Description

Ambac’s primary Countrywide/BofA trial is scheduled to start on September 7, so it’s (probably) now or never on assessing Ambac’s current situation and taking a new investing position.  I recommend it, for those whose backgrounds incline them to do so. Including prejudgment interest, Ambac has roughly $5 billion in claims against BofA and another $1 billion against Nomura, $6 billion total.  Its recorded book value for those litigation assets is $1.5 billion.  If Ambac could recover half its true litigation asset value in a verdict or settlement, that’s an excess $1.5 billion above book value, or $30 per share on a $15 stock. That highly simplistic SWAG is within my new bottoms-up range of base-case excess value, $12-33 per share.

I wrote up Ambac in late 2019 because the payoff date in its primary BofA lawsuit finally looked imminent, with trial scheduled for mid-2020, and the rest of Ambac’s situation had simplified to create a fairly clean special situation around that one lawsuit.  Covid destroyed the set-up by delaying the trial date for more than two years.  Ambac has had several significant developments during the extended wait, some good, some bad, with the bad worth modestly more and the net intrinsic change fully reflected in the 20% stock price drop since then.  Here’s my take on the developments:

BAD YET ACTUALLY SNEAKY-GOOD - DLJ decision & Ambac book-value write-down

In mid-March 2022, New York’s highest court issued a decision in U.S. Bank v. DLJ Mortgage Capital.  Contrary to earlier lower court decisions, the high court ruled that plaintiffs suing under the “notice” provision of standard RMBS contracts (similar to some of Ambac’s) can recover damages only for individual loans on which they gave the seller timely, loan-level notice of defects.  They cannot rely, when using the contract’s notice mechanism, on “general notice” that the entire loan trust pool was littered with defects.  Ambac’s stock fell 23% on the day this news hit, because Ambac gave specific timely notice on about 7,000 loans out of the 375,000 loans at issue in its primary lawsuit.  Ambac announced that, although it thought the DLJ decision would not preclude its recovery on any loans, it would be materially marking down its litigation asset book value to account for the risk.  Its 1Q financials lowered the value from $1,730m to $1,502m.

The stock price fully recovered in 2Q, which was the correct response to 2Q’s news on the DLJ issue.  Ambac made several persuasive and mutually exclusive arguments to the court as to why the DLJ ruling does not affect its claims – so persuasive that Ambac’s odds of winning those arguments looked extremely high, in a vacuum:

1. Unlike U.S. Bank (a trustee), Ambac has a separate recovery avenue under its insurance agreement, called “reimbursement”: it entitles Ambac to reimbursement for “any payment made under the Policy arising as a result of Countrywide’s failure to substitute for or deposit an amount in respect of any defective Mortgage Loan.”  No notice is explicitly required.
2. The DLJ opinion states that it applies only to the “notice” avenue of recovery and not the  “discovery” avenue; a plaintiff can recover if the bank itself discovered material defects in loans.  Importantly, the bank is deemed to have discovered defects if it either literally knew of them or if it “should have known” of them, through willful ignorance or, more to the point, knowing that the entire loan trust was a toxic pool of terrible underwriting that it itself had performed and that was blowing up before its eyes. This avenue is particularly powerful against Countrywide in Ambac’s primary case because (a) it’s Countrywide, we all know what Countrywide did; (b) Countrywide originated 85% of the loans in these pools; and (c) Countrywide served as originator, trust structurer, sponsor, and loan servicer, so it was intimately involved in these toxic loans throughout their lifetimes.
3. Countrywide refused to provide, when asked, the individual file data that would have allowed Ambac to find specific defects (a condition precedent to Ambac providing notice).  If your counterparty knowingly hinders a condition precedent to notice, it waives the notice requirement.
4. Countrywide refused to repurchase all but a few loans that Ambac did notice, as required under the contracts; notice isn’t required when providing it would have been “futile.”

I say that the odds of winning these arguments seemed high “in a vacuum” because judges reject seemingly-strong arguments all the time.  But we don’t have to wonder about the judge here; he held a status conference on these issues back in May.  He sounded highly predisposed to accept Ambac’s position, ruled that Ambac could submit all of its “defect discovery” evidence (which is going to make Countrywide look very bad), and confirmed that Ambac can prove damages via loan sampling (which the DLJ decision had placed in modest doubt). 

Management has not re-marked its RMBS value higher despite this new information.  This decision was prudent.  The first event was a definitive and possibly-applicable high court decision, while the second was non-definitive comments and interim rulings at a trial court status hearing.  We mostly know what this judge thinks about these issues, but we don’t know what a theoretical future appellate panel will think.  The result of this asymmetry, though, is that Ambac’s RMBS value may be under-marked by as much as $200m, relative to where it had been marked pre-DLJ.

