In medicine, “zebra” describes an instance when the doctor diagnoses an unlikely but eye-catching condition instead of a less noteworthy but more probable one that also fits the symptoms. Professor Theodore Woodward of Maryland University coined the expression when he admonished his students “When you hear hoofbeats, think of horses not zebras.”
As the zebra phenomenon illustrates, even experts tend to judge more memorable events as more likely – a bias known as availability – and to remember unusual events more clearly than more mundane ones. Uncorrected, such predispositions can lead to bad prescriptions, whether in medicine or further afield.
Woodward’s rule-of-thumb came to mind as policymakers around the world reacted to the announcement of Libra, the Facebook-led digital currency that is to launch in the first half of 2020. To say that their responses struck a note of caution is an understatement. Representative Maxine Waters called on Facebook to halt the project. Her Senate ally Sherrod Brown warned that “we cannot allow Facebook to run a risky new cryptocurrency out of a Swiss bank account.” The international reactions were perhaps less shrill but not much warmer, with French Finance Minister Bruno Le Maire openly worrying that Libra might dispute states’ currency monopolies.
Hearings before the House Financial Services Committee, which Rep. Waters chairs, and the Senate Banking Committee, of which Sen. Brown is the Ranking Member, are due to take place on July 17 and 16, respectively. Policymakers here and elsewhere, however, appear to already have made up their minds about Libra.
In so doing, I fear that the gatekeepers of the global financial system are mistaking horses for zebras: focusing on improbable risks and discounting the tangible but less headline-grabbing benefits of a global low-cost payments application such as Libra’s founders envisage. Because Facebook is leading the effort and marketing Libra as a cryptocurrency, the official responses are informed more by connotation (risk, monopoly, market power, data abuses), than by any actual knowledge concerning what Libra will do.
Ironically, the fears being expressed about Libra reflect the implicit belief that it may prove to be an extremely important and popular innovation. The new venture will take advantage of an existing user base among the partner organizations of easily more than three billion. Thanks to network effects, the prospects for Libra’s further growth should it become popular at all appear great. Those potential users include tens of millions of people who do not currently hold bank accounts, but who are increasingly likely to own smartphones. To condemn Libra, simply because it appears capable of capturing a wide network of users, before it even has a chance of proving its worth would foreclose better provision of banking and payment services to millions of lower-income consumers.
Are there in fact sufficient grounds for intervening to put a stop to or otherwise limit the spread of Libra? In this and a forthcoming post, I examine four types of risk that Libra skeptics have warned against: systemic risk, monopoly risk, security risk, and discrimination risk. My purpose is not to defend, which is as yet untested. I only wish to show that a close look at the alleged risks of Libra makes them appear more remote than the fears expressed by certain high-profile officials suggest. They’re zebras, not horses.
In today’s essay I consider Libra’s potential to present a systemic risk to the banking system. In the follow-up post, I will discuss the likelihood that Libra will create significant monopoly, discrimination, and national security risks.
Systemic risk
Mark Carney, Governor of the Bank of England, has said that “Libra, if it achieves its ambitions, would be systemically important.” A cryptocurrency commentator has suggested that Libra “wants to be too big to regulate.” Matt Stoller, one of Big Tech’s loudest critics, recently asked in the New York Times: “What happens if all users want to sell their Libra currency at once, causing the Libra Reserve to hold a fire sale of assets?”
So fresh is the memory of the last crisis in regulators’ minds, and so large looms the size of the technology giants, that many unquestioningly accept claims such as the ones cited above. But a different picture emerges if we attempt to answer Stoller’s question by reference to what we know so far about Libra.
For every unit of Libra that it issues, the Libra Association will hold a combination of bank deposits and “short-term government securities in currencies from stable and reputable central banks.” These assets will form the Libra Reserve, which by its promise of redemption will, Libra’s backers hope, lend the digital currency credibility and thus stability of purchasing power.
My colleague Larry White recently explained the Libra white paper’s ambiguity regarding the convertibility of Libra into the Reserve’s component assets. He concludes that Libra will resemble a government-bond money-market fund, which is liquid and relatively stable (and often treated by holders and regulators as a cash equivalent) but whose net asset value does fluctuate.
Despite being backed by various national currencies, relative to the value of any one national currency (whether or not that currency forms part of the Libra Reserve), Libra’s exchange rate will be free-floating rather than fixed. That fact may hinder adoption, because users may be wary of holding an unstable medium of exchange. But it also casts doubt on the claim that Libra poses systemic concerns. Would a simultaneous decision by users to sell their Libra holdings pose a systemic risk?
Larry, in his piece, suggests not. Systemic problems in banking typically arise when people realize the bank has less in liquid assets than in demandable liabilities such as deposits. Every depositor then has reason to rush to cash out before others have done so. The bank has promised to satisfy all depositor claims one-for-one, and it does so until the money runs out. Hence the incentive to “run” on the bank.
