Monday, July 12, 2010

Bank of America comes clean – well sort of …

Bank of America has finally admitted that it understated the quarter end assets and liabilities for the years 2007 to 2009.  It does not (yet) admit that similar transactions took place in many other years and it does not spell out the effect of these transactions on BofA’s need to carry capital.  To quote BofA’s local paper:

Bank of America Corp. has told securities regulators that it made six quarter-end transactions from 2007 to 2009 that were not in "strict compliance" with accounting rules.

In correspondence with the Securities and Exchange Commission, the Charlotte bank said the so-called "dollar roll" transactions were designed to meet internal balance sheet limits. The bank said it does not believe the transactions had a material impact on its financial statements, according to a May 13 letter posted by the SEC late Friday.

I wrote a post stating that BofA had long been reducing its quarter-end balances in March this year so this should not surprise regular readers.  Nor will not surprise regular readers of the Huffington Post and many other places where my article was reprinted.  I think the WSJ also had a poke at the story after my blog post.  Alas the story died down as BofA issued denials only to retreat from those denials in a (then private) letter to the SEC. 

BofA note that the transactions did not change reported profit.  I agree.  The transactions were however designed to shrink reported quarter-end balance sheet and hence reduce the apparent need to hold capital.  One of the reasons why BofA was short capital when the crisis came was that they did things like this to reduce the stated need for capital and they ran capital close to the “apparent” minimums. 

Anyway there are things that bug me about BofA’s admission.  Firstly at senior management it appears that they did not even know they were doing this.  The company denied the bleatingly obvious in the aftermath of my original blog post.  I do not think they were directly lying – the better explanation is that they simply did not know.  Moreover they now state that the transactions “were designed to meet internal balance sheet limits”.  In other words some internal part of the bank was using more balance sheet – hence more capital – than it was permitted under internal risk controls and entered into quarter end transactions to hide it. 

Lets put this more directly.  BofA imposes internal risk controls (usually called limits).  BofA staff enter convoluted transactions to avoid having to meet those limits.  Head office does not know – and only in response to an SEC subpoena (following a blog post a nondescript fund manager in Australia) do they conduct a review and find these transactions.  This is – it seems – worse than the transactions itself.  What it demonstrates is that BofA does not police its own risk control rules until forced to by SEC subpoena.  Put that way you have to ask “who in BofA will be forced to resign?”

More pertinently – this could be spotted by a (very) careful reading of annual and quarterly reports from Australia.  Everything needed to demonstrate that there was something strange about quarter ends could be done by someone with the published annual reports and quarterly summaries.  If anyone on BofA’s board carefully read the accounts they would have spotted the same thing.  (I guess that this demonstrates that the entire BofA board did not or was not capable of undertaking such a careful reading of their own accounts.)

My original post did this for 2006.  In 2006 as I showed the quarter end assets were substantially less than the assets averaged over the quarter.  In some quarters the difference is 46 billion dollars (substantially more than the 10 billion admitted to in 2007-2009).  The same incidentally is true of 2005 and I think (though I have not rechecked) that I first spotted this in 2004.  So far BofA has not come clean about those years.  But then again we now know that head office did not know that within the bank parties were entering transactions designed to thwart internal balance sheet limits – so if BofA cares to check they will find that the problem exists over many years. 

My estimate is that – as a result of this transaction – BofA’s looked like it required about 2 billion dollars less capital than it should have been carrying had it stated its balance sheet fairly.  2 billion in capital is significant – but is remains small in the overall problems that BofA had during the crisis.  This is – in an accounting sense – a second-order issue.  But as a statement about BofA’s control culture it is not good.  [The culture of hiding risk taking however should – in a post-crisis environment – be relatively easy to address…] 

The unnamed counterparty

These transactions were done – according to press articles – with counterparties unknown.  It is passé these days to charge the prostitute but not to charge the John.  The press however would often prefer to report on the (high profile) John than the prostitute.  Either way I wish this were corrected.

The transactions designed to hide quarter-end assets and debt were described as “roll transactions”.  I guess I am new at this game but I had not previously heard that jargon.  Anyway – with a “roll transaction” there has to be a counterparty who is willing to prostitute their balance sheet and allow the “assets” (at least temporarily) to be stuffed in.  The willing whore in this case was almost certainly Japanese – at least for some quarters.  Japanese banks typically had average balances of securities LOWER than end period balances (Mizuho is an exception).  I once meticulously went through the quarterlies of MUFJ and found exactly this trend.  In their SEC filings MUFJ tends to report its capital (or at least it did in those days) as average capital to average assets.  (I guess that makes sense when their end-period assets are stuffed with assets parked from American banks.)

