Description

***I was unable to paste the graphs/tables - please use the following link to access the writeup***

https://www.dropbox.com/s/mho58qsecajjf9f/AMBC_Long_Pitch_VIC.pdf

Recommendation: Long Ambac Financial Group Common Stock (AMBC)

Note: Ambac Financial Group has 45.0mm common shares outstanding at a current price of $20.80, giving the company a total market capitalization of $936.0 mm. Additionally, there are 5.05 mm warrants outstanding at a strike of $16.67, expiring on April 30, 2023. At the current price of the common stock, the diluted share count is 46mm (net of share repurchases with cash from warrant conversion) and the diluted market capitalization is $956.8mm. The common stock and warrants are listed on NASDAQ and began trading under the symbols “AMBC” and “AMBCW,” respectively, on May 1, 2013. Throughout the writeup, AFG refers to Ambac Financial Group (Holdco) and AAC refers to Ambac Assurance Corporation (Opco).

*We apologize in advance for the length of this writeup, but felt it necessary given the number of moving parts in the situation.

Thesis

We believe that Ambac is a highly attractive post-reorg equity with ~100% upside to fair value in the base case and an upside case in which the investment will be a home run. This upside is accompanied by a palatable downside case in line with current trading prices. In our view, Ambac has substantially over-reserved for losses on insured RMBS and will likely recover more from litigation claims than it has booked, implying that reported adjusted book value significantly understates fair value. The complexity surrounding these contingent claims and the stigma associated with a post-reorg, RMBS/municipal bond insurer has deterred investor interest. As the improved performance of the underlying RMBS collateral is recognized, and Ambac announces the proceeds from its litigation, the market will be forced to revalue the equity. With management’s compensation package on the verge of being finalized, we anticipate that improved transparency and positive commentary will help close the gap between market and intrinsic value.

Introduction

Investors don’t like uncertainty. If you read any of the sell-side research on Ambac or talk with market participants, you will hear the constant refrain that the Ambac situation is “filled with uncertainty” and “investors should wait until there is further clarity.” This is exactly why there is an opportunity to own an asset with asymmetric risk/reward and numerous defined catalysts for value realization.

The complexity of the Ambac situation is clearly discouraging investors from the name and has, in our view, led to a substantial misvaluation of the post-reorg equity. Investors cannot seem to grasp: 1. Structure of the rehabilitation plan, claims settlement, and Segregated Account; 2. Losses on insured RMBS; 3. Value of the litigation claims; 4. Value of the NOLs; 5. Exposure to Detroit and Puerto Rico; 6. The complicated financial statements (filled with a litany of confusing accounting). The natural response to this is to put minimal emphasis on the financials and embed a discount in the market value to compensate for uncertainty.

The lack of clarity coming from Ambac’s management team is further depressing investor sentiment. Two recent examples: 1. Ambac did not hold a second quarter earnings call; 2. The company waited until October 10th to post a deck detailing its Puerto Rico exposure, even though Puerto Rico bonds began selling-off in mid-June. The expectation is that management will finalize their compensation package by the end of 2013. With this package set, we believe that management will move to improve transparency and tout the company’s upside potential. We anticipate management will reverse RMBS reserves, give further clarity on litigation claims, and potentially comment on Ambac’s intention to restart its bond insurance business.

We believe that Ambac’s ultimate value will be based on adjusted book value, and we use this methodology as the basis for our analysis. To calculate adjusted book value, we start with Ambac’s reported adjusted book value, post Fresh Start accounting, and make further adjustments based on our estimates of certain contingent assets and liabilities. Essentially, Ambac is a big box of cash and financial assets, with the goal of keeping as many of those assets inside the box by the end of the run-off of the insurance book (the fewer the claims paid, the more assets left in the box). There is no rational reason why the market should not value this pool of assets at 1.0x fair value, with potential upside to adjusted book value coming in the form of a restarted insurance business.

