An integrated approach to a safer and more resilient banking system

Keynote address by Mr Agustín Carstens, General Manager of the BIS, at the conference on "Exploring Conventional Bank Funding Regimes in an Unconventional World" co-sponsored by the Federal Reserve Banks of Dallas and Atlanta, Dallas, 18–19 July 2024. 

Let me first thank Lorie Logan for inviting me to speak at this event, which covers many of the key current policy issues in banking that are faced not only in the United States but also in many other jurisdictions.

The banking turmoil in Switzerland and the United States in March 2023 highlighted the importance of sound bank governance and risk management, strong bank supervision and regulation, appropriate mechanisms for liquidity provision by central banks, effective deposit insurance and adequate resolution regimes.

I want to emphasise today that we should not think about these dimensions in isolation, but in a holistic and integrated manner. In doing so, I will focus on liquidity risk, but not limit myself to it. Liquidity risk is inherent to banking because of the maturity and liquidity transformation banks conduct.

While liquidity problems can occur for many reasons, more often than not they reflect fundamental concerns about solvency and business models. This became evident during the March 2023 banking turmoil.

One of the key triggers of the failure of Credit Suisse and Silicon Valley Bank (SVB) was the market's surprise realisation that they needed to raise substantial amounts of additional capital but could not do so. The need, and inability, to raise capital reflected the underlying problems in these banks. In the case of Credit Suisse, poor management and governance over many years led to a growing perception among investors that its business model was unsustainable. Large withdrawals by wealth management customers in late 2022 were both a signal and a consequence of a weak business model. This was followed in March 2023 by the bank being cut off in funding from wholesale and derivatives markets. In the case of SVB, unhedged exposure to rising interest rates led to large losses that threatened its solvency, while the bank did not hold large liquidity buffers. These shortcomings reflect the fact that SVB had weak risk management and inadequate governance. Rapid deposit withdrawals ensued.

So, doubts about bank fundamentals often manifest themselves as liquidity problems. Of course, even sound financial institutions may face liquidity shortfalls if certain markets cease to function properly.

Improved approaches to prevent or manage liquidity distress would make the failures of banks – and other financial institutions – less likely. They can also help prevent contagion across markets. They can further reduce the cost of failures, for example, by facilitating the provision of central bank liquidity support for solvent banks or the orderly unwinding, sale or resolution of insolvent ones.

The starting point should be banks' own governance and risk management practices. But history tells us that, while this is necessary, it is not sufficient.

Therefore, prudential supervision and regulation remain key. During the Swiss and US episodes, for example, bank supervisors were not able, for various reasons, to take forceful and prompt supervisory enforcement actions when these banks exhibited clear weaknesses in governance and risk management. As I have argued previously, supervision needs to be strengthened to ensure that it can identify banks' weaknesses at an early stage and act forcefully to ensure that banks address them.¹ To do this, supervisors will need to have operational independence, to strengthen their forward-looking culture and to continuously seek to improve their capabilities by allocating more resources to develop expertise and leverage on innovative technology. At the global level, it is urgent to invest more resources in banking supervision.

In terms of prudential regulation, the implementation of Basel III reforms is delivering a clear enhancement in banks' resilience in terms of both solvency and liquidity. To be sure, Basel III is not designed to prevent all bank failures, nor is it applied to all banks. For example, SVB was not subject to the Liquidity Coverage Ratio (LCR) regulation introduced by Basel III, and unrealised losses on that bank's available for sale securities (AFS) were not deducted from regulatory capital as prescribed by the Basel standards. Nevertheless, by raising the overall levels of capital and liquidity in the banking system, the Basel Framework enhances banks' resilience, supports confidence and limits contagion. Therefore, it is important for Basel III implementation to be completed in all major jurisdictions.

In addition to better regulation and supervision, the lender of last resort function of central banks can also be improved. While it is appropriate and important that central banks can perform the lender of last resort function, it is neither cost-free nor risk-free – for central banks, for the financial system or for society at large. Excessive reliance on it undermines market discipline, contributing to a build-up of vulnerabilities. And it exposes central banks, and hence taxpayers, to potentially large liabilities.

Finally, the reach of deposit insurance is another important element of the framework. A large share of uninsured deposits was an important driver of the large-scale and rapid deposit withdrawals from the failed US banks. After the collapse of SVB and Signature Bank, all depositors were protected under a systemic risk exception. As those banks were subject to lighter regulatory requirements than those applied to larger banks, the systemic risk exception was used to deal with the failure of banks that had not been deemed systemic in life. In the light of this, there seems to be merit in discussing the circumstances under which these banks become systemic and how we can ensure that they are subject to sufficiently stringent supervisory and regulatory requirements when they are alive.

All this points to the need for a holistic and integrated approach to containing liquidity risk. The various measures to address liquidity risk I mentioned previously all have their costs and limitations. So we should not rely on any single one alone. Instead, we need to understand their interplay and design a proper mix, while recognising that the best combination of approaches will probably depend on specific characteristics of each jurisdiction.

Let me now move on to how we can improve on the current toolkit for the management of liquidity risks and the provision of liquidity in times of stress.

One question that tends to be raised is whether adjustments to liquidity regulations are called for. In the light of the events of March 2023, a review of the LCR and Net Stable Funding Ratio (NSFR) requirements has some merit. But this should be secondary to the need to implement all aspects of Basel III in full and consistently, and as soon as possible. That said, I am pleased to see that the Basel Committee will explore policy options related to liquidity risk over the medium term. Beyond that, there is scope for considering how to improve the usability of liquidity buffers. But we need to be realistic about what we can hope to achieve here. We cannot expect banks to maintain buffers large enough to weather all extreme liquidity withdrawal scenarios.

