Remarks by Vice Chairman Travis Hill at the Mercatus Center on “Banking’s Next Chapter? Remarks on Tokenization and Other Issues”

March 11, 2024 / Source: FDIC

Thank you to the Mercatus Center for having me today. I am going to devote most of my remarks to a discussion of tokenization and the future of our financial system, and then will conclude by briefly discussing a few discrete issues related to bank failures.

Tokenization

Money and payments have been evolving for as long as they have existed. From general commodities to precious metals¹ to cash to credit cards, the methods that society has used to store and transfer value have changed dramatically over time, and each major upgrade to the monetary architecture has introduced both new benefits and new risks. Similarly, our payment, clearing, and settlement infrastructure – the plumbing at the heart of the financial system – has evolved considerably over the past several decades.

As most of us are reminded daily, electronic and digital forms of money and payments now predominate. Credit card, ACH, and other noncash payments have increased rapidly in recent years,² and the share of households that utilize mobile banking as the primary method to access their accounts continues to rise.³ Furthermore, the vast majority of what people generally think of as “money” is in reality ledger balances sitting on databases maintained by commercial banks.⁴

Against this backdrop, I am going to focus on one specific innovation that has been the subject of a tremendous amount of research and development in recent years: tokenizing commercial bank deposits and other assets and liabilities.⁵ At its most basic level, tokenization transforms the way ownership of assets is recorded and enables far-reaching new functions, as described further below. Although tokenized assets, including tokenized deposits, may reside on blockchains and other distributed ledgers, they should not be confused with blockchain-native bearer assets like Bitcoin and Ether (generally referred to as crypto-assets or cryptocurrencies). Instead, tokenization involves a representation of “real-world assets”⁶ on a distributed ledger, including, but not limited to, commercial bank deposits, government and corporate bonds, money market fund shares, gold and other commodities, and real estate.⁷

Last year, the Bank for International Settlements declared that “the monetary system stands at the cusp of another major leap.“⁸ Many other organizations and commentators have similarly expressed optimism that tokenization may have a transformative impact on banking and payments.⁹ This process is still in the early stages, with many open questions, but development is happening fast and occurring across the globe.

Potential Benefits, Challenges, and Risks

A number of financial institutions around the world have been actively exploring the possibility of tokenizing assets to improve the way we transfer value. Notwithstanding recent advancements, our payment and settlement systems remain slow and inefficient for many users, with delayed settlements for large classes of transactions (such as cross-border transactions and bond issuances) and numerous intermediaries, each adding costs. Tokenization and distributed ledgers have the potential to overcome many of these obstacles by operating 24/7/365 and introducing settlement finality in real time. But tokenization offers much more than just a shiny version of Zelle or Venmo. In particular, it offers programmability, the ability to hard-wire on the ledger future transfers of value that automatically self-execute based on the occurrence of future conditions; atomic settlement, or the simultaneous exchange and settlement of payment and delivery, including among multiple parties;¹⁰ and immutability of the shared ledger that can serve as a transaction record and reliable audit trail.¹¹

We already see powerful examples of how tokenization is beginning to deliver tangible benefits, such as the introduction of intraday-repo¹² and dramatic increases in settlement times for multi-currency bond issuances.¹³ While the existing use cases have focused on institutional customers, in the future, the benefits could expand to retail; to give one example, programmability may be able to simplify the home buying process by eliminating the need to place funds in escrow prior to closing.¹⁴

Similarly, atomic settlement introduces enormous efficiencies to the functioning of key markets. For example, settlement lags in the foreign exchange market – in which approximately $7.5 trillion in trading occurs each day¹⁵ – could be eliminated,¹⁶ and there is potential for lag times in the trade finance industry – which typically relies on document exchanges among a large number of entities – to be greatly reduced.¹⁷

Of course, as with most things, there are also challenges and risks. On the technological and operational side, there remain many open questions. For example, if tokenization plays a central role in our future financial system, will there be a small handful of unified ledgers on which all transactions occur, or will many institutions maintain their own blockchains? To what extent will these platforms be interoperable with one another, so that customers using different blockchains can transact seamlessly with each other in a safe and secure manner? From a legal perspective, what additional work is needed to clarify the extent to which ownership and other rights associated with a given asset attach to and move with the token?¹⁸

These and many other critical questions will be answered – one way or another – as financial institutions, developers, regulators, and other stakeholders continue developing the technology. Meanwhile, global standards are being established, directly or indirectly, and with many non-U.S. jurisdictions actively engaged in this area,¹⁹ the United States risks ceding influence at this critical stage.

