Key Takeaways: The BCBS 2023 Banking Turmoil and Liquidity Risk Progress Report
Last month, the Basel Committee on Baking Supervision (BCBS) published its progress report on the 2023 banking crisis (March-May 2023). Some of the critical findings in this report align with what I wrote in my blog explaining the collapse of Silicon Valley Bank in July last year. Especially the confirmation that the accounting treatment of HTM securities was not the central issue in the collapse of SVB, compared to other vital issues.
Below are The Crucial Points From The Report and My Take:
1. The Bank Could Generate Liquidity Irrespective of HQLA Accounting Classification:
When the Silicon Valley Bank collapsed, the media and many banking professionals attributed the leading cause to the HQLA securities’ accounting treatment. Especially securities classified as held-to-maturity (HTM) are reflected on the balance sheet at amortized cost. It was said back then that SVB’s HTM portfolio had significant unrealized losses, and selling these securities would have brought those to light. I mentioned in my blog last year that I disagreed with the relevance of the accounting treatment to the SVB’s failure. Instead, the main issue was SVB holding illiquid off-the-run MBS securities, which dropped significantly in value when interest rates rose. Instead of holding these MBS securities (which are negatively convexed and extremely sensitive to rate increases), if SVB held only US Treasuries, its HTM portfolio wouldn’t have dropped significantly. Hence, it was not the HTM classification but the SVB’s choice of assets.
One theory was that HTM designation shielded others from seeing these losses. While HTM classification may have shielded these losses from the public for a few days, that does not avoid the real issue. I have discussed this in detail in my blog and below at a high level.
The Basel report acknowledges that banks can monetize HTM assets to generate liquidity not only via outright sales but also via repo transactions; hence, as long as the bank has quality assets, it can monetize them by paying lower haircuts. There are operational capabilities banks need to have in place to use the repo market, and US regulation has specific requirements for this.
2. The Limitations of Outflow/Run-off Rates:
The above table in the Basel report is extremely useful for understanding the speed of deposit outflows from the stress banks last year and during the 2008 financial crisis. If we focus on last year’s banking stress, SVB, Credit Suisse, and Signature Bank noted outflows far above LCR rates in just one or two days. FRC (First Rubublic) is showing artificially better results of 90 days for deposit outflows as regulators tried to save it (with the help of JP Morgan and Citi, etc.), which delayed its demise by three months. If this effort had not been made, I am sure FRC deposits would have been left at a similar speed as others. In general, last year’s banks facing stress experienced far higher and faster deposit outflows than calibrated in the Liquidity Coverage Ratio rules. In reality, if there is a bank run in today’s digitized age, ~100% is the only correct outflow rate, and this table highlights that.
Another aspect highlighted in the Basel report is that internationally active banks can face different stressed outflows in different locations. So, whichever outflow rate is lower than 90 -100% a bank uses in its stress test (especially for uninsured deposits) won’t protect it during stress or bank run situations. Hence, I believe banks must take all other essential steps highlighted in this Basel report to avoid a bank run-like situation at all costs. Of course, banks should try to calibrate the outflow rates as sensibly as possible in their internal stress tests.
3. Increase Monitoring Frequency:
One of the findings mentioned in this report is that banks should increase the monitoring frequency of their stress metrics such as LCR (Liquidity Coverage Ratio), NSFR (Net Stable Fundign Ratio), ILST (Internal Stress Test), etc., to daily or even intraday during the time of stress and business as usual times. l also mentioned in my blog that the tailoring rule in the US doesn’t make sense, and the smaller banks should be treated with the same standards as the larger peers. The tailoring rule currently allows smaller banks to report some of these metrics at a lower frequency and also has concessions regarding monitoring thresholds.
4. Monitoring of Concentration Risk:
Another conclusion is to implement the monitoring of concentration risks. Last year, The bank stress highlighted this issue as SVB, for example, concentrated on the private equity clients and Signature Bank to the crypto firms, both of which were extremely sensitive to negative news. Concentration to the specific products also needs to be monitored as some of these banks had concentration toward uninsured deposits or short term liabilities. SVB also had too much concentration to MBS in its HQLA portfolio compared to most of its stable peers.
