Getting Gemütlichkeit with Genworth's LTCI

Genworth has shat the bed! This calls for something special: a complete, unedited, typo-ridden, stream of unconsciousness, celebration of the fallibility of human judgement (...hopefully not my own.)

The company reported over $500M each in reserve charges and goodwill writeoffs in Q3. What's more, the circumstances of the charges indicate a large GAAP active life charge in Q4.

The pending charges are the probable reason that Genworth was at one point down over 50% in two days. My current guess is that this pretax GAAP charge will be greater than "Fitch's current rating expectations assume an additional $500 million to $1 billion of pre-tax GAAP charges." 

Rating agencies took disparate views on the news. Fitch put the company on negative watch while expecting more charges. S&P took the opposite position, downgrading the company and noting "Any additional material reserve strengthening could result in a downgrade." Moody's was least committal, placing the company on rating watch negative while it "will use the review period to evaluate the results of Genworth's review of the margins on its LTC active life reserves."

Moody's also quoted it's SVP, Scott Robinson, who said "Genworth has taken prudent actions to protect its capital position and has the capacity to absorb the announced reserve charge, which was higher than we expected. While we will gain insight into the company's long-term care reserve margins during the review process, we believe the company remains exposed to further, significant deterioration in its legacy block of business."

Indeed. The policies in the active life reserve display much better characteristics than those in the claim reserve (other than not being in claim e.g. tighter underwriting, benefit length, daily max, attained age, etc.). However the active life reserve is split into two buckets for GAAP reserve loss recognition testing (LRT) margin, the older of which is markedly worse than the newer. The GAAP LRT margin on this pre-10/3/1995, "PGAAP" group of policies was only $100M on a $2.5B reserve (both are after tax). The newer "HGAAP" block of policies has a $2.9B GAAP LRT margin on a $13.4B reserve.

The Q3 claim reserve strengthening was roughly 15%. Claim reserves are best estimates, which are used to the active life testing margin process to establish the economic balance sheet margin. That margin is then stress tested to come up with the LRT margin. The bad news is that the actual best estimate of the present value of future claims on the active life book was $37.8B at the end of last year. The good news is that the increase to this number should be substantially less than 15% due to the shorter tail and better risk profile of this book. It is harder for claim severity to increase when duration and daily benefits have tighter caps.

The assumption changes appears to expose PGAAP book to considerable negative margins, while the HGAAP book appears to maintain a decent margin. The increase in PGAAP expected future claims will be higher than for HGAAP. The ratio of expected future claims to active life reserve is certainly higher for PGAAP. We can only guess, but perhaps it is as high as 3.5 versus the aggregate ratio of 2.3. Using a 15% increase on this assumption infers a GAAP charge of roughly $1.3B before any remedial action.

As noted, active life reserves are not just best estimates of future losses, they also also include expected future premiums. Increased loss expectations may increase the companies expected future premium rate increases. This led Genworth to note in its earnings report that "the company is developing related management actions, that it expects will offset much, or possibly most, of the reduction on margins from the claim reserve review." So maybe this pretax GAAP charge number we are playing with gets down to $1B.

But Q4 GAAP troubles could be compounded by additional goodwill writeoff of as much as the remaining $300M. The company noted in a November 6th press release responding to credit rating changes that "These changes in ratings or outlook are expected to reduce sales in some of its products." Lower expected sales was one reason for the Q3 writeoff and there is no indication that the company anticipated the rating and outlook changes.

All this GAAP analysis raises one of the central questions surrounding investment management: "who really gives a shit?" Now that we know the actuarial assessment has changed the economics outlook for the book of business, who cares how the bean counters catch up with reality? For one, we will gain more, albeit incremental insight into the book with every quarter and every review. For another, if we switch to statutory accounting, the bean counting determines dividend capacity and could even require regulators to intervene in the subsidiary's operations. 

There are big differences in GAAP and statutory margin tests. A big one for Genworth is how the book is divided. Rather than PGAAP and HGAAP books, Genworth's statutory tests divide the company between subsidiaries: GLIC, GLICNY, and BLAIC (a Bermuda reinsurance entity for GLIC and GLICNY). GLIC is the largest unit with the coziest margins. GLICNY is smallest and has already established an asset adequacy reserve. This reserve was actually reduced by $40M at the end 2013, but the take away for investors is that this margin is nil. This is where the likely statutory charge will come from in Q4.