THE ORDINARY-GOOD

The Harborview fraud case survived summary judgment and is scheduled for trial

In addition to the primary Countrywide/BofA case, Ambac has another lawsuit (referred to as Harborview) against Countrywide/BofA for a separate set of dodgy loans. Countrywide was not as involved in those loans as in the primary case and is not a party to Ambac’s insurance contract for those loans.  Therefore Ambac has sued only for fraud.  BofA had moved for summary judgment on all claims.  On August 29 (this past Monday), the trial court ruled in Ambac’s favor and set a trial date for January 2024, “or earlier if the court’s calendar permits.”  Ambac is claiming $400m in damages.  Unlike the old fraud claims in the primary lawsuit, these claims are not duplicative of any contract claims, so they are fully excluded from Ambac’s stated RMBS asset book value.  I expect that Ambac and BofA would include a Harborview settlement as part of any settlement of their primary case.

One of my colleagues happens to know the Harborview judge fairly well and says he’s a great judge for Ambac to have in that case. The judge will be predisposed to “think like a juror” and favor Ambac, given Countrywide’s poor behavior throughout the RMBS debacle and likely poor behavior with these specific loans.

Ambac also has the First Franklin case against BofA involving yet more loans. I know almost nothing about it, other than that it is "progressing towards summary judgment." I assume it would also get wrapped into a global settlement between Ambac and BofA.

The prejudgment interest meter keeps running

My original prejudgment interest calculations assumed a mid-2020 resolution.  Two extra years of prejudgment interest at ~9% interest on ~$3 billion in claims is a lot of potential extra value.  Most of us have learned since my first write-up that, in the contract cases, Ambac receives a contractually-set prejudgment interest rate of prime plus a spread compounded, rather than the New York statutory rate of 9% simple non-compounded.  Because the contractual rate has worked out to “around” 9% non-compounded, I have continued to use 9% non-compounded in my spreadsheet; being more precise than that wouldn’t materially change the calculations.  The actual number could be 8% or 7%; as with so many details in this case, the other uncertainties skew the potential outcomes so much that it’s not worth nailing down this one.

Initial success in developing new business lines

Ambac has a boatload of NOLs, which reside partly at the holdco level.  Partially to enable use of these NOLs and partially to have a platform to redeploy expected RMBS proceeds, Ambac holdco has been growing three new business lines via organic investment and smaller acquisitions: specialty property & casualty insurance, insurance distribution (managing general agencies and managing general underwriters), and a grab-bag of corporate-level minority investments in “insurtech and insurance services.”  I have always assumed these expansions would be NPV-neutral and have pretty much ignored them, because the new business prospects are a very distant second to the litigation asset in valuing the equity.  That said, the 2Q earnings release and presentation suggest these efforts are gaining traction and look promising.

The remaining insurance liabilities keep falling

Puerto Rico is now entirely resolved and its tail risk eliminated.  Based on that resolution, in 1Q when Ambac reduced its RMBS book value by ~$200m, it also decreased its Puerto Rico net reserve values by ~$200m, for a net book value change of roughly zero.  Ambac’s total legacy insurance exposure, and in particular its adversely-classified exposure, have continued to fall steadily:

THE BAD

Primary-lawsuit fraud claims were rejected on appeal

In January 2020 the appellate court in the primary BofA case affirmed the dismissal of Ambac’s fraud claims as duplicative of the contract claims.  When the decision came out, I was sanguine about this development, because I never ascribed much direct settlement value to those fraud claims.  Their compensatory damages would have been largely duplicative of the contract claims, so they would add only punitive damages, which a jury might choose not to award at all.  Over time I have viewed this outcome more negatively, for two indirect reasons. First, although the fraud claims didn’t change the base-case expected damages much, dismissing them eliminates BofA’s tail risk of a giant punitive damages award.  It therefore makes BofA more willing to roll the dice with a trial outcome and therefore reduces Ambac’s settlement value (i.e., increases the likely % discount from the base-case trial outcome).  Second, it converts the trial from a jury trial to a bench trial.  A jury would likely have been more inclined to punish Countrywide for its pervasively bad behavior, whatever the legal technicalities.  This dynamic also reduces the settlement value even though it doesn’t change the theoretical base-case damages.

The Puerto Rico liabilities grew and grew

Although Puerto Rico is now resolved and although the final adjustment was positive, up until that point Puerto Rico had been the gift that keeps on giving, in the darkly ironic sense.  Thanks to Covid and the resulting recession spurring the Puerto Rico authorities to be even more intransigent, Ambac spent two years steadily adding more reserves for Puerto Rico, which steadily sapped its adjusted book value.

