Tuesday, November 27, 2007

AMBAC: Episode II

My Thanksgiving special on AMBAC got quite a bit of attention, and I have unfortunately been swamped and unable to give too many good responses to the many quality comments.

I got several requests for some details about how I got to my loss assumptions. Here are my exact loss projections for all of AMBAC's CDO's.
CDO losses are best estimated using cash flow models. So I built a crude cash flow model, moderated certain assumptions based on each CDO type and vintage, and ran each assuming 15-25% in subprime defaults. I had to make an assumption about how much was in first lien vs. second lien paper, but given my knowledge of how ABS CDOs were constructed recently, I assumed there was more home equity paper than not.

For what its worth, UBS came up with almost the same figure for losses from AMBAC's CDO portfolio.

That is only one of several assumptions that could drastically change the results. In addition:

  • How well foreclosed loans recover. Non-agency MBS deals were very heavily weighted in California paper. I think its safe to say that subprime loans in the hottest markets may experience more negative home price appreciation and therefore worse recovery. But it could be that AMBAC was more cautious about loading up on the hottest markets. Hard to tell.
  • How effective mods are. Various programs by banks and the FHA could have a large impact on Senior CDO performance. Remember that very small changes in loss rates cause very large changes in Senior CDO performance. Its the structured squared effect.

Readers EJ and LastToKnow have both pointed out that the 2007 vintage CDO ^2 listed in AMBAC's disclosures actually have older ABS collateral. I haven't confirmed this myself, but if true, that too could diminish losses materially.

I also got a note saying something to the effect that I didn't understand the ratings agency criteria and that I should suppose that I understand their models better than they do. I never pretended that I understood the ratings agency models (which are proprietary). Furthermore, I do understand the fact that losses in insured ABS will occur over time, which I mentioned in the original post. However, if its known that a given ABS tranche is non-performing, the ratings agencies will consider this in giving their rating. Some seem to claim that just because AMBAC or MBIA's claims will be paid out over time, that there is nothing to worry about. That would imply that a given insurer could become like a ship adrift with no crew, still afloat but bound to eventually hit rocks and sink. Once an ABS pool has gone bust, the ratings agencies are going to count those losses against the insurer's capital.

Now onto today's news. AMBAC and others claimed that reinsurance could solve their problems at a Bank of America conference. I'm sure that they can raise substantial capital that way, but only by buying reinsurance on parts of their lucrative municipal portfolio. It also sounds like they plan on using run off and retained earnings to bolster their capital over time.

Only time will tell whether they will eventually have to go to the market with an equity offering, either preferred or common. If AMBAC is told they need a relatively small number in additional capital, reinsurance is a no-brainer. It may be that insurers get away with slowly increasing capital as losses mount. If so, that would be very bullish for the stock.

However both AMBAC and MBI stock were sharply lower on the day, I suspect due to concern that re-insuring the safest part of their portfolio isn't ideal. Of course, doing a dilutive equity offering isn't such a great option either, so I don't know what shareholders were expecting. Perhaps Citi and Freddie Mac's efforts to raise new capital are causing AMBAC and MBIA shareholders to realize just how expensive keeping their AAA rating is going to be.

As far as insured municipal bonds go, I think things are looking better. The companies seem to have a plan for retaining their rating. We'll see how things unfold, but I think it looks promising.

Those that are short monoline credits are going to be disappointed, I think. They will most likely be able to raise capital and keep their rating, and eventually their bond spreads will tighten.

Stock holders are looking at a bumpier road. The stocks are trading at massive discounts to book value. So if buying reinsurance solves the capital problem, it would seem likely that the stock price would move toward book value. However, raising capital will be expensive, no matter what path is chosen. And there remains the possibility that they need to raise equity in a dilutive fashion.

We must be cautious.

Disclosure: No positions in any bond insurer, although I own many insured municipal bonds.


kernelbleeper said...