Any distinct (non-geographic) characteristics between the books of GLICNY versus GLIC are not immediately clear to your blogger, but it is safe to say that GLICNY book is generally better than the claim reserve book and PGAAP book. GLICNY accounts for less than 10% of total active life reserves and probably accounts for less than $3.8B of the $37.8B aggregate PV of claims and expenses.

Being a New York subsidiary GLICNY follows that state's mandates for margin testing. This includes the unique disallowance of any expected but not yet approved rate increases. This nullifies managements main lever in offsetting increased loss reserves. Still the size of the likely statutory charge in Q4 will be a few hundred million or less. While this could drive unassigned surplus and therefore dividend capacity to or below zero for 2014, management has precluded dividends anyways. 

To put this in perspective, the company would have to post a about a $3B pretax reserve charge to take its RBC ratio down to 200% where the company would be required to submit a plan on how it intends to increase the ratio. $3.6B might get the company to a level at which regulators can take action (principally restricting new business). $4.2B is the level at which regulators have the option to take control of the company and either liquidate or - as is more commonly the case - rehabilitate it. At about a $5B the regulators would be required to take control of the life subsidiaries. 

Needless to say, none of those charges are coming out of the NY sub alone anytime soon. The aggregate book meanwhile would need expected losses to increase at twice the rate of the claim reserve book, with no management action to get to a level where Genworth needs to file an company action plan. Its more likely that the claim book loss expectation increase will be more than twice the rate of active life book loss expectation increase and that the latter will be reduced by management action.



Such reserve charges look increasing unlikely given that two of the large drivers of LTC  mispricing are approaching their limits. When underwritten, the oldest books of business were assumed to lapse at a rate of 6.5% and earn a yield of 6.75%. The lapse rate in the 2013 statutory CFT margin is 0.45%. Genworth uses the actual forward rate curve to predict future investment returns. A final 10 year yield that is 220 bps lower would decrease margins at the company about $2B. A final 10 year yield below 1.5% would be needed to to completely erode the 2013 CFT margin. A ten year of approximately 0.3% is needed to completely erode the 2013 economic balance sheet margin on its own. Given where the forward rate is now, this level of adjustment in interest rates will not be coming this year.

What about reserve increases in future years? This is a more valid concern, but Genworth's life insurance subsidiaries book over $500M of pretax profits annually across all busisness lines. Additionally the present value of future profits on new LTC business accretes to margin analysis as it is booked. Combined these add over $700M to margins annually. Meanwhile the oldest, worst business runs off.

Of course, many believe that it's impossible to make money in longterm care insurance. The most common reasons sited are the information advantage of policyholders, the contradictory incentives of insurer and insured, and the difficulty of necessary forecasting. There may be no other form of insurance that suffers so broadly from all three conditions. 

There may be no business today that suffers from such sparse competition and a negative sentiment. This bodes well for future returns. Genworth has shortened its tail risk, tightened underwriting, and increased prices for expected returns above 20% on new business. So prices for new business have risen not just to reflect more conservative assumptions, but also to make more money per unit of new risk. Meanwhile, Genworth has proven the regulatory appetite to allow for significant price increases on existing business to offset extreme deviations from industry forecasts.

While this is a LTC focused post, there just has to be some comment on valuation. The yield on the jr. subordinated 6.15s of 36 spiked above a 10% ytw from around 7% before the earnings blunder. Genworth has $4.4B of carrying value of outstanding debt (excluding debt at in international MI) with $1.1B of cash at the holding company and $3.3B value of its interest in publicly traded international mortgage insurance affiliates. This shows an impressive faith in management's ability to lose money.

Genworth's US MI unit, GMICO is worth $2.9B if valued at 2/3 (an approximation of relative rate of NIW) the enterprise value of larger MGIC. There are several ways to get to the current enterprise value (with debt at carrying value) from here. One way would value all of Genworth's other businesses at 2x 2013 earnings except for LTC which will need a 0x multiple to get us to the 8.8B EV. 


Taking a different view, management has estimated they can achieve a long term RoE of 9%. If they achieve that target, buying at 1/3 of book value will earn a return on market equity of 27%. Of course, what type of consensus makes for 27% return on your investment?

