banks,” in doing so. Many such banks were established during the 1950s, 60s, and 70s. But by the late 1980s enthusiasm for them had diminished considerably, and small wonder: though such banks were more likely than commercial banks to lend non-financial businesses, and they certainly helped to boost particular industries, no study has yet found strong evidence that they’ve contributed significantly to any nation’s overall economic growth. On the contrary: according to several surveys (e.g., Barth, Caprio, and Levine (2001); La Porta, Lopez-De-Silanes, and Shleifer (2002)), instead of promoting economic growth and development, the presence of these banks lowered it. One reason for that outcome appears to be that any substantial nationalization of banking tends to reduce financial-sector growth and innovation. A 2001 World Bank report put it starkly: “Whatever its original objectives,” the report said, “state ownership of banks tends to stunt financial sector development, thereby contributing to slower growth.”
Of course it’s possible that the Fed would do a better job allocating credit than other official banks have done. But the experience of so many other official efforts warrants skepticism. It’s true that the studies I’ve mentioned find that the administration of credit by government bureaucrats is less counterproductive in more developed economies than in less-developed ones. But that finding may only reflect the more limited role of bureaucratic credit administration in such countries, and the fact that it has been more often devoted to addressing relatively obvious market failures.
Yet evidence of substantial gains still eludes researchers, while (as Ed Kane observed some years ago, when government credit-allocation schemes were more common in the United States than they are now), evidence of adverse, unintended effects of government-administered credit programs, and of many being rife with corruption, is all too abundant. As Kane explains, in practice bureaucratic credit allocation
works far less predictably than proponents presume and has long-run effects quite different from its short-run impacts. Popular conceptions of how government economic programs work are marred by a naive tendency to project the effects of government actions wholly in line with the intentions of their sponsors. Voters (and legislators) are prone to wishful thinking about what governmental good intentions can accomplish in the market economy.
It seems to me that your proposal rests on just the sort of popular and optimistic, but mistaken, conceptions Kane warns against; that is, it fails to heed his advice that policymakers weigh their programs’ “intended improvements on targeting dimensions of macroeconomic performance” against the inevitable “unintended inequities and inefficiencies introduced by forcing credit flows out of their customary channels.”
Note that Kane isn’t rejecting arguments for bureaucratic (as opposed to profit-driven) credit allocation tout court. Nor am I. What I reject is the suggestion that a wholesale Fed takeover of credit now being allocated by commercial banks would be beneficial. While government credit programs of relatively limited scope (for instance, some of those studied by Calomiris and Himmelberg, 1993) may pass Kane’s test, it’s exceedingly unlikely that any plan for comprehensive nationalization of bank credit, or something close, would do so.
Even limited, targeted programs often fail to achieve their mission. Consider the Reconstruction Finance Corporation (RFC). You say it “successfully led a massive nationwide capital mobilization campaign to aid Depression-struck sectors of the American economy,” and treat it as a model for your proposal’s “National Investment Authority.” In fact, like the Fed’s concurrent 13(b) business lending program, and its more recent Main Street Lending Program, the RFC’s C&I loans ended up falling far short of its statutory capacity—a “lackluster” outcome that economic historian Joseph Mason attributes to the fact that the RFC’s loans “were grossly overpriced compared to bank rates on short-term business loans during the period.” Nor did help from the RFC necessarily prove a boon to specific industries it targeted. For example, as Mason and Daniel Schiffman report in another study, instead of contributing to railroads’ long-term viability, the assistance the RFC granted to them appears to have inspired in their managers “a shallow view toward short-term survival that entailed drastically reduced expenses [sic] on maintenance of capital goods like equipment and rights of way.”
If the record of government credit allocation doesn’t offer a compelling argument for having the Fed takeover bank lending and otherwise increase its role in credit allocation, neither do the Fed’s emergency interventions in 2008 and 2020. Although it is certainly true that during these episodes the Fed found itself more heavily involved than ever before in supporting particular markets (and, in 2008, particular firms), most of these extraordinary efforts were aimed at keeping financial markets liquid during an emergency, not at supplying credit to firms and markets that would not have received it had those markets been functioning normally. And to the extent that the Fed appeared to prop-up insolvent businesses, its efforts were rightfully criticized.
It therefore seems to me disingenuous to present plans for assigning the Fed a greatly enlarged and perpetual role in allocating credit as mere “straightforward extensions” of the Fed’s extraordinary undertakings during two unparalleled crises. And although it’s true that those undertakings have whet both the public’s and politicians’ appetite for still greater Fed intervention, it doesn’t follow that they reflect “ongoing changes in the role of a modern central bank not only as the nation’s primary money-modulator but also…as its credit-allocator.” Whether the changes in question are in fact “ongoing,” or merely temporary (if repeatable) responses to exceptional emergencies, remains to be seen. There is neither anything inevitable about the “extensions” you favor, nor any basis for your claim that they are “indispensable…for ensuring the functioning of the modern economy.”
