AFS, HTM, and Unintended Consequences

Federal Reserve Vice-Chair Michael Barr has proposed requiring banks over $100 billion to incorporate unrealized losses on available-for-sale securities in regulatory capital.  Taking these unrealized losses into account better reflects the banks’ actual loss absorbing capacity.  However, if the experience with the banks already subject to this requirement is any indication, it will have the unintended consequence of leading bank to designate more securities as held-to-maturity.  Regulators also need to take more action to discourage such transfers from AFS to HTM.

In a July 10, 2023 Speech to the Bipartisan Policy Center, Barr discussed capital initiatives underway at the Federal Reserve.  These include making a wider range of banks subject to tougher capital standards.  Excessive growth was a key feature of recent bank failures and for bank failures from time immemorial.  Higher capital requirements constrain such growth.  One rationale for excluding large but not enormous banks from higher requirements is that they did not pose a systemic risk to the financial system.  Reactions to recent bank failures make it difficult to still make that claim with a straight face.  Unrealized gains and losses on AFS securities are reflected in equity capital under GAAP as accumulated other comprehensive income (AOCI).  However, only a small number of very large banks must reflect these gains and losses in regulatory capital.  Barr proposed to extend that requirement to banks over $100Bn.

I agree with Barr’s claim that including unrealized AFS gains and losses would “better reflect banking organizations’ actual loss-absorbing capacity.”  Barr goes on to assert that “banks that were required to reflect unrealized losses on AFS securities in regulatory capital managed their interest rate risk more carefully.”  I’m not so sure about that.  These banks largely responded by gaming the rules to report more investments as held-to-maturity.  The table below looks at trends in HTM designations for those banks already required to include unrealized AFS losses in capital.

We aren’t dealing with trivial amounts here.  Increases in HTM (and corresponding decreases in AFS) run into the hundreds of billions, as shown below:

While HTM designations make capital less volatile, they also hinder a bank’s ability to respond to a liquidity crunch or to rising interest rates.  A bank that classifies securities as HTM supposedly has an intent and ability to hold those securities until maturity.  Selling these securities taints the entire HTM portfolio.  The remaining HTM securities must be classified as AFS with a moratorium on future HTM classifications for two years.  GAAP allows a very limited set of exceptions.  These exceptions do not include portfolio rebalancing to mitigate interest rate risk or securities sales to meet liquidity needs. 

This trend toward more HTM designations started prior to the recent run-up in rates.  In a 2018 research paper, Andreas Fuster and James Vickery looked at how banks subject to the AOCI requirement attempted to lessen capital volatility.  Banks mainly responded by classifying more securities as HTM.  There was little if any effort to mitigate the interest rate risk of investment portfolios.  I’ve previously likened responding to higher rates by designating securities as HTM to closing your eyes before getting hit.  The actions by the large banks are disappointing but not surprising.  I’ve encountered many, many instances over the years where banks chose accounting form over economic substance.  I’ll discuss this topic in more depth in an upcoming post.

But what about the regulators?  Current regulations and written guidance provide little to discourage HTM transfers.  The definition of “high-quality liquid assets” does not consider accounting classification.  In other words, HTM securities can qualify as HQLA even though there are severe restrictions on their sale.  Banks may still borrow against HTM assets, but a borrowing provides less liquidity than a sale.  For example, First Republic had access to extensive borrowing facilities, but its low yielding assets were insufficient to cover increased borrowing costs.  Liquidating assets as it lost deposits would have made more sense.

OCC noted aggregate unrealized losses on both AFS and HTM securities in its Fall 2022 Semiannual Risk Perspective.  However, the Comptroller’s Handbook for Capital and Dividends only touches on unrealized losses in terms of reporting requirements.  The Comptroller’s Handbooks for Interest Rate Risk and for Liquidity do not discuss the safety and soundness implications of HTM designations.  The FDIC’s Risk Management Manual of Examination Policies notes that adequate capital for safety and soundness may differ significantly from minimum leverage and risk-based standards.  However, the Manual only discusses unrealized losses on securities to reporting requirements.  

The Federal Reserve’s Examination Manual comes closest to touching on risks of HTM securities.  The Manual indicates that examiners should consider unrealized gains when assessing Capital Adequacy but does not specifically discuss the approach to unrealized losses.  The Liquidity section states that “accounting conventions can also affect the market liquidity of assets” and that HTM securities provide less liquidity than those designated as AFS or as trading.

It makes sense to require all banks over $100 billion to reflect unrealized AFS losses in regulatory capital.  However, regulators need to take steps to discourage redesignation of securities as HTM.  Liquidity rules are a good place to start.  HTM securities should not be considered HQLA, or their inclusion should be severely restricted.  Regulatory guidance should also make clear that examiners will consider significant unrealized losses, regardless of their classification, when evaluating capital adequacy.  That won’t require a full marking to market.  Supervisors would retain some flexibility to consider individual circumstances.  But the revised guidance could send a clear message that a change in accounting designation would not be a substitute for actual risk reduction.