Wall St. is of the notion that Fitch, Moody's et al will require a lot of additional capital to preserve the AAA ratings. The co's believe reinsurance would do the trick. Would Fitch & Friends go for the reinsurance angle?

kernelbleeper said...

Quickly following up -

As far as reinsurance, what I'm afraid is the companies placing all the chips down on that bet and it not working. In my opinion, they should be enacting the "depression" scenario and securing capital to satisfy that particular drama.

AMBAC's website shows an investor conference call with FBR Capital Mkts tmw 11-28 at 10am.

Unknown said...

Here's the slides from Ambac's conference on Nov. 27.


It's far from obvious, looking at page 9, that they will hit large realized losses on their portfolio. Though its not clear how the collateral quality underlying the CDO squared deals compares with their other collateral.

The "ship adrift" is not the right analogy for why the time course of losses matter. Besides the delay in when capital is needed, it's important point ithat they are gaining cash flow all the time from the other 80% of their portfolio that is not RMBS. That other cash flow is explicitly given as part of what the rating agencies take into account when they compute capital adequacy.

Anonymous said...

I really enjoy your work. As a shareholder of AGO but someone who does not know a fraction of what you do, I would like to bring up some points as to why I believe AGO will survive intact and undiluted and possibly gain share.

1) If you look at the chart of the Subprime RMBS securities they own as of 9/30, you will see that although they insured $5.7 billion during 2006 and 2007, only $2.5 billion of the securities were actually issued in 06 and 07. The rest appeare to be 05 vintages wrapped post issuance. $600 Million of the 06 vintages were wrapped in Q3 of 07 under presumably tighter underwriting conditions.

2) Average current subordination on the 06 vintages with face values above $100 million ($1.5 billion total outstanding) is 42.7% with the low being 35.8%.

Average subordination for the lower principle amounts ($25MM face 06 and a $44MM 07 vintage) is $29.26% with a low of 20.8%.

Their 2005 vintages have average subordination of 49.69% with a low of 31.4%.

3) While I don't expect this trend to continue, it seems that subordination increased (sometimes very slightly and sometimes significantly) on every single subprime RMBS from June to September 2007.

Given this information, I am cautiously optimistic that AGO pulls through without additional capital needed. Let me know if/where I am going wrong on this.

Thanks in advance for taking the time to read this.

Anonymous said...

BTW I am of the opinion that MBI and possibly ABK will offer far better long term appreciation than AGO given their stock prices today. But because I am incapable of valuing the CDOs and the ^2s, I am afraid to jump in. I will leave exceptional results (upside or downside) to those much more intelligent than I am.

Accrued Interest said...

AGO has happened into the best market for a new municipal insurer in the last 15 years. Until very recently, anyone trying to break into the muni insurance market was at a serious competitive disadvantage over the most established players. The Big 4 had all the reputation, AGO, XLCA and CIFG didn't so they had to underprice the Big 4. While XLCA and CIFG got involved in a lot of subprime stuff, AGO looks safer than anyone other than FSA. And AGO is getting all kinds of free press from all these research reports, conference calls, etc. on monoline insurers.

Now, the valuation question is a fair one. It might be that MBI and ABK are so f'ing cheap that performance in those stocks winds up being much better. Who knows. But AGO has good timing.

Accrued Interest said...

BTW I put up a new poll on which insurer might go down. I don't know about anyone else, but I like to see what my readers think.

kernelbleeper said...

What do you think of Ackman's comments? He is stating that all the insurers will "blow up", and making public his short position.

The headlines all over ABK Yahoo news.

kernelbleeper said...

NEW YORK, Nov 28 (Reuters) - Bond insurer MBIA Inc (MBI.N: Quote, Profile, Research) could be insolvent as soon as the second quarter of 2008 if it were unable to access additional capital, according to a slide at a presentation by Pershing Square's William Ackman.

Ackman estimates MBIA will incur $2.2 billion of losses in the fourth quarter, and rival bond insurer Ambac Financial Group Inc (ABK.N: Quote, Profile, Research) will incur $4.2 billion of losses.

Accrued Interest said...

Well, he's got a repuation for sensationalism. His loss estimate is similar to mine, interestingly, but I feel more confident they'll be able to raise capital. He might profit anyway if the capital raised dilutes shareholders.

kernelbleeper said...

Curious interview with Ackman on CNBC. I will try to state the points as close as I can from memory.

1) He is short MBIA & AMBAC

2) He is donating profits from his short positions to charity (he mentioned 500 million but did not say whether they were bond insurer specific).

3) His short position targets the holding companies not the insurance end.

4) He would like to see the bond insurers survive.

5) He said his "blow up" comment was taken out of context.

6) He wants to see more transparency

Anonymous said...

Accrued Interest,

The CEO of Ambac said the following today Nov. 28:

The CDO-squared portfolio has 48% fixed rate loans.

The average FICO score of the CDO-squared portfolio is 644;

The vintage distribution of the CDO-squared portfolio is:
2007: 2%
2006: 33%
2005: 48%
2004 and earlier: 17%

He also said all of their HELOC, closed second mortgage and Alt-A is prime.

He is not worried at all about the Alt-A portion.

The portion he is most worried about (and where they have booked some reserves) is HELOC.

He is very upset about the UBS loss estimate.

Knowing this, do you still maintain your loss estimes?

Accrued Interest said...


Throwing down the gauntlet eh? The 33% in 2006 vintage will still wind up with big losses, but its possible I'm too harsh. I'm going to rework my analysis.

Let's say you cut the losses in half. They still probably need to raise capital, but perhaps only about $1 billion, and probably not all in one year. That would be a much easier hurdle to leap eh?

Analysis is ongoing. Appreciate the back and forth. Its what makes up all better.

Anonymous said...

I'm trying to think with a clear mind about Ambac: is the stock a buy or a sell?

1. Like many other highly leveraged financial institutions, Ambac assumes tail risk that can cause it to blow up, if a sort of doomsday scenario of deep economic depression occurs.

2. Is such a depression likely (with the FED and the treasury department standing by idly)? No.
Is is possible? Yes.

3. They have $6B in capital and another $6B scheduled to come in the door from current contracts.
That means if they lose AAA status, close up shop and run off their business, they can absorb losses in the order of $10B before their liquidation value falls below their current ~2B$ stock value.

4. Current loss estimates from the mortgage crisis range between very little (Ambac itself) and $4B (such as by Accrued Interest and some others). At the current stock price that leaves a margin of safety of another $6B in losses they can absorb.

5. On the upside, if the mortgage crisis works out okay for Ambac, they retain AAA status, and they go back to earning 15% ROE on their $6B in capital, this makes for a fantastic investment.

6. Weighing these 5 points I rate this a buy that I give 1% of my capital.

cak said...

Quoted in Fortune, eh? Who are you, masked man?

This is the slide show from today's FBR conference. Most of the pertinent material is at the end. Pgs 22 & 23 are interesting (or amusing).


Accrued Interest said...

The problem is that if I blogged with my real name then my writing would implicity represent the views of my firm, in which I'm a minority partner. But our firm likes to present with a very measured tone, and is loathe to talk about short-term events. Plus then every one else would want a say in what I write. It would make the blog very very very boring. So I blog on my own time and anonymously.

Accrued Interest said...

I'll give this to AMBAC. They are going out of their way to tell us everything. Got to be a good sign, right?

Anonymous said...


They are giving us everything but the names of Mezz SF CDOs where the risk is highest, that doesn't cut it for me.

What are your loss assumptions on the CRE collateral? I think this is an area of risk that is completely overlooked by the street right now. The underwriting behavior for these loans looks little different from that employed in RMBS. The CMBX is clearly indicating a level of risk that is disturbing (and yes I know the caveats about everyone taking one side in an index trade, but a spread as big as we are seeing there says something material is brewing - this is way beyond techicals).

Anonymous said...

AI, don't know if you saw Mish's piece on AMBAC/MBIA solvency.