Situation Overview

Prior to the financial crisis, Ambac wrote insurance for a wide variety of credits, including RMBS, municipal bonds, and student loan ABS.  Starting in 2007, many of the RMBS credits Ambac insured began to default, causing Ambac to incur significant losses.  Although Ambac managed to raise $1.5 billion of capital in March of 2008, in March of 2010, the Office of Insurance of the State of Wisconsin forced AAC (Opco) to establish a Segregated Account for approximately $35 billion of structured finance products with high probability of default.  The Segregated Account of AAC subsequently entered rehabilitation, with the goal being an orderly run-off of policies and claims. Pursuant to the Segregated Account Rehabilitation Plan, permitted policy claims are settled with a cash payment of 25% of the claim amount and the issuance of Surplus Notes for the remaining 75%. These Surplus Notes accrue interest at a rate of 5.1%, providing Ambac with a highly attractive long-term financing structure. In 2010, AAC also commuted most of its asset-backed CDO exposure for $2.6 billion in cash and $2 billion in newly issued Surplus Notes.  Despite these actions, AFG (Holdco) announced in June of 2010 that it would need to restructure through bankruptcy to reduce its debt burden and settle an outstanding IRS claim.

After a protracted bankruptcy process, AFG emerged this past May with 45 mm common shares, 5.05 mm warrants at a $16.67 strike price, and significant NOLs.  As currently constructed, AFG is a holding company whose principal asset is AAC, itself comprised of the AAC General Account and the AAC Segregated Account (in rehabilitation).

Analysis of Key Value Drivers

Positive Adjustments to Book Value

Loss Reserves on RMBS

The crux of Ambac’s misvaluation lies in its gross loss reserves, particularly with regard to RMBS.  Specifically, Ambac has taken gross loss reserves of $5.843 billion against the RMBS credits it has insured. As Ambac described, the “adverse development in loss reserves established in prior years” was “primarily due to the continued deterioration of the collateral supporting structured finance policies, including RMBS.” It is key to note that Ambac did not Fresh Start its loss reserves upon emergence. To analyze what represents fair value for loss reserves, we need to develop a strong view on what has happened with the insured bonds and the underlying mortgage collateral.  Since Ambac reports a book with all of its exposures, we can look at the implied market prices for the RMBS credits that Ambac has insured.

Note: To see data yourself, access http://www.ambac.com/authenticate.asp?file=Single_Risk_Exposures_by_CUSIP.xlsx

On a cumulative basis, Ambac has $17.673 billion in outstanding RMBS credits that it has insured.  With loss reserves of $5.843 billion, Ambac is implying a haircut of 33.06% on these insured credits.  Ambac breaks out its par outstanding and reserves by both vintage and underlying collateral type, as seen in the tables below.

As the tables illustrate, Ambac has substantial exposure to vintages in the boom years of 2005-2007, which led them to take very significant reserves against their exposures when the RMBS market collapsed.  If we think about the process of taking reserves, Ambac notes that its reserves represent “a probability-weighted average of all possible outcomes.”  However, history shows that Ambac’s reserves have lagged the market’s understanding of expected future losses. 

Let’s take a moment to discuss our methodology for understanding the market’s expectation of future losses.  Mortgage-backed securities depend on the cash flows stemming from the underlying collateral, which is a pool of mortgages (in this case, residential mortgages).  When the market expects worse employment statistics and lower housing prices than were assumed in pricing the trusts, RMBS prices trade down as the underlying mortgage cash flows have become riskier and potentially insufficient to service the bonds, increasing the chance of bond impairment.  Here, you might ask, isn’t the purpose of bond insurance? In short, yes. Assuming that the insurance counterparty is solvent enough to make good on insurance claims, the RMBS should not trade down, as the insurance company is supposed to step in to make the payments.  However, in the case of Ambac, history suggests that the market does not necessarily trade in this fashion.

