Just when you thought the Radian story couldn't get any more interesting, over the weekend, Barron's released an article outlining how Radian was in worse shape than it looked. I wrote a similar article a few months ago.
Out of the many points it made, Radian has refuted suggestion that Radian Guaranty only has $2.1 billion of cash and investments to pay its claims. Radian is saying that their other mortgage insurance subsidiaries, collectively, have cash and investments of $3.1 billion and that it is unfair to focus solely on Radian Guaranty.
Radian has a point. $600 million, or 20%, of Radian's loss reserves are at these companies, so it is unfair to look at Radian Guaranty as though it only has $2.1 billion in investment assets to support its future claims. Nonetheless, once the illiquid surplus stakes are removed from its balance sheet, Radian Guaranty (and by extension its other mortgage insurance companies) still has a poor liquidity position relative to its peers. Investors shouldn't lose sight of this key fact.
Furthermore, Radian makes its operating company liquidity position appear better by stretching out its claims process - I discussed this in my previous article. If Radian was paying claims as fast as MGIC (MTG), a competitor, it would have already paid an extra $400 million. If we are to assess the liquidity position of Radian Guaranty fairly, we should also deduct these delayed claims ((1-20% of claims paid by other insurers) * $400 million = $320 million).
The following table sums up the true liquidity position of Radian Guaranty as of 2Q12 - deducting their illiquid equity stakes from the calculation. Note the adjustment to delinquencies includes a reduction for pending claims that should have already been paid (7,856), as well as an adjustment of 20% to reflect the fact their other MI subs will pay part of the claims.
|Radian Guaranty liquidity|
|Radian Guaranty investment assