Not related to GSEs, but Accrued Interest make an interesting point:
"For example, there are 1,082 tax-exempt municipals bonds rated below investment-grade by Moody's. Of these, only 307 have traded any time this year. Now I grant that there is some correlation among junk-rated muni spreads, but how comfortable would you feel about the valuation of some struggling nursing home deal in Wisconsin by examining the trading level of a convention center in Texas?"
Don't forget that a lot of municipals (investment-grade or not) are really low-quality CRE debts or mortgages in disguise. The projects only barely made sense in the first place because of cheap tax-exempt financing.
But because they were able to get cheap debt and keep refunding (refinancing), they stayed just afloat.
"Its a little spooky to consider that Freddie Mac can't get a good value on their securities, that their dealer evaluations varied so much. That's the more important point in analyzing Freddie's balance sheet."
That Accrued Interest piece sucked. It was written by someone who doesn't understand accounting and who doesn't understand how to value illiquid securities.
As a former junk bond trader, I can say with 200% certaintly that "dealer quotes" on illiquid securities are significantly different than if FRE went to a dealer and said "give me a firm bid for $20 million of these bonds."
A "firm bid" is where these securities should be marked, not where the dealer gives a "quote" based on ratings, yield spreads, and other fixed income variables that are used to "mark to model."
This sytem is completely broken and I would bet anyone on this board any amount of money that AT LEAST 50% of FNM's and FRE's Level 3 assets are AT LEAST 50% overvalued.
As a former junk bond trader, I can say with 200% certaintly that "dealer quotes" on illiquid securities are significantly different than if FRE went to a dealer and said "give me a firm bid for $20 million of these bonds."
And if no one is offering "firm bids" because no one wants to buy illiquid securities today? Isn't that kind of the point?
"The hysteria over Freddie's Level 3 assets is misplaced. Thoughtful analysis as to why Freddie Mac felt compelled to move their assets into Level 3 is what's needed. Its a little spooky to consider that Freddie Mac can't get a good value on their securities, that their dealer evaluations varied so much. That's the more important point in analyzing Freddie's balance sheet."
I'll re-paste this part of Accrued's post.
Although I agree that we don't know what the value of these assets are, this is preciesely the point.
I get nervous especially when on the conference call they just blithely say "yeah, don't worry, we're just gonna slide these over here... no big deal, just a simple accounting thing ya know".
when in reality NOBODY KNOWS what the heck these things are worth. And despite that, it always seems to me that when things are slid over into Level 3 they tend to keep a high % of their original "worth".
then those calculations are used to estimate how "safe" the company is, how much higher we can push their capital reserve requirements, and how much more mortgages we'll allow them to take with their GSE label.
It gets old hearing people say "don't worry about this. you don't know that these aren't worth 100% of what we say they are worth" when the people making the model don't know that either.
I would say that I'm no more "wrong" calling all that trash worth $0 than they are calling it worth $157b.
After all, it was these great models that predicted that housing could never fall in value, right?
and there's sort of a mind game going on here too... psychological (but not in the same way as Shiller's article was psycho)...
these companies say "we have $157b of assets that are in Level 3"
so the premise is that they are worth $157b. then the naysayers have to show proof of their disagreement. since there is NO data, the naysaysers can't show proof, and thus the given value ($157b) stands.
this would be different if you said:
"we have assets of unknown value. we estimate they may be worth as much as $157b"
now the naysayers only need say "we've also shown through our model that the assets are worth $x" and you have 2 equal claims.
Tanta, the point is that these financial firms have Level 3 assets way overmarked on their balance sheets and the real problem is many multiples worse than the most people understand, including the good writers on this site.
To say that the price mark of most Level 3 assets on every bank balance sheet even remotely resembles reality is just plain ignorance.
And yes, contrary to the Accrued Interest piece, FRE is insovlent BUT for the implicit backing of the taxpayer. If it comes down to that, this whole country is insolvent.
Part of requiring the Level 3 disclosure was to allow investors to consider how much they want to trust asset valuations based on models, especially in a market like this. So if you want to discount the valuation of Level 3 assets, the new disclosure allows you to do so.
One way of looking at this is that it would have been MORE misleading for them to have left this stuff in Level 2. That would have been to express more confidence in the valuation than was warranted.
Had they left the stuff in Level 2, they would have been saying (implicitly) that they had solid inputs for establishing the price. By moving it to Level 3, they were saying that they had some real reason to doubt the solidity of those inputs. They did not change the inputs they used; they are, they say, using the same average dealer quotes to come up with the value that they did before. They are simply branding those dealer quotes "model-based" by shifting the assets into Level 3.
Very often I find myself puzzled by people who accuse businesses of "concealing" something that was only brought to their attention by a disclosure in the latest financial statement.
Tanta, the point is that these financial firms have Level 3 assets way overmarked on their balance sheets
And that's worse than having "overmarked" Level 2 assets?
I am trying to point out that if you have assets you fear could be overmarked, you're supposed to shift them into Level 3. Where the rest of us have ability to keep our eye on them.
I realize you won't be happy until FRE and everybody else writes that stuff way down. Fine. But what has that got to do with the Level 3 designation?
I find myself puzzled by people who give banks and securities firms the benefit of the doubt on their accounting and price marking system, especially since I know, FIRST HAND (as in, I have been involved personally giving accounts "pricing quotes") that are SIGNIFICANTLY and INTENTIONALLY much higher than if I had to give them a real market clearing level.