Another important aspect regarding the design of prudential liquidity regulation relates to the accounting treatment of instruments considered as high-quality liquid assets (HQLA) – particularly whether the classification of instruments as held to maturity may affect banks' ability or willingness to monetise them if needed. In the case of SVB, its securities recorded as held to maturity for accounting purposes did not need to be marked to market, even if they were considered part of a liquidity buffer. But when those bonds were to be sold to help meet outflows, the bank needed to recognise losses. This sudden loss recognition completely destabilised the bank, as it pushed it into a situation of capital insufficiency in the worst possible scenario, ie when liquidity needs were exacerbated. Clearly these situations need to be avoided.

Beyond liquidity regulation, I would like to highlight the interplay of four measures: liquidity provision through central bank facilities, business governance, internal liquidity risk management and supervision. I am focusing on these four measures given the weaknesses in them that surfaced in March 2023. A key goal is to increase ex ante resilience, rather than just ex post liquidity assistance in case of stress.

Let me start with the recent proposals to ensure that banks can access central bank liquidity in crisis times through the pre-positioning of collateral. Broadly speaking, there seem to be two proposals on the pre-positioning of collateral. One is to pre-position against all runnable liabilities, including all deposits and short-term debt – called the "pawnbroker for all seasons" by Mervyn King and Paul Tucker. The other one is to pre-position against all runnable liabilities other than insured deposits, as recently proposed by the Group of Thirty.

There are several arguments in favour of these proposals. They could ensure that institutions can seamlessly access central bank liquidity when needed, at less financial risk to the central bank than with current arrangements. The availability of access to central bank liquidity can boost confidence and reduce the risk of runs, making it less likely that the facilities will be needed. Pre-positioning requirements may also help steer banks away from excessive risk-taking. Indeed, pre-positioning is already in place in some jurisdictions, although not linked directly to runnable liabilities.

But pre-positioning of collateral also has drawbacks. Most notably, when it becomes compulsory in relation to runnable liabilities, the market may expect banks to maintain the minimum amount of collateral at all times and even more during stress periods. As a result, the pre-positioned collateral may become less usable. Moreover, calibrating haircuts is difficult, especially under systemic liquidity stress. In addition, the central bank's collateral pre-positioning policy may induce banks to change their asset profile and thus influence credit allocation, which is something central banks generally try to avoid.

In weighing up the costs and benefits of pre-positioning, we should not view it in isolation. Instead, we should consider it in the context of bank governance practices. In principle, pre-positioning requirements could form part of the toolkit to address weaknesses in bank governance. For example, banks that fall short in some aspects of governance or liquidity management practices could be subject to more stringent pre-positioning requirements. This would create an incentive for financial institutions to adopt measures to lower their pre-positioning requirements by reducing the overall liquidity risk profile or runnability of their liabilities. By considering the interplay between several measures to lower liquidity risk – central bank liquidity facilities, supervision and internal risk management – we could make it less likely that banks will need to call upon such assistance from the central bank, thereby strengthening ex ante resilience. 

Another way to help ensure that banks have usable collateral available in times of stress is to ensure banks' preparedness to tap central bank liquidity facilities. The events of March 2023 illustrated the importance of such preparedness. For example, SVB had not tested its ability to borrow from the discount window and lacked adequate collateral and operational frameworks. In this case, enhanced supervision could help bolster the effectiveness of central bank liquidity facilities. In particular, supervisors and central banks could regularly engage in detailed liquidity stress scenario discussions with banks, as is done in some jurisdictions. Such preparedness exercises would require banks to document the types of assets available for pre-positioning, complete the necessary legal preparations, and demonstrate their operational readiness to use central bank liquidity against collateral. As a further benefit, these measures would also help improve banks' governance, particularly their internal control and risk assessment mechanisms.

As these two examples show, in thinking about how to strengthen the management of liquidity risks we should not consider alternative measures in isolation. In particular, we should look to combine stronger supervision and greater emphasis on internal governance and liquidity management, with an improved system of central bank liquidity provision.

Let me say just a few words on non-bank financial intermediaries (NBFIs). Basel III only applies to banks. In the case of NBFIs, regulatory progress has been wanting. The interconnections between banks and NBFIs have increased over time. For example, rapidly growing private credit funds have liquidity lines with banks. Therefore, when NBFIs face liquidity problems, they will draw down the liquidity lines from banks, which can put pressure on banks. Furthermore, over the past five years severe disruptions in government bond markets in major advanced economies highlighted the role of various types of NBFI in propagating funding and market liquidity shocks across the financial system. Therefore, work by international standard setters to bolster NBFIs' resilience by limiting liquidity mismatches and leverage needs to be pursued with vigor.

Let me conclude.

Enhancements of the regulatory and supervisory frameworks under Basel III have brought about safer and more resilient banking systems globally. Nonetheless, the banking turmoil in March 2023 and the call on central banks to step in to provide emergency liquidity assistance show us that there is scope for improvement.

The priority is to fully implement Basel III. Beyond that, it is time to think about how we can enhance liquidity risk management, governance and central bank liquidity provision. In doing so, we need to take a holistic approach.

Pre-positioning of collateral can be an important element in this effort, but modalities matter and potential adverse effects need to be avoided.

Thank you very much for your attention.

¹        See A Carstens, "Investing in banking supervision", speech at the European Banking Federation International Banking Summit, Brussels, 1 June 2023.