Stress Testing Nine Banks--It's All about Unemployment

[Pages:24]RESEARCH

April 22, 2009

Paul J. Miller, Jr., CFA . 703.469.1252 . pmiller@ Bob Ramsey, CFA . 703.312.1760 . bramsey@ Jessica Halenda . 703.312.9537 . jhalenda@

Financial Institutions

Important disclosures can be found on pages 23 - 24 of this report. Industry Update

Stress Testing Nine Banks--It's All about Unemployment

Summary and Recommendation

FBR has constructed it own stress test ahead of the planned release of the government's stress test parameters this Friday, April 24. We tested nine commercial banks under coverage, using 10%, 12%, and 14% unemployment rate scenarios. We conclude that, if unemployment peaks at 10%, roughly consistent with the government's stress test, most of the big banks will be able to earn through it. On the other hand, if unemployment is closer to 12%, which FBR believes is more realistic, their viability without additional capital is more questionable. FBR surveyed 62 buy-side clients and found that 41% expect unemployment to peak between 10% and 11% and that 39% expect unemployment to peak between 11% and 12%. A matter of significant debate among investors about the government's stress test is: will all the banks "pass," reinstating confidence in the market, or will certain institutions "fail" and be forced to raise additional capital at dilutive levels? We expect banks to be categorized into buckets, with some allowed to return TARP funds, some required to hold them, and some required to raise additional capital. In order for the test to have any credibility, we expect that a material number of the 19 tested institutions will have to raise additional capital; but the Treasury's goal is not to "fail" any of the tested institutions. The tests could result in considerable dilution for common investors, likely from forced conversion of preferreds and TARP and additional capital raises. One issue that we have with the government's stress test is that it will likely focus on Tier 1 capital, without adequate emphasis on the proportion of Tier 1 capital that is composed of tangible common equity. We do believe the government will require tangible common equity equal to at least 3% of risk-weighted assets, but we consider 3% just too low in any economic environment.

Key Points

? FBR unemployment survey. FBR surveyed over 250 buy-side clients and had responses from 62. Overall, 79% of the respondents expect unemployment to peak between 10% and 12%, with a slightly higher percentage (40%) expecting a peak in the 10% to 11% range. Less than 7% of respondents believe unemployment will peak below 10%, and 15% expect a peak above 12%. Further, most respondents expect a peak by mid 2010. Only 11% expect unemployment above 11% by year-end.

? Upcoming events: Treasury stress test results expected May 4. The Treasury Department announced in February that the 19 largest bank holding companies would be subject to a stress test to evaluate their capital adequacy. The test analyzes bank earnings and losses from both on- and off-balance-sheet exposures in both a baseline scenario and a more severe set of economic scenarios for 2009 and 2010. The Treasury hopes that the stress test will restore confidence in financial institutions. Expectations are that the government will release additional details of its methodology on Friday, April 24 and that, on May 4, it will release stress test results, including both the amounts of capital deficiencies and capital plans of individual companies. Companies will be able to raise capital through the Capital Assistance Plan if they are unable to do so in private markets.

? Value-add: our own version of the stress test. Of the 19 institutions taking the government's test, we apply the test to the nine commercial banks under coverage. We found that most banks are OK with minimal additional capital in our 10% unemployment scenario; but we are concerned about the declining proportion of tangible common equity relative to Tier 1 capital. The Federal Reserve has generally required voting common equity to be the "dominant" element within a bank holding company's Tier 1 capital. We believe that tangible common equity should make up a substantial majority of Tier 1 capital, ideally over 75%; and, historically, this has been the case for most financials. In the near term, we would not be surprised if the government were to allow forbearance on tangible common equity, but we continue to assert that stronger tangible common equity levels are necessary for these companies to be "investable" again.

? Is a global stress test appropriate? In applying universal loss assumptions to the nine commercial banks under coverage, we found that some companies' test losses, notably BBT and RF, are much higher than we currently model. Further, their losses to date have trended below the industry as a whole. This likely reflects the fact that all loan portfolios are not created equal and that, historically, losses have been the worst at the largest banks. In an effort to avoid any bias, we have chosen not to make any adjustments to loss estimates by individual company. We acknowledge that this may result in test results that are overly harsh for some. We understand that the government's stress test will allow some variation in estimated losses.