This whole system is imbued with unspeakable fraud and corruption, and the "price quote marking" system is a big part of that.
To dismiss that as just being part of the accounting mechanism to give the market some "clarity" is complete ignorance and lack of knowledge about how the sytem works in reality.
if you're arguing "it doesn't matter if we call these level 3 vs level 2" then I agree with this.
but I think most of us are instead wondering aloud HOW the firms value these Level 3 assets. there seems to be a general trend of optimism in valuation when it comes to Level 3.
I would be less worried if there was more transparency to their model.
for instance: it's one thing if they got these quotes that were all over the place:
$157b
$105b
$0
$200b
$375b
and then they said "yeah, this stuff is worth $157b on level 3"
it's another if they get quotes that they just don't "like":
$100b
$85b
$65b
$90b
then they say "yeah this stuff is worth $157 on level 3"
And one more point, it's not about putting some GAAP labels like Level 2 and Level 3 on assets and deciding what's good and what's bad. I would be most Level 2 assets are significantly over-marked as well.
It's not about GAAP accounting standards at all. Since when does GAAP bear any resemblance to market reality?
I'm using the Level 3 asset classification because that's the easiest one to say "they're not even remotely marked close to reality."
The problem compounds even more when you layer in the issue of Level 2 assets.
The bottom line is that the underlying MARKET economic reality of a substantial portion of the assets in the entire banking system bears no resemblence whatsoever the facade that GAAP accounting, the SEC and our policy makers are allowing to be erected.
Seems to me that AI missed the guarantees. You have to add them in when figuring potential losses. He also takes a "one or the other" approach to figuring total losses. What about both the Level III and the mortgages on the books suffering losses at the same time, and then, for good measure, throw in losses on the guarantees? I think the "losses have to be four times their estimates" ignores the "all things go wrong at once" possibility. Now throw in losses on their complex interest rate derivatives if the bond market goes into a tizzy. Why would the bond market go into a tizzy? Well, because the government's bail out of Fannie and Freddie comes into view.
So, basically, I found AI's analysis to be incomplete.
I know, FIRST HAND (as in, I have been involved personally giving accounts "pricing quotes") that are SIGNIFICANTLY and INTENTIONALLY much higher than if I had to give them a real market clearing level.
You know this. I know this. Freddie effin' Mac knows this.
Therefore, Freddie Mac decides that since your quotes are probably worthless, and it has to use those quotes to value its assets, it should declare them to be Level 3.
I still don't get your point. By putting this stuff in Level 3 FRE is basically saying that dealer quotes are not reliable. You are saying that dealer quotes are not reliable. You disagree with FRE where?
if you're arguing "it doesn't matter if we call these level 3 vs level 2" then I agree with this.
That is not at all what I am saying!
I am saying that this stuff no doubt belongs in Level 3. It would have been hard to defend FRE leaving them in Level 2. Leaving them in Level 2 would have involved FRE implicitly saying that the valuation is more reliable than it really is.
I think you want to believe that FRE put these assets in Level 3 just so it could make up valuations. OK, fine.
But if they had left those assets in Level 2, everyone would be claiming that they did so in order to make up valuations!
What about both the Level III and the mortgages on the books suffering losses at the same time, and then, for good measure, throw in losses on the guarantees?
It's been over a week since I looked at the actual financials, you know, but I thought I remember reading that the guaranty obligation was already in the Level 3 category? Did I misread that?
for instance: it's one thing if they got these quotes that were all over the place
That is EXACTLY what they said, in the conference call and in the financials. The dealer quotes had such wide variation that they felt using the average, which they do, was kinda shaky. So they moved the stuff to Level 3.
You are forgetting that if you can get dealer quotes on something, it usually is fine to call it Level 2. Freddie used to call it Level 2. Then they decided that the dealer quotes were too squirrelly.
I'm not trying to argue that the stuff is "properly" valued; I haven't the faintest damned idea what it's worth. I don't think anyone else does, either. Which is why I'm more suspicious about everyone's Level 2 assets than I am about their Level 3 assets! I think the point here is that the credit markets are so bad that everything is either 1 or 3; there is no 2. And FRE has almost no Level 1 assets, so that means what it means.
Are assets moved to level 3, so that actual losses do not have to be immediately REALIZED since THE HOLDING ENTITY can not accurately state that level 3 assets are not over valued or undervalued by a specific dollar amount at a particular point in time.
Isn't this an easy way to defer losses today in hopes that they might appreciate in value tomorrow. After all most companies can only declare loses up to a certain dollar amount before they become publically insolvent.
Are assets moved to level 3, so that actual losses do not have to be immediately REALIZED since THE HOLDING ENTITY can not accurately state that level 3 assets are not over valued or undervalued by a specific dollar amount at a particular point in time.
Man. Now I have to respond to fervent criticism on two blogs!
Ron:
The post specifically calls out dealer quotes as being systematically too high. Everyone in the bond business knows this. FYI. But as another comment says, what's the alternative?
I think I'm with Tanta here. The hysteria over Level 3 per se misses the point. And for what its worth, my understanding is that any property on which they have foreclosed is part of their L3 assets. That's where most of their losses are coming.
The losses have to be 16x their current expectations. I merely comment that seems unlikely not impossible.