Since the RMBS collateral is correlated along the lines of housing prices, in the event of substantial impairment, Ambac would likely not be solvent enough to make good on the claims. Given this logic and having spoken with market participants, we believe that Ambac-wrapped RMBS trade primarily on collateral recoveries and ascribe minimal value to the insurance. To further gauge this assumption, we constructed a representative index of non-insured RMBS with vintages, collateral-types, and underwriters similar to Ambac’s book. The graph below shows the relatively tight correlation between the market trading prices of Ambac’s book and our non-insured RMBS index, indicating that collateral performance is the primary driver of Ambac-wrapped RMBS. As a result, we believe we can look at market prices of Ambac-wrapped RMBS to gauge Ambac’s future insurance losses.

Using Bloomberg and speaking with traders on the Street, we were able to price 83% of Ambac’s outstanding RMBS exposure, including a diversified sampling with vintages across all relevant time periods and collateral types.  The trends since 2007 show that Ambac’s reserves have lagged, then mirrored, and finally exceeded the market implied haircut (Market Haircut = 1 – Bond Price) on Ambac-wrapped RMBS securities. Specifically, if we look at 2007 based on our sample book, Ambac had reserved 6.18% of par while the market implied losses of 6.34% (i.e. bonds traded at 93.66).  By 2008, Ambac’s reserves were only up to 10.96%, while the market’s implied losses had exploded to 52.10%.  As the following table illustrates, Ambac’s reserves significantly lagged the market’s implied haircut through 2011.

After 3 years of recovery in the RMBS market, by Q2 2012, the market implied haircut had essentially come in line with Ambac’s book reserves.  After Q2 2012, a dramatic disconnect between market pricing and Ambac’s reserves emerged. We believe this discrepancy stems from management’s incentive to be excessively conservative with reserves. The decision to over-reserve has the dual effect of creating a low hurdle for future “outperformance” and depressing the stock price prior to the granting of management’s stock-based compensation plan.

To estimate future reserve reversal, we look to the current gap between market implied and book haircuts. As the table and graph above illustrate, the current market implied haircut is 16.11%. This represents a massive gap 16.95% gap on $17.673 billion par value between Ambac’s current reserves and market implied reserves.  To illustrate the substantial impact on valuation from this disconnect, we calculated the implied gain on reversing loss reserves if Ambac’s actual haircut is 15%, 20%, 25%, and 30%, as seen in the table below.

Using the market’s implied haircut, the gain on loss reserves would be more than $3.1 billion, representing ~3.0x Ambac’s current market cap.  Even if we are to assume the market is undervaluing Ambac’s potential haircut by 10% (perhaps because of insurance value), the implied gain on loss reserves would still be more than $1.4 billion. Ambac will be forced to reverse these loss reserves over time, proving an upward catalyst in terms of book value and stock price.

Litigation Claims

Although Ambac has taken substantial gross loss reserves on its RMBS portfolio, Ambac appears well-positioned to recover some of that value via litigation recoveries. When Ambac insured RMBS, the originating institutions signed agreements with Ambac verifying that “underlying loans were not materially fraudulent or were the product of material underwriting defects, in breach of express representations and warranties” (Assured Guaranty vs. Flagstar Bank, Civ. 02375). Several recent court decisions have found that a wide array of originators breached their signed representations and warranties, leading to recoveries for counterparties like Ambac that were harmed by the false representations.

With regard to breaches of representations and warranties, Ambac records its expected recoveries as “Subrogation Recoveries” (contra-liability to Loss Reserves), currently marked at $2,395 mm. These recoveries stem from Ambac’s outstanding cases against mortgage originators. As the table below shows, monoline insurers have been exceptionally accurate in forecasting subrogation recoveries, coming out slightly conservative in the three most prominent instances.

*Note that MBIA outlined its expected subrogation recovery from Flagstar and Bank of America (BAC) as a combined value, and the actual subrogation recovery reflects the combined settlement amount.

