The recent failure of Silicon Valley Bank (SVB) unexpectedly put on spotlight the negative developments and unfolding risks of financial instruments with hidden losses, specifically bonds. Hidden losses or unrealized losses in bonds were not a root cause nor reason for banks’ trouble, but ultimately, they gave the final kick to bank`s collapse. Even though the regulatory framework for banks between USA and Europe have crucial differences, making SVB scenario highly unlikely to fully replicate in Europe, markets in EU can also face hidden losses being materialized when a crisis hits. Dynamic changes in terms of monetary policies, anticipated levels of interest rate and everchanging geopolitical movements will test the resilience of the financial system and eventually lead to a rise in credit and liquidity risks for all players in financial markets. Bonds with hidden losses could play a role in how they might manifest.

Silicon Valley Bank (SVB) was a specialized bank for the tech startup industry, having massive growth between 2019 and 2022. At the end of 2022, it was the 16^th largest bank in the USA. This growth resulted in SVB having a significant amount of deposits and assets, where excess liquidity funds were used to buy US treasury bonds and other long-term debts. These kinds of bonds are considered to have relatively low risk and returns and are an excellent source for risk diversification. However, bond prices tend to have an inverse relationship to interest rates: once the interest rate goes up, the bond price goes down.

And that’s what occurred in spring 2023, the inverse relationship burst out.

As the US Federal Reserves increased interest rates in response to high inflation, SVB’s bonds declined in value and became riskier investments. As this was happening, some of the SVB customers demanded higher returns on their deposits and started withdrawing their deposits, while some of them withdrew the money due their own financial troubles. To accommodate these large withdrawals, SVB decided to sell USD 21 billion from the bond portfolio at a loss of USD 1,8 billion. This was the beginning of the downfall: the realized loss came as breaking news on social media and depositors panicked causing a classic bank run. Panic escalated since most of the deposits, roughly 88%, were uninsured where many startups had millions that surpassed the maximum insured limit of USD 250 thousand. From there, it took 36 hours for the second biggest fail of a bank in USA history.

At the prominent speech held by Kerstin af Jochnick “Banks after the end of low for long” ⁽¹⁾, Member of the Supervisory Board of the ECB, held at the 28^th Annual Financials CEO Conference in London in September 2023, she compares the regulation for banks in Europe and United States and explains how those differences played a role in the latest turmoil and what was the impact of the hidden losses.

She pointed that the business model of SVB features high dependance on volatile funding, notably uninsured deposits, and material unrealized losses in the securities portfolio, due to USA regulation which allows local banks to have an option not to reflect unrealized losses in their respective financial statements. Indeed, SVB selected an option not to recognize the unrealized gains and losses on the debt securities available for sale in the accumulated other comprehensive income in capital, something not provided in EU regulation. In addition, EU banks are subject to liquidity ratios whereas the troubled banks in the United States did not have this prudential requirement in place.

As a lesson to be learned from this latest turmoil across the ocean, the speech highlighted that ECB increased scrutiny of bank’s unrealized losses within assets valued at amortized cost, as a source of hidden losses for banks in EU. In that sense, an ad-hoc analysis ⁽²⁾ of unrealized losses on EU banks’ bond holdings was performed in July 2023, as targeted in the risk assessment conducted by the EBA in collaboration with competent authorities, which aims at understanding the potential evolution of unrealized losses on EU banks’ debt securities held at amortized cost.

The analysis revealed that banks directly supervised by the ECB have around EUR 70 billion of unrealized losses, net of hedging, on debt securities held at amortized cost. Overall, this amount is considered not to be material where the system can absorb the loss without large disturbance, both in capital and liquidity terms. Indeed, as evidenced by data published in this ad-hoc report, the overall amount of such unrealized losses was relatively contained, at EUR 73 billion as of February 2023, compared with USD 620 billion for US banks at year-end 2022.

The issue of hidden losses locked in bonds at amortized cost across the globe was discussed in Global Financial Stability Report issued by IMF in April 2023 ⁽³⁾. In this report, IMF staff estimates that the impact on regulatory ratios of unrealized losses in held-to-maturity portfolios for the median bank in Europe, Japan and emerging markets would likely to be modest.