HERE's the bottom line: I don't care what kind of GAAP bin the assets are classified in. My point is that the balance sheets of ALL U.S. banking/financial firms are fraudulently mis-marked - by a SUBSTANTIAL amount.
Here's another real life example which bankers don't understand. BAC recently took a much bally-hooed $6 billion "provision" for loses. They have $603 billion of risky assets which include: $165mm Trading Acct (could be a lot of Treasuries but also would include "kitchen sink" crap traders are hiding), $223 billion debt securities which would include all kinds of asset-backed crap "held for investment" because "we couldn't stuff any more into dumb-shit clients' portfolios," $50 billion derivitives - hmmmmm, and $165 billion "other," - hmmmm cubed.
So essentially BAC took a provision of 1% against those assets. That would be a penny. So if Pimpco came into BAC and said "make me a market on this SIV-backed money market security, you have me marked at par," BAC would say, "oh based on our loss provision I can pay you 99." RIIIIIGHT! ROFLMAO.
Even with the write-downs BAC has taken to date, to think that any part of the $600 billion in "assets" is market even within a country mile of where it should be marked is naive, gullible, and outright stupid.
This is total and complete fraud being perpetrated on the investing public and the global financial system.
And the writers, commentators and editors of blogs and all other forms of media are either being complicit or just don't understand the depth of the issues.
I know of recent bids emerging from vulture funds bidding a penny to a nickel for HELOC securities. Does anyone in their right mind believe BAC has its HELOC inventory marked even remotely near that level? Even remotely near 50 cents on the dollar?...
"I'm not trying to argue that the stuff is "properly" valued; I haven't the faintest damned idea what it's worth. I don't think anyone else does, either."
In this situation the holder of the assets gets to make the decision whether it sits at whatever level and price. This system seems to create problems discussed here and makes evaulating from an investor point of view almost imposssible other then to sit on the sideline or look for investments that offer better transparency. It would seem that a range of bids and from whom should be part of the financial disclosure.
The Guarantee Obligation is far bigger than the Level III assets, which are mostly subprime ABS.
Here's a quote from their 1q restuls:
"The company estimates that its total credit guarantee portfolio increased at an annualized rate of about 10 percent to approximately $1.8 trillion year-to-date through April 30, 2008'"
I guess I don't understand how AI would not include the G.O. when estimating losses, as it is far bigger than the Retained Portfolio.
What am I missing? Isn't Freddie on the hook for G.O. losses?
A financial security is only worth what the next best buyer is willing to bid with REAL capital for size. Our whole system is riddled with banking assets that have no bid.
And there has been no bid on most of these assets since August. So at what point do you admit that the models are wrong and that clearing prices are correct? Twelve months? Otherwise we are Japanese.
I said:
"for instance: it's one thing if they got these quotes that were all over the place"
Tanta replied:
"That is EXACTLY what they said"
Yes. I know that's what they SAID. but this is why I'd be interested in seeing the bids. again look at my examples.
if the quotes were "all over the place" but all really LOW this means something different IMO than the quotes being all over the place low and high.
I guess my question is this: using Level 3 accounting, how would one account for an asset that only gets bids extremely LOW.
so on $157b assets, what if it only gets:
$80b, $75b, $60b, $0, $100b bids.
do you get to mark that to $157b then?
because maybe, just maybe, those assets ARE worth in the $60-100b range.
it's just like the house across the street from me that is now in foreclosure. The owner was asking $400k for it. the bids he got were in the $250-325k range. He clearly didn't like the "irrational bids that were all over the place". Thus, foreclosure. but based on MY model (what houses have actually sold for around here in that condition) the place was worth around $320k. so he marked to model based on his wishing price.
FWIW:
I agree with you that these assets should have been placed in Level 3 over Level 2.
but it doesn't make me happy that we're valuing that company based on those assets at all.
I know there's no other choice, or so everyone keeps telling me. I guess what this means is that the American Financial system is severly broken, and has no chance to be fixed.
The company estimates that its total credit guarantee portfolio increased at an annualized rate of about 10 percent to approximately $1.8 trillion year-to-date through April 30, 2008'
The "credit guarantee portfolio" is not the same thing as the "guarantee obligation."
$1.8 trillion is the maximum possible exposure to credit losses they could have if every single loan in every single guaranteed security went belly-up and there were no collateral recoveries or shared recourse/indemnifications etc. of the seller/servicer. There will of course never be a time at which 100% of loans in 100% of securities experience 100% losses.
The "guarantee obligation" on the books is the net expected realized loss, not the maximum possible exposure. (The "guarantee asset" reflects the g-fee income that offsets that obligation.)
I don't see the problem with just using expected credit losses, which is more or less what the guaranty obligation is. As AI says, even if you want to assume catastrophic losses, you don't assume 100% on the guaranteed portfolio.
So at what point do you admit that the models are wrong
Well, some of us aren't Japanese, not that there's anything wrong with that. But to use an example from a previous spread, if I had a Picasso that no one bid on, I wouldn't declare it's value to be zero. The point that I would re=value at is that point in time at which I'm forced to recognize some value for my assets.
My personal issue with with level 3 assets is over the last year or so it's been the general rule that assets stuck in level 3 by financial institutions have proved to be worth much less than the estimated value. I believe Freddie Mac should take that into account when valueing the assets.
They should do something like take what they think the assets are really worth and multiply that amount by say 0.8, to get a conservative real value.