FBR CAPITAL MARKETS & CO.

1001 Nineteenth Street North . Arlington, VA 22209

FBR CAPITAL MARKETS & CO.

Institutional Brokerage, Research and Investment Banking

The Market Expects Unemployment to Exceed the Treasury's "Stressed" Scenario

The government's stress test is based on unemployment of 8.9% by the end of 2009 and on a peak of 10.3% in 2010. The table below provides the baseline scenario and alternative economic scenarios that the Treasury will use to evaluate the bank holding companies.

Real GDP growth

Baseline

2009

2010

-2.0%

2.1%

More Adverse

2009

2010

-3.3%

0.5%

Unemployment rate ?

8.4%

8.8%

8.9% 10.3%

Housing price declines ?

-14.0% -4.0% -22.0% -7.0%

? Annual average unemployment ? Based on Case-Shiller 10-City composite index, change 4Q over 4Q

Source: FBR Research, Treasury Department

We thought that this sounded more like a base case than a stressed scenario. To find out what the "consensus" expectation is, we surveyed approximately 250 buy-side clients and received 62 responses about unemployment expectations. We found that the "consensus" for peak unemployment is substantially higher than the government's stress test for the 19 largest banks.

Survey results: unemployment to peak between 10% and 12%. Overall, 79% of the respondents expect unemployment to peak between 10% and 12%, with a slightly higher percentage (40%) expecting a peak in the 10% to 11% range, rather than in the 11.1% to 12.0% range (39%). Less than 7% of respondents believe unemployment will peak below 10%, and 15% expect a peak above 12%.

Peak Level Unemployment

50%

Where do you think unemployment will peak?

40%

30%

20%

10%

0% Less than 10%

10-11%

11.1-12%

Source: FBR Research, QuestionPro

Greater than 12%

Page 2

FBR CAPITAL MARKETS & CO.

Institutional Brokerage, Research and Investment Banking

Additionally, 47% of respondents expect the peak to be reached in 2Q or 3Q 2010. Very few respondents (6%) expect unemployment to peak in 4Q 2010 or later. Alternatively, about 24% expect unemployment to peak in 4Q 2009.

Unemployment Peak Quarter

In which quarter do you think unemployment will

50%

peak?

40%

30%

20%

10%

0% 2Q09 3Q09 4Q09 1Q10 2Q10 3Q10 4Q10 1Q11 2Q11

Source: FBR Research, QuestionPro

Approximately 42% of participants expect unemployment to rise to 10.1% to 10.5% by year-end. Only 35% of respondents expect a year-end unemployment rate below 10%, and only 11% expect unemployment above 11% by year-end.

Year-End Unemployment

50%

Where do you think unemployment will be at the end of 2009?

40%

30%

20%

10%

0%

8.5-9% 9.1-9.5% 9.6-10%10.1-10.5% 10.6-11%11.1-11.5% 11.6-12%

> 12%

Source: FBR Research, QuestionPro

Taking the Stress Test One Step Further

We have said that most financial institutions are not adequately capitalized to handle 12% unemployment. Fortunately, only 15% of respondents expect unemployment to meet or exceed that level. To better understand how leveraged the banks are to various unemployment scenarios, we have created our own "stress test," which provides loan loss expectations for the banks, given three scenarios for peak unemployment (10%, 12%, and 14%) and associated peak-to-trough home price declines of 35%, 40%, and 45%. Following discussions with former chief risk officers, we have taken a thoughtful approach to cumulative losses over the next two years, incorporating 90+ day delinquencies, roll rates to default, and severities by loan product, with increasing defaults, roll rates, and severities in higher unemployment scenarios.

Of the 19 institutions taking the government's test, we apply the test to the nine commercial banks under coverage. We found that most banks are OK with minimal additional capital in our 10% unemployment scenario, but we are concerned about the declining proportion

Page 3

FBR CAPITAL MARKETS & CO.

