Editors' note: This column is part of the Vox debate on "Lessons from Recent Stress in the Financial System"
Before March 2023, bank runs seemed to be of historical interest only. They had played a fatal role in the Great Depression of the 1930s, but the introduction of deposit insurance and the actions of central banks as lenders of last resort seemed to have done away with that problem. The financial crisis of 2007-2009 saw primarily runs on money market funds and breakdowns of bank funding through money markets.
By contrast, the March 2023 runs on Silicon Valley Bank (SVB) in California and on Credit Suisse (CS) have received much attention. The run on SVB had systemic effects in providing information about mid-size US banks overall. Equally, Credit Suisse was one of 30 institutions that the Financial Stability Board had designated as ‘globally systemic’.
Many economists think of runs as resulting from a fragility that is an essential feature of banking and that requires government support. The 2022 Nobel Memorial Prize was given for research “on banks and financial crises” with the finding that “avoiding bank collapses is vital.”
In the aftermath of the runs on SVB and Credit Suisse, many have called to stabilise banks by extending the coverage of deposit insurance and other government guarantees. Similar calls were made during the 2007-2009 crisis and again during the Covid crisis starting March 2020. Government guarantees and other supports expanded greatly in these crises already.
Calls to expand government guarantees are misguided. The breakdowns of SVB and Credit Suisse had little to do with the kind of fragility that Diamond and Dybvig (1983) studied in the research recognised by the Nobel Committee. In the Diamond-Dybvig analysis, runs are due entirely to depositors’ self-confirming prophecies about each other. Information about the value of the banks’ assets and their other liabilities plays no role.
By contrast, the runs on SVB and Credit Suisse were triggered by announcements that alerted depositors and other investors to deeper problems affecting the solvency of the banks. The failures of bank executives and supervisors to address these problems must be the starting point for any policy discussion.
The two cases are somewhat different. In the remainder of this column, we discuss Credit Suisse. SVB will be the subject of a companion column.
The final run on Credit Suisse started on Wednesday, 15 March 2023. On that day, the Swiss National Bank announced that Credit Suisse satisfied all capital and liquidity requirements and that it would provide liquidity support of up to CHF50 billion.
This announcement did not stop the run.
On Sunday 19 March, the Swiss authorities announced that UBS, the other Swiss megabank, would take over Credit Suisse for CHF3 billion in UBS stock. Moreover, certain Credit Suisse contingent securities (‘cocos’) treated by regulators as Alternative Tier 1 Equity (AT1), would be wiped out, while equity shareholders would retain CHF3 billion. The Swiss government would provide UBS with a second-loss guarantee of up to CHF9 billion against losses on Credit Suisse assets; this guarantee would set in after UBS had borne a first loss of CHF5 billion. The Swiss National Bank would provide both banks with liquidity support of up to CHF100 billion, with a loss guarantee from the Swiss government, in addition to the usual liquidity provision against collateral. Shareholders of the banks would not have the right to vote on these matters. The government invoked emergency executive powers to take these steps.
These actions raise several questions. First, why did the authorities avoid using the vaunted resolution procedures that had been introduced after 2008?
The total loss absorbing capacity (TLAC), which consists of equity, AT1 securities, and so-called bail-in debt meant to absorb losses to avoid government bailouts, amounted to CHF110 billion, almost 20% of reported total assets. It should have been easy to wipe out equity and AT1 securities and to convert bail-in debt into equity, thereby re-establishing solvency. In the past decade, authorities and many economists claimed that adequate total loss absorbing capacity would eliminate the too-big-to-fail problem, so we would see no more bailouts. Why was total loss absorbing capacity not used?
Second, did the authorities believe that Credit Suisse was solvent? If they believed that the bank was suffering from only a pure liquidity problem, why was the initial liquidity support from the central bank so limited? And why did the authorities end up taking so many measures that could be justified only if there were serious solvency concerns, for example the second-loss guarantee for UBS and the fiscal backstop for the Swiss National Bank, in addition to the write-down of AT1 securities? If the authorities doubted the bank’s solvency, why did the Swiss National Bank try to stop the run by a supposedly reassuring statement on 15 March, combined with an announcement of limited support? There is some incongruity between the announcement of 15 March and the measures taken on 19 March.