When it comes to risks, as the resolution authority, we at the FDIC are always focused on bank runs and resolvability. Instant settlement has the potential to further increase the speed and intensity of runs.²⁰ Programmability may make it easier for customers to automatically remove funds following negative news about a bank, which could exacerbate runs. We also need to consider how to ensure that some form of “off switch” exists to ensure tokens stop moving immediately at the point of a bank failure.

On the other hand, tokenization can also mitigate a range of risks. Atomic settlement reduces the risk of loss in the time between payment and delivery,²¹ and the risk that transactions do not occur altogether because a party lacks trust that delivery will occur after payment, risks that can be magnified during market breakdowns.²² Programmability can allow a bank to respond to liquidity stresses immediately and automatically, moving liquidity when and where it is needed. And a more resilient IT infrastructure can significantly reduce payment errors.²³

Illicit finance presents another set of challenges that many are currently working to address. For example, to comply with existing “know your customer” and related obligations, it may be possible to embed a “whitelist”²⁴ within a token (i.e., only authorize the token to be transferred to pre-approved wallets), or, on a permissioned ledger, to restrict access at onboarding. In addition, the transparency and immutability of the ledger can help regulators and law enforcement agencies obtain accurate and verifiable data on past transactions, and seize assets from criminals.²⁵

Regulatory Approach and Need for Clarity

Another challenge is the regulators. In 2021, the banking agencies developed and published a roadmap for providing clarity to the public on a range of legal and policy questions related to digital assets.²⁶ The objective was to set policy across the board in a consistent manner. Staff members across the three agencies made meaningful progress in late 2021 and early 2022, but nothing was ever released publicly, and instead, the agencies established processes under which institutions must engage with their regulator on an individual basis before engaging in any activities related to digital assets.²⁷ At the FDIC, this process applies not just to “crypto,” but also includes “participating in blockchain- and distributed ledger-based settlement or payment systems, including performing node functions.“²⁸

I appreciate the need for regulators to be deliberative and careful in approaching these issues. We should do our homework and make sure we understand the implications of new technologies that can reshape banking. And I recognize the value in being cautious regarding the extent to which the FDIC-insured banking system engages with the crypto economy.

But there are significant downsides to the FDIC’s current approach, which has contributed to a general public perception that the FDIC is closed for business if institutions are interested in anything related to blockchain or distributed ledger technology. The confidential nature of the existing process means there is little public information on what types of activities the FDIC might be open to, if any. I have heard of some cases in which requests submitted by institutions have gone unanswered for long periods of time, while other institutions have spent months responding to a long stream of information requests, diverting attention away from developing new technologies and systems. While the largest banks are able to hire consultants and staff in Washington, D.C. to read the tea leaves to discern what might be approved, the message being heard by the vast majority of the industry could be interpreted as don’t bother trying.

I recognize that sometimes it can be difficult for regulators to issue broadly applicable policy in areas where the technology is evolving quickly, but I think our goal should still be to provide as much clarity as is feasible regarding what is permissible and what we consider safe and sound. To the extent that we need to maintain a bank-by-bank approval process, it is critical that we provide feedback to institutions in a timely way. Even if the answer is “no,” it is better to give institutions a clear answer with an explanation than to leave them in indefinite limbo when there is no prospect of a “yes” answer. I also think it is important that we allow experimentation and testing – in particular in cases where there is no material risk to the institution – without requiring a lengthy approval process.

Furthermore, it would be helpful to provide certainty that deposits are deposits, regardless of the technology or recordkeeping deployed, and if there are reasons to distinguish some or all tokenized deposits from traditional deposits for any regulatory, reporting, or other purpose, the FDIC should, following an opportunity for public comment, explain how and why. And finally, the agencies need to distinguish between “crypto“ and the use by banks of blockchain and distributed ledger technologies. I do not think banks interested in the latter, insofar as it simply represents a new way of recording ownership and transferring value, should need to go through the same gauntlet as banks interested in crypto.

Custody

On a related note, in March 2022, the Securities and Exchange Commission (SEC) issued Staff Accounting Bulletin 121 (SAB 121), which provides that an entity that safeguards crypto-assets should recognize such an asset on its balance sheet as both an asset and a liability.²⁹ This treatment sharply departs from how custodians account for all other assets held in custody, which are generally held off-balance sheet and treated as the property of the customer, not the custodian. On-balance sheet recognition triggers the full panoply of capital, liquidity, and other prudential requirements only for bank custodians, which makes it prohibitively challenging for banks to engage in this activity at any scale. It is worth asking whether it is in the public interest for one crypto exchange to provide custody services for most of the market in approved Bitcoin exchange-traded products, while highly regulated banks are effectively excluded from the market.³⁰ Additionally, the SEC’s definition of “crypto-asset” is extremely broad and could be read to capture not just blockchain-native assets but also tokenized versions of real-world assets.³¹ I think this is a clear example of why it is generally constructive for agencies to seek public comment before publishing major policy issuances, and at a minimum believe it would be helpful to clarify that SAB 121 does not apply to the wider universe of tokenized assets beyond blockchain-native assets.

Tokenization Coda

In 1958, after a number of other banks had tried and failed to introduce similar products, Bank of America famously launched its first credit card in Fresno, California, and within a year had distributed cards to “every nook and cranny in California.”³² This initial effort was disastrous, and Bank of America suffered substantial financial losses, along with reputational damage.³³ But the bank persevered and was making money on the business by 1961,³⁴ while credit cards would go on to revolutionize how millions of Americans pay for things. But it is worth asking – would regulators today ever allow a bank to work through those types of setbacks in developing a new way for customers to make payments?

We should certainly be humble³⁵ when it comes to making any sorts of predictions about what the future financial system will look like, but as financial institutions and developers around the world continue to develop blockchain and distributed ledger technologies, a poor regulatory approach to these issues presents substantial opportunity costs for bank customers and the U.S. economy, discourages institutions from investing in the future, and cedes influence to non-U.S. jurisdictions.

Bank Failures

I am now going to pivot and briefly discuss a few other topics. As we consider the changing landscape of banking and payments, we should remain mindful of our core mission of minimizing the risk and severity of disorderly bank failures. In that vein, I’m going to pose three questions for policymakers to consider:

1. Should External Auditors Issue “Going Concern” Opinions About Banks?

Under U.S. auditing standards, the external auditor of a company is generally required to assess the company’s future prospects as a going concern, and publicly disclose if it concludes it is more likely than not that the company will be unable to meet its obligations when they come due over the next year.³⁶ But unlike any other industry, to an unparalleled degree, because of the economic and social costs of bank failures, the U.S. has built an enormous apparatus to minimize the risk of banks failing³⁷ and a special resolution regime to maximize the likelihood that such failures occur in an orderly way. Meanwhile, a public going concern opinion is often described as the “death knell“ for an institution and the “nuclear option” for the auditor.³⁸

Today, auditors of stressed banks face a Sophie’s Choice-like dilemma. Auditors do not want to repeat the perceived errors of Silicon Valley Bank (SVB) and Signature Bank, in which an auditor issued clean audit opinions shortly before their failures³⁹ and is now subject to lawsuits.⁴⁰ But auditors also appear to recognize the impact that publicly issuing such an opinion can have on public confidence in a bank, and do not want to be blamed as the proximate cause of the bank’s failure. So perhaps unsurprisingly, over the last year, we’ve had numerous examples of auditors discussing issuing going concern opinions behind the scenes, and no examples⁴¹ of going concern opinions actually being issued.⁴² The fundamental problems are that bank failures are generally hard to predict up until the very end,⁴³ and publicly predicting them can be a self-fulfilling prophecy.

