Introduction and summary
Some advocates and policymakers have argued that the experience of the banking system and bank regulatory and supervisory system during the COVID-19 financial meltdown proves these systems are resilient and strong. This view, if it drives future policy deliberations, could lead us in the wrong direction.
Consider the following quote from a policymaker: “The post-crisis regulatory framework is strong, as evidenced by how well it fared against a severe real-life stress test—the COVID event.” In the same speech the policymaker acknowledged an alternative view, noting “some have argued that the time of COVID was not a true test of the system, because of the unprecedented level of fiscal and other support provided to the real economy during that time, from which the financial sector indirectly benefited.” But this more nuanced view of bank performance during the pandemic is not the dominant narrative.1
Indeed, to the best of our knowledge, the Minneapolis Fed is the main proponent of the contrasting view regarding bank strength during the COVID-19 financial event. We provided evidence in spring 2021 that banks avoided large losses from the COVID-19 financial shock because of the fiscal support the government provided. We will summarize that work shortly.
On that note, we will go further and review the specific factors that policymakers, bank lobbyists, and others cite to support the claim that banks were a source of strength for the economy, leading it out of the depths of the COVID-19 financial shock. That claim is not simply that banks avoided a 2007-2008 Great Financial Crisis–type of bank-specific bailout. The narrative also claims that banks affirmatively led the recovery based on their financial strength.
Before discussing the evidence, we should be clear why a balanced assessment of this period matters. After all, the COVID-19 shock is in the rearview mirror. It matters because the story we tell about banks’ financial strength in 2020 has implications for important policy choices leaders will make going forward.
Policymakers will need to make decisions in the short run about the capital regime banks will face. Should banks face higher capital requirements when times are good to push against exuberance and absorb losses when times are bad? Should policymakers lower certain capital requirements (the so-called supplementary leverage ratio) because banks argue the requirement is too high and constrains their activities? Should steps be taken to reduce the year-by-year variation in the stress test results? Should the Federal Reserve reinstate aspects of the stress test for the biggest banks to make the tests tougher?
Reviewing the narrative is also important for setting longer-run policy. A growing set of research—which we highlight later—finds that bank equity requirements should be substantially higher than they are today if we are to avoid future bailouts. Understanding what occurred during the COVID-19 financial crisis is important to informing long-run bank policy.
The rest of this article is organized as follows. We briefly summarize the main points from our original research on government support for banks during COVID-19 in the next section.2 Then, we review the main claims used to support a narrative in which banks were in very strong financial condition, supporting and leading the economy out of the COVID-19 financial shock. Finally, we explain why we find these claims to be overstated, misleading, or missing critical context. We close with a summary.
Banks benefited significantly from government support during the COVID-19 financial crisis
Unlike the Great Financial Crisis of 2007-2008, during the 2020 pandemic, the banking sector did not receive an industry-targeted bailout. Instead, the government and the Federal Reserve responded with a series of extraordinary actions that bolstered the aggregate economy and helped consumers, homeowners, and businesses.
Did banks benefit from this support? Yes, and in a prior analysis, we quantified the benefit at roughly $325 billion.3
In summary, the fiscal responses indirectly assisted banks, because the cash that Congress pushed out to the American people allowed firms and families to keep repaying their loans so that banks did not take losses. We quantified the value of this indirect support by estimating what loan losses would have been without the interventions and then measuring the difference between the estimated losses and the actual losses. Specifically, we constructed a simple model that generated projected amounts of future loan losses for a specified change in the unemployment rate under historical conditions. Comparing the results from the model with the actual losses recorded by banks suggests that the government actions could have prevented $230 billion of loan losses over three quarters if the unemployment rate had risen to 15 percent.4 We also described three alternative measures of the implied support and noted that they were relatively similar to what our simple model generated, bolstering confidence in the model.
In addition, the actions taken by the Fed to support financial markets allowed many banks to earn record trading profits during the last three quarters of 2020 instead of continuing to accumulate large losses, as many of them did during the first quarter of the year. Market disruptions like those that occurred in March 2020 could easily have developed into a more sustained challenge, leading to significant losses for banks with large capital market operations or even the failure of a key counterparty. Quantifying the losses from such outcomes is difficult, but the official Fed supervisory stress test offers some insight. It includes losses from a “global market shock”5 that is similar to what might have transpired in 2020 absent the Fed’s actions. Losses from such an event in the official stress test were projected to be $95 billion in the 2020 exercise.
These estimates together generate the $325 billion in benefits to banks.