Based on past history of successful forecasting and conversations with those following the litigation, we believe Ambac’s representation and warranty recoveries will be largely in line with the $2,395 mm figure it has booked.

Even though these recoveries are booked in the financials, the market has still reacted extremely favorably to successful past litigation. For example, after Countrywide (BAC) announced a settlement with MBIA for $1.6 billion, the market cap increased by 46% or about $850 mm the next day, even though recoveries were only about $200 mm above what had been booked. Given the uncertainty with which the market has viewed prior litigation claims, we believe that future settlements/judgments in Ambac’s currently filed litigation will provide a strong catalyst.

The most notable current litigation case is against BAC, who we believe is strongly incentivized to settle. The Gibbs & Bruns settlement (BAC against BNY Mellon) was settled for 8.5 cents on the dollar for RMBS claims. This low amount can be attributed to BAC not being deemed directly responsible for the RMBS losses; instead, the financial crisis was blamed. If BAC were not to settle with Ambac and proceed to trial and possibly lose, the Gibbs & Bruns settlement could potentially be reversed and BAC would risk losing a significantly higher amount than if they just settled with Ambac. Justice Eileen Bransten is presiding over the Ambac-BAC case; this is important as she also presided over the MBIA-BAC case. The MBIA settlement came down to the fact that the First Department New York court ruled that causation for liability is no longer required in order to claim incurred losses. This ruling is crucial because the crux of BAC’s argument was that the losses incurred were due to the financial crisis. Therefore, we expect BAC to settle and not risk going to trial.

Ambac will likely obtain additional litigation recoveries on fraud claims it has filed against mortgage originators.  The key element to understand is that fraud claims cannot be booked in the financials.  Thus, Ambac’s fraud claims, such as a $300 mm claim against JP Morgan, represent additional upside to the marked R&W subrogation recoveries.  These fraud claims have been successful recently, as evidenced by Countrywide being found fraudulent for its sale of RMBS to Fannie and Freddie.

Our conversations with participants following the cases suggest $500 mm to $1,000 mm of additional litigation recoveries.  These recoveries will likely stem from the aforementioned JP Morgan fraud claim, additional fraud claims such as those against First Franklin/Merrill Lynch, and further R&W filings for parts of the insurance book that have not yet had subrogation taken.  We underwrite to a base case of $750 mm in additional litigation recoveries, with a low case of $500 mm and a high case of $1,000 mm.

Repurchased AAC-Wrapped RMBS

In managing its risk exposure, Ambac has repurchased, at a substantial discount to par, bonds that AAC insured.  Specifically, Ambac has par value of $1.42 bn in repurchased AAC-wrapped RMBS bonds (see row 497 of http://www.ambac.com/authenticate.asp?file=Invest_Sched.xlsx). Ambac repurchased these bonds for $921.5 mm (65 cents on dollar) and discloses the fair value as $1.08 bn (76 cents on dollar).

Based on our conversations with various professionals and our understanding of the relevant accounting, Ambac has taken reserves against these bonds that remain on its books. Since Ambac now owns these AAC-wrapped bonds, any reserves should be completely reversed, allowing Ambac to recognize a gain. Since we have already adjusted the RMBS loss reserves to fair value, we only add back the difference between Ambac’s 76 cent mark on the bonds and par value. In all three cases, we add back 24% on $1.42 bn of par which yields a $341 mm increase to book value.

NOL Valuation

Ambac has $5,151.4 mm of NOLs (~$4,000 mm held within AAC and ~$1,200 mm within AFG) which are fully reserved against on its balance sheet. Given that Ambac is not currently writing new insurance, the NOLs initially appear to have limited value. However, there clearly is a possibility that Ambac receives approval to restart its insurance business; in this case, the NOLs will have substantial value. Furthermore, management has indicated that they would potentially raise capital and purchase a business similar to AAC to utilize the NOLs. In trying to put a value on the NOLs, we analyze them under two primary scenarios: 1. Restart of the insurance business in 2015; 2. Raise capital and purchase assets (high and low case).  We also weight the possibility that Ambac is unable to unlock any value from its NOLs.