2. To What Extent, if Any, Should Poorly-Rated Banks Have Reduced Access to the Discount Window?

The Federal Reserve fulfills its lender-of-last-resort function by offering ongoing liquidity to banks through the discount window, under terms consistent with the famous Bagehot dictum to lend freely on good collateral at a penalty rate.⁴⁴ Banks with poor supervisory ratings,⁴⁵ very weak capital ratios, or otherwise deemed to not be in sound financial condition are only eligible to borrow from the discount window on secondary credit, which is typically overnight and applies a higher penalty rate and higher haircuts to collateral than primary credit.⁴⁶ According to the FDIC’s post-failure report, the final nail in First Republic Bank’s coffin was its shift to secondary credit, at which point the bank lacked sufficient liquidity to meet its obligations.⁴⁷ The knowledge that this was the potential consequence of a CAMELS downgrade was a regular topic of conversation in the weeks leading up to the bank’s failure. In addition, more generally, just the potential that a bank might be limited to secondary credit, even if the institution is not actively borrowing from the window, can impact various assessments made by the bank’s auditor, including the bank’s ability to hold securities⁴⁸ and its prospects as a going concern.

Which banks get moved to secondary credit at times appears… curious. Some banks are poorly rated (and thus ineligible for primary credit) because of compliance or operational weaknesses, without any evidence suggesting a deterioration in financial condition. Perhaps at a minimum, either the Federal Reserve’s presumption that 4- or 5-rated banks are ineligible for primary credit should only apply when the poor rating is a result of unsound financial condition,⁴⁹ or alternatively perhaps supervisors should only downgrade a bank to a composite 4 or 5 for reasons related to financial condition.⁵⁰ Meanwhile, in other cases, banks that are not poorly rated get moved to secondary credit, with little information available regarding how such decisions get made, but yet a distressed First Republic remained on primary credit for several weeks until its ratings downgrade.

3. To What Extent Should the Federal Home Loan Banks (FHLBs) Lend to Struggling Banks?

Anecdotally, it has seemed over the past year that the FHLBs have been quicker to restrict, and less predictable in restricting, lending to banks experiencing stress.⁵¹ I appreciate that the FHLBs are poorly positioned to serve a lender-of-last-resort-type function,⁵² and that there are deeper questions regarding the FHLBs’ footprint that are worth asking, but I encourage policymakers to think holistically about the implications of cutting off banks from the FHLBs when stress occurs. The ultimate costs if the institution subsequently fails are likely to be borne by the FDIC rather than the FHLBs,⁵³ and it is worth remembering that once a bank has reached that stage, its options to meet liquidity needs are likely to be limited, with all the alternatives potentially also costly to the Deposit Insurance Fund (DIF).⁵⁴ In addition, unpredictability around FHLB access can present challenges for banks and supervisors, for example when establishing and assessing contingency funding plans.⁵⁵

Bank Failures Coda

Putting this all together, one can imagine a sequence of events like the following: a bank experiences an adverse event, resulting in its FHLB cutting off funding, resulting in its primary regulator downgrading its CAMELS ratings, resulting in the Federal Reserve moving the bank to secondary credit, resulting in the bank’s external auditor concluding the bank lacks the capacity to hold its HTM securities to maturity, resulting in the bank’s tangible common equity ratio plummeting, resulting in the external auditor issuing a public going concern opinion, which sparks a loss of confidence in the bank that leads to its abrupt and disorderly failure.⁵⁶ As a practical matter, this final outcome is often avoided because individuals along the line prioritize avoiding an unnecessary bank failure, but we should not assume that will always be the case.

If a bank is insolvent or does not have a viable future, authorities should move swiftly and decisively to put the institution into receivership. But weak banks can and often do survive and recover,⁵⁷ and we should be thoughtful in considering policy choices that may further cripple wounded institutions and reinforce the procyclicality of our current system.