I'm very suspicious that Fannie Mae reported a much larger portfolio loss than Freddie Mac when there's no discernable difference between the two other than how they do their accounting.
AI implies that the losses are derived from assets, not guarantees, by stating that the asset value has to go down by 4% for Freddie senior debt to get hit. He then says this "factor of 16" miss on loss estimates is unlikely. This is just plain wrong. The total liability is $2.6tr, so a 1% loss would result in the equity being wiped out. That's 4x Freddie's 25bp estimate, and certainly in the realm of possibility.
Then, again, there's the issue of how Freddie's "black box" hedging will perform if rate volatility jumps. I think no one (including the company) really understands the risk, but that doesn't mean it should be ignored.
Level 3 assets don't mean you get to ignore data to arrive at the answer you want, it just means that you are using unobservable valuation inputs that respresent a market participant's assumptions "developed based on the best information available in the circumstances (FAS157 A25)."
So if there are dealer quotes out there that show that the assets are worth ~80B instead of ~157B, I think the burden of proof would fall to FRE to demonstrate why their information is in fact the best information available.
And I think Tanta and AI are 100% correct: I'd MUCH rather have someone disclose an asset (priced or mispriced) as a Level 3 if they have unobservable inputs. To AI's point, you can discount however you'd like. If you infer that all non-L3 assets are priced perfectly because of the disclosure, that's your choice.
Both these companies are clearly going to be insolvent by the end of the year, but everyone knows that
Congress will do anything to keep them afloat, because if Fannie and Freddie go under, the entire global
financial system will melt down, said Christopher Whalen, a founder of Institutional Risk Analytics, an
independent research firm. These companies earnings dont matter. Their accounting hardly matters.
People buy the stock because they believe the federal government will bail them both out if things get really
bad. http://www.institutionalriskanalytics.com/pdf/NYT_051508.pdf Institutional Risk Analytics
That is a ridiculous statement that's been making the rounds lately. Who the hell would buy a stock assuming there will be a govt. bailout? Bear Stearns was taken out at 1/8th of their price on 2/29. Who sees that and concludes a bailout is good for shareholders???
Accrued Interest - thanks for a great effort to explain some important points over on your blog post.
I walked into a shitstorm of ignorance on a related thread here at CR a week ago following Freddie's conf. call.
It started by my trying to make the simple point that 'level' designation is not a measurement of value. I was immediately shouted down by a mob of morons, natch.
Level 3 assets are going to be valued at estimated discounted future cash flow. Placing these securities as L3 only indicates that they can't be reliably or readily marketed. It's a form of confession. The risk lies with the probability inputs to the cash flow analysis.
Ron, keep in mind that every publicly traded manufacturing and retail business in the US has substantially overvalued fixed assets on their balance sheet when comparing the reported values to market values (what do you think all those store shelves at the mall are worth at auction? They are valued at cost - straight-line depreciation).
Level 2 inputs allow for situations where "price quotations vary substantially either over time or among market makers" (FAS157 28b).
If FRE's case isn't an example where divergent dealer quotes can be allowed as a Level 2 input, can someone provide a reasonable example where divergent quotes are a Level 2?
If not, whats the purpose of FAS157 28bs allowance of divergent price quotations? Or was this a case where the divergence of Level 2 inputs led FRE to use Level 3 inputs and, because the Level 3 inputs had the more material weighting, the entire asset was deemed Level 3 (per FAS 157 guidance that an asset gets sent to the lowest bucket that has significant inputs).
Yes I am a long retired fund manager, yes I live on the other side of the Atlantic, yes I first read the Financial Times in the Fifties. All these comments just show how far down the slippery slope we have travelled.
There was a time when there was no such distinction such as level 2 and level 3 assets. All assets had to be valued at lower of cost or market value and market value meant what they could be really sold for on an arms length basis: no slippery "explanations" would get past the auditors.
The present situation is a fiasco. I am with RonBurgandy.
Ancient... Yeah way back in like... 2006. FAS 157 is brand spanking new. Anyway, FAS 157 doesn't change anything about what kinds of assets are marked-to-market, merely increases the disclosure as to how they were valued. That seems to be the big misunderstanding. That FAS 157 is allowing companies to hide stuff, when in fact, they are just as able to hide stuff as before.
Begbie: My impression is that FRE made the call to move them to Level 3 because of the degree of divergence. I mean, if you get quotes of $50, $70, and $30, you should conclude that the valuation isn't based on anything readily observable.
AI: I agree with you, but it does call in to question what the adverb "substantially" means in the context of FAS157 28b. Given the FRE context, I'm assuming "substantially" means "not enough that you conclude nobody knows what they're doing".
Blech....feels like another bright line in the works to me.
The technicalities are way over my head, but I'm impressed with the way Tanta and Accrued Interest are colleagues rather than rivals. Good work, everyone!
Not related to GSEs, but Accrued Interest make an interesting point:
"For example, there are 1,082 tax-exempt municipals bonds rated below investment-grade by Moody's. Of these, only 307 have traded any time this year. Now I grant that there is some correlation among junk-rated muni spreads, but how comfortable would you feel about the valuation of some struggling nursing home deal in Wisconsin by examining the trading level of a convention center in Texas?"
Don't forget that a lot of municipals (investment-grade or not) are really low-quality CRE debts or mortgages in disguise. The projects only barely made sense in the first place because of cheap tax-exempt financing.