In our valuation scenario where Ambac restarts its insurance business in 2015, we looked to the historical cash taxes and effective tax rate of the company prior to the financial crisis. Using the average 2001-2007 cash taxes of $171.2mm divided by the effective tax rate of 26.5%, we get $646.2mm as the NOL utilization amount going forward. In the case where Ambac raises capital to purchase AAC-like assets, we constructed a high and low case with varying assumptions on size of equity raise, leverage, and unlevered ROI. After evaluating these scenarios, we came to a probability weighted NOL value of $256.5mm. The valuation output is summarized in the table below.

In thinking about the value of NOLs within the context of the larger story, we feel that many analysts have overstated the potential impact of NOLs on valuation.  In particular, the idea of purchasing a new business to utilize the NOLs doesn’t derive as much value as many analysts seem to believe.  Instead, the situation where the NOLs really benefit valuation is if Ambac receives permission to restart its insurance business, a situation in which value would likely be so high that the impact of NOLs would be minimal on a home run investment. While the $5,151.4mm headline NOL number appears to represent a lot of value, the NOLs are not a key driver of the Ambac misvaluation story.

Insurance Intangible Asset and Goodwill

In making adjustments to GAAP book value, Ambac backs out $2.136 billion of value for insurance intangible asset and goodwill ($1.621 bn for insurance intangible, $514.5 mm for goodwill). Understanding this valuation adjustment is an exercise in understanding Ambac’s Fresh Start accounting.

In applying Fresh Start, Ambac “adjusted the historical carrying value of its assets, liabilities, and non-controlling interests to fair value,” with reorganization value in excess of fair value of “identified tangible and intangible assets” being attributed to goodwill (see table below).  

Given that Ambac sees goodwill as deriving value from neither tangible nor intangible assets, we agree with Ambac’s decision to back goodwill out of book value.

However, we believe that Ambac should not have backed out the insurance intangible asset from book value.  The insurance intangible asset “represents the fair value adjustment for financial guarantee insurance and reinsurance contracts.” Due to certain accounting rules, “insurance and reinsurance assets and liabilities continue to be measured in accordance with existing accounting policies and an intangible asset is recorded representing the difference between the fair value and carrying value of these insurance and reinsurance assets and liabilities.”

The insurance intangible asset relates to a number of B/S accounts: premium receivables, reinsurance recoverable on paid and unpaid losses, deferred ceded premium, subrogation recoverable, losses and loss expense reserve, unearned premiums, and ceded premiums payable. Since we are making fair value adjustments to the loss reserve, we do not want to double count and therefore only add back a portion of the insurance intangible. We compute this amount as the total amount of the intangible ($1,621.6 mm) less our RMBS loss reserve reversal in the low case ($1,418 mm), yielding a $203.6 mm increase in book value. In the low case, we have essentially assumed that Ambac’s fair value adjustment of $1,621.6 mm is not overly conservative and simply split this amount into two separate add-backs. In the base and high case, we use higher estimates for RMBS loss reserve reversal and therefore adjust book value upwards by more than the insurance intangible.

Negative Adjustments to Book Value

Student Loan Reserves

While the gross loss reserves for RMBS credits appear materially overstated (to the aforementioned tune of $1.4-3.0+ billion), we believe that student loss reserves may be understated.  During the boom lending years, for-profit education loans proliferated, and the current market perception of many of these loans is quite negative.  Specifically, using the same methodology as we used for RMBS credits, we were able to price 52% of the AAC-wrapped student loan ABS bonds.  On these bonds, the implied market haircut is 26.58% (vs. Ambac’s book haircut of 18.52%).  However, a substantial portion of the book that we have not priced ($1.35 billion) comes from bonds secured by student loans originated by National Collegiate Trust (First Marblehead).  The average National Collegiate bonds we priced have an implied haircut of 42.7%.  If we were to assume the unpriced National Collegiate bonds look the same, our implied market haircut for the book reaches 31.5%, suggesting Ambac may be under-reserved in a downside scenario.