One might dismiss this estimate by noting that the point of the government COVID-19 response was to improve the financial condition of households and firms, including banks. So who cares how big or small the amount is? Indeed, why does it matter at all as long as banks do not receive direct financial support from the government? However, focusing on the form of government support could lead us to miss what is occurring right before us. In 2008, the government could have avoided injecting funds into select large banks if it had directed the funds to households and firms that owed banks money. The banks would have been paid, avoiding failure-inducing losses. But such “direct to borrower” government support, had it occurred in 2008, would have been just as much of a bailout as the direct-to-bank version that actually occurred.
The case for banks as a source of strength during the COVID-19 financial crisis
Banks benefited from massive government support provided to the economy in response to COVID-19. Despite this fact, a narrative has emerged that claims banks are the pillars on which the economic rescue stood, rather than beneficiaries of government bailouts. One banking analyst sums this narrative up: “During a significant shock to the system, the banks kind of were able to be leaned upon to support rather than the area that is being bailed out. It is just fundamentally a different place for them.”6 Or consider the view the former acting Comptroller of the Currency stated during congressional testimony in late 2020, “Just as banks have been a source of strength for citizens, businesses and communities during the pandemic, they will lead the national recovery as well.”7
What support do advocates for this narrative bring to bear?
Some policymakers and banking observers focus on two points. First, they note that the banking industry had accumulated—in their view—large amounts of capital and liquidity before the onset of the pandemic.8 Second, they point to results from multiple rounds of Federal Reserve stress tests run during the COVID-19 financial downturn. Specifically, the outcomes of the tests are used to reject the claim that government support for banks increased bank strength.9
These factors are roughly the same that a banking advocacy group points to.10
In terms of the Federal Reserve stress tests, this banking advocacy group argued that “both  stress tests neutralized the effects of governmental support, including unemployment insurance provided in the CARES Act and the support to small businesses through the Paycheck Protection Program. … The stress test results demonstrated the covered banks had sufficient capital to absorb those losses.”
They also point out the changes in lending and capital during 2020: “The events of 2020 provided clear evidence of the resilience of U.S. banks. Banks funded the economy with extraordinary levels of lending and rode out the crisis without any major losses and with capital ratios rising, not falling.”
We now turn to review these points. We argue that they are removed from their key context, which makes them misleading when taken at face value.
Evaluating the evidence for the source of strength narrative
The “source of strength” argument is threefold.
- Capital levels were high and rose at banks during the COVID-19 financial meltdown;
- Bank lending increased during this same meltdown; and
- Banks passed several rounds of very rigorous stress tests.
The problem with this narrative is that each point contains material omissions:
- Bank capital levels rose during the COVID-19 shock because the Federal Reserve took actions to make that happen; moreover, bank capital levels are actually not high relative to what would be needed to prevent future bailouts.
- Bank lending increased primarily because government programs took the risk of loan losses from banks, and for other reasons unrelated to bank management decisions.
- Performance of banks under the COVID-19 stress tests reflects certain assumptions that made the tests easier, and a fair number of banks would have failed had the assumptions been more realistic.
A full telling of the narrative is not one where banks are obviously a source of strength. We provide more details in the rest of this section.
1. Bank capital levels rose because of government action—and they are still too low
In 2020, both capital levels and capital ratios rose at large banks. Aggregate capital (measured using “common equity tier 1” capital, or CET1) at the 21 largest domestic banks included in the Minneapolis stress test model increased by nearly 8 percent during the year. The corresponding CET1 capital ratio climbed over 80 basis points to reach 12.4 percent by year-end. This increase in capital was not the result of banks raising additional equity from investors. Instead, capital levels rose because the Federal Reserve prevented these large banks from spending billions of dollars buying back their own stock. The Fed prohibited share repurchases and restricted dividends beginning in the third quarter of 2020.11 Absent these actions, we believe CET1 would have fallen in the second half of the year and capital ratios also would have declined.
Why do we take this view? The 21 firms bought back roughly $77 billion of their own stock in the second half of 2019 but only $600 million during the second half of 2020, when the Fed put its restriction in place.