5. Monitoring Risk Dynamics Throughout The Group:
One key takeaway from the Credit Suisse failure is the importance of monitoring risk dynamics throughout the group. The Basel report highlights that the efficient allocation of liquidity and collateral at the group level to individual entities is a crucial factor. The distribution of resources could alleviate local liquidity pressures or spread them throughout the group. In the US, there are specific regulations in place to avoid this.
6. HQLA Availability:
One of the issues mentioned in the report is that Credit Suisse had its HQLA set aside to cover daily operational and intraday liquidity needs. Hence, a large part of the CS HQLA portfolio was not genuinely available to cover the 30-day LCR stress period. The report suggests some changes in local regulations to address this issue. However, this issue is addressed in the US now as the new FR 2052a guidance issued in 2021 asks banks to identify HQLA securities pledged to cover intraday needs or any other operational requirements, and these securities are then not counted towards the HQLA calculation in the LCR numerator. However, it’s still essential for banks to have the ability to identify such earmarked HQLA securities daily to ensure accurate calibration of the HQLA balance.
7. Other Tools Should Complement LCR/NSFR:
The report highlights that other tools must complement LCR/NSFR and also questions the risks LCR covers. There is scope to calibrate the LCR/NSFR outflow rates further, but I am not sure, as I mentioned in #2 above, how much that will help. The failed banks had specific issues, and at least the US version of the LCR is very robust, but these smaller banks were given some concessions. Their interpretation and implementation of these rules had gaps, as I covered in my detailed blog earlier. There may be specific gaps, including intraday pledged securities in the HQLA, which Basel regulation needs to fix, but I mentioned that the US rule already addressed this in 2021. It’s correct that banks need robust intraday liquidity risk, trapped liquidity monitoring, and other capabilities. This report speaks mostly about LCR/NSFR and how it did not effectively stop the run. However, I think the internal stress test should be the binding constraint (not LCR/NSFR), and banks have the freedom to calibrate and be more conservative in run-off rates specific to their business. This report, however, does not highlight the importance of ILST in addition to the regulatory metrics.
Material Additional Risk Factors Not Covered By LCR:
In addition to the key findings above, the report discusses some key risks not covered by LCR. Ideally, I think the banks should consider these in their internal liquidity stress tests if the regulation is not covering these:
- Procyclical Increase in Prepositioning Requirements:
Some CCPs or counterparties can increase the collateral requirements with short notice. If you recall, Robinhood faced this issue during the Gamestop saga, and the report highlights the issues faced by Credit Suisse, in the case of CS, CCPs, and some counterparties that increased the collateral requirements during stress. Incoming payments from the CS counterparties were delayed; however, CS continued the regular payment schedule to avoid any negative signaling to the counterparties. Banks’ ILST methodology must address this, as regulation can’t be this dynamic. - Repo Market Functioning:
The repo market itself may stop functioning smoothly during severe market stress and can prove an unreliable source of contingent liquidity. Several efforts are underway in the US to address this issue, as I discussed in my previous post. Central clearing of treasuries, prepositioning of collateral at the discount window, etc., to name a few. However, these efforts mainly focused on the US Treasury security market. The FRB allows banks to post collateral other than treasuries in the US, but the haircut will depend on the FRB vendor data. Hence, the ineffective functioning of the overall repo market may be a risk that is hard to manage. The only asset class that may be largely immune to this soon is the US treasury market due to all the efforts agencies have made for quite some time now. - Lack of Effective Contingency Funding Plan (CFP):
The experience of US banks suggests that they lacked effective Contingency Funding Plans. The lack of operational capability to monetize HQLAs was a major driver, as per recent speeches from FRB Vice Chairman Barr; US banks have made progress in implementing operational capabilities to access the discount window in 2024. I covered this in my blog here.