As Larry explains, such an incentive is absent in the case of a mutual fund such as the Libra Reserve, because any loss in net asset value is proportionately distributed among Libra holders, whether they cash out soon after the depreciation or not. Libra holders will therefore lack any incentive to “run” on the Libra Reserve.
Given that a drop in Libra’s value, however large, would be unlikely to cause a massive redemption by Libra holders, could a large-scale redemption (whatever the reason for it) endanger financial stability anyway? After all, users’ exchanging Libras for cash would mean banks’ having to pay out on the deposits that formed part of the Libra Reserve. It would also mean proportionate declines in the price of the government securities included in the Reserve, if the new supply was not met by offsetting demand.
I believe a rapid and massive sell-off of Libra could not, on its own, set off a systemic crisis. Consider a scenario in which there is $100,000,000 worth of Libras in circulation, of which holders in the U.S. suddenly wish to exchange $20,000,000 worth for U.S. dollars. Assume that 10 percent of the Libra Reserve is made up of bank deposits denominated in U.S. dollars, euros, yen, and pound sterling; and that the remaining 90 percent consists of (comparably) safe bonds issued by rich-country governments.[1] The specific share of each currency and country bond matters little for our purposes.
The $20,000,000 redemption would cause the Libra Association to shrink its Libra Reserve by 20 percent, selling Libra’s component assets (presumably) in proportion to their share in the Reserve (and not in proportion to the currencies into which holders wanted to exchange it). So, as holders sought to sell Libra, the authorized resellers managing the Reserve would sell deposits and bonds in equal proportion for every unit of Libra that was redeemed.
Could that process cause a liquidity crisis among the banks who held Libra Reserve deposits? Further, could that liquidity crunch turn into a full-fledged solvency crisis if the Libra redemption was large and fast enough? There are reasons to think either of those crises unlikely. First, deposits will not be held by one bank in one country but by many banks in many countries. They would thus form a small liability on any individual bank’s balance sheet. Second, Libra users are unlikely to hold the redeemed amount in cash and will instead re-deposit it at a bank, statistically returning a significant proportion of deposits to the banks affected.[2] Because banks, unlike what Libra has promised to do, keep only a fraction of their deposits in reserves, it may well be the case that the total amount of bank deposits grows following a large Libra redemption.
Third, bank deposits are typically government-insured up to very generous limits – $250,000 in the United States. While deposit insurance presents moral hazard and there is evidence that more generous insurance schemes encourage greater risk-taking, for our purposes deposit insurance would largely eliminate the incentive by both Libra holders and non-holders to run on their banks following a large-scale Libra redemption.
Fourth, bankers who were asked to pay out on Libra Reserve-related deposits would have ample access to short-term emergency liquidity from their central banks. As in other instances when they act as lenders of last resort, central banks would run the risk of mistaking insolvent for illiquid institutions. But that distinction would not depend on Libra. Rather, it would depend on individual banks’ (and the financial regulators’) prudential management. Unless banks had solvency problems of their own, a mass Libra deposit redemption would not cause a systemic crisis.
What about the redemption’s impact on bond markets? Could Libra holders rushing to cash out cause a sovereign debt crisis? Again, it is doubtful that they could, absent other reasons to “run” on a country’s bonds.
According to the Libra Association’s white paper, the Libra Reserve will consist of a diversified portfolio of historically (relatively) stable bonds. When Libra holders redeem, the capital outflow will not be concentrated on one country’s bonds. Rather, Libra will sell government securities from a number of jurisdictions in proportion to their share in the Reserve. Moreover, it is unclear that there would be an outflow at all, since the jurisdictions whose securities are likely to be chosen also serve as safe havens for investors in times of uncertainty. As with bank deposits, the very issuers of the bonds affected by a large Libra redemption would probably see those funds return as direct bond purchases.
In other words, a massive redemption caused by factors unique to Libra would not seem enough to cause a wider systemic crisis. Things might turn out differently if the Libra Association went back on its commitment to purchase stable assets on a one-for-one basis for every unit of Libra it issued. However, the prospect of a rapid loss of confidence in Libra as a result of such a turnabout, and the potential for customer lawsuits, provide strong disincentives against the Libra Association’s reneging on its promise of full backing. Still, given the commitments that the Libra Association has made so far, the grounds for expecting Libra to present a new systemic risk are lacking.
[1] I assume the share of bonds in the Reserve to be much greater than that of bank deposits, because deposits typically pay lower interest rates, resulting in an opportunity cost for the Libra Association.
[2] I assume, I hope realistically, that the Libra Association will hold deposits at large and reputable banks across jurisdictions with comparably stable and reliable financial markets. These are the same institutions at which Libra holders who cashed out would bank. [I suspect they might prefer to keep deposits at TBTF banks.]