We calculate what capital ratios would have looked like if firms had bought back the same amount of stock in the second half of 2020 as they did in 2019 (see Table 1). Under this assumption, banks would have effectively reduced the fourth-quarter 2020 CET1 level by roughly $76 billion (the difference between the $77 billion paid out in third and fourth quarters of 2019 and the $600 million paid out in third and fourth quarters of 2020) to $1,057 billion. Under this scenario, the corresponding capital ratio would stand at 11.5 percent—exactly where it was at the end of 2019.
|21 Firm Total||2019Q3||2019Q4||2020Q1||2020Q2||2020Q3||2020Q4|
|CET1 ($ billions)||1,045||1,050||1,042||1,065||1,094||1,133|
|RWA ($ billions)||8,921||9,107||9,448||9,087||8,963||9,167|
|Capital Ratio (CET1/RWA)||11.7%||11.5%||11.0%||11.7%||12.2%||12.4%|
|Share buybacks ($ billions)||37.2||39.9||31.1||0.4||0.3||0.3|
|CET1 with 2019-level buybacks in Q3 & Q4||1,057|
|Corresponding capital ratio||11.5%|
It is possible that banks would not have repurchased shares in the second half of 2020 absent Fed restrictions on this activity. And the largest banks did halt their repurchases for part of the first quarter and during the second quarter of that year. However, these banks were clear that the suspensions were only temporary. The restrictions the Federal Reserve placed on repurchases speaks to the uncertainty of what large banks would have done if left to make their own decisions.
As to the general claim that banks came into the COVID-19 crisis with high levels of capital: The question is, high relative to what? Capital levels are certainly higher than they were before and immediately after the Great Financial Crisis of 2007-2008. But they are too low relative to the capital needed to prevent a future bank bailout, according to a growing body of economic research.
We discuss the detailed evidence supporting this view in the The Minneapolis Plan to End Too Big To Fail.12 A number of other researchers (Barth and Miller, 2018; Begenau and Landvoigt, 2021; Egan et al., 2017; Firestone et al., 2019; Passmore and von Hafften, 2019; and Perri and Stefanidis, 2017) have also found that capital ratios and leverage ratios should be at or above the levels described in the Minneapolis Plan. Treasury Secretary Janet Yellen addressed the tension between higher capital levels and the corresponding cost of intermediation in a 2018 speech when she was Federal Reserve Chair concluding that “even now, U.S. capital requirements are closer to the lower, not the higher, end of those that could be justified by cost-benefit analysis.
2. 2020 loan growth reflects government support for bank loans
Bank lending did increase significantly as the pandemic swept across the nation in the spring of 2020. Total loans grew by nearly 8 percent from the end of February to mid-May—the fastest quarterly rate of growth seen since at least the early 1970s. But virtually all the increase was in lending to businesses (commercial and industrial loans [C&I], commercial real estate loans, and loans to non-bank financial firms). Lending to households and consumers—just over half of the industry loan portfolio—rose modestly at the start of the pandemic but then began falling in April, and by the end of the year was almost 2.5 percent lower than it had been in February.13
The increase in business lending did not reflect new lending in the way most observers would use that term. Much of the initial growth during March was likely due to firms simply drawing down pre-existing commitments that banks were contractually obligated to provide. For example, outstanding C&I loans made by banks surged nearly 16 percent during the first quarter of 2020. But the total amount of unused commitments for the category fell by almost 13 percent. As a result, aggregate C&I credit—which includes the loans that banks had already made plus the unused commitments—only rose by roughly 1.6 percent in the period, or just 40 basis points more than the average quarterly increase in aggregate C&I credit over the prior five years.14 Certainly, banks making good on their existing loan promises is better than banks reneging. But it is not clear this development is a source of strength. After all, we saw this same pattern during the first stages of the 2007-2008 Great Financial Crisis, which some also viewed as a good thing, but which was followed by bank failure.15
Banks did begin to rapidly increase their business lending during the second quarter of 2020, which pushed aggregate domestic lending up to almost $10.5 trillion. However, this increase was driven by a new type of loan set up by Congress, in which banks took on virtually no risk: government-guaranteed Paycheck Protection Program (PPP) loans. Under this program, small businesses and nonprofit organizations could obtain government-guaranteed loans to help them cover payroll costs and other operating expenses—loans that could eventually be forgiven under certain conditions. The loans were administered primarily by banks, which earned revenue for processing the applications. By the end of June, banks had made more than $485 billion in PPP loans and potentially earned $19 billion in fees from distributing the initial funds.16
Importantly, when the PPP loans are removed from the calculation, we find that total bank loans as of June 2020 would be just below $10 trillion—which was exactly the size of the industry loan portfolio at the end of 2019. The government, not banks, deserves credit for this increase in lending.
3. Bank performance in the stress tests does not support the source of strength claim
Bank performance in the stress tests is not consistent with a source of strength claim for two reasons. First, the tests contained several favorable assumptions that created the appearance of strong performance even when that may not have been the case. Second, and more importantly, a material number of banks tested did not do well on it, and likely would have failed if prior stress testing rules had been maintained.
