Fannie Mae faces $25bn in derivatives lossesBy Stephen Schurr in New YorkPublished: March 9 2004 19:31 | Last Updated: March 9 2004 19:31Fannie Mae paid a net $25.1bn on derivatives transactions in under four years - nearly all of which may represent losses that cannot be recouped, in turn depressing future earnings.<>The potential scale of the liabilities, which have yet to be recognised in the company's earnings or in the minimum capital adequacy required by its regulator, raise fresh doubts about the financial health of the mortgage finance giant.<>Regulation of Fannie Mae and its sibling Freddie Mac is rapidly moving up the agenda in Washington, amid concerns that the two goverment-sponsored entities have grown so big that they pose a systemic risk to the US financial system. The two entities own or guarantee mortgages totalling $4,000bn.
From the FT article:The total cash flow hedge losses in AOCI are the sum of the realized and unrealized derivative gains and losses. Therefore, one can arrive at the realised losses by subtracting the unrealised amount from the total. Fannie Mae does not disclose the unrealised figure. However, the vast majority of its $1,072bn derivatives portfolio gets classified as cash flow hedges.Fannie Mae breaks down the notional value of its various derivatives positions. The Financial Times concluded the best method of estimating the amount of realised cash flow hedge losses is to apportion gains and losses by notional value.When net values are broken down by proportion in this manner, the unrealised cash flow hedge losses would total $1.1bn.At the end of the third quarter, Fannie Mae's AOCI was negative $24.757bn on a pretax basis (on an after-tax basis, the total was $16.092bn). If AOCI is the sum of realised and unrealised derivatives positions, then subtracting the $1.1bn would mean Fannie Mae's total realised derivatives losses would be $23.653bn (or, on an after-tax basis, $15.375bn).http://tinyurl.com/2j3pyI have a question of methods. Are you comfortable with FT's decision to apportion according to notional value? I think you can see that apportionment as Table 25 on page 74 of the following PDF. Notice how notional and net fair value figures change from 2001 to 2002 for each entry:http://www.fanniemae.com/ir/pdf/sec/2003/f10k03312003.pdfI don't know enough to have an opinion in that regard, and am more asking an inquisitive question about FT's methods than posing an opinion in the form of a question.My other question is, aren't unrealized losses expected to compensate for realized gains for other parts of their balance sheet, assuming they've properly hedged?Those questions asked, what strikes me is the sheer monumental challenge of trying to hedge with derivatives such a massive portfolio of liabilities and assets. There was a recent story in the press about Fannie's outdated IT. Here it is:Fannie Mae relies extensively on manual financial editor systems that "carry significant risk of error" and should address the problem, Fannie's regulator told the company this week. The Office of Federal Housing Enterprise Oversight said in a letter dated Tuesday that it was concerned about Fannie's lack of fully automated systems for certain accounting functions and its reliance on more than 70 manual systems for reporting its financial results. http://www.washingtonpost.com/wp-dyn/articles/A10911-2004Feb26.htmlI mean, on a $1 trillion asset base, $25 billion becomes little more than a 2.5% hedging error.These are complicated questions, I've not quite the tools to dissect, and I'm not sure what I'm supposed to think.Provacative posts here today, by the way,Jimi
In Buffett's last letter to shareholders, he talks about Berkshire's substantial losses last year in unwinding GenRe's hedge book. A hedge book that, on paper, apparently didn't look bad. A hedge book that they could have continued to carry forward without reporting any losses last year. But a hedge book that's costing an arm and a leg to unwind.Riiinggg!tr? [Jimi?]. Warren here. Yeah, I've got an open deritive here on 2023 Argentinian swap rates and I was wondering if you'd be interested in buying to close?"
I mean, on a $1 trillion asset base, $25 billion becomes little more than a 2.5% hedging error.But with total stockholder equity of only $16.8 billion, it's bit of a bummer for shareholders isn't it? Leverage of 55:1 isn't real forgiving of 2.5% errors. It isn't even very forgiving of 0.5% errors. Todd
"I mean, on a $1 trillion asset base, $25 billion becomes little more than a 2.5% hedging error.""But with total stockholder equity of only $16.8 billion, it's bit of a bummer for shareholders isn't it? Leverage of 55:1 isn't real forgiving of 2.5% errors. It isn't even very forgiving of 0.5% errors. "Todd All true. However, the underlying portfolio has an LTV of 60% - 65%. So, things have got to get pretty ugly before FNM needs to tap the reserves, and then Shareholder equity. Of course, the risk is in "the tails". If the FNM derivative usage is truly "hedge-based", the risk-profile on the interest rate side is truly improved. IF FNM is taking directional positions in their derivative portfolio, all bets are off.
Fannie Mae whistleblower Roger Barnes, in the following report collapses the entire accounting scandal into one sentence:"The FNMA director responsible for modeling, reportingand accounting for purchase premia/discount amortization, Jeffrey ...Juliane indicated that he was prepared to generate any results desired..." http://financialservices.house.gov/media/pdf/100604rb.pdf.This and other nuggets recently perused confirms my worst suspicions and provides hard evidence that nothing generated by any of Fannie's models can be trusted. Fannie Mae also appears to be a pioneer in the use of 'negative amortization'.
Best Of |
Favorites & Replies |
Start a New Board |
My Fool |