Institutional Brokerage, Research and Investment Banking

of tangible common equity relative to Tier 1 capital. We believe that tangible common equity should make up the substantial majority of Tier 1 capital, ideally over 75%; but we think the government could give some near-term forbearance on tangible common equity. On the other hand, if unemployment is closer to 12%, which FBR believes is the more realistic level, their viability without additional capital is more questionable. Our full results by company can be found at the end of this note.

Methodology

? How quickly the losses are realized will be as important as the absolute amount. If the losses we model over the next two years actually are realized over the next three, the banks will have more earnings to help offset the losses. The level, speed, and peak in losses are all important, as are a bank's capital, reserves, and acquisition-related loan marks. Risk officers will usually utilize a two-year cumulative loss model to assess needed capital levels.

? We have used managed card receivables and managed tangible assets for BAC, AXP, and COF. While losses on securitized loans will not flow through net charge-offs, they affect the parent's overall profitability, and so we consider managed card portfolios as the best way to consider the loss potential.

? We model losses based on unpaid principal amounts of loans outstanding and give companies credit for any purchase accounting marks that they have taken as a type of shadow loan loss reserve. In certain situations, we have made assumptions about the amount of purchase accounting marks that remain or the amount of unpaid principal outstanding that seems reasonable, based on company disclosures.

? If a company defines its Alt-A exposure, we use its categorization; but most companies do not. We classify most residential mortgage loans as "prime," absent a better definition by the company. We recognize that certain of these portfolios include a fair amount of loans that could be classified as Alt-A.

? If a company provides details on its residential construction exposure, we use its disclosures; otherwise, we make an estimate based on regulatory filings.

? Our analysis looks at loan losses and does not include losses or recoveries on a company's securities portfolios (beyond any modeled gains or losses run through the income statement).

? Our analysis does not adjust for taxes. We have based our analysis off of pre-provision, pretax earnings. In many of our stressed scenarios, the companies lose money, given the uncertainty about a company's ability to use deferred tax assets and the discretion that regulators and accountants have in maintaining them on the balance sheet.

? Preferred dividends. We have adjusted our pretax, pre-provision earnings for grossed-up preferred dividends to present them on a pretax basis, along with other items. Preferred dividends will be a drain on a company's ability to retain earnings unless preferreds are converted, which would boost tangible common equity ratios.

? Unrealized securities losses. We calculate tangible common equity net of AOCI, and so we have not reversed it in our analysis of tangible common equity; however, in our analysis of Tier 1 capital, AOCI is excluded, consistent with regulatory methodologies.

? Loan loss reserves. The amount of Tier 1 capital and tangible common equity at the end of 2010 depends on the size of the company's allowance (and provisions). We assume that each company's allowance at the end of 2010 is equal to 70% of annual losses over the two-year period, which is consistent with our expectation of losses being lower in 2011. If the allowance is larger or smaller, capital ratios will be impacted.

? Balance sheet growth. We assume no change in balance sheet size, and our capital ratios are calculated off of current risk-weighted assets and tangible assets. To the extent that companies continue to de-leverage, smaller balance sheets will have a favorable impact on capital levels. Some of this benefit will be mitigated by lower net interest income earned off of smaller balance sheets, which will result in smaller pre-provision, pretax earnings and less retained earnings.

Page 4

FBR CAPITAL MARKETS & CO.

Institutional Brokerage, Research and Investment Banking

Bank of America

Capital Cushion After 2 Years 2-Year Cumulative Loss Estimates

TCE + ALLL ($)

ALLL (estimate) TCE ($) TCE/Tangible assets TCE/RWA Tier 1 Capital + ALLL ($) Tier 1 Capital ($) Tier 1 Capital/RWA

Unemployment Scenario

10%

12%

$90,227

$136,527

$89,896

$43,294

$31,580 $58,317

2.5% 3.4% $191,373 $159,794

$47,784 -$4,491

-0.2% -0.3% $144,771 $96,986

9.4%

5.7%

14% $192,998

-$13,498

$67,549 -$81,048

-3.5% -4.8% $87,979 $20,429 1.2%

40,000

10% Unemployment

35,000

30,000

25,000

20,000

15,000 10,000

Pre-tax, Pre-provision earnings

5,000

1Q08

2Q08

3Q08

4Q08

BAC Net Charge-Offs

12% Unemployment

14% Unemployment

Historical

1Q09 2Q09 3Q09 4Q09 1Q10 2Q10 3Q10

4Q10

$ in Millions

$ in Millions

BAC Capital Cushion (TCE+ALLL+PAA Marks)