We first consider some of the key favorable assumptions embedded in the tests. For example, in the special sensitivity analyses conducted in June of 2020, the starting points for the individual banks were measured from the fourth quarter of 2019 and not from the first quarter of 2020. This meant that the analyses were using relatively strong bank-level positions rather than the weaker ones that existed once the pandemic began. And the tests that were released at the end of 2020 were based on bank data from the end of June, when financial markets and conditions had begun to stabilize. Results from the tests would likely have been worse had first-quarter 2020 been used as the starting point for the exercises.
The Fed did attempt to capture some of the important real-world dynamics by increasing corporate loan balances and market risk-weighted assets (RWA) in the first quarter of its forecast period (so the projected values for the banks mirrored what the industry experienced in March). However, no adjustment was made for the rest of the loan portfolio categories even though they were also growing slightly over the period. And no additional adjustments were made to any of the later projection periods, either. Maintaining a constant level of market RWAs for the remaining projection period also seems implausible given what transpired in financial markets during first-quarter 2020. Holding the balance sheet constant over the rest of the projection period—and only allowing a portion of the loan portfolio to grow in one quarter—likely resulted in better results from the stress tests.
Finally, it is important to highlight the results from the harshest scenario used in the sensitivity analyses—the “W-shaped” recession in which a second COVID-19 event leads to another surge in unemployment. This test revealed that eight of the 33 banks were projected to have a minimum CET1 ratio of 4.8 percent or lower. It seems reasonable, in such an outcome, that these banks could also experience deposit outflows and other liquidity shocks as counterparties become concerned about their condition, leading to potential disruptions like the ones that happened in the 2007-2008 Great Financial Crisis. The previous stress tests had banks failing if their capital ratios fell too low (below specified ratios), and if that regime had continued, almost a quarter of the banks would have failed in the COVID-19 event tests. That is not a validation of the strength of large banks.
Some argue that large banks led the economy out of the COVID-19 financial shock. This view is based on three claims that are suspect.
First, in contrast to the view that bank capital is high and growing, we find that bank capital is insufficient to avoid bailouts in the future, and it was government action that led to the accretion of banks’ capital.
Second, the suggestion that large bank lending was the source of strength for the economic recovery after the pandemic shock in early 2020 fails to account for the massive government role in facilitating such lending growth via the PPP, among other factors.
Third, and finally, the repeated argument that bank performance in stress tests confirms their strength ignores the many ways in which these tests were weakened and the failure of many banks to pass even with these favorable conditions.
Yes, banks did not get firm-specific bailouts à la 2009, but they did benefit from generalized government support.
In sum, we do not see the case, based on banks’ performance during the early stages of COVID-19, for relaxing bank regulation or supervision. Instead, we should strive for a robust regime that further reduces the likelihood taxpayers will need to bail out the largest banks.
1 See https://www.federalreserve.gov/newsevents/speech/quarles20211202a.htm.
2 See Feldman and Schmidt (2021).
4 We noted in the original analysis that government forbearance programs, which allowed borrowers to avoid making payments on their loans, could have resulted in lower loan losses than those estimated by our model. However, some data suggest that the total amount of potential losses related to forbearance may be limited. Black Knight Inc. (a provider of software, data, and analytics for mortgage and home equity lending) reported that 6.9 million homeowners entered into forbearance plans during the 11 months leading up to Feb. 23, 2021. As of that date, 3.7 million of these owners had begun making payments on their loans or had paid the mortgage off in full through a refinance or home sale. The total number of homeowners with a mortgage was roughly 53 million at the time.
5 A detailed description of the global market shock component of the official exercises can be found in the Stress Test Scenarios document published by the Federal Reserve at the start of each testing cycle, typically in mid-February (https://www.federalreserve.gov/newsevents/pressreleases/files/bcreg20220210a1.pdf).
6 See Haggerty (2021).
7 See Brooks (2020, p. 2). For a similar tone, see remarks by the chair of the Federal Deposit Insurance Corporation, Jelena McWilliams, or “A Selected Compilation of U.S. Banks’ Accomplishments in 2020,” from the Bank Policy Institute.
8 For example, former Vice Chair Quarles noted that “banks entered the COVID event with high levels of capital and liquidity, and served as a source of strength to the economy in a time of need” (https://www.federalreserve.gov/newsevents/speech/quarles20211202a.htm).