Still if some trader at BofA (or someone else wanting to skirt BofA’s internal controls) wants to park assets at quarter end – and to pay good money for that privilege – then who am I to suggest that otherwise lowly profitable Japanese banks should say no?  

 

John

 

Disclosure:  I have never thought that we should use the blog to “talk our book”.  We will occasionally explain why we own things (which I guess is talking our book) but we are happiest discussing what is wrong with our positions.  Our biggest position remains long Bank of America (and it is more-or-less the only stock I suggest when people want a stock tip).  BofA might argue that with friends like us they don’t need enemies.  Maybe that is true – but perhaps they need board members that can read accounts. (I am offering…) 

Also – for the SEC – note that BofA shareholders have suffered much already because BofA took too little capital into the crisis.  What you should be seeking from BofA is not penalties – it is a process for fixing their internal controls so that head office can ensure that divisions are not entering transactions designed to thwart internal controls.  Surely that is far more important than a rap over the knuckles? 

Hey – the SEC should not need to seek this.  BofA should just do it.  I anticipate being a shareholder in a decade – so this matters to me. 

16 comments:

Anonymous said...

John

I think this is being over egged as the amount was very small in the total scheme of things.

Why would any firm try to something that had .1% effect on capital as it doesn't make sense.

J

John Hempton said...

In 2006 it changed the balance sheet by about 40 billion dollars - not 10 billion.

The key issue here is that some INTERNAL division of BofA was being remunerated on their RETURN ON EQUITY. If they reduced their balance sheet by 40 billion they reduced their capital needs by 2 billion.

They probably only employed 5 billion of capital - so this was the difference between 5 and 7 billion capital employed.

I should not need to calculate how much difference that might make to their remuneration.

The issue is NO LONGER the aggregate number - it is that BofA's internal controls suck.

That latter issue is under-egged.

J

Anonymous said...

John - You highlight one of the most interesting aspects of the financial crisis: the complete failure of many company boards of directors. It's easy and cheap to dismiss board members as overpaid and incompetent, but you add some real substance to that claim.

At the failed banks and insurance companies boards have turned out to be sheepish, non-critical spectators. No questioning of numbers (why is one team of 200 odd guys in a satellite office making a substantial amount of your profits?), no questioning of risks (why are all your profits mortgage lending related?), no questioning of financials or controls. It is not always true, but the omen (and need for careful consideration) was often an outside or unusual profit (Kerviel, Cassano). If any board would have spent some serious time looking at the 10 biggest profit centers in each of these firms they would have found out what is going on.

Oh well - if only commenting on blogs could get me an overpaid seat on some board.

Anonymous said...

John:

The roll-over was around $10 billion that I am aware of.

I don't accept your capital calculation that easily as different types of assets require smaller or larger capital haircuts.

Do you know that type of securities that were rolled and the amount capital required to anchor this sum?

J

Anonymous said...

John,

have you looked at the TARP warrants available on BAC?

How do you feel about the risk/reward in these warrants as opposed to the stock?

John Hempton said...

The bank was pretty close on 20 times gross levered when the amount rolled was nearly 50 billion -

20 times gross leverage for the holding company is - or at least was at the time - a level which if you breached you got a regulatory visit.

The 2 billion estimate is more-or-less sound...

--

As for the warrants - I own a few - but mostly the common...

J

Anonymous said...

John

Looks like your proposed bet with Shawn Richard would have been a winner.

Anonymous said...

You make the point that this was easy to spot (from Australia) and I am consistently impressed by your analytical effort and ability but would you have thought to look at this issue absent the Lehman repo 105 (manipulation of a true sale) issue?

John Hempton said...

I knew about it in 2006 - so yes - pre Lehman 105. I never thought to write about it till post 105.

J

狂猪 said...

Hi John,


I have a different take on BofA's investment outlook relative to it's peers (for example XLF). However, let me first thank you for your generosity of sharing your analysis of the US banks system during the darkest hour of the financial crisis. It was some of the best analysis I've read.

