The FDIC’s Chief Risk Officer has released a postmortem report of the First Republic bank failure. A close reading of the report indicates that supervisors largely overlooked weaknesses in interest rate risk and liquidity risk management at First Republic. The FDIC’s supervision of First Republic appears to be a classic case of the halo effect, where the bank’s reputation and strengths in some areas led supervisors to underestimate critical risks and weaknesses.
The FDIC’s review of the First Republic failure follows the Federal Reserve’s internal review of supervision at Silicon Valley Bank. The tone and the contents of the two reports are decidedly different, however. The Fed characterized its internal review of the SVB failure as “unflinching.” That was largely the case with the review describing regulatory standards for SVB as “too low” and that supervision “did not work with sufficient force and urgency.” The tone of the FDIC’s review of First Republic was milder, with softening phrases like “could have been more forward looking” and “opportunities to take a more holistic approach.” First Republic’s status as the second biggest bank failure in U.S. history should merit a more self-critical tone.
Although First Republic was a bit larger than SVB, the SVB failure received more attention and supervisory missteps have been cast in a harsher light. This may be because the SVB collapse started the financial contagion and First Republic was its biggest victim (to date). However, weaknesses at First Republic made the bank especially vulnerable to the contagion, dooming both public and private efforts to prop it up.
Halo Effects
The First Republic review did not provide primary source materials such as supervisory letters. We don’t have a lot of insight into exactly what FDIC communicated to bank management. We can, however, compare the evolution of the bank’s financial condition and risk profile with supervisory actions and ratings. The ratings don’t hold up well to any kind of scrutiny. Until March 31, 2023, its composite rating was “2” or satisfactory. The FDIC rated both Management and Liquidity (!) as “1,” the highest rating. Sensitivity to Market Risk was rated “2.”
This is where the halo effect comes in. The post-mortem notes that “First Republic had historically been viewed by FDIC officials as a generally sound and well-managed institution” with pristine asset quality.” Moreover, FDIC officials found bank management to be “responsive to supervisory feedback and recommendations, quick to remediate findings, and easy to deal with.” That’s all great but can also create blind spots. I’ve seen many times over the years where a bank’s strong reputation caused supervisors to overlook red flags. Reputation should count for something, but it shouldn’t be the end all and be all.
Let’s look at First Republic’s supervisory history. The FDIC’s oversight and assessments of liquidity risk and interest rate risk were at times downright perplexing. Bear in mind this First Republic was the second largest bank under the FDIC’s direct supervision. (FDIC is a backup federal supervisor for all federally insured banks.)
Liquidity
The FDIC rated Liquidity at First Republic at “2” as of year-ends 2018, 2019, and 2020. FDIC then upgraded Liquidity for 2021 to a “1,” the highest rating. How did this happen? Liquidity risk consists of two essential elements: the stability of liabilities and the liquidity of assets. The postmortem doesn’t provide much detail on the rationale for the upgrade, but it appears the FDIC underrated the risks of the former and overrated the mitigating effects of the latter.
First Republic increased uninsured deposits from $80.3 billion in 2021 to 116.7 billion in 2022. Uninsured deposits stood at only $59 billion as of year-end 2020. While the failures of SVB, Signature, and First Republic have highlighted the dangers of uninsured deposits, this is not exactly a new concept. Historically, large deposits are the first to head out the door at the first sign of trouble.
First Republic supposedly mitigated this risk by increasing liquid assets. But let’s look at the numbers. According to its Uniform Bank Performance Report, short-term assets (primarily bank deposits) did increase during 2021 from 3.41% of assets to 7.09%. However, the December 2021 level was still well below the peer group median of 17.86%, placing the bank in the 25th percentile for this category. The investment portfolio consisted primarily of municipal bonds, long-term and not especially liquid securities. Moreover, First Republic designated munis primarily as held-to-maturity, giving the bank little flexibility to rebalance the portfolio. By the end of 2022, First Republic’s muni portfolio had a fair value worth less than 83% of carrying value.