9 In the same speech, the Vice Chair observed that “such an argument ignores, however, that the Federal Reserve did not take such support for granted during the throes of the crisis itself, and as a result we ran multiple stress tests throughout the COVID event, with three separate and distinct scenarios, along with a ‘sensitivity analysis,’ which itself included three additional hypothetical recessions. Each of these stress tests assumed no additional fiscal or other measures to support the economy, and demonstrated that, even without such support, the banking industry would have fared very well.” Similar comments were made in an earlier 2021 speech: “Over the past year, the Federal Reserve has run three stress tests with several different hypothetical recessions and all have confirmed that the banking system is strongly positioned to support the ongoing recovery” (https://www.federalreserve.gov/newsevents/pressreleases/bcreg20210624a.htm).
10 Specifically, the Bank Policy Institute (2021).
11 As part of the release of the 2020 stress test results and the additional sensitivity analyses on June 25, 2020, the Federal Reserve required large banks to suspend share repurchases, capped dividend payments to the amount paid in the second quarter of 2020, and further limited dividends to an amount based on recent earnings. These restrictions were gradually lifted later in 2021.
12 See Federal Reserve Bank of Minneapolis (2017).
13 Loan growth rates are based on data from the H.8 release (Assets and Liabilities of Commercial Banks in the United States).
14 Calculations are based on aggregated data from the quarterly reports of condition and income for commercial banks (i.e., the Call Reports).
15 Ivashina and Scharfstein (2010) show that the $100 billion increase in C&I loans that occurred from September to mid-October 2008 was not driven by an increase in new loans and was due, instead, to firms utilizing existing credit lines (https://www.sciencedirect.com/science/article/pii/S0304405X09002396).
16 See https://www.umass.edu/news/article/new-research-brief-finds-banks-have-cashed.
Bank Policy Institute. (2021, October 18). “I Swear the Sun Rose in the West”: Denying the Obvious About Bank Performance in 2020. Retrieved from https://bpi.com/i-swear-the-sun-rose-in-the-west-denying-the-obvious-about-bank-performance-in-2020/
Barth, J.; and Miller, S. (2018). Benefits and Costs of a Higher Bank Leverage Ratio. Journal of Financial Stability, 38, 37-52. Available at https://doi.org/10.1016/j.jfs.2018.07.001
Begenau, J.; and Landvoigt, T. (2021). Financial Regulation in a Quantitative Model of the Modern Banking System. Available at SSRN 3095550.
Brooks, B. (2020, November 10). Statement of Brian P. Brooks, acting Comptroller of the Currency, before the Committee on Banking, Housing, and Urban Affairs. Retrieved from https://www.occ.gov/news-issuances/congressional-testimony/2020/ct-occ-2020-148-written.pdf
Egan, M.; Hortaçsu, A.; and Matvos, G. (2017). Deposit Competition and Financial Fragility: Evidence from the US Banking Sector. American Economic Review, 107(1), 169-216.
Federal Reserve Bank of Minneapolis. (2017.) The Minneapolis Plan to End Too Big to Fail. Minneapolis, MN: Author. Retrieved from https://www.minneapolisfed.org/~/media/files/publications/studies/endingtbtf/the-minneapolis-plan/the-minneapolis-plan-to-end-too-big-to-fail-final.pdf
Feldman, R., and Schmidt J. (2021, May 26). Government fiscal support protected banks from huge losses during the COVID-19 crisis. Retrieved from https://www.minneapolisfed.org/article/2021/government-fiscal-support-protected-banks-from-huge-losses-during-the-covid-19-crisis
Firestone, S.; Lorenc, A.; and Ranish, B. (2019). An Empirical Economic Assessment of the Costs and Benefits of Bank Capital in the US. Federal Reserve Bank of St. Louis Review, 101(3), 203-230. Available at https://doi.org/10.20955/r.101.203-30
Haggerty, N. (2021, May 11). It’s not 2008, but bank CEO hearings could still get testy. American Banker. Retrieved from https://www.americanbanker.com/news/its-not-2008-but-bank-ceo-hearings-could-still-get-testy
Ivashina, V., and Scharfstein, D. (2010). Bank lending during the financial crisis of 2008. Journal of Financial Economics, 97, 319-338. Available at https://doi.org/10.1016/j.jfineco.2009.12.001
Passmore, W.; and von Hafften, A. (2019). Are Basel’s Capital Surcharges for Global Systemically Important Banks Too Small? International Journal of Central Banking.
Perri, F.; and Stefanidis, G. (2017). Capital Requirements and Bailouts. Federal Reserve Bank of Minneapolis Staff Report 554.
Yellen, J. (2018, September 21). The Tenth Anniversary of the Financial Crisis. Keynote address at the Griswold Center for Economic Policy Studies fall symposium, Princeton University, Princeton, NJ.