This may look like child's play to veterans of the mortgage insurance meltdown, but it still is uncertain how the market will react to the Q4 charges. Even though many called for $500M in reserve charges in Q3, the end results still drove many to despair. In the meantime we'll be keeping our pockets open and hoping for GNW to be a penny stock again soon.

HAMP: Data that Doesn't Foretell Doomed

Technicians, chartists and that lot are wizards. Whether you believe in them is up to you. But we just have that feeling that the bottom may have been put in for the solvent MI group: especially Radian, MGIC while Genworth offers less volatility and potential return.

Call it a double bottom, capitulation, or just a long awaited piece of good news. The recent market movements have made for an attractive opportunity amongst the three basic asset classes: redeploying capital into mortgage insurance assets and out of bond insurance assets and especially other assets. Not so much as changed in the bond insurance world, and that's why those assets are an attractive relative source of funds. Other assets - for those who believe in holding such things - remain the most attractive source of funds.

That's why this had to be written now.

What needs to be written is that the one (or the one) source of cure data outside of mortgage insurers supports the case for current reserve levels - or lower. Don't get us wrong: the next few quarters will in all likelihood continue to be ugly, with high loss provisions for the group as the delinquent inventory ages and new defaults remain stubborn. But that doesn't lend credit to suggestions that reserves should be 75% of risk in default for loans over 3 month delinquent (a la Barron's.)

The source is HAMP, see the program's press releases here or visit The Dragon's compilation of the latest two monthly reports here.

The data is broken out with more granularity than the MI's reports and cover two categories: loans from canceled HAMP trials and loans from loans not qualifying for HAMP. In all, 2.1M million loans are covered in the canceled loan and non-qualifying loans. The WSJ estimated their are about 6.3M delinquent loans.

For non-qualifying loans, roughly 0.9M of 1.55M or 58% of disposed loans were either current, paid off, modified or on a payment plan. 57% of 0.6M canceled loans fell into the same categories. Roughly 15% of canceled and non-qualifying loans were categorized as action pending or action not allowed. The remainder were in categories that would likely result in a claim - foreclosure pending, foreclosure completed, and short sale/deed-in-lieu.

If this data were extrapolated to the delinquent inventory at MI companies, they would look comically over-reserved. But this is a biased sample. Those loans not counted might tend to be those with the most hopeless borrowers or worthless properties, not to mention foreclosure process difficulties. On the other side of the scale, these numbers don't account for the 90% higher rate of still-current loans out of 657k permanent non-canceled HAMP modifications.

There's more: this data has no accounting for rescission and denials which might benefit claim payment rates to the tune of 15% or more. All of a sudden 50% of delinquent inventory covered by loss reserves doesn't sound so bad. Volatility be damned: we've used this opportunity to load up for the long haul.

The Sum of Genworth

Genworth's MI unit is the second highest rated U.S. mortgage insurer covered by Moody's at Ba1 (after AIG's United Guaranty at Baa1). This junk rated unit has been a considerable drag on the rating and stock price of the entire corporate family, except for the 40% of Genworth MI Canada that trades on the Toronto Exchange. The value of that entire unit based on Wednesday's closing price is 2.45B in Loonies or about US$2.6B. The 60% owned by Genworth US is worth $1.56B. Genworth's wholly-owned Australian MI unit produced more income than the 60% Canadian interest in each of the last 5 quarters but lets say it worth the same amount - $1.56B.

That would make sense except Genworth US is valued at $3.9B, meaning that the remaining international units are worth nothing, the life and retirement division is worth 2.5 times 2010 earnings, and the US MI unit loses everything and the approximately $300M or so of stock in the Canadian MI subsidiary.

The obligations of the US MI unit are not guaranteed or supported by any other Genworth entity. At least theoretically, it is bankruptcy-remote. For the US MI unit's equity to be worth nothing, the entire default inventory would have to go to claim - with no natural cures, modifications or rescission. That would wipe out the unit's equity which is roughly equal to the unreserved default inventory. From there, the unit would have to experience enough claims from new defaults to wipe out over $500M in annual earnings before taxes and claims. Regular readers will guess that we don't think that will happen.

Then there are the other divisions. It is easy to see why Genworth spun off part of its Canadian MI unit. It's the same reason that GE spun off Genworth years ago. The parts are worth more than the sum.