But because they were able to get cheap debt and keep refunding (refinancing), they stayed just afloat.
and this:
"Its a little spooky to consider that Freddie Mac can't get a good value on their securities, that their dealer evaluations varied so much. That's the more important point in analyzing Freddie's balance sheet."
Sorry for the OT, but seems relevant to everything that's been going on...
Cooked numbers
That Accrued Interest piece sucked. It was written by someone who doesn't understand accounting and who doesn't understand how to value illiquid securities.
As a former junk bond trader, I can say with 200% certaintly that "dealer quotes" on illiquid securities are significantly different than if FRE went to a dealer and said "give me a firm bid for $20 million of these bonds."
A "firm bid" is where these securities should be marked, not where the dealer gives a "quote" based on ratings, yield spreads, and other fixed income variables that are used to "mark to model."
This sytem is completely broken and I would bet anyone on this board any amount of money that AT LEAST 50% of FNM's and FRE's Level 3 assets are AT LEAST 50% overvalued.
Period.
As a former junk bond trader, I can say with 200% certaintly that "dealer quotes" on illiquid securities are significantly different than if FRE went to a dealer and said "give me a firm bid for $20 million of these bonds."
And if no one is offering "firm bids" because no one wants to buy illiquid securities today? Isn't that kind of the point?
"The hysteria over Freddie's Level 3 assets is misplaced. Thoughtful analysis as to why Freddie Mac felt compelled to move their assets into Level 3 is what's needed. Its a little spooky to consider that Freddie Mac can't get a good value on their securities, that their dealer evaluations varied so much. That's the more important point in analyzing Freddie's balance sheet."
I'll re-paste this part of Accrued's post.
Although I agree that we don't know what the value of these assets are, this is preciesely the point.
I get nervous especially when on the conference call they just blithely say "yeah, don't worry, we're just gonna slide these over here... no big deal, just a simple accounting thing ya know".
when in reality NOBODY KNOWS what the heck these things are worth. And despite that, it always seems to me that when things are slid over into Level 3 they tend to keep a high % of their original "worth".
then those calculations are used to estimate how "safe" the company is, how much higher we can push their capital reserve requirements, and how much more mortgages we'll allow them to take with their GSE label.
It gets old hearing people say "don't worry about this. you don't know that these aren't worth 100% of what we say they are worth" when the people making the model don't know that either.
I would say that I'm no more "wrong" calling all that trash worth $0 than they are calling it worth $157b.
After all, it was these great models that predicted that housing could never fall in value, right?
and there's sort of a mind game going on here too... psychological (but not in the same way as Shiller's article was psycho)...
these companies say "we have $157b of assets that are in Level 3"
so the premise is that they are worth $157b. then the naysayers have to show proof of their disagreement. since there is NO data, the naysaysers can't show proof, and thus the given value ($157b) stands.
this would be different if you said:
"we have assets of unknown value. we estimate they may be worth as much as $157b"
now the naysayers only need say "we've also shown through our model that the assets are worth $x" and you have 2 equal claims.
Tanta, the point is that these financial firms have Level 3 assets way overmarked on their balance sheets and the real problem is many multiples worse than the most people understand, including the good writers on this site.
To say that the price mark of most Level 3 assets on every bank balance sheet even remotely resembles reality is just plain ignorance.
And yes, contrary to the Accrued Interest piece, FRE is insovlent BUT for the implicit backing of the taxpayer. If it comes down to that, this whole country is insolvent.
I think we are missing this point:
Part of requiring the Level 3 disclosure was to allow investors to consider how much they want to trust asset valuations based on models, especially in a market like this. So if you want to discount the valuation of Level 3 assets, the new disclosure allows you to do so.
One way of looking at this is that it would have been MORE misleading for them to have left this stuff in Level 2. That would have been to express more confidence in the valuation than was warranted.
Had they left the stuff in Level 2, they would have been saying (implicitly) that they had solid inputs for establishing the price. By moving it to Level 3, they were saying that they had some real reason to doubt the solidity of those inputs. They did not change the inputs they used; they are, they say, using the same average dealer quotes to come up with the value that they did before. They are simply branding those dealer quotes "model-based" by shifting the assets into Level 3.
Very often I find myself puzzled by people who accuse businesses of "concealing" something that was only brought to their attention by a disclosure in the latest financial statement.
Tanta, the point is that these financial firms have Level 3 assets way overmarked on their balance sheets
And that's worse than having "overmarked" Level 2 assets?
I am trying to point out that if you have assets you fear could be overmarked, you're supposed to shift them into Level 3. Where the rest of us have ability to keep our eye on them.
I realize you won't be happy until FRE and everybody else writes that stuff way down. Fine. But what has that got to do with the Level 3 designation?
I find myself puzzled by people who give banks and securities firms the benefit of the doubt on their accounting and price marking system, especially since I know, FIRST HAND (as in, I have been involved personally giving accounts "pricing quotes") that are SIGNIFICANTLY and INTENTIONALLY much higher than if I had to give them a real market clearing level.
This whole system is imbued with unspeakable fraud and corruption, and the "price quote marking" system is a big part of that.