Interest Rate Risk
First Republic had a much larger concentration of residential mortgages than did most other large banks. Mortgage loans and pass-throughs represented nearly half of total assets, compared to a peer group median of less than 20%. As the bank grew, so did its mortgage portfolio. Total assets increased by nearly $65 billion from year-end 2019 to year-end 2021. Residential mortgages accounted for nearly half of the growth. The bank had also been shifting the composition of the mortgage portfolio from 5/1 ARMs to 10/1 ARMs and 30-year fixed rate mortgages, which accounted for 75% of new originations by December 2021.
The postmortem indicates (on page 6), “First Republic’s primary interest rate risk mitigation strategy relied on continual growth to produce a consistent volume of loans priced at current interest rates. Each examination report restated First Republic’s growth and repricing strategy as key to mitigating interest rate risk.” In other words, the bank sought to mitigate IRR by doubling down on the risk. Such a strategy is dubious under the best of circumstances but makes especially little sense for a mortgage portfolio. The borrower’s refinancing option creates an adverse selection problem where the highest yielding mortgages prepay with the bank stuck with the low yielding leftovers. Despite the sharp rise in interest rates since 2020, mortgage yields increased by only 7 basis points from 2021 to 2022 and were 11 basis points below 2020 levels. First Republic did not hedge its interest rate risk, considering its deposits as a natural hedge.
First Republic’s IRR position deteriorated rapidly during 2022. The projected decline in economic value of equity (EVE) from a +200 bps shock went from 12% in March 2022 to 117% by December 2022. Part of the deterioration reflected a decline in baseline EVE of nearly 60%. It’s also likely First Republic revised key assumptions to reflect higher rate sensitivity for the bank’s deposits.
FDIC supervisors not only accepted First Republic taking on outsized interest rate risk, but also reacted slowly when things started to fall apart. For example, the report notes (page 26) that, “The dedicated examination team became aware of the second quarter breach in August 2022 and learned from Board meeting minutes received in November 2022 that the bank had decided not to take action, but to monitor the situation.” [Emphasis added.] This strikes me as an extremely passive approach to supervision. Haven’t they heard of phones or email?
Supervisory Actions
The FDIC primarily addresses bank weaknesses through Supervisory Recommendations (SRs) and Matters Requiring Board Attention (MRBAs). The FDIC issued zero MRBAs to First Republic between 2018 and 2022. The lack of any MRBAs for a bank of First Republic’s size is unusual. From 2010 to 2014, 3,400 FDIC Reports of Examination cited MRBAs, including 48% of 1 and 2 rated banks. Signature Bank had 13 outstanding MRBAs, though 8 related to anti-money laundering issues.
20/20 Hindsight
I’ve found that many of the critics of the Fed’s supervision of SVB engaged in Monday morning quarterbacking. What may seem obvious now was not at the time with critiques relying on selective amnesia to suggest otherwise. The First Republic postmortem goes too far in the other direction. The report uses the term “20/20 hindsight” twice, “unprecedented” three times, and “unexpected” four times.
In many ways, however, SVB was much more of a black swan than was First Republic. The previously largest bank run in history (WaMu) was $16.7 billion over seven days. SVB lost $42 billion in a single day and an additional $100 billion was expected out the door had regulators not intervened. It’s hard to conceive of any bank surviving a run of that magnitude.
First Republic is a different story. While fallout from the SVB failure no doubt hurt the bank, First Republic still had access to cash through the Bank Term Funding Program and through a $30 billion lifeline provided by a consortium of other large banks. Unfortunately, First Republic’s asset yields were so low that it was no longer viable at anything approaching market funding rates.
First Republic’s balance sheet structure was especially vulnerable to the confluence of sharply rising interest rates and depositor flight. None of these events were expected, but the Titanic didn’t expect an iceberg either. Most bank failures arise from events that are neither welcome nor expected. Effective risk management means positioning the bank so that it can withstand and rebound from these adverse, unexpected events. Neither bank managers nor FDIC supervisors fully appreciated First Republic’s vulnerability.