However, the latest bank collapse alerts all financial market participants that these losses could be problematic when combined with weaknesses in individual banks’ asset and liability management, even though on systemic levels appears to be immaterial.

The same concerns apply for the EU banking industry. The concerns remain about vulnerabilities that may be hidden in bonds at amortized cost and present uncertainty of what combination of adverse events can unfold to unfavorable scenario that would lead an individual bank to insolvency. The last events reminded how funding can be diminished at a speed rate. It also reminded the necessity to better understand behaviors of both individual and corporate customers in the new digital era and how procurable information just a scroll away can influence the deposit base to vanish once confidence is shaken. This concern was also recognized through the SSM supervisory priorities for 2024-2026 ⁽⁵⁾ provided by ECB, where the priority no.1 is to Strengthen resilience to immediate macro-financial and geopolitical shocks. One of the vulnerabilities of this priority is shortcomings in assets and liabilities management framework along with the blend of risks coming from macro-financial and geopolitical developments. Banks are expected to manage those vulnerabilities as top of their agenda in the overall risk management.

To compliment regulators’ expectations, under the radar should be placed the latest development shared by Mrs. Kerstin in her speech. For the prudential purposes, she stated that it is under consideration if bonds held at amortized cost can continue to qualify as high-quality liquid assets for the purpose of banks’ liquidity ratios. This concern lies in the fact that when the bank’s liquidity needs arise, and bonds at amortized cost are sold to close the gap, possible suffering of losses might potentially further destabilize the bank.

So how does it work?

Impact on liquidity measures

For prudential purposes, regulators use two liquidity ratios: Liquidity Coverage Ratio (LCR) when assessing banks liquidity position for 30-day horizon, and Net Stable Funding Ratio (NFSR) when assessing banks liquidity position in one year horizon.

LCR measures the portion of assets that can be quicky converted to cash (referred to as high quality liquid assets or HQLA) to meet bank’s expected cash outflows for 30 days, and it is expressed as a ratio that must equal or exceeds 100%. Usually, bonds issued by EU governments are regarded as high quality liquid assets. Subject to a certain regulatory condition, if a bank values them at amortized cost, then that amortized value is used for determining the level of LCR ratio. Similar logic applies when assessing the NSFR ratio.

Mrs. Kerstin shares the notion that for prudential reasons, when assessing the level of LCR and NSFR, those bonds should be valued at market value, although a bank keeps them at amortized cost in its books. Depending on the asset and liability structure of the bank, such a change in prudential requirements can have a material effect on the levels of LCR and NSFR.

So, what lies ahead? Banks can expect closer supervision of their funding structure, imposture of additional challenges in the assessment and management of interest rate and credit spread risks, as well as review of the calibration of their ALM models in order to better reflect changes in the customer behavior, especially the one that can be triggered from the new movements in the level of interest rates. Banks would be expected to stress on their own the possible range of haircuts on bonds at amortized cost recognized as HQLA, and how it could affect its regulatory liquidity ratios and its risk appetite in liquidity terms. It could be argued, for the prudential framework, that lower value of those bonds should be used for ILAAP purposes, liquidity stress testing and for determining availabilities of recovery options for the Recovery plan, as well as the availability of possible liquidity emergency funds and banks own counterbalancing capacity.

Particular focus should be put to banks that are dealing with shortcomings in their funding structure (such as undiversified funding structure, increase in funding costs, funding concentrations, maturity challenges with funding, etc.) or currently are under supervisory measures over funding issues.

Impact on capital measures

Furthermore, one cannot exclude the possibility of the bank’s economic capital for ICAAP purposes to be additionally challenged by regulators by encouraging banks to recognize possible losses on bonds at amortized cost on the spot. In case of such prudential expectations, the economic capital of the bank would further shrink due to losses as a difference between the amortized book value and the market value. In addition, it would affect the banks’ capital planning, both for regulatory and internal capital, while adjusting stress test scenarios to reflect the new expectations. Adjusting the value of the bonds at amortized cost could become a new benchmark for reverse stress testing. These assumed shifts would affect the perceived value of those bonds by having higher risk rather than lower and eventually become costly for banks to measure and manage. Additionally, one cannot shake of notion that at some point those bonds would receive a higher risk weight in RWA assessment.