In thinking about specifics, it is important to take note of a key dynamic in expected student loan losses: the difference between for-profit and not-for-profit issuers.  As Ambac describes:

“Collateral for the For-Profit Issuers consists of private loans which do not have any federal guarantee as to defaulted principal and interest. Collateral for the Not-For-Profit Issuers consists of both FFELP and private student loans. Private loan defaults have been on the rise since the credit crisis in 2008 began. Elevated unemployment rates, combined with high student loan debt levels will continue to put pressure on borrower’s ability to pay their loans.”

As a result of the differences in collateral and government backing, Ambac has taken large reserves on its for-profit issuers and substantially smaller reserves for its not-for-profit issuers.

Based on our analysis, we view Ambac’s book haircut of 18.5% as perhaps 8% under-reserved in the base case, implying an additional $463 mm in student loan losses. In underwriting the low and high case, we include additional losses of $600 mm and $200 mm, respectively. Ambac notes that in what they describe as “the highest stress scenario,” these losses might reach $753 mm, indicating that our adjustments are quite conservative.

Puerto Rico Exposure

After Detroit filed for bankruptcy, investors looking for the next municipal crisis turned to Puerto Rico.  As the graph below illustrates, Ambac’s share price has closely followed the weakness of the S&P Municipal Bond Puerto Rico Index (SAPIPR), plummeting 45% from a high of $27.25 to a low of $15.03 in 5 months.

The market’s reaction to the Puerto Rico situation shows that investor understanding of Ambac’s obligations is limited. The Excel file we mentioned in the RMBS section also contained all of Ambac’s Puerto Rico exposure (the Excel was published on June 30th). Yet, when the Barron’s Puerto Rico article broke, the market acted as if it had absolutely no understanding of Ambac’s Puerto Rico exposure.  It took Ambac’s publication of its exposure in a slide deck to calm investors; the stock rallied 15% in the two days after Ambac published its deck, even though the information had already been public.

The table below shows that out of Ambac’s $2.5 billion of Puerto Rico exposure, only 10% is Commonwealth General Obligation (GO).  The remaining 90% is backed by pledged revenue streams from sources like toll receipts, and sales, gas, rum, and hotel occupancy taxes.  Ambac’s total Puerto Rico exposure is summarized in the table below.

Aside from the sufficient debt service coverage ratios highlighted in the table above, revenue-backed bonds are safer for the following reasons:

Sales Tax Revenue Bonds (COFINA): $808mm exposure

• Backed by 2.75% of the first dollars collected of Puerto Rico’s 7% sales tax
• Sales tax fed by a large base of products and services, excluding the volatile energy and automobile sectors
• Monies segregated from the General Fund
• Exempt from the Constitution’s claw-back provision

Highway and Transportation Authority: $760mm exposure

• Backed by certain toll receipts, taxes on gasoline and other fuel, vehicle license fees, and investment income
• Required to maintain a debt service reserve equal to the lesser of the maximum debt service on Highway Revenue bonds or 10% of the principal amount of bonds outstanding
• Toll revenues (27% of revenues) exempt from the Constitution’s claw-back provision

Infrastructure Financing Authority – Rum Tax: $574mm exposure

• Backed by Federal excise taxes on rum
• First  $117 million of rum taxes segregated into a designated fund
• May pursue a budgetary appropriation of funds if less than $117 million is received

Convention Center District Authority – Hotel Tax: $137mm exposure

• Backed by hotel occupancy taxes (e.g. casinos, resorts, hotels, motels, inns, etc.)
• Tax collections segregated into a separate account every month, which transfers funds to the trustee within 3 business days

The GO bonds are admittedly riskier considering Puerto Rico’s shrinking economy.  The commonwealth has been suffering from a recession for 7 straight years and has a 15% unemployment rate.  It currently carries $70 billion of outstanding debt and a heavy $17,000 per capita debt burden.  In addition, Puerto Rico depends on access to capital markets in order to balance its budgets.  With the current high-yield environment, Puerto Rico is avoiding any new issuances since it has sufficient funding to support operations until at least June 30, 2014, but a sustained loss of market access could pose a risk in the future. 