At any rate, here is my take on the investment prospect of BAC relative to XLF. I suspect BAC is unlikely to significantly outperform or under-perform XLF going forward. Statistic's sampling theory dictates sample mean is a good estimator of population mean. I'd argue BAC is a reasonable sample (good enough) for use to estimate the performance of the US financial sector (the population). As such, the probability is low for BAC's performance to drift very far apart from XLF's performance.

Sampling theory depends on the use of good probabilistic sample. BAC is one bank and normally a sample of one isn't a good probabilistic sample. However, it is wrong to think of BAC as just one bank. BAC has local bank outlets throughout the US. All local business factors that affects other community banks affects BAC. BAC also have global presences. As such global business factors also affect BAC. The performance of BAC is really the average performance of all these local and international banking operations. Similarly, the performance of XLF is the average of performance of many US financial firms.

I am not arguing BAC is an ideal probabilistic sample. All I am arguing is BAC is good enough as a representative sample for the US financial system. I would take this line of reasoning a step further and theorize that the bigger a company is the more representative it is of the business sector it is in. As such, the future performances of a very large corporation is unlikely to drift apart very far from the business sector it is in.

What about WaMu? It is big and it collapsed. Whereas the US financial system is still around. Clearly these two drift far apart. The key to keep in mind is the above discussion is in terms of probability. Low probability event can still happen. And you know the low probability event that took down WaMu:)

My view seemed to be supported by the stock performance of XLF, BAC, JPM, WFC and C for the last 12 months, 6 months and 3 months. Prior to July 2009, there was a significant advantage to owning BAC. In other words, the mis-price of BAC relative to XLF seems to have been corrected a year ago. I don't know this for sure. John, if you have more insights on the hidden value of BAC I love to hear it.

Besides XLF outperformed BAC, JPM and WFC slightly during the last 12 months, there were 2 other benefits. First, XLF was less volatile. Second, with XLF, idiosyncratic risk was completely eliminated. I probably value the second benefit more than most. The way I see it, if BAC goes out of business, it's business will most likely get taken over by other firms in the XLF. In another word, the profit potential and therefore long term shareholder value stays in XLF.

Anyway, these are the reasons why going forward I think XLF is way better than BAC. However, just to emphasis, since this is a probabilistic look at BAC relative to XLF, it cannot account for any significant hidden value in BAC's operation. This is my analysis' biggest weakness.

狂猪 said...

Hi John,


I have a different take on BofA's investment outlook relative to it's peers (for example XLF). However, let me first thank you for your generosity of sharing your analysis of the US banks system during the darkest hour of the financial crisis. It was some of the best analysis I've read.

At any rate, here is my take on the investment prospect of BAC relative to XLF. I suspect BAC is unlikely to significantly outperform or under-perform XLF going forward. Statistic's sampling theory dictates sample mean is a good estimator of population mean. I'd argue BAC is a reasonable sample (good enough) for use to estimate the performance of the US financial sector (the population). As such, the probability is low for BAC's performance to drift very far apart from XLF's performance.

Sampling theory depends on the use of good probabilistic sample. BAC is one bank and normally a sample of one isn't a good probabilistic sample. However, it is wrong to think of BAC as just one bank. BAC has local bank outlets throughout the US. All local business factors that affects other community banks affects BAC. BAC also have global presences. As such global business factors also affect BAC. The performance of BAC is really the average performance of all these local and international banking operations. Similarly, the performance of XLF is the average of performance of many US financial firms.

I am not arguing BAC is an ideal probabilistic sample. All I am arguing is BAC is good enough as a representative sample for the US financial system. I would take this line of reasoning a step further and theorize that the bigger a company is the more representative it is of the business sector it is in. As such, the future performances of a very large corporation is unlikely to drift apart very far from the business sector it is in.

What about WaMu? It is big and it collapsed. Whereas the US financial system is still around. Clearly these two drift far apart. The key to keep in mind is the above discussion is in terms of probability. Low probability event can still happen. And you know the low probability event that took down WaMu :)

My view seemed to be supported by the stock performance of XLF, BAC, JPM, WFC and C for the last 12 months, 6 months and 3 months. Prior to July 2009, there was a significant advantage to owning BAC. In other words, the mis-price of BAC relative to XLF seems to have been corrected a year ago. I don't know this for sure. John, if you have more insights on the hidden value of BAC I love to hear it.