Given all that, bonds at amortized cost could potentially become a source of great difference between the accounting principles and the prudential expectations and requirements.

Why did the bank choose amortized cost as the source of their valuation for the bond portfolio at first place? Bank usually choose this option for bonds when they are interested only into receiving the full nominal value at maturity, with collecting coupons along the way, without the intend to sell prior to maturity. This type of business model for bonds to be kept until maturity, is recognized in terms of classification as “hold to collect” under IFRS 9. Once the business model (intent) is chosen, the sale of those bonds prior to maturity is considered to be incidental to the business model, and the reclassification to different types is allowed only in rare cases where bank should prove fundamental change of its business activities. Significant changes in value of financial assets, or in our case, significant change in the price of bonds, are not a recognized reason for changing from amortized cost to other options, such as FVOCI (Fair Value through Other Comprehensive Income) or FVPL (Fair Value through Profit or Loss).

The same stands for CSSF that expects that ⁽⁴⁾, in practice, the conditions justifying a change in the business model will rarely be met. A mere change in the entity’s intentions about the holding period of certain financial instruments (i.e. sale prior to maturity) is unlikely to trigger the requirements of the IFRS 9 regarding reclassification. Further, the CSSF is of the opinion that changes in the market conditions (i.e. change in bond price) cannot, by themselves, justify a reclassification of financial assets, even if the related impacts on prudential ratios are significant. Accordingly, the CSSF considers that it is highly unlikely that the entity could conclude and demonstrate that a change in the business model has occurred and that the underlying assets have to be reclassified.

The above stated elements just remind us that banks cannot simply change the classification of their current bond portfolio from amortized cost to FVOCI or FVPL to better manage and mitigate possible risks. Therefore, the risk of an unfolding scenario that can be devastating for banks lies within the maturity of the bond portfolio, and other combining elements that may arise in everchanging macroeconomic environment.

What options do banks have?

Banks do buy bonds, especially government backed, to diversify and manage their risks, to manage their liquidity risk and to enhance their profitability. They usually represent the classic case of low risk investments, with the lowest default risk. Banks in the end do receive the full nominal value and the market movements do not affect those kinds of bonds. But this holds true until the bank itself is the source of the adverse events and adverse market perception, where the core nature of these instruments that are considered low risk, gives an additional, negative burden on banks overall liquidity position and makes bank overall bounce back capacity difficult to execute.

Additionally, if market regulators via their prudential instruments expect and encourage banks to match the value of the bonds at amortized cost as if they are valued at market value for risk management purposes within the prudential framework, are we on the verge for standard setters to rethink the usage of amortized cost to be solely intended for loan portfolio?

As a final thought, we are stressing the notion that the distance between low risk instrument unfolding to high risk instrument might be the matter of hours in our modern macro-financial environment. Latest banks’ troubles prove that low risk instruments seek attention and measurement on their own. It’s up to banks to display scenarios in all directions when it comes to those bonds, ensuring that all risks and losses are properly considered and accounted for. The incentive to reassess the measurement of the bonds at amortized cost also comes from regulators who are currently reconsidering the treatment of those losses under the prudential framework and supervisory expectations. If validated, several aspects of the overall risk management could be affected, as discussed in the article. The underlying reason is to set up an environment where hidden losses will be either accounted for before they become materials or identified in their early stages. In a medium term horizon, one might argue that the banks choice in business model for those instruments will be measured and mitigated by rethinking and recalibrating the risk framework and appetite for the hidden losses captured in those instruments.

⁽¹⁾ Banks after the end of “low for long” (europa.eu)

⁽²⁾ Ad-hoc analysis of unrealized losses on EU banks’ bond holdings.pdf (europa.eu)

⁽³⁾ Global Financial Stability Report, April 2023: Safeguarding Financial Stability amid High Inflation and Geopolitical Risks (imf.org)

⁽⁴⁾ Additional guidance for credit institutions related to specific reporting aspects (cssf.lu)

 ⁽⁵⁾https://www.bankingsupervision.europa.eu/banking/priorities/html/ssm.supervisory_priorities202312~a15d5d36ab.en.html