BofA might successfully strand the legal damages at the Countrywide level

Here is what I wrote about this issue three years ago:

The tail-risk downside here is that the parties fail to reach settlement and BAC wins at trial. The primary source of that risk seems clear: BAC’s legal argument that, if Countrywide is found liable, those liabilities do not extend from the subsidiary Countrywide to its parent BAC.  Ambac could theoretically win a multi-billion judgment against an insolvent Countrywide that never pays. The strongest argument that BAC is likely to be found liable in addition to Countrywide – and, more tangibly, that BAC's defense doesn't destroy the settlement value – is that BAC had the same defense available against MBIA and Assured Guaranty. If BAC can escape Countrywide’s liability to Ambac, then it could have escaped liability to the other two as well. No one is better situated than BAC to assess this legal argument’s strength, and BAC necessarily decided it was weak enough to justify paying MBIA and Assured Guaranty $1.7 billion and $1.6 billion early in those cases’ lives.  BAC’s parent/subsidiary defense is also automatically baked into Ambac’s book-value estimate of its settlement proceeds and therefore into our estimate of those proceeds.

No new facts or arguments have emerged to suggest how likely it is for BofA to succeed on this argument, if it went through to judgment and appeal.  The pure passage of time, though, is its own signal.  As this case has dragged on for another three years without BofA choosing to settle, it seems modestly more likely that BofA believes it has a decent chance of winning this argument, and therefore will continue playing extra-hardball in settlement negotiations.

NEW LITIGATION VALUATION

Here’s how all these points roll together to revise my prior outcome-range calculations:

• Start with the new, lower RMBS asset book value
• Eliminate the possibility of punitive damages in the primary case
• Add a new line for the Harborview fraud claim.  Remember, fraud damages are given zero value in the RMBS book asset.  I assume a $100m damages award on the $400m claim. These damages would probably (but not definitely) receive prejudgment interest also.
• Increase the boost in “inherent lawyer conservatism” from 10% to 20%, to account for Ambac’s DLJ-ruling mark-down without a later mark-up.  Raising that percentage increases the $ boost from $179m in the initial write-up to $297m now.  I.e., I am effectively adding $100m out of a possible $200m for new added conservatism from the DLJ ruling.
• Add two more years of prejudgment interest.
• Increase the “reasonable range” of settlement discounts from 30-50% to 40-60%, based on the loss of the fraud claims / jury trial and my increased concern about the parent/subsidiary liability-stranding issue.
• Change the share count to subtract shares that Ambac has repurchased; add shares from assumed exercise of the warrants.  The new count is 45.7m currently-diluted + 4.9m warrants = 50.6m fully diluted.  Warrant exercise would add $82m of cash = $1.61 per share.

In summary: My reasonable range of base-case upside to stated book value per share is $12-33 per share.  It could possibly be higher, could be lower.  Note that this base case assumes a settlement outcome where the lawyers are specifically factoring in all of the tail-risk upsides and downsides from a trial outcome when deciding what to pay or accept.  I.e., it already embeds a discount for downside risks.  Book value is currently $17 per share, and the stock is trading at $15; I have no idea whether the current stock price embeds much of my expected excess value from the base case.  True tail risk downside in a disastrous outcome of the Ambac Assurance opco going to zero is limited by the holdco’s net assets outside of AAC, which are worth at least $5 and possibly $10.  (I’ll take cfavenger’s word for that top end.)

TIMING TO RESOLUTION IS HIGHLY UNCERTAIN

Ambac and BofA could settle tomorrow, or they could gut it out and settle a year from now.  Here is the litigation timing set-up:

• Trial commences September 7
• Trial is scheduled to last at least five weeks
• The judge would probably take somewhere between a month and several months to to issue his verdict and decision
• The losing party, or both parties assuming a partial victory, could appeal.  Briefing would take several months, then a month or two until oral argument, then an appellate decision 2-6 six months after argument.

A classic settlement point is just before trial, when the parties largely know what evidence will be submitted and can still avoid the expense and time commitment of the trial itself.  Another classic point is after the parties rest their cases but before a court decision, when the parties know how things played out and have some clues from the judge, but they can still “pay” to avoid some uncertainty.  However, I have seen (and been involved in) settlements that occurred after an appellate court partially affirmed and partially reversed a trial decision, when the appellate court said “go back and re-try these few issues.”

Catalyst

Trial finally spurs a settlement.  Maybe.