To dismiss that as just being part of the accounting mechanism to give the market some "clarity" is complete ignorance and lack of knowledge about how the sytem works in reality.
tanta;
if you're arguing "it doesn't matter if we call these level 3 vs level 2" then I agree with this.
but I think most of us are instead wondering aloud HOW the firms value these Level 3 assets. there seems to be a general trend of optimism in valuation when it comes to Level 3.
I would be less worried if there was more transparency to their model.
for instance: it's one thing if they got these quotes that were all over the place:
$157b
$105b
$0
$200b
$375b
and then they said "yeah, this stuff is worth $157b on level 3"
it's another if they get quotes that they just don't "like":
$100b
$85b
$65b
$90b
then they say "yeah this stuff is worth $157 on level 3"
And one more point, it's not about putting some GAAP labels like Level 2 and Level 3 on assets and deciding what's good and what's bad. I would be most Level 2 assets are significantly over-marked as well.
It's not about GAAP accounting standards at all. Since when does GAAP bear any resemblance to market reality?
I'm using the Level 3 asset classification because that's the easiest one to say "they're not even remotely marked close to reality."
The problem compounds even more when you layer in the issue of Level 2 assets.
The bottom line is that the underlying MARKET economic reality of a substantial portion of the assets in the entire banking system bears no resemblence whatsoever the facade that GAAP accounting, the SEC and our policy makers are allowing to be erected.
Seems to me that AI missed the guarantees. You have to add them in when figuring potential losses. He also takes a "one or the other" approach to figuring total losses. What about both the Level III and the mortgages on the books suffering losses at the same time, and then, for good measure, throw in losses on the guarantees? I think the "losses have to be four times their estimates" ignores the "all things go wrong at once" possibility. Now throw in losses on their complex interest rate derivatives if the bond market goes into a tizzy. Why would the bond market go into a tizzy? Well, because the government's bail out of Fannie and Freddie comes into view.
So, basically, I found AI's analysis to be incomplete.
I know, FIRST HAND (as in, I have been involved personally giving accounts "pricing quotes") that are SIGNIFICANTLY and INTENTIONALLY much higher than if I had to give them a real market clearing level.
You know this. I know this. Freddie effin' Mac knows this.
Therefore, Freddie Mac decides that since your quotes are probably worthless, and it has to use those quotes to value its assets, it should declare them to be Level 3.
I still don't get your point. By putting this stuff in Level 3 FRE is basically saying that dealer quotes are not reliable. You are saying that dealer quotes are not reliable. You disagree with FRE where?
if you're arguing "it doesn't matter if we call these level 3 vs level 2" then I agree with this.
That is not at all what I am saying!
I am saying that this stuff no doubt belongs in Level 3. It would have been hard to defend FRE leaving them in Level 2. Leaving them in Level 2 would have involved FRE implicitly saying that the valuation is more reliable than it really is.
I think you want to believe that FRE put these assets in Level 3 just so it could make up valuations. OK, fine.
But if they had left those assets in Level 2, everyone would be claiming that they did so in order to make up valuations!
What about both the Level III and the mortgages on the books suffering losses at the same time, and then, for good measure, throw in losses on the guarantees?
It's been over a week since I looked at the actual financials, you know, but I thought I remember reading that the guaranty obligation was already in the Level 3 category? Did I misread that?
for instance: it's one thing if they got these quotes that were all over the place
That is EXACTLY what they said, in the conference call and in the financials. The dealer quotes had such wide variation that they felt using the average, which they do, was kinda shaky. So they moved the stuff to Level 3.
You are forgetting that if you can get dealer quotes on something, it usually is fine to call it Level 2. Freddie used to call it Level 2. Then they decided that the dealer quotes were too squirrelly.
I'm not trying to argue that the stuff is "properly" valued; I haven't the faintest damned idea what it's worth. I don't think anyone else does, either. Which is why I'm more suspicious about everyone's Level 2 assets than I am about their Level 3 assets! I think the point here is that the credit markets are so bad that everything is either 1 or 3; there is no 2. And FRE has almost no Level 1 assets, so that means what it means.
Are assets moved to level 3, so that actual losses do not have to be immediately REALIZED since THE HOLDING ENTITY can not accurately state that level 3 assets are not over valued or undervalued by a specific dollar amount at a particular point in time.
Isn't this an easy way to defer losses today in hopes that they might appreciate in value tomorrow. After all most companies can only declare loses up to a certain dollar amount before they become publically insolvent.
Are assets moved to level 3, so that actual losses do not have to be immediately REALIZED since THE HOLDING ENTITY can not accurately state that level 3 assets are not over valued or undervalued by a specific dollar amount at a particular point in time.
No.
Man. Now I have to respond to fervent criticism on two blogs!
Ron:
The post specifically calls out dealer quotes as being systematically too high. Everyone in the bond business knows this. FYI. But as another comment says, what's the alternative?
I think I'm with Tanta here. The hysteria over Level 3 per se misses the point. And for what its worth, my understanding is that any property on which they have foreclosed is part of their L3 assets. That's where most of their losses are coming.
The losses have to be 16x their current expectations. I merely comment that seems unlikely not impossible.
Man. Now I have to respond to fervent criticism on two blogs!
And we certainly appreciate you doing so!
HERE's the bottom line: I don't care what kind of GAAP bin the assets are classified in. My point is that the balance sheets of ALL U.S. banking/financial firms are fraudulently mis-marked - by a SUBSTANTIAL amount.