Besides allowing the Fed to use its monopoly of deposits to fund loans, your plan would allow it to “channel a significant portion of funds … into large-scale purchases of securities issued by various public instrumentalities for purposes of financing of critical public infrastructure projects.” It would also allow the Fed to use its Quantitative Easing capacity to augment its lending and security purchases, or to make “helicopter drops” of funds into people’s Fed accounts. And it would be able to do these things, not only in order to “stimulate economic activity” during severe macroeconomic crises when interest rates have fallen to their effective lower bounds, but at any time, in its effort to “ensure better utilization of the national economy’s productive capacity.”
You claim that, under your plan, despite “proactively managing the economy-wide flow of credit, the Fed would not be making any direct investment decisions, especially at the level of individual projects or entities.” I’m unable to see how this can be so. Though banks could take part in administering Fed credit by borrowing from the NDW, the Fed would ultimately be responsible for deciding which banks to lend to, which assets to purchase, and which personal FedAccounts to credit through helicopter-money grants. So while it may be technically true that “the creation of FedAccounts does not really have to affect the asset sides of banks’ own balance sheets,” it isn’t the case that your particular plan would leave banks’ asset holdings unchanged, or (to be more specific) that its “proposed NDW mechanism will enable banks to continue their lending activities” if by this you mean continue to engage in the same lending activities they would engage in if they could go on funding their loans with customer deposits.
The same goes for the Fed’s security purchases under your plan. Besides being allowed to purchase practically any securities (as opposed to the more limited set it is presently authorized to purchase by section 14 of the Federal Reserve Act), it would enjoy complete discretion in deciding which of these specific assets to purchase and in what quantities—something you correctly call “an inherently political act.” It would, for example, be free to decide what share of its purchases to devote to funding the proposed National Investment Authority (NIA) and various other public “instrumentalities.” It follows that, even though NIA rather than Fed bureaucrats would decide how to dispose of funds the Fed invests in it, it would not be the case that public investment decisions would be left entirely “to the Treasury and the newly created NIA.”
It follows that your plan would in fact “fundamentally redefine the public-private balance of power in the finance franchise.” It would also fundamentally redefine the Fed-Treasury balance of power, and ultimately that between the Fed and Congress, by allowing the Fed to lay claim to some part of Congress’s “power of the purse”—a change that would have far-reaching Constitutional implications. It is, I think, owing to this enhancement of the Fed’s power to determine how public resources are used, and not simply because we oppose “dramatically increasing the role of public allocation of capital,” that Charles Plosser and others (myself among them) worry that allowing central banks substantial leeway to alter the size and composition of their balance sheets can render them “dangerously powerful and vulnerable to political manipulation and abuse.”
I’ve written at such length regarding what I consider risks your plan poses because in your paper you say so little about them. For example, you simply assume, with no reference to supporting theories or evidence, that credit formerly allocated by commercial banks will be invested more productively by the Fed, as if there were no substantial risk of it being invested less productively. (The same goes for your claim that, by short-selling particular securities through your proposed “OMO Plus” program, the Fed would be able to correct “artificially inflated” asset prices, as if identifying such prices were easy, even though the general consensus is that it is quite difficult.)
Nor does your proposal grapple with important counterarguments in the few instances in which it acknowledges them. For example, at one point you take note of the “frequently voiced” concern that plans for “digitizing central bank money will render central banks dangerously powerful and vulnerable to political manipulation and abuse.” You also note how any call for turning central banks into “large-scale investors in financial assets triggers familiar warnings about governments ‘crowding out’ private investment or ‘picking winners and losers’ in ostensibly private markets.” But instead of taking such concerns seriously, you write them off as being “rarely substantiated by reference to anything more specific than deeply internalized skepticism toward the government as an economic actor.” I’ve tried to show that there are plenty of specific reasons for worrying about a plan that would “end banking as we know it” and vastly increase the Fed’s role in credit allocations.
Many have said that concerns like mine shouldn’t play a part in assessing your fitness for the position of Comptroller of the Currency, because that position wouldn’t allow you to put your plan into effect. I appreciate the point. I also understand that we need regulators who are not about to cater to bankers’ interests when these interests run contrary to the public’s welfare, and you would certainly qualify on that score. Yet I can’t help thinking that someone placed in charge of a quarter of the nation’s banks, and 63 percent of U.S. bank deposits, ought to recognize the valuable services banks perform, and recognize them well enough to be willing to oppose any plan that would prevent them from continuing to perform those services. For this reason, I very much hope that you’ll take my earnest, if perhaps sometimes mistaken, criticisms of your proposal to heart.
Center for Monetary and Financial Alternatives
The Cato Institute
First published at Alt-M