140,000

10% Unemployment

12% Unemployment

120,000 100,000

Can't fall below TCE+ALLL Ratio of 4%

80,000

60,000

40,000

20,000

-

(20,000)

(40,000) 1Q08

2Q08

3Q08

4Q08

1Q09

2Q09

3Q09

14% Unemployment 4Q09 1Q10 2Q10

Source: FBR Research and company documents

3Q10

4Q10

Page 5

FBR CAPITAL MARKETS & CO.

Institutional Brokerage, Research and Investment Banking

Bank of America

Bank of America Amount Outstanding (Mil) Loan Mix (%)

PRIME

234,121 21%

90+ delinquent Roll to Default Annual Default

2 years Default Rate

3.5% 50% 1.8% 3.5%

Severity 2 years Loss Rate

40% 1.4%

Estimated Credit Loss Loss Mix (%)

3,278 4%

ALT-A

ESTIMATES (10.0% Unemployment, -35% HPA Peak-Trough)

Consumer Subprime Option ARM Construction

HELOC

Managed Card

Other Consumer

33,000 3%

8% 60% 4.8% 9.6%

2,000 0%

15% 70% 10.5% 21.0%

23,000 2%

20% 75% 15.0% 30.0%

0%

15% 85% 12.8% 25.5%

162,575 15%

5% 90% 4.5% 9.0%

218,031 20%

11.1% 90%

10.0% 20.0%

104,141 9%

6% 90% 5.4% 10.8%

50% 4.8%

1,584 2%

60% 12.6%

252 0%

65% 19.5%

4,485 5%

75% 19.1%

0%

95% 8.6%

13,900 15%

95% 19.0%

41,384 46%

90% 9.7%

10,123 11%

C&I

262,000 24%

3% 95% 2.9% 5.7%

85% 4.8%

12,694 14%

Bank of America Amount Outstanding (Mil) Loan Mix (%)

PRIME

234,121 21%

90+ delinquent Roll to Default Annual Default

2 years Default Rate

Severity 2 years Loss Rate

7.0% 60% 4.2% 8.4% 9,833.08 42% 3.5%

Estimated Credit Loss Loss Mix (%)

8,260 6%

ALT-A

ESTIMATES (12% Unemployment, -40% HPA Peak-Trough)

Consumer Subprime Option ARM Construction

HELOC

Managed Card

Other Consumer

33,000 3%

13% 70% 9.1% 18.2% 3,003.00 55% 10.0%

2,000 0%

25% 75% 18.8% 37.5% 375.00 60% 22.5%

23,000 2%

25% 78% 19.5% 39.0% 4,485.00 70% 27.3%

0%

17% 87% 14.8% 29.6% 80% 23.7%

162,575 15%

10% 90% 9.0% 18.0% 14,631.75 95% 17.1%

218,031 20%

12.0% 92%

11.0% 22.1% 24,070.62

97% 21.4%

104,141 9%

8% 90% 7.2% 14.4% 7,498.15 90% 13.0%

3,303 2%

450

6,279

0%

5%

-

27,800

46,697

13,497

0%

20%

34%

10%

C&I

262,000 24%

6% 95% 5.7% 11.4% 14,934.00 86% 9.8%

25,686 19%

ESTIMATES (14% Unemployment, -45% HPA Peak-Trough)

Bank of America Amount Outstanding (Mil) Loan Mix (%)

PRIME

234,121 21%

ALT-A

Consumer Subprime Option ARM Construction

HELOC

Managed Card

Other Consumer

33,000 3%

2,000 0%

23,000 2%

-

162,575

218,031

104,141

0%

15%

20%

7%

90+ delinquent Roll to Default Annual Default

2 years Default Rate

7.5% 65% 4.9% 9.8%

15% 70% 10.5% 21.0%

25% 75% 18.8% 37.5%

30% 85% 25.5% 51.0%

20% 88% 17.6% 35.2%

16% 90% 14.4% 28.8%

17.0% 92%

15.6% 31.3%

10% 90% 9.0% 18.0%

Severity 2 years Loss Rate

45% 4.4%

60% 12.6%

70% 26.3%

75% 38.3%

85% 29.9%

95% 27.4%

97% 30.3%

90% 16.2%

Estimated Credit Loss

10,272

Loss Mix (%)

