June 12, 2020

### Assessing the Resiliency of Non-DFAST Banks to a Financial Shock

Kevin F. Kiernan and Cindy M. Vojtech1

1. Summary

Every year the Federal Reserve Board conducts stress tests on large bank holding companies (BHCs) to ensure that those institutions will remain healthy enough to lend to households and businesses even in a significant downturn. This note analyzes the resiliency of the banking industry by also stressing banks that are not subject to that annual Dodd-Frank Act stress test (DFAST). This is a more complete stress shock than Vojtech (2019) because it captures banks of all sizes and stresses the entire bank. The earlier analysis focused on BHCs larger than $3 billion and on a specific vulnerability: commercial real estate. Also note that this stress shock exercise was completed prior to disruptions caused by the COVID-19 pandemic. If larger losses than DFAST are used, more banks would fall below regulatory thresholds. Non-DFAST banks make up roughly 25 percent of banking assets. While it is unlikely for a systemic risk to originate from this minority, it is important to understand potential vulnerabilities in this part of the banking system because widespread failures among this group could materially affect regional economic performance and exacerbate a downturn. In addition to the size disparity, there are significant differences between DFAST and non-DFAST banks in terms of balance sheet composition and loan portfolio composition.2 Loans make up a larger share of total assets at non-DFAST banks than at their DFAST counterparts (figure 1). There is also variation within loan holdings (figure 2). Non-DFAST banks hold more of their portfolio in commercial real estate (CRE) loans and less in commercial and industrial (C&I) loans. Thus, banks in this sector may be susceptible to common shocks, making their performance highly correlated. Almost all credit card loans are held by DFAST banks while almost all farm loans are held by non-DFAST banks. On the other hand, the two bank groups hold similar shares of their loan book in mortgages. ##### Figure 1: Balance Sheet Composition (September 30, 2019) ##### Figure 2: Loan Portfolio Composition (September 30, 2019) Common equity tier 1 (CET1) capital is the highest quality of capital in the regulatory framework. As shown in figure 3, non-DFAST banks, on average, fund more of their balance sheet with such capital than DFAST banks. However, even those higher capital buffers did not protect this sector from bank failures during the Global Financial Crisis (GFC). This figure also illustrates the dramatic increase in capital following the GFC with DFAST banks more than doubling their capital ratios. This significant increase in capital at the largest banks is partially driven by new regulations such as stress testing. ##### Figure 3: Regulatory Capital Those larger capital stocks offer strong protection from a variety of stresses. After applying the loss rates generated by the DFAST 2017, 2018, and 2019 exercises, our analysis shows that the vast majority of banks would have enough capital to weather a large negative shock without dipping below regulatory minimums. A small number of banks in our sample breach the minimum 4.5 percent CET1 threshold, and these banks represent less than 2 percent of banking system assets. However, using a higher threshold, the banks falling below 7 percent CET1 (the minimum plus the required capital conservation buffer) represent 4 to 11 percent of industry assets depending on the severity of the DFAST loss rates. The easiest method of restoring capital for banks that fall below 7 percent is to reduce lending, which could have adverse effects on the communities they serve. The regulatory agencies recently finalized rules allowing community banks (banks with less than$10 billion in consolidated assets) to qualify for simplified regulatory requirements if their regulatory leverage ratio is above 9 percent.3 About one-third of the banks that fall below the 4.5 percent CET1 minimum in this hypothetical exercise are banks that currently qualify for simplification under this new regulation. However, these banks make up a very small portion of the industry, less than 4 percent of community banking assets, and community banks have 7 percent of industry assets.

Finally, geographic analysis of the stress losses shows that some states' deposits are more concentrated in banks falling below the 7 percent threshold. Those states could face more credit contraction in an economic downturn should banks need to shore up their capital positions. For example, averaging across DFAST results and using the most conservative loss projection method, 17 states have 20 percent or more of their deposit base exposed to banks that fall below a 7 percent CET1 ratio.