All is not bad, though.  Recent reforms put in place by newly elected Gov. Alejandro Padilla offer signs of improvement.  For example, the pension reform plan reduces contractual obligations that the Employee Retirement System has to make to public employees while increasing the member contributions and retirement age, which will help reduce the $1.3 billion deficit.  Furthermore, he has expanded the tax base to include services like consulting and B2B activity while enforcing stricter measures against tax evasion.

Investors should be cognizant of the difference between the situations in Puerto Rico and Detroit.  Whereas Detroit had no satisfactory plan to address their fiscal issues and relied on an Emergency Manager, Puerto Rican leaders are expressing that they have every intention of honoring the commonwealth’s debt and avoiding a restructuring or default.  In an October 15, 2013 webcast, Governor Padilla remarked, “We will do everything, and I repeat, everything that is necessary for Puerto Rico to honor all its commitments.  These are not just constitutional obligations but also moral obligations.”

This segues into a discussion about the legal protections that the Puerto Rican Constitution has in order to protect bond holders in the off chance of a non-payment.  Article VI Section 8 states:

“In case the available revenues including surplus for any fiscal year are insufficient to meet the appropriations made for that year, interest on the public debt and amortization thereof shall first be paid, and other disbursements shall thereafter be made in accordance with the order of priorities established by law.”

In other words, the public debt of Puerto Rico has a priority claim on resources and must be paid prior to other obligations if there are insufficient funds for debt service.

Base Case
To estimate the losses from Puerto Rican debt exposure, we still worry that despite Governor Padilla’s aggressive efforts to meet obligations, the GO bonds will not be paid in full.  In our base case, we make the assumption that the GO obligations are completely wiped out. Given that all of the revenue bonds have DSCR in excess of 1.4x, we assume that these bonds recover in full.  While the ultimate distribution of haircuts may vary, we see $250 mm in losses as reasonably conservative given the fear-mongering about the municipal bond market post-Detroit. 

High Case
A more bullish case does not call for a complete loss on the GO bonds.  Improvement in the economy and Puerto Rico’s focus on repaying its “moral obligations” lead to a scenario where the GOs only take a modest haircut.  Given that current yields are in the 7% range (10-11% including tax-free benefits) on these long-dated obligations, we see scenarios where GO holders take a 20% haircut as reasonable.  In the high case, losses are assumed to be $50 mm. 

Low Case
A bearish case calls for not only a total loss on GO bonds but also on some of the less secure revenue-backed bonds. We view bonds issued by Infrastructure Financing Authority (rum tax) and Convention Center District Authority (hotel tax) as higher risk.  The bonds backed by COFINA and the Highway and Transportation Authority are safer, as they have more consistent streams of revenue and are protected from the Constitution’s claw-back provision.  With approximately $500 mm in revenue bonds with DSCR of 1.4x, we could see losses of 20% if the recovery does not take hold. As such, we see Ambac losing $350 mm on Puerto Rico in a low case.

Detroit Exposure

In February, the state government took control of Detroit’s fiscal matters and appointed Kevyn Orr as the emergency manager. Orr attempted to reach a settlement with creditors in which they would receive 10 cents on the dollar, but creditors refused this proposal. After this failure to settle, Orr recommended that the city file for bankruptcy. Detroit filed for Chapter 9 Bankruptcy on July 19th, citing liabilities greater than $18 billion. Orr has chosen to group the unlimited tax GOs with the limited tax GOs despite their different rights.