On the first question, the Swiss finance minister and the head of Finma, the Swiss supervisory authority, said that bank resolution was unsuitable because the Credit Suisse crisis involved a ‘loss of trust’ and a run.
Why had loss of trust not been considered when the rules for bank resolution were drawn up? Before addressing this question, we first discuss the fatal lack of trust.
A loss of trust usually occurs for a reason. Whereas the theoretical analysis of Diamond and Dybvig suggests that trust can disappear simply because each depositor expects other depositors to lose trust, in reality loss of trust in a bank is most often due to bad news, and it is grounded in serious concerns about the bank’s solvency.
In this case, the 2022 Annual Report of Credit Suisse, published on 14 March, was a major bearer of bad news (Credit Suisse 2023). This report shows a remarkable discrepancy between the so-called consolidated accounts of Credit Suisse Group and the accounts of the Credit Suisse Group Holding Company, which show the values of equity in the operating companies rather than the values of the operating companies’ assets.
The holding company shows a large loss of over CHF24 billion due to write-downs on the equity of the operating companies.
These equity positions and the write-downs on them do not appear in the consolidated group accounts on which public discussion has focused. For these accounts, management explained that the ‘fair values’ of all reporting units exceeded their previously reported book values and no write-downs were necessary. Management also explained that this assessment might change if the exodus of customers were to continue.
The auditors’ certifications of the accounts warned that management had failed to design a good method for assessing risks of mistakes in valuations. Management itself conceded that “internal control over financial reporting was not effective” (p. 50).
The 2022 Annual Report also gives numbers for the earlier exodus of depositors and clients that occurred in October 2022. In the fourth quarter of 2022, deposits declined by CHF138 billion, from CHF371 billion to CHF233 billion. Credit Suisse’s assets under management declined from CFH1.6 trillion on 1 January 2022 to CHF1.3 trillion; CHF100 billion of this decline occurred in the last quarter. The report conveys no sense of the existential threat to Credit Suisse from this shrinkage. It sketches a restructuring strategy with a cost reduction target of 15% and anticipated restructuring costs of CHF2.9 billion, which was clearly insufficient. In 2022, hardly any costs seem to have been cut. As revenues declined from CHF22 billion to CHF15 billion, operating expenses declined only by CHF1 billion, from CHF19 billion to CHF18 billion.
Credit Suisse management’s slow reaction to the bank’s crisis warranted a forecast of persistent losses. There is no trace of these anticipated losses in the reported accounts. Without a substantial change of strategy, however, such losses would surely lead to bankruptcy. For depositors worried about the safety of their funds, the fact that the accounts do not show the anticipated losses is irrelevant. With expectations of change for the worse, it just seemed safer to run.
Such concerns may explain the ambivalence of the authorities on whether Credit Suisse had only a liquidity problem or also a solvency problem. They may also explain why UBS required so much protection against potential losses. But why did the authorities not intervene sooner or more effectively?
The decline of Credit Suisse had begun earlier. Credit Suisse had paid large fines to US and UK authorities since 2014. In the spring of 2021, it transpired that Credit Suisse was heavily involved in the Archegos and Greensill bankruptcies, with losses of CHF4.7 billion just on Archegos. An outside investigation that Credit Suisse had initiated showed that its employees had overlooked many warning signals and concluded that the disaster was due to “a lack of competence” and a “weakness of risk culture,” with employees prioritising their own bonuses over concerns for the bank’s risks (Credit Suisse 2021).
This report shows Credit Suisse investment bankers routinely violating rules and instructions.
In 2022, Credit Suisse Group reported operating losses in all departments except for the Swiss bank. Throughout the year, its share prices declined. In December 2022, Credit Suisse raised CHF 3.9 billion in new equity, far from enough to compensate for the year’s losses. The overall picture is one of incompetence and recklessness.
