The massive unrealized losses in the HTM portfolio at SVB that arguably triggered the liquidity run had been known for a long time.
As of September 2022, their 10-Q clearly showed a $15.9 Bn loss on $93.2 Bn portfolio (-17%, Fair Value $77.3 Bn). Dec. 2022 figures showed a similar loss ($15.2 Bn on a $91.3 Bn HTM portfolio).
Unrealized losses of $17.1 Bn as of Dec. 2022 on HTM would wipe out 89% of Tangible Common Equity (TCE).
What caused the problem?
Deposit Concentration: Due to its business model and origin, SVB's deposit base was highly concentrated in a few "related" sectors (VC and Tech related companies), with an estimated ($151 Bn out of $175Bn, 86%) in uninsured deposits.
Very large HTM portfolio: As of Dec 2022 their total assets were around 214 Bn, with their HTM portfolio making up ~43% of total assets. This to accommodate a very large deposit base ($175 Bn out of total assets of $213 Bn) and very fast growth over the last couple of years.
Their balance sheet reflects a massive bet on long term rates: with $88 Bn of the $93 Bn in the HTM portfolio with a maturity beyond 10 years (not taking into consideration prepayments in a mainly Agency MBS/CMBS portfolio), see bottom table on initial picture.
No stop-losses: Unrealized losses $17.1 Bn as of Dec. 2022 on HTM would wipe out 89% of TCE. As of Dec.2021 their portfolio had only a ~1% unrealized loss (Fair Value $97 Bn on $98 Bn HTM portfolio). In addition, whatever hedging (derivatives designed as hedging instruments) they had in place in 2021 had been discontinued by September 2022.
Source: SVG Financial Group Form 10-Q, Sept.2022
Balance sheet structural risk management is all about assumptions
Investment in long term fixed rate instruments is commonly used to stabilize net interest income margin, mainly investing stable, non-rate sensitive deposits and prevent margin compression in a low rate environment.
Low cost non-maturity deposits (NMD) provide a significant source of income and value for banks, however they are not marked-to-market. This creates a significant mismatch vs. the asset side (here the investment portfolio). The use of the HTM category to "smooth out" the future income of the NMD seems to makes sense, but what happens if those deposits disappear (bank run)?
Interest rate risk meets liquidity risk:
Under this scenario in order to meet your liquidity needs (deposit outflows) banks will need to liquidate the portfolio, and unrealized losses will materialize.
The underlying assumption, future revenue on NMD balances, will no longer hold, there are no balances to earn the spread on (away from a on-going concern). Suddenly you do not have the capacity to hold these assets to maturity (HTM portfolio).
Over the weekend the Federal Reserve Bank announced a term lending facility (Bank Term Funding Program, BTFP https://www.federalreserve.gov/newsevents/pressreleases/monetary20230312a.htm), offering loans up to one year in term for eligible depository institutions (banks, savings associations, credit unions, etc,) pledging U.S. Treasury, agency debt and mortgage-backed securities, that will be valued at par. This should provide some breathing room for some of these institutions.
Need to rethink ALM framework
From an ALM perspective this raises the question how you should adjust your IRR/Liquidity bets as your capacity to hold on to the HTM portfolio diminishes as unrealized losses start to pile up (think about an "all-in" OCI view).
Smaller institutions should question their reliance on metrics designed and used for G/D-SIBs as their risk profile is clearly different)
For sure these events will (or should) trigger a fundamental review of ALM practices. Keep tunned.
If you would like to learn about our cost-efficient IRRBB and Liquidity Risk managements solutions, please visit us at www.ibsmsolutions.com or message me directly.
Dear Karl Rubach, CFA - thanks for delivering some tangible background and facts. To add to your conclusion, there is a metric, namely the economic value of equity (EVE), showing the risks associated with asset-liability mismatch early on. Already back in March 2022, the metric pointed to the extreme risks associated with an increase in rates (USD 5.4 bn for a 200bp upward shock, see screenshot). In my opinion, we should work on refining this metric. In Europe, an EVE outlier test is the new standard, and raises supervisory attention, if the change in EVE over Tier 1 is exceeding 15%. That was the case for SVB.
I think, due to the unrealized losses for the HTM positions not affecting Tier 1, this "Delta EVE/Tier 1"- ratio was not showing the severity of the situation sufficiently. The ratio remained high during the year but did not deteriorate further. Do you share my opinion or am I missing a fact in my quick analysis, Karl Rubach, CFA?
Karl— nice summary—was an egregious failure of basic Risk management 101… borrow short and lend long and 🙏🏻? they needed your services!! Unfortunately they will not be held accountable for the damage they have created not only to their shareholders and clients but to the markets in general…and dare I ask where we’re the rating agencies and regulators?
Well articulated, Karl. Regarding your point about the intersection of interest rate and liquidity risks, many ALM practitioners, myself included, tend to separate their assumptions and management strategies for each parameter, as if they were independent of each other. However, the situation faced by SVB serves as a real-world reminder of how a liquidity run can abruptly truncate interest rate duration. Failure to recognize the interplay between these two risks can result in significant gaps in the overall balance sheet risk management.
Retired
1yWell said Karl; nice to see you positioned to help the industry at this critical juncture.
Dozent at ZHAW School of Management and Law/ Universität Konstanz
1yDear Karl Rubach, CFA - thanks for delivering some tangible background and facts. To add to your conclusion, there is a metric, namely the economic value of equity (EVE), showing the risks associated with asset-liability mismatch early on. Already back in March 2022, the metric pointed to the extreme risks associated with an increase in rates (USD 5.4 bn for a 200bp upward shock, see screenshot). In my opinion, we should work on refining this metric. In Europe, an EVE outlier test is the new standard, and raises supervisory attention, if the change in EVE over Tier 1 is exceeding 15%. That was the case for SVB. I think, due to the unrealized losses for the HTM positions not affecting Tier 1, this "Delta EVE/Tier 1"- ratio was not showing the severity of the situation sufficiently. The ratio remained high during the year but did not deteriorate further. Do you share my opinion or am I missing a fact in my quick analysis, Karl Rubach, CFA?
Treasurer, S.V.P.--Expanding the US markets knowledge of successfully navigating China's economy and banking System
1yKarl— nice summary—was an egregious failure of basic Risk management 101… borrow short and lend long and 🙏🏻? they needed your services!! Unfortunately they will not be held accountable for the damage they have created not only to their shareholders and clients but to the markets in general…and dare I ask where we’re the rating agencies and regulators?
Head of Revenue Management, RBC Clear | MD | Revenue Management| Cash Management | Pricing| Treasury & Balance Sheet Mgmt| Digitalization| Payment| Funding & Liquidity
1yWell articulated, Karl. Regarding your point about the intersection of interest rate and liquidity risks, many ALM practitioners, myself included, tend to separate their assumptions and management strategies for each parameter, as if they were independent of each other. However, the situation faced by SVB serves as a real-world reminder of how a liquidity run can abruptly truncate interest rate duration. Failure to recognize the interplay between these two risks can result in significant gaps in the overall balance sheet risk management.
VALORISA RD - Economist/Banker/Treasury -Board Member
1yReally good points… food for thought… and the urgent need being more proactive and professional in financial markets!