5%

Source: FBR Research and company documents

4,158 2%

525

8,798

0%

5%

-

44,481

66,154

16,871

0%

23%

34%

9%

C&I

262,000 24%

8% 95% 7.6% 15.2%

88% 13.4%

35,045 18%

CRE

Resi Construction

Total

62,300 6%

3% 65% 2.0% 3.9%

50% 2.0%

1,215 1%

9,900 1%

1,111,068 100%

12% 85% 10.2% 20.4%

5.9% 82% 4.8% 9.7%

65% 13.3%

76% 8.1%

1,313 1%

90,227 100%

CRE

Resi Construction

Total

62,300 6%

5% 70% 3.5% 7.0% 2,180.50 60% 4.2%

2,617 2%

9,900 1%

1,111,068 100%

15% 87% 13.1% 26.1% 1,291.95 75% 19.6%

8.8% 84% 7.4% 14.8%

78% 12.3%

1,938 1%

136,527 100%

CRE

Resi Construction

Total

62,300 6%

7% 70% 4.9% 9.8%

65% 6.4%

3,969 2%

9,900 1%

1,111,068 100%

18% 90% 16.2% 32.4%

11.7% 86%

10.1% 20.3%

85% 27.5%

79% 17.4%

2,726 1%

192,998 100%

Page 6

FBR CAPITAL MARKETS & CO.

Institutional Brokerage, Research and Investment Banking

Wells Fargo

Capital Cushion After 2 Years 2-Year Cumulative Loss Estimates TCE + ALLL ($) ALLL (estimate) TCE ($) TCE/Tangible assets TCE/RWA Tier 1 Capital + ALLL ($) Tier 1 Capital ($) Tier 1 Capital/RWA

Unemployment Scenario

10% $67,270 $97,176 $23,544 $73,632 5.8% 6.7%

$145,243 $121,698

11.1%

12% $111,014 $53,438 $38,855 $14,583 1.1% 1.3% $101,504 $62,649 5.7%

14% $155,184 $9,370 $54,314 -$44,945 -3.5% -4.1% $57,436 $3,121 0.3%

35,000 30,000 25,000 20,000 15,000 10,000 5,000

1Q08

10% Unemployment

WFC Net Charge-Offs

12% Unemployment

14% Unemployment

Historical

Pre-tax, Pre-provision earnings

2Q08 3Q08 4Q08 1Q09 2Q09 3Q09 4Q09 1Q10 2Q10 3Q10

4Q10

$ in Millions

120,000

WFC Capital Cushion (TCE+ALLL+PAA Marks)

10% Unemployment

12% Unemployment

14% Unemployment

100,000

80,000

$ in Millions

60,000

Can't fall below TCE+ALLL Ratio of 4%

40,000

20,000

1Q08

2Q08

3Q08

4Q08

1Q09

2Q09

3Q09

4Q09

1Q10

2Q10

Source: FBR Research and company documents

3Q10

4Q10

Page 7

FBR CAPITAL MARKETS & CO.

Institutional Brokerage, Research and Investment Banking

Wells Fargo

Wells Fargo Amount Outstanding (Mil) Loan Mix (%)

90+ delinquent Roll to Default Annual Default

2 years Default Rate

Severity 2 years Loss Rate

Estimated Credit Loss Loss Mix (%)

PRIME

129,800 14%

3.5% 50% 1.8% 3.5%

40% 1.4%

1,817 3%

ALT-A

ESTIMATES (10.0% Unemployment, -35% HPA Peak-Trough)

