The Sad State of Interest Rate Risk Disclosures

Bloomberg Tax reports that Silicon Valley Bank omitted economic value of equity (EVE) metrics from its December 2022 10-K after including EVE for the previous ten years.  Perhaps more notable, however, is that few large banks report EVE externally.  Even fewer provide sufficient detail on key assumptions to allow for a meaningful assessment of interest rate risk measures.

Since 1997, the SEC has required public companies to make quantitative and qualitative disclosures of market risk in their public filings.  These may include value-at-risk disclosures for trading activity and interest rate risk for the banking book.  IRRBB disclosures typically focus on the sensitivity of net interest income (earnings-at-risk) and, occasionally, EVE.  The SEC has provided little in the way of specific guidance on which metrics to use.

SVB’s 2021 10-K showed a high and increasing level of EVE exposure.  SVB projected a 27.7% decline in EVE under a +200bps rate shock.  NII was projected to benefit from rising rates.  The 2022 10-K continued to mention EVE as a risk measure but no longer provided any estimates.  IRR disclosures were limited to earnings-at-risk, which showed a modest exposure to falling rates.

SVB’s decision to discontinue external reporting of EVE and management’s failure to even explain the change presented clear red flags.  The bank failed just 14 days later.  However, few large banks even disclose EVE externally and most IRR disclosures focus on relatively modest rate shocks, as shown in the chart below:

Perhaps surprisingly, the recently-failed Signature Bank’s 2022 10-K included both earnings at risk and EVE sensitivity measures and covered a wider range of rate scenarios than did the other large banks.  Unfortunately, Signature’s better practices in IRR disclosures pretty much ended there.  The 10-K includes a standard disclaimer that “the computation …is based on numerous assumptions, including relative level of interest rates, asset prepayments, deposit decay and changes in repricing levels of deposits to general market rates, and should not be relied upon as indicative of actual results.”  Okay, so what did Signature assume about, say, deposit decay and repricing? Beats me. The disclosure provides zero detail.

Turning to still solvent banks, Truist Financial Corp. only includes NII sensitivity and rate changes of only 100bps.  Truist Financial also assumes a gradual rather than instantaneous change in rates, meaning much less rate stress the market experienced over the past couple of years.  However, the bank at least discloses a key assumption.  The deposit beta represents the projected change in deposit rates for a 100bps increase in overall market rates.  The beta assumption is a major driver of IRR metrics and can vary considerably from bank to bank.  Truist assumes a beta of roughly 50% on its non-maturity deposits.  The disclosure also projects the impact on NII of decreases in non-interest deposits of $20 billion and $40 billion. 

Like Truist, most large banks do not disclose EVE.  (Some banks disclose that they use EVE but keep mum on the details.)  Disclosures typically assume moderate rather than severe rate changes.  Unlike Truist, the other large banks don’t even provide modest details on deposit assumptions.  They may note that the projections depend on certain key assumptions but provide no supporting details.  Market risk disclosures that exclude a key metric and are based on black box models, where key, bespoke assumptions are kept secret are unlikely to be of much value.

The rationale for market risk disclosures is twofold.  First, accurate and complete disclosures give investors a better idea of what they may be getting into.  These disclosures can also be a source of “market discipline.”  The idea here is that regulators and supervisors get assistance courtesy of Adam Smith’s invisible hand.  The Bank of International Settlements sees “market discipline” as complementing capital standards and supervisory review by imposing “strong incentives on banks to conduct their business in a safe, sound and efficient manner.”

Why aren’t IRR disclosures more useful?  Without much guidance from the SEC or much push from their prudential regulators, banks are largely on their own when it comes to IRR disclosures.  That usually means looking at what similar banks disclose and leaving it at that.  Megabanks look to other megabanks and large regionals look to other large regionals.  Provide just enough but not too much.  That’s even true with more expansive IRR disclosures.  The two major custodial banks, State Street and BNY Mellon, report IRR in a manner more similar to each other than to other large banks.

It’s also not clear whether there was much clamor for improved IRR disclosures.  Investors tend to have short time horizons and don’t focus a lot on tail events.  I’m aware of at least a few cases over the years where public disclosures indicated excessive interest rate risk at certain banks but generated little market reaction. 

Market discipline also sounds good in theory, but the record is more mixed in practice.  As we know from Tulipomania in the 1600s to subprime mortgages in the early 2000s, markets don’t always ensure prudent behavior.  However, the recent, high profile bank failures might make investors more attuned to interest rate risk and more interested in improving risk disclosures. 

The European Banking Authority’s IRR disclosure requirements provide a good model for US regulators.  Article 448 requires reporting of both EVE and NII sensitivity; specifies rate scenarios; and requires disclosure of key modeling assumptions.  In addition, banks must describe how they measure, mitigate, and control IRR; describe key modeling assumptions used by their internal models; and the average and longest maturity they assign to non-maturity deposits.  While some may view the EBA’s approach as overly prescriptive, the regulation itself runs only about 500 words.  Not too much to ask for more consistent and transparent IRR reporting.