For the authorities such developments pose the question of how to handle banks that are unsuccessful in their markets. In other areas of the economy, unsuccessful firms end up leaving the market. If they start out with enough equity, this process may take time, but at some point, creditors, particularly banks, will advise them to change their strategies or lose their funding. Market exits are useful because they reduce the waste of resources.
With banks, creditors are too weak and uncoordinated to exert such influence. They can only withdraw their funding, as happened with Credit Suisse. If government guarantees were to eliminate the incentives of depositors and other bank creditors to leave, the waste of resources from unsuccessful activities would be perpetuated, ultimately at the expense of taxpayers.
Supervisors tend to limit their interventions to verifying that capital and liquidity requirements are satisfied. The so-called Pillar 2 of Basel rules and the associated laws permit intervening more actively, but supervisors usually shy away from interfering with the banks’ activities. In the case of Credit Suisse, in autumn 2022, supervisors warned the bank to increase its liquidity buffers, but they do not seem to have reacted to the bank’s substantial and continued losses even though these losses were threatening its viability. On 5 April 2023, the head of the supervisory authority declared that interference with the bank’s strategy was none of the supervisor’s business (Amstad 2023).
In the same declaration, the head of the supervisory authority also said that the measures taken were preferred to a recovery and resolution procedure because imposing losses on the bail-in debt holders might have triggered a systemic crisis. This explanation is at odds not only with previous references to lack of trust as the key problem, but also with the total loss absorbing capacity (TLAC) narrative for dealing with the too-big-to-fail problem. According to this narrative, the holders of bail-in debt know that they can be made to share losses in resolution and must be prepared for such losses, with compensation through higher promised returns. The Swiss authorities’ statements and actions invalidate this narrative.
We have previously warned that the reforms of bank resolution since the crisis of 2007-2009 are not viable (Admati 2016, Hellwig 2014, Admati and Hellwig 2024 and forthcoming). It is time for authorities worldwide to abandon the fiction that reforms undertaken since 2008 have eliminated the too-big-to-fail problem.
Existing arrangements for resolving systemic institutions without disruptions and without bailouts have several fatal weaknesses.
First, there is no satisfactory mechanism to prevent fragmentation along jurisdictional lines once a resolution or bankruptcy procedure is triggered. Such fragmentation destroys operations that are integrated across the legally independent subsidiaries in the group, for example for cash management or IT systems. In the case of Credit Suisse, the derivatives subsidiary in London, which is systemically important, would have been seriously at risk. Fragmentation could be avoided if the management of resolution were left to a single authority, but such a ‘single point of entry’ approach is politically unacceptable to those countries whose authorities would have to step aside.
Second, existing rules seldom provide for funding a bank while the resolution procedure is implemented. The Swiss authorities’ concerns were legitimate in this regard, as EU authorities had seen in the case of Banco Popular Español in 2017. In the US, the Treasury may provide loans, and in the UK the Bank of England may use its funds with Treasury approval and potential backup. In the euro area and in Switzerland, however, there is no provision for funding. The ECB has made clear that they cannot step in if there are doubts about solvency.
Third, in many jurisdictions legal provisions about fiscal backstops are unclear.
In the US, taxpayers effectively come in if loans from the Treasury cannot be repaid. In Switzerland, the government’s guarantees to UBS and to the Swiss National Bank required an extraordinary executive act of the government under emergency law.
The Swiss deposit insurance system is based on self-regulation with contributions from the industry, without a fiscal backstop.
Finally, the authorities may not have the institutional and human resources needed for handling bank resolution. Thus in 2017, the Italian government asked Banca Intesa to undertake the liquidation of the bad assets of Banca Popolare di Vicenza and Veneto Banca. Similarly, the Swiss authorities preferred that UBS be in charge of Credit Suisse rather than dealing with resolution themselves.
Even for UBS, the task of reorganising and partially winding down Credit Suisse seems daunting.
The Credit Suisse case illustrates the failure of post-2008 attempts to eliminate the too-big-to-fail problem. Policymakers, regulators, supervisors, and economists must acknowledge this failure and turn to effective reforms. The problem of resolving institutions with systemic activities in multiple jurisdictions appears insoluble. Perhaps, therefore, we should not have financial institutions that are systemically important in several jurisdictions.
