Introduction
The UBS purchase of Credit Suisse on March 19th raises a number of issues for counterparty risk management. The speed of the transaction also raises questions; what did the Swiss regulators know (or fear) that the rest of the industry didn’t see (know)? The specific counterparty risk issues include:
- The efficacy of resolution of a major GSIB bank during a crisis.
- The rapid deterioration of a bank’s liquidity coverage ratio due to 30-day net cash outflows exceeding rule-based projections.
- The need for more transparency on the composition of deposits e.g., percentage uninsured.
- The role technology and social media played in exacerbating the banking crisis.
Everyone Has A (Resolution) Plan
One would think that the crisis at Credit Suisse would have been the opportunity to apply the Resolution Plans for winding down the bank but this didn’t happen. Days after the purchase of Credit Suisse the Swiss National Bank chair Thomas Jordan said “Resolution in theory is possible under normal circumstances, but we were in an extremely fragile environment, with enormous nervousness in financial markets in general… resolution in those circumstances would have triggered a bigger financial crisis, not just in Switzerland but globally.” What are “normal circumstances” when these resolution plans could be implemented? It is hard to imagine that the environment would not be “fragile” and “nervous” if a large GSIB is tumbling towards insolvency.[1] [2]
How Liquid Are You?
Credit Suisse’s Liquidity Coverage Ratio (LCR) of 144% was well above the regulatory threshold at the end of Q4 2022. This sounds pretty good so where’s the problem? The LCR is a fraction with High Quality Liquid Assets (HQLA) in the numerator and 30-Day Net Cash Outflows (NCO) in the denominator. The NCO is where the problem begins. The major banks’ LCR Disclosure document, which is submitted quarterly, has specific detail on the calculation of the NCO. Quoting from the Goldman Sachs 2022 submission: “The LCR Rule defines NCOs as the net of cash outflows and inflows during a prospective stress period of 30 calendar days. NCOs are calculated by applying prescribed outflow and inflow rates to certain assets, liabilities and off-balance- sheet arrangements.” The bank takes each specified category and applies the rule-based haircut to it. For example, under the category “Deposit outflow from retail customers and counterparties” Goldman has $263 billion that under the rules-based haircut contributed $48 billion in NCO. This included “Stable retail deposit outflow” of $49 billion haircut down to $1.5 billion. These are the haircuts that the rule applies but how will they compare to the deposit outflows in a real crisis? Goldman in the introduction to the NCO section says, “Due to the inherently uncertain and variable nature of stress events, the firm’s actual cash outflows and inflows in a realized liquidity stress event may differ, possibly materially, from those reflected in the firm’s NCO’s.”
Let’s consider for a moment the similarity of two banks, that on the surface do not seem at all similar: Silicon Valley Bank and Credit Suisse. What they had in common was large, chunky, uninsured deposits. It was the rapid increase in the outflow of these deposits that made them “differ, possibly materially” from the NCO forecast (and its 97% haircut of stable retail deposits). It would seem that the NCO haircut methodology needs reexamination. Perhaps a distinction is needed between haircuts for insured and uninsured (large) deposits.
The other issue is the HQLA and how accessible are these assets in a crisis. This was further complicated at Credit Suisse because some of the HQLA and 144% LCR were trapped in the ring-fenced Swiss entity. In general however it would seem to make sense that the composition and near-term accessibility of the HQLA (e.g., the first seven days of the crisis) also need to be tested and verified.
That Was Then This Is Now
The next area of concern is what I would call Intra-Quarter Risk. The financial metrics that the public sees only come at the end of each quarter. This is problematic when a crisis erupts half-way into the next quarter. How reliable are those quarter-end metrics? One would think that the regulators see the banks’ capital and liquidity metrics more often (weekly?) during the quarter. Were these intra-quarter submissions by Credit Suisse what made the Swiss regulators uneasy? Again, were the regulators seeing a rapid deterioration of the LCR and significant divergence from the NCO calculations? This is a blind spot for those responsible for counterparty risk management.
Meme 2.0
The next contributing factor to this crisis, and one that does not have an easy answer, is technology and social media. When SVB started to get into trouble some of the large uninsured depositors took to social media urging fellow depositors to get out now. Call it a techno fueled run on the bank.[3] If everyone just stayed calm everything would probably have been alright. No one wants to get caught being the last one out so the craziness spreads. This brings to mind another financial saying: The market can be irrational longer than you can be liquid. Social media was an accelerant on the fire that was SVB. It quickly spread to First Republic, other regionals and finally Credit Suisse.[4] The other thing is that the technology we all have today allows us to move money quickly with a few keystrokes on an app on our smart phone. Very convenient but very dangerous in a bank run.
Where Do We Go From Here?
We may never really know what went down during that weekend when UBS was forced to buy Credit Suisse but there are implications that need to be considered. These are some of the takeaways from the recent banking crisis:
- Since Dodd Frank and Basel III we have taken the regulatory metrics at face value. The farther out from quarter-end the more likely the regulatory metrics like LCR will become stale. Unless there is some kind of systemic event like the past month these metrics are probably still generally where they were at the previous quarter’s end. In a crisis however, there should be less confidence that they are valid. Metrics like the LCR (and the NCO forecast) can deteriorate much more quickly than modeled and quickly burn through the HQLA. Large uninsured deposits may need more conservative haircuts in the NCO.
- Following on the last point, will regulators require more disclosure on deposits: percent uninsured; average size of uninsured deposits? Is it possible to get more detailed information from the financial statements on retail deposits? The large banks may decide to be more transparent about deposits in the coming Q1 2023 earnings reports.
- The banks are all sitting on large unrealized losses on the securities holdings. Having to realize some of these losses to raise cash to meet depositor withdrawals is what tripped the crisis at SVB. It is not a high probability that the GSIB’s will get caught in this capital loss/liquidity trap but it is still worthwhile to know the exposure. This information may be available in the footnotes to the quarterly financial statement.
- Do the regulators believe in the banks’ resolution plans? When the next crisis occurs, like the one we’ve just experienced, will the regulators bail out the big (and not so big) banks like they did in 2008 rather than test resolution?
The recent banking crisis presents an opportunity to evolve our thinking and approach to counterparty risk management. What is the way forward for managing counterparty risk in an effective and meaningful way?
It is probably reasonable to assume that if a banking crisis occurs, either isolated or systemic, the previous quarter’s regulatory metrics will not be reliable indicators of where the bank is at that moment. Therefore I would suggest, in addition to reducing your financing exposure (excess cash and fully paid securities should already be at your custodian) calling the bank’s treasurer and head of funding to see what color than can provide. This was done frequently by the larger funds during the 2008 financial crisis.
There will probably also be additional regulation in the aftermath of this crisis that will help monitor some of these new risks. Hopefully it will be constructive and realistic and not just a “check the box” exercise for bank compliance. That said, regulators can only fight the last war. No one can predict the next crisis but we can learn from this last one and prepare as best we can.
[1] This reminds me of the quote by the famous boxer Mike Tyson: “Everyone has a plan until they get punched in the mouth.”
[2] The major banks have entire departments dedicated to developing, maintaining and submitting these resolution plans. Reading the quote from Mr. Jordan might make them reconsider their career options.
[3] Similar to the meme attack on stocks by retail investors in early 2021, this coordinated social media activity smacks of market manipulation.
[4] They tried to drag Deutsche Bank into it with a single late Friday CDS trade but the sand bags held.