Consumer Subprime Option ARM Construction

HELOC

Managed Card

Other Consumer

0%

8% 60% 4.8% 9.6%

25,000 3%

15% 70% 10.5% 21.0%

117,500 13%

20% 75% 15.0% 30.0%

0%

15% 85% 12.8% 25.5%

110,000 12%

5% 90% 4.5% 9.0%

23,000 3%

11.1% 90%

10.0% 20.0%

96,700 11%

6% 90% 5.4% 10.8%

50% 4.8%

0%

60% 12.6%

3,150 5%

65% 19.5%

22,913 34%

75% 19.1%

0%

95% 8.6%

9,405 14%

95% 19.0%

4,366 6%

90% 9.7%

9,399 14%

Wells Fargo Amount Outstanding (Mil) Loan Mix (%)

90+ delinquent Roll to Default Annual Default

2 years Default Rate

Severity 2 years Loss Rate

Estimated Credit Loss Loss Mix (%)

PRIME

129,800 14%

7.0% 60% 4.2% 8.4%

42% 3.5%

4,579 4%

ALT-A

ESTIMATES (12% Unemployment, -40% HPA Peak-Trough)

Consumer Subprime Option ARM Construction

HELOC

Managed Card

Other Consumer

0%

13% 70% 9.1% 18.2%

25,000 3%

25% 75% 18.8% 37.5%

117,500 13%

25% 78% 19.5% 39.0%

0%

17% 87% 14.8% 29.6%

110,000 12%

10% 90% 9.0% 18.0%

23,000 3%

12.0% 92%

11.0% 22.1%

96,700 11%

8% 90% 7.2% 14.4%

55% 10.0%

60% 22.5%

70% 27.3%

80% 23.7%

95% 17.1%

97% 21.4%

90% 13.0%

-

5,625

32,078

0%

5%

29%

-

18,810

0%

17%

4,926 4%

12,532 11%

ESTIMATES (14% Unemployment, -45% HPA Peak-Trough)

Wells Fargo Amount Outstanding (Mil) Loan Mix (%)

PRIME

129,800 14%

ALT-A

Consumer Subprime Option ARM Construction

HELOC

Managed Card

Other Consumer

-

25,000

117,500

0%

3%

13%

-

110,000

0%

12%

23,000 3%

96,700 11%

90+ delinquent Roll to Default Annual Default

2 years Default Rate

7.5% 65% 4.9% 9.8%

15% 70% 10.5% 21.0%

25% 75% 18.8% 37.5%

30% 85% 25.5% 51.0%

20% 88% 17.6% 35.2%

16% 90% 14.4% 28.8%

17.0% 92%

15.6% 31.3%

10% 90% 9.0% 18.0%

Severity 2 years Loss Rate

45% 4.4%

60% 12.6%

70% 26.3%

75% 38.3%

85% 29.9%

95% 27.4%

97% 30.3%

90% 16.2%

Estimated Credit Loss

5,695

-

Loss Mix (%)

4%

0%

Source: FBR Research and company documents

6,563 4%

44,944 29%

-

30,096

0%

19%

6,979 4%

15,665 10%

C&I

253,500 28%

3% 95% 2.9% 5.7%

85% 4.8%

12,282 18%

C&I

253,500 28%

6% 95% 5.7% 11.4%

86% 9.8%

24,853 22%

C&I

253,500 28%

8% 95% 7.6% 15.2%

88% 13.4%

33,908 22%

CRE

Resi Construction

136,000 15%

3% 65% 2.0% 3.9%

50% 2.0%

2,652 4%

9,700 1%

12% 85% 10.2% 20.4%

65% 13.3%

1,286 2%

CRE

Resi Construction

136,000 15%

5% 70% 3.5% 7.0%

60% 4.2%

5,712 5%

9,700 1%

15% 87% 13.1% 26.1%

75% 19.6%

1,899 2%

CRE

Resi Construction

136,000 15%

7% 70% 4.9% 9.8%

65% 6.4%

8,663 6%

9,700 1%

18% 90% 16.2% 32.4%

85% 27.5%

2,671 2%

Total

901,200 100%

5.9% 82% 4.8% 9.7%

72% 7.5%

67,270 100%

Total

901,200 100%

8.8% 84% 7.4% 14.8%

75% 12.3%

111,014 100%

Total

901,200 100%

11.7% 86%

10.1% 20.3%

77% 17.2%

155,184 100%

Page 8

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download