Here's another real life example which bankers don't understand. BAC recently took a much bally-hooed $6 billion "provision" for loses. They have $603 billion of risky assets which include: $165mm Trading Acct (could be a lot of Treasuries but also would include "kitchen sink" crap traders are hiding), $223 billion debt securities which would include all kinds of asset-backed crap "held for investment" because "we couldn't stuff any more into dumb-shit clients' portfolios," $50 billion derivitives - hmmmmm, and $165 billion "other," - hmmmm cubed.
So essentially BAC took a provision of 1% against those assets. That would be a penny. So if Pimpco came into BAC and said "make me a market on this SIV-backed money market security, you have me marked at par," BAC would say, "oh based on our loss provision I can pay you 99." RIIIIIGHT! ROFLMAO.
Even with the write-downs BAC has taken to date, to think that any part of the $600 billion in "assets" is market even within a country mile of where it should be marked is naive, gullible, and outright stupid.
This is total and complete fraud being perpetrated on the investing public and the global financial system.
And the writers, commentators and editors of blogs and all other forms of media are either being complicit or just don't understand the depth of the issues.
I know of recent bids emerging from vulture funds bidding a penny to a nickel for HELOC securities. Does anyone in their right mind believe BAC has its HELOC inventory marked even remotely near that level? Even remotely near 50 cents on the dollar?...
"I'm not trying to argue that the stuff is "properly" valued; I haven't the faintest damned idea what it's worth. I don't think anyone else does, either."
In this situation the holder of the assets gets to make the decision whether it sits at whatever level and price. This system seems to create problems discussed here and makes evaulating from an investor point of view almost imposssible other then to sit on the sideline or look for investments that offer better transparency. It would seem that a range of bids and from whom should be part of the financial disclosure.
Tanta,
The Guarantee Obligation is far bigger than the Level III assets, which are mostly subprime ABS.
Here's a quote from their 1q restuls:
"The company estimates that its total credit guarantee portfolio increased at an annualized rate of about 10 percent to approximately $1.8 trillion year-to-date through April 30, 2008'"
I guess I don't understand how AI would not include the G.O. when estimating losses, as it is far bigger than the Retained Portfolio.
What am I missing? Isn't Freddie on the hook for G.O. losses?
A financial security is only worth what the next best buyer is willing to bid with REAL capital for size. Our whole system is riddled with banking assets that have no bid.
That's the bottom line.
I was using their credit loss number, which includes the GO portfolio. Well, in as much as I understand anything about FRE's accounting, that is.
I couldn't find the quote quickly which mentions their credit loss estimation. I know it was highlighted on several street research reports.
I know its completely presumptuous on my part but accrued interest would make a terrific (guest) commentator on this blog.
For example:
corporate bonds
And there has been no bid on most of these assets since August. So at what point do you admit that the models are wrong and that clearing prices are correct? Twelve months? Otherwise we are Japanese.
I said:
"for instance: it's one thing if they got these quotes that were all over the place"
Tanta replied:
"That is EXACTLY what they said"
Yes. I know that's what they SAID. but this is why I'd be interested in seeing the bids. again look at my examples.
if the quotes were "all over the place" but all really LOW this means something different IMO than the quotes being all over the place low and high.
I guess my question is this: using Level 3 accounting, how would one account for an asset that only gets bids extremely LOW.
so on $157b assets, what if it only gets:
$80b, $75b, $60b, $0, $100b bids.
do you get to mark that to $157b then?
because maybe, just maybe, those assets ARE worth in the $60-100b range.
it's just like the house across the street from me that is now in foreclosure. The owner was asking $400k for it. the bids he got were in the $250-325k range. He clearly didn't like the "irrational bids that were all over the place". Thus, foreclosure. but based on MY model (what houses have actually sold for around here in that condition) the place was worth around $320k. so he marked to model based on his wishing price.
are the financial firms doing this too?
FWIW:
I agree with you that these assets should have been placed in Level 3 over Level 2.
but it doesn't make me happy that we're valuing that company based on those assets at all.
I know there's no other choice, or so everyone keeps telling me. I guess what this means is that the American Financial system is severly broken, and has no chance to be fixed.
The company estimates that its total credit guarantee portfolio increased at an annualized rate of about 10 percent to approximately $1.8 trillion year-to-date through April 30, 2008'
The "credit guarantee portfolio" is not the same thing as the "guarantee obligation."
$1.8 trillion is the maximum possible exposure to credit losses they could have if every single loan in every single guaranteed security went belly-up and there were no collateral recoveries or shared recourse/indemnifications etc. of the seller/servicer. There will of course never be a time at which 100% of loans in 100% of securities experience 100% losses.
The "guarantee obligation" on the books is the net expected realized loss, not the maximum possible exposure. (The "guarantee asset" reflects the g-fee income that offsets that obligation.)
I don't see the problem with just using expected credit losses, which is more or less what the guaranty obligation is. As AI says, even if you want to assume catastrophic losses, you don't assume 100% on the guaranteed portfolio.
So at what point do you admit that the models are wrong
Well, some of us aren't Japanese, not that there's anything wrong with that. But to use an example from a previous spread, if I had a Picasso that no one bid on, I wouldn't declare it's value to be zero. The point that I would re=value at is that point in time at which I'm forced to recognize some value for my assets.
My personal issue with with level 3 assets is over the last year or so it's been the general rule that assets stuck in level 3 by financial institutions have proved to be worth much less than the estimated value. I believe Freddie Mac should take that into account when valueing the assets.