2. Data

##### Table 3: Resiliency for Community Banks: BHCs with Total Assets Below $10 billion and a Leverage Ratio Greater than 9.0 percent DFAST year: 2017 2018 2019 Estimation method: (1) Avg. (2) OLS (3) QREG (4) Avg. (5) OLS (6) QREG (7) Avg. (8) OLS (9) QREG Below 4.5 percent CET1 ratio Community BHCs [number] 23 20 40 39 41 101 18 20 60 Percent of CB assets [percent] 0.8 0.8 1.3 1.2 1.3 3.4 0.4 0.6 1.7 Below 7.0 percent CET1 ratio Community BHCs [number] 117 146 278 236 266 492 87 154 342 Percent of CB assets [percent] 4.3 5.4 9.9 10.6 10.1 20.7 3.3 5.7 13.2 Number of community banks >9.0 percent leverage ratio Start 3,423 3,423 3,423 3,423 3,423 3,423 3,423 3,423 3,423 End 1,811 1,345 1,096 1,571 1,132 861 1,975 1,372 1,056 Difference 1,612 2,078 2,327 1,852 2,291 2,562 1,448 2,051 2,367 Source: Federal Reserve Board, Form FR Y-9C, Consolidated Financial Statements for Bank Holding Companies; Consolidated Reports of Condition and Income, FFIEC 031/041; If a bank falls below the 7.0 percent threshold, it is using its conservation buffer. As a result, that bank would face restrictions on payout policies which would decrease the possibility of the bank falling below the regulatory minimum of 4.5 percent. For the banks that fell below 4.5 percent under our hypothetical scenarios, we rerun the stress tests and turn off dividends. Table 4 summarizes the results. Stopping dividend payments stops about half the banks from breaching regulatory minimums. ##### Table 4: Summary of Resiliency Tests Without Dividends BHCs that fall below 4.5 percent CET1 ratio DFAST year: 2017 2018 2019 Estimation method: (1) Avg. (2) OLS (3) QREG (4) Avg. (5) OLS (6) QREG (7) Avg. (8) OLS (9) QREG Original BHCs [number] 69 66 131 115 128 274 56 66 175 Percent of banking assets [percent] 0.5 0.3 0.6 0.9 0.6 1.6 1.8 0.8 1.3 No dividend BHCs [number] 39 32 52 53 55 116 32 32 74 Percent of banking assets [percent] 0.2 0.1 0.2 0.4 0.2 0.5 1.5 0.6 0.7 Community Banks Original Community BHCs [number] 23 20 40 39 41 101 18 20 60 Percent of CB assets [percent] 0.8 0.8 1.3 1.2 1.3 3.4 0.4 0.6 1.7 No dividend Community BHCs [number] 10 7 13 13 13 34 8 7 21 Percent of CB assets [percent] 0.1 0.1 0.3 0.2 0.3 0.8 0.1 0.1 0.5 Source: Federal Reserve Board, Form FR Y-9C, Consolidated Financial Statements for Bank Holding Companies; Consolidated Reports of Condition and Income, FFIEC 031/041; Another way to assess the resiliency of BHCs is to understand the sources of their vulnerabilities. Namely, what loan portfolios are driving losses? Figure 5 shows loan loss composition for the DFAST BHCs and the non-DFAST BHCs across the three estimation methods. The loan loss rates from DFAST 2017, 2018, and 2019 are applied to the loan types, and composition graphs are generated using the average dollar loss estimate for each loan type across the three DFAST exercises. As highlighted in Vojtech (2019), CRE is much more important for the non-DFAST BHCs (purple). Approximately half of non-DFAST loan losses come from CRE. As expected, given less exposure to credit card loans, non-DFAST BHCs have a much smaller percent of their losses from credit cards (light blue). The composition of losses is about the same across the three estimation methods. However, the dollar losses range from an average of$193 billion to $243 billion. ##### Figure 5: Composition of Average Loan Losses DFAST 2017, 2018, and 2019 4. Geographic Analysis A final way to assess the resiliency of the entire financial system is to understand what geographic areas are more vulnerable to a severe shock. Using FDIC Summary of Deposit data, we match those BHCs that fall below the 7.0 percent CET1 threshold to their deposits in each state. Those states that have a higher share of deposits in such banks are likely to have more financial strain. The shares of such deposits in each state averaged across the three years of DFAST exercises in this analysis are summarized in table 5. For brevity only two forecast methods are shown—average and quantile—that generally capture the forecast range, somewhat mild to severe. Using the average method, 11 states have more than 10 percent of their deposit base affected (column 2). However, under the much harsher assumption using a quantile regression at the 75th percentile, 17 states have 20 percent or more of their deposit base exposed to BHCs with CET1 ratios estimated to be less than 7.0 percent. Note that while some of these states have high exposures, they represent a small portion of the national deposit base. Those shares are listed in columns 3 and 6. ##### Table 5: Forecasts Averaging Across DFAST 2017, 2018, and 2019 Percentage points (1) (2) (3) (4) 31.3 