As Mr. Market has taken a short-seller's research and made significant adjustments to Bank of America in the recent week, I see opportunity in the shares. Opportunity exists because Branch Hill Capital has significantly over-shot its research. Research slides here.
Let's start with the basics:
|Stock Price (10/15 Close)||$ 11.98|
|Book Value/Share (6/30 10-Q)||$ 23.24|
|Tangible BV/Share (6/30 10-Q)||$ 12.14|
|Tangible BV/Share (6/30 with CCB Value)||$ 12.86|
|EPS Mean Analyst Est. 2010 (Yahoo Finance)||$ 0.91|
|EPS (6/30 10-Q) Actual||$ 0.55|
|2009 EPS||$ (0.29)|
|2008 EPS||$ 0.54|
|2007 EPS||$ 3.29|
At the recent closing price, BAC is trading at a discount of 1.32% to tangible book value, 6.84% to tangible book value after the inclusion of the China Construction Bank interest, and 48.45% of book value. The stock trades at a PE of 13 for 2010 Mean Estimates and 8 times 2011 estimates.
This is the basic analysis. Trading at a discount to book and a low multiple to future earnings means a great deal doesn't it? However, the major issue weighing down the stock is the unforeseen risk with mortgage put backs required by GSEs, monoline insurers, and private label MBS.
Branch Hill Capital Projects the following losses based on its own research coupled with Compass Point Research:
|Loss Projection||21.8 billion||7.2 billion||$45 billion||74 Billion|
|Put Back Projection||36.6 billion||12 billion||Not provided|
Branch Hill fortunately provides its assumptions and rough analysis to provide the figures. Let's start with the Monolines because this is the easiest to counter given the recent 10-Q from BAC. Branch Hill uses a arbitrary multiple of 3 to estimate the ultimate put back claim from the Monolines based on the recent 10-Q. Let's go to the 10-Q regarding the $4 billion mentioned:
So $2.3 billion of $4 billion were reviewed and it is believed there is not a defect which would indicate a breach of contract. This is an astounding 57.5% of the claims! Now there are another $9.8 billion underreview but no purchase request has been filed. It is not exactly clear the remaining $1.7 billion are valid claims, but needless to say taking a completely arbitrary multiple of 3 against the higher figure then assigning another arbitrary loss of 60% against this figure misleads the reader to believe higher losses can be expected.At June 30, 2010, the unpaid principal balance of loans related to unresolved repurchase requests previously received from monolines was approximately $4.0 billion, including $2.3 billion that have been reviewed where it is believed a valid defect has not been identified which would constitute an actionable breach of representations and warranties and $1.7 billion that is in the process of review. At June 30, 2010, the unpaid principal balance of loans for which the monolines had requested loan files for review but for which no repurchase request has been received was approximately $9.8 billion
Moving onto the Private Label figure the loss projection is again misleading. Here is a link to a Compass Research Note dated August 16, 2010 that is referenced in the slides by Branch Hill. The link is a pdf document. Compass Research uses the lawsuit of FHLB San Francisco v. Credit Suisse Securities (NYSE:USA) LLC, et al.) and makes the following indication:
This is likely a strench to extrapulate from the FHLB suits to the entire industry. It also does not provide any potential credit to the counter parties on their viewpoint on the documentation and exceptions. This underpinnes Branch Hill's calculation of losses.In the worst case scenario, we assume that the rescission requests identified in the FHLB suits are indicative of the total potential
pool of loans that could be rescinded industry-wide. While we cannot opine on whether or not the suit’s rescission percentage will
ultimately be proven accurate, we believe that the data set forth in each particular suit is substantial enough to establish a worst case
Unfortunately as much as I have dug online as an individual investor, I have not located any definitive data on what the success rate is on private label rescissions/put backs and/or the current recission rate across financial institutions. Compass provides the following table in its research:
|FHLB Recission Rate||43.2%||49.1%||54.5%|
|Alt A Success Rate (1)||50.0%||60.0%||75.0%|
|Alt A Severity Ratio (2)||50.0%||55.0%||60.0%|
|Subprime RMBS Success Rate||80.0%||80.0%||80.0%|
|Subprime RMBS Severity Ratio||50.0%||55.0%||65.0%|
(2) Higher % equals larger losses as banks lose more from loans of the provided vintage
So because they are subprime, it is assumed they will be worse than Alt-A, and therefore more likely to be successfully put back to the banks. The figures also reflect high recission rates to be applied across a large span of assets. So Compass/Branch Hill assumes that GSE's will put successfully put back 20% with a loss of 60% and in private label mortgages it will be nearly double in terms of success? I doubt this logic will hold true. Why would the underwriting be that much worse by the same lenders to reflect the material discrepancy? Furthermore, only $33 million has been submitted for recission on Private Label MBS. More will come, but $33 million is nothing when you compare it to Branch Hill's $45 billion loss figure.
Moving to GSEs, reference Oppenheimer & Co.’s Chris Kotowski. As stated on Barrons.com:
Moreover, Kotowski asserts, since there’s a delay of 12 to 20 months between the time a loan becomes delinquent and the time that Fannie or Freddie request a repurchase, it’s important that “problem flows,” meaning, new loans showing up as an issue, are actually declining. “The level of GSE put-back requests should have hit their peak/plateau somewhere between Q1 2010 and Q3.”
Based on the total expected repurchases of Fannie and Freddie, $27 billion, and B of A’s share, B of A could be facing a total loss of $3.153 billion, Kotowski estimates.
Read his research for further information.
Regarding the loss assumptions, the S&P Case Shiller national average has lost 27.33% from the peak in Q2 2006 to Q2 2010. Even if you pad some numbers for costs of foreclosing, the properties did not lose 50% to 65% of their value plus any principal reduction received during the performing years of the loan. This is just silly. If they were land loans, I would understand the discount. I have looked and not located, but would love to see the hard data on what the actual loss has been.
As it is getting late, the summary is this:
- BAC is undervalued - I see $17 as an easy 12 month target
- Branch Hill capital is pushing paper to benefit its disclosed short on the stock. The numbers do not add up and the assumptions within the report do not stand up to the facts.
- BAC with current reserves of $3.9 billion is adequately protected in the near-term and has the earning capability to cover additional exposure in the long-term.
Disclosure: Long BAC