The mega-banks that have emerged over the past 30 years are enormous, extraordinarily complex, and opaque. Episodes like Archegos with Credit Suisse or the London Whale with JPMorgan Chase suggest that these institutions are too big and too complex to manage, as well as too dangerous for financial stability, global or national.
Given the impossibility of resolving such institutions without imposing severe costs on the financial system or on taxpayers, it is essential to avoid getting close to a risk of failure. Hence the authorities must be able to step in and intervene much sooner than the Swiss authorities did. They should engage in prompt corrective action when they can still address dangerous developments without immediately creating solvency concerns and causing runs.
Doing so would be possible if equity requirements were much higher than they are at present.
This call for much more equity is not just about increasing the capacity to absorb losses. Many economists believe that, in view of the bankers’ resistance to higher equity requirements, certain kinds of debt (cocos with high trigger before resolution, total loss absorbing capacity in resolution) might also absorb losses instead of equity. Whether funding by such debt is conducive to financial stability is controversial.
Our argument here, however, is concerned with the governance of supervision and its implications for the governance of corporate management.
Facing adverse developments, both supervisors and managers procrastinate. Losses leading to a violation of equity requirements would require the supervisors to step in and demand that the bank avoid payouts to shareholders and raise more equity. If this intervention occurs at a much higher level of required equity than banks have today, it would not automatically create a concern about solvency and trigger a run.
Raising new equity would impose more of the downside risk on shareholders and bank managers and prevent them from shifting it to others. The ability to raise equity provides a market-based ‘stress test’, a kind of evidence about the market’s assessment of the bank’s prospects that is more informative than the stress tests used by regulators today.
The need to bear more of the downside risk rather than just the upside gains would give shareholders stronger interest in challenging flawed management strategies.
Admati, A R (2016), “The Missed Opportunity and Challenge of Capital Regulation”, National Institute Economic Review 235 (1): R4-R14.
Admati, A R, P M DeMarzo, M F Hellwig and P Pfleiderer (2013), “Fallacies, Irrelevant Facts, and Myths in the Discussion of Capital Regulation: Why Bank Equity is Not Socially Expensive”, Working paper.
Admati, A R, P M DeMarzo, M F Hellwig and P Pfleiderer (2018), “The Leverage Ratchet Effect”, The Journal of Finance 73(1): 145-198.
Admati, A R and M Hellwig (2013, expanded edition forthcoming, January 2024) The Bankers New Clothes: What’s Wrong with Banking and What to Do about It, Princeton University Press.
Amstad, M (20230, Address at a media event on 5 April 2023 in Bern.
Brunetti, A (2023), “Big banks must become globally resolvable – or significantly ‘smaller’”, VoxEU.org, 1 May.
Credit Suisse (2021), Report on Archegos Capital Management, Special Committee of the Board of Directors.
Credit Suisse (2023), Annual Report 2022.
Diamond, D W and P H Dybvig (1983), “Bank Runs, Deposit Insurance, and Liquidity”, Journal of Political Economy 91(31): 401-419.
Financial Stability Board (2021), “Evaluation of the Effects of Too-Big-to-Fail Reforms”, Final Report, accessed 3 May 2023.
Hellwig, M F (2014), “Yes, Virginia, There is a European Banking Union! But It May Not Make Your Wishes Come True”, In Proceedings of the 42nd Economics Conference, Austrian National Bank, Vienna: 157-181.
Hellwig, M F (2021), “Twelve Years after the Financial Crisis—Too-big-to-fail is still with us”, Journal of Financial Regulation 7: 175–187
Wallace, N (1988), “Another Attempt to Explain an Illiquid Banking System: The Diamond and Dybvig Model with Sequential Service Taken Seriously”, Federal Reserve Bank of Minneapolis Quarterly Review 12(4): 3-16.
Wang, Z (2023), “CoCo Bonds: Are They Debt or Equity? Do They Help Financial Stability? — Lessons from Credit Suisse NT1 Bonds”, ECGI blog, 6 April.