Besides XLF outperformed BAC, JPM and WFC slightly during the last 12 months, there were 2 other benefits. First, XLF was less volatile. Second, with XLF, idiosyncratic risk was completely eliminated. I probably value the second benefit more than most. The way I see it, if BAC goes out of business, it's business will most likely get taken over by other firms in the XLF. In another word, the profit potential and therefore long term shareholder value stays in XLF.

Anyway, these are the reasons why going forward I think XLF is way better than BAC. However, just to emphasis, since this is a probabilistic look at BAC relative to XLF, it cannot account for any significant hidden value in BAC's operation. This is my analysis' biggest weakness.

狂猪 said...

Hi John,


I have a different take on BofA's investment outlook relative to it's peers (for example XLF). However, let me first thank you for your generosity of sharing your analysis of the US banks during the darkest hour of the financial crisis. It was some of the best analysis I've read. A signficant portion of my portfolio is in financial partly because of reading your work.

At any rate, here is my take on the investment prospect of BAC relative to XLF. Statistic's sampling theory dictates sample mean is a good estimator of population mean. I'd argue BAC is a reasonable sample (good enough) for use to estimate the performance of the US financial sector (the population). As such, the probability is low for BAC's performance to drift very far apart from XLF's performance. Otherwise, BAC wouldn't be a good estimator.

Sampling theory depends on the use of good probabilistic sample. BAC is one bank and normally a sample of 1 isn't a good probabilistic sample. However, it is wrong to think of BAC as just one bank. BAC has local bank outlets throughout the US and around the world. BAC is a collection of many banks. The performance of BAC is really the average performance of all these local and international banking operations.

BAC is definitely not an ideal probabilistic sample. All I am arguing is BAC is good enough as a representative sample for the US financial system. In general, I'd argue that the bigger a company is the more representative it is of the business sector it is in. Statistics make it very difficult for very large corporation to out perform the business sector it is in. It is easy to see this must be true if we consider the extreme case of a monopoly. To achieve sector outperformance, we have to look for small businesses.

One key point to keep in mind is the above discussion is in terms of probability. Low probability event can still happen. You know the low probability event that took down WaMu too well:)

To be continue ... (blogger.com has a length limit)

狂猪 said...

Continuing from previous comment...

My view seemed to be supported by the stock performance of XLF, BAC, JPM, and WFC for the last 12 months, last 6 months and last 3 months time frames. Prior to July 2009, there was a significant advantage to owning BAC. After that BAC has under performed slightly. In other words, the performance difference of BAC relative to XLF seems to have corrected a year ago after the financial panic over large US banks dissipated. Distortion due to panic tends to correct quickly. Unless there is a signficant hidden value in BAC's book, I think this trend is likely to continue. John, if you have more insights on the hidden value of BAC that is _not enjoyed_ by other banks I'd love to hear it.

Besides XLF slightly outperformed BAC, JPM and WFC during the last 12 months, there were 2 other benefits of owning XLF. First, XLF was less volatile. Second, with XLF, idiosyncratic risk was completely eliminated. I probably valued the second benefit more than most. The way I see it, if BAC dies, it's business will most likely get taken over by other firms in the XLF. In another word, the profit potential and therefore long term shareholder value stays in XLF.

Anyway, these are the reasons why going forward I think XLF is way better than BAC. However, just to emphasis, since this is a probabilistic look at BAC relative to XLF, it cannot account for any significant hidden value in BAC's operation that is not enjoy by it's competitors. This is my analysis' biggest weakness. John, your deep insight into BAC's financial condition can overrule by my view above.

Jacob said...

John

What do you think of Citi's report this morning?-hiding leverage via classifying repo agreements as sales....

Thanks

Wenger J Khairy said...

Hi John,
Just want to thank you for your excellent write up on the GSE.

I am in the midst of going thru Parts I - XI but had to drop this comment

You rock man


Wenger J. Khairy
Essendon supporter and Malaysian Socio Political Blog Writer

Banking News said...

Big banks manage all things

General disclaimer

The content contained in this blog represents the opinions of Mr. Hempton. Mr. Hempton may hold either long or short positions in securities of various companies discussed in the blog based upon Mr. Hempton's recommendations. The commentary in this blog in no way constitutes a solicitation of business or investment advice. In fact, it should not be relied upon in making investment decisions, ever. It is intended solely for the entertainment of the reader, and the author.  In particular this blog is not directed for investment purposes at US Persons.