Ambac has $170 mm in Detroit exposure: $77.6 mm in unlimited tax GOs and $92.7 mm in limited tax GOs. The impairment risk on the unlimited tax GOs has increased substantially given Orr’s proposed treatment. Due to these adverse conditions, we are assuming that Ambac will lose 100% on its insured Detroit credits in the low case, 90% in the base case, and 80% in the high case.

Summary Valuation

As noted in the introduction, we believe Ambac’s ultimate value will be based on adjusted book value.  In the 6/30 10-Q, Ambac reports adjusted book value as -424.9mm; the following table takes Ambac’s reported number and makes the adjustments described in the above sections to come to our ultimate valuation.

An alternative, but less rigorous methodology for estimating Ambac’s adjusted book value is to start with Statutory Capital and then scale it up by an adjustment factor, as determined from comps like MBIA and Assured Guaranty (using comps is questionable since their books were not subject to Fresh Start). This valuation serves as a rough check on the adjusted book value detailed above using a different set of accounting standards unique to insurance companies. As the table below shows, Statutory Capital has risen substantially over the last several quarters and yields a reasonably similar valuation to our base case from above.

Restarting the Insurance Business

We have not handicapped the likelihood and value of Ambac writing new bond insurance, but we believe this possibility offers meaningful additional value to investors in an upside scenario.

The fundamental drivers for the bond insurance business look highly positive going forward. Prior to 2008, seven AAA rated monoline insurers accounted for the vast majority of bond insurance (both ABS and municipal). The financial crisis crippled the financials of the bond insurers due to losses on ABS insurance. Currently, only Assured Guarantee is writing new insurance, as it still has an adequate Insurance Financial Strength (IFS) rating. Ambac, MBIA, FGIC, Syncora, and CIFG are all currently in run-off and lack the necessary ratings/statutory capital to write new business. As a result, Assured Guarantee has a virtual monopoly on the highly profitable, relatively low risk municipal bond insurance business.

Ironically, although Detroit and Puerto Rico are risks to the legacy portfolios of the monoline insurers, they have spurred investor demand for municipal bond insurance. The historical assumption that municipal bonds have no credit risk is being questioned. Moreover, as interest rates rise and debt service coverage is eroded, investors will be incented to demand insurance. We believe that these two factors will drive the municipal bond insurance market to grow at a healthy clip.

S&P requires that financial insurers have a minimum of ~$500mm in Statutory Policyholders’ Surplus to garner an A rating or better. In general, an A rating is necessary to write new insurance policies. While Ambac’s surplus sits at $393.7mm as of Q2, it has increased substantially in recent quarters (see above) and will likely exceed $500mm in the not-so-distant future. We expect that Ambac will ask to be rated and attempt to restart its insurance business. With only Assured Guarantee and potentially MBIA in the market, we see this possibility as hugely accretive to valuation.

Parting Thoughts

A Few Comments on the Warrants

As stated in the preliminary note, Ambac has 5.05mm in warrants outstanding at a strike of $16.67, expiring on April 30, 2023.  At current prices of $12.39, the warrants represent a levered bet on Ambac seeing considerable upside to current levels.  Given how many moving pieces there are and the levered nature of the capital structure, we believe the additional upside the warrants could provide does not justify the additional risk.  We see the risk-adjusted returns as superior for the common equity at current levels.

The Final Word

We believe that Ambac offers numerous ways to win and has a palatable downside case that we view as relatively unlikely to occur. The complexity and uncertainty associated with Ambac has led to a market dislocation, leaving us with an equity that has asymmetric risk/reward and defined catalysts for value realization. As transparency improves and the contingent assets/liabilities are realized, investors will re-price the equity to fair value. While there are many moving parts in the Ambac situation, we hope this writeup has been helpful in describing and quantifying the key value drivers.

Catalyst