They should do something like take what they think the assets are really worth and multiply that amount by say 0.8, to get a conservative real value.
I'm very suspicious that Fannie Mae reported a much larger portfolio loss than Freddie Mac when there's no discernable difference between the two other than how they do their accounting.
Tanta,
AI implies that the losses are derived from assets, not guarantees, by stating that the asset value has to go down by 4% for Freddie senior debt to get hit. He then says this "factor of 16" miss on loss estimates is unlikely. This is just plain wrong. The total liability is $2.6tr, so a 1% loss would result in the equity being wiped out. That's 4x Freddie's 25bp estimate, and certainly in the realm of possibility.
Then, again, there's the issue of how Freddie's "black box" hedging will perform if rate volatility jumps. I think no one (including the company) really understands the risk, but that doesn't mean it should be ignored.
Level 3 assets don't mean you get to ignore data to arrive at the answer you want, it just means that you are using unobservable valuation inputs that respresent a market participant's assumptions "developed based on the best information available in the circumstances (FAS157 A25)."
So if there are dealer quotes out there that show that the assets are worth ~80B instead of ~157B, I think the burden of proof would fall to FRE to demonstrate why their information is in fact the best information available.
And I think Tanta and AI are 100% correct: I'd MUCH rather have someone disclose an asset (priced or mispriced) as a Level 3 if they have unobservable inputs. To AI's point, you can discount however you'd like. If you infer that all non-L3 assets are priced perfectly because of the disclosure, that's your choice.
Both these companies are clearly going to be insolvent by the end of the year, but everyone knows that
Congress will do anything to keep them afloat, because if Fannie and Freddie go under, the entire global
financial system will melt down, said Christopher Whalen, a founder of Institutional Risk Analytics, an
independent research firm. These companies earnings dont matter. Their accounting hardly matters.
People buy the stock because they believe the federal government will bail them both out if things get really
bad.
http://www.institutionalriskanalytics.com/pdf/NYT_051508.pdf
Institutional Risk Analytics
Great thread and comments. Thanks to all!
Harsh Reality:
That is a ridiculous statement that's been making the rounds lately. Who the hell would buy a stock assuming there will be a govt. bailout? Bear Stearns was taken out at 1/8th of their price on 2/29. Who sees that and concludes a bailout is good for shareholders???
Accrued Interest - thanks for a great effort to explain some important points over on your blog post.
I walked into a shitstorm of ignorance on a related thread here at CR a week ago following Freddie's conf. call.
It started by my trying to make the simple point that 'level' designation is not a measurement of value. I was immediately shouted down by a mob of morons, natch.
Level 3 assets are going to be valued at estimated discounted future cash flow. Placing these securities as L3 only indicates that they can't be reliably or readily marketed. It's a form of confession. The risk lies with the probability inputs to the cash flow analysis.
Ron, keep in mind that every publicly traded manufacturing and retail business in the US has substantially overvalued fixed assets on their balance sheet when comparing the reported values to market values (what do you think all those store shelves at the mall are worth at auction? They are valued at cost - straight-line depreciation).
I was immediately shouted down by a mob of morons, natch
If you weren't such a shill for Countrywide I'd feel sorry for you.
Quick question:
Level 2 inputs allow for situations where "price quotations vary substantially either over time or among market makers" (FAS157 28b).
If FRE's case isn't an example where divergent dealer quotes can be allowed as a Level 2 input, can someone provide a reasonable example where divergent quotes are a Level 2?
If not, whats the purpose of FAS157 28bs allowance of divergent price quotations? Or was this a case where the divergence of Level 2 inputs led FRE to use Level 3 inputs and, because the Level 3 inputs had the more material weighting, the entire asset was deemed Level 3 (per FAS 157 guidance that an asset gets sent to the lowest bucket that has significant inputs).
Yes I am a long retired fund manager, yes I live on the other side of the Atlantic, yes I first read the Financial Times in the Fifties. All these comments just show how far down the slippery slope we have travelled.
There was a time when there was no such distinction such as level 2 and level 3 assets. All assets had to be valued at lower of cost or market value and market value meant what they could be really sold for on an arms length basis: no slippery "explanations" would get past the auditors.
The present situation is a fiasco. I am with RonBurgandy.
Ancient... Yeah way back in like... 2006. FAS 157 is brand spanking new. Anyway, FAS 157 doesn't change anything about what kinds of assets are marked-to-market, merely increases the disclosure as to how they were valued. That seems to be the big misunderstanding. That FAS 157 is allowing companies to hide stuff, when in fact, they are just as able to hide stuff as before.
Begbie: My impression is that FRE made the call to move them to Level 3 because of the degree of divergence. I mean, if you get quotes of $50, $70, and $30, you should conclude that the valuation isn't based on anything readily observable.
AI: I agree with you, but it does call in to question what the adverb "substantially" means in the context of FAS157 28b. Given the FRE context, I'm assuming "substantially" means "not enough that you conclude nobody knows what they're doing".
Blech....feels like another bright line in the works to me.
"no slippery "explanations" would get past the auditors."
Ancient Brit | 05.20.08 - 3:47 pm
Heh, heh.
The technicalities are way over my head, but I'm impressed with the way Tanta and Accrued Interest are colleagues rather than rivals. Good work, everyone!