Anonymous said...

Ambac said they are willing to provide the individual CUSIPs underlying their deals (except CDO-squared).
I don't see it on their website, so I suppose you have to ask them for it?
Somebody with a big database of info about these CUSIPs could run very detailed loss scenarios...

cak said...

Great blog.

New to the site. Working in reverse, has anyone had a chance to crunch Pershing's paperwork from May '07?
AMBAC showed slides today of "us vs. the shorts", the short I'm presuming to be Ackman. As he has annointed himself Caligula, er..king of the shorts, it would be interesting to see how Pershing's analysis stacks up.

Accrued Interest said...

On Mish:

I think he's being simplistic. The $29 billion JP Morgan mentioned isn't just at AMBAC and MBIA, but in the whole industry. Plus the insuers don't have to pay this out all at once, so they won't be suddenly bankrupt.

Its funny because I feel like I'm caught in the middle of this debate. I've been quoted all over the place as someone predicting a very high level of losses. And yet I'm reasonably sure AMBAC and others will be able to raise adequate capital. So while the losses will be very painful, the AAA will remain.

It seems like I'm either not bearish enough or bullish enough for anyone's tastes. But I really think that's how it plays out.

As far as CRE, I really think these losses will be manageable, even in a pretty bad recession scenario. I think they have enough subordination to be safe. I have the same view of their credit card and other ABS portfolios.

Anonymous said...

Thanks for the great blog.

Band Uncle said...

Nice work AI. You have tackled the job of trying to quantify possible losses quite admirably. I think there are a few other factors to consider. At the extreme, if the govt manages a refi based mortgage bailout this could be a windfall to the ABK and other senior tranche mortgage investors. Second, with the 10yr at these levels, existing resets will not be as onerous as many have predicted. Finally, I am beginning to look at the monolines as a possible real estate play. If the sub tranches fail and the insurers become obligated to pay claims, their subrogation rights will give them ownership/control of the underlying loans and possibly the collateral as well. I would have to believe that there wold be serious investor interest in US housing at prices represented by ABK's exposure. For these reasons, as well as the importance and profitability of this business, I believe ABK will attract capital without destructive dilution. As you have pointed out, if ABK does incur exposure it will not be instantaneous, e.g., like a run on a bank, but will occur over a lenghthy period of time as the mortgage defaults work their way through the system. This should give them the flexibility to work out their problems. These people really know this business. I have found them to be better "bankers" in terms of credit and risk analysis--especially stress testing--than the IBs and rating agencies. They really do underwrite with the intention of not incurring any measurable risk. Resolution is not likely to manifest overnight, e.g., a white knight capital infusion or acquisition, but instead will take the time to fully realize the extent of the housing crisis. Finally, basic economics should preserve the monolines--the cost of recapitalization appears to be much less than the cost of repricing insured debt and developing a new method for selling bonds for unknown muni issuers.

Eyal Bar said...

Some of the comments here are very surprising to me. AMBAC built a company which now likely requires an UNANTICIPATED capital injection. So obviously these guys have an anticipation problem, i.e. their predictions are not the best. Therefore, how can we believe when, for example, they say that ALT-A is nothing to worry about?

Bad Uncle,

Many smart people believe the 10yr is where it is very temporarily. Also, your comment about insurers as a real estate play lacks sense in my view. If anything is to be understood from this crisis, it's that excess leverage led to hard times. You think reasonable people will want to play real estate via the super-duper leverage of owning equity or equity-linked investments in credit insurers?? You can't even get a 95 LTV loan out there!!! And you shouldn't be able to, especially not for 2nd/vacation/investment home, which is basically what you'd be doing by investing in credit insurers.
I think it's true that MIs have RELATIVELY an economic incentive to do things that add value and scream of prudence and logic, as opposed to the many volume-driven actors in this theatric play that we are watching. But even they have stock options etc'.

Monolines as a real estate play? No, I'd rather buy the building itself.

Band Uncle said...