0.1 43.6 1.3 21.8 1.1 39.7 0.1 17.7 1.3 36.8 0.4 13.4 3.2 28.0 0.6 13.3 0.2 27.1 0.8 12.9 0.3 27.1 0.7 12.7 0.7 25.4 0.7 12.6 0.2 25.2 1.1 11.4 0.8 25.0 0.2 10.8 0.2 24.2 0.2 10.3 0.6 23.8 1.8 9.9 0.7 22.7 0.3 9.6 2.7 22.6 1.2 9.4 1.3 21.9 2.9 8.9 0.1 21.5 3.2 8.9 1.2 20.1 0.8 8.8 0.4 20.0 2.0 8.4 0.6 19.0 0.1 8.3 0.5 18.1 0.5 7.4 3.2 17.9 1.1 Source: Federal Deposit Insurance Corporation, Summary of Deposits (June 2018); Federal Financial Institutions Examination Council, Form FFIEC 031, Consolidated Reports of Condition and Income for a Bank with Domestic and Foreign Offices, and Form FFIEC 041, Consolidated Reports of Condition and Income for a Bank with Domestic Offices Only (December 31, 2018); Federal Reserve Board, Form FR Y-9C, Consolidated Financial Statements for Bank Holding Companies; References Hirtle, Beverly (2016). "Bank Holding Company Dividends and Repurchases during the Financial Crisis," Federal Reserve Bank of New York Staff Reports, no. 666, March 14, revised April 2016, https://www.newyorkfed.org/research/staff_reports/sr666.html. Vojtech, Cindy M. (2019). "Assessing the Resiliency of the Banking Industry to a Commercial Real Estate Price Shock," FEDS Notes. Washington: Board of Governors of the Federal Reserve System, May 30, 2019, https://doi.org/10.17016/2380-7172.2380. Appendix ##### Table 6: Robustness Tests for Banking System ##### Panel A: 80 percent of Baseline Nine-quarter PPNR DFAST year: 2017 2018 2019 Estimation method: (1) Avg. (2) OLS (3) QREG (4) Avg. (5) OLS (6) QREG (7) Avg. (8) OLS (9) QREG Below 4.5 percent CET1 ratio BHCs [number] 90 95 187 143 191 369 68 93 242 Percent of banking assets [percent] 0.8 0.5 0.9 1.0 1.2 3.2 1.9 0.9 1.9 Below 7.0 percent CET1 ratio BHCs [number] 435 480 796 701 787 1,189 320 503 919 Percent of banking assets [percent] 7.5 7.8 10.8 11.0 11.8 14.8 8.4 9.9 14.4 ##### Panel B: 120 percent of Baseline Nine-quarter PPNR DFAST year: 2017 2018 2019 Estimation method: (1) Avg. (2) OLS (3) QREG (4) Avg. (5) OLS (6) QREG (7) Avg. (8) OLS (9) QREG Below 4.5 percent CET1 ratio BHCs [number] 61 57 96 94 100 212 54 58 127 Percent of banking assets [percent] 0.4 0.2 0.3 0.6 0.3 0.9 1.8 0.7 0.9 Below 7.0 percent CET1 ratio BHCs [number] 248 264 436 394 440 738 170 272 527 Percent of banking assets [percent] 1.6 1.4 3.1 2.6 3.1 6.5 2.4 2.5 5.7 Source: Federal Reserve Board, Form FR Y-9C, Consolidated Financial Statements for Bank Holding Companies; Consolidated Reports of Condition and Income, FFIEC 031/041. ##### Table 7: Average DFAST Loss Rates by Loan Category Percentage points 2017 2018 2019 2.2 2.7 1.4 4.5 4.9 2.6 6.4 7.3 6.3 7 8.3 6.4 13.7 14.4 16.8 5.9 5.5 4.7 3.6 4 3.6 5.8 6.4 5.7 Source: Federal Reserve Board, Form FR Y-9C, Consolidated Financial Statements for Bank Holding Companies; Consolidated Reports of Condition and Income, FFIEC 031/041. ##### Table 8: Sample Sizes for Loss Rate Regressions Firm count Loan Category 2017 2018 2019 Jr. lien mortgage & HELOC 25 24 10 Commercial real estate 28 29 15 Other consumer loans 25 27 11 Other loans 29 29 16 Note: Sample is restricted to firms with greater than 2.5 percent of total loans in each category except for Jr. Lien Mortgages and HELOC where the threshold is 1 percent. Source: Federal Reserve Board, Form FR Y-9C, Consolidated Financial Statements for Bank Holding Companies. 1. The analysis in this note is based entirely on publicly available information. We thank William Bassett, Jose Berrospide, Andrew Cohen, and Robert Sarama for comments. Return to text 2. For the summary statistics in the first part of this note, the DFAST group is defined using the BHCs subject to the 2018 test. The non-DFAST group is defined as all other banking institutions. The 2018 DFAST BHCs are: Ally Financial, American Express, Bank of America, Bank of New York Mellon, Barclays, BB&T, BBVA Compass, BMO Financial, BNP Paribas USA (formerly BancWest), Capital One, Citigroup, Citizens Financial, Deutsche Bank, Discover, Fifth Third, Goldman Sachs, HSBC, Huntington, JPMorgan, KeyCorp, M&T Bank, Morgan Stanley, MUFG, Northern Trust, PNC Financial, RBC USA, Regions Financial, Santander, State Street, SunTrust, TD Group, U.S. Bancorp, and Wells Fargo. All BHCs above$3 billion in consolidated assets are captured through a mandatory filing, the FR Y-9C form. All other BHCs and standalone banks are captured using Call Report (FFIEC 031 and 041) data. Return to text

5. We do not account for deferred tax assets (DTAs) because we are not calculating a quarter-by-quarter estimate of net income. If DTAs were included, they would offset losses and perhaps keep some banks from falling below regulatory capital thresholds. Return to text