Eyal Bar,

The essential underlying credit support of a RMBS is the the collection of mortgages. The MIs interests are the most senior of the mortgage tranches and therefore will have preferential access to and control of the collateral. If things do go badly and the RMBS' default, they will effectively control the properties at a cost approximating 80% or less of the original loan (less interest payments). This could be a situation much like that when the RTC sold off loans in bundles, good, bad and indifferent, at pennies on the dollar. I know these securities are fairly complex but there is a lot of notional here, and I sense that a lot of this notional is being discounted pretty heavily, note ETrade's sale at 27 cents on the dollar. I look at the MI assets (true AAA investments, unique highly profitable franchise, and subrogation rights on the bad policies) and I see some excellent opportunities. Don't forget that Sears and McDonalds have been looked at as RE plays.

Anonymous said...


Love to hear your opinion on how the second lien/HELOC business will impact the mgt insurers. Seems like this could be a real problem. Afterall, in MBIA's case the amount of HELOC and 2nd lien is 5x the amount of sub-prime.

In a second lien mgt, the lender has the 0-20% loss (although others are possible 5-20%, 10-25%, etc...). Let us assume that the borrower was a prime borrower, but that between the slightly inflated assesed value and the HPD (BTW I hate the term negative HPA) and forecloser costs, let us say that the value is 80% of original assesed value. Should the borrower default, that seems like a 100% severity.

Since it seems at least in MBIA's case that they insured only at the AAA or Super-AAA, which would have say 20-40% subordination, is it possible to hit these levels given that the severity will be very high ~100%? How many "liar loans" in a particular insured issue do you need? While on average I think that these loans were a small percentage of overall loans made(12% for 2006), some cdo out there is going to have less and some is going to have more.

Seems like this is not that different from mortgage insurance (PMI, RDN,...). They have taken huge writedowns, why not the monolines? Is it simply just the old arguement that MTM does not count for us?

Net of all this is a challenge to you AI - you did the #'s on the CDO's - at least for AMBAC. How about a shot at the numbers on the much larger 2nd lien and HELOC's.

Accrued Interest said...

Well I thought about $2.5 billion for the RMBS position. And most of that came from HELOCs, because as you point out, HELOCs probably have pretty high severity.

Remember that a lot of AMBAC's $50+ billion RMBS exposure is pre 2003 vintage, and its mostly paid down. You'd figure that by the time 90% of a HELOC deal is paid off, that the remaining 10% isn't likely to default. And if it is, the borrower pool was pretty strong to begin with.

Now what we don't know, and what AMBAC is trying to convince us, is that they did better credit work than average. Meaning that they insured higher quality MBS deals compared with others. Now, I ignored this possibility when doing my analysis. But it could be a big deal. About 50% of subprime purchase loans were stated income in 2006. But AMBAC might have avoided pools with higher SI. Maybe their pools are only 5% SI? 10%? Suddenly their portfolio looks a lot better.

kernelbleeper said...

Back to that 2007 vintage, as stated in Moody's AAA affirmation last night.

Ambac – Affirm Aaa Rating with Stable Outlook.
In Moody’s opinion, Ambac’s current capitalization is adequate for its rating level in both the base case and stress case described above. Moody’s does note that there exists meaningful uncertainty with regard to the ultimate performance of the firm’s insured mortgage-related portfolio in the current environment, particularly with respect to the three insured 2007 vintage CDO-squared transactions, which exhibit significant modeled losses in our stress scenario. Nonetheless, the affirmation of Ambac’s Aaa insurance financial strength rating with a stable outlook reflects Moody’s view that the company’s capital position is adequate to deal with such uncertainty, in the context of likely further capital strengthening measures and considering the company’s robust franchise in the financial guaranty insurance sector. Ambac’s announcement that it has entered into a commitment to cede $29 billion in par exposure to Assured Guaranty Re Ltd. (insurance financial strength at Aa2) is considered a positive in that regard.

Anonymous said...

is it possible that monolines are "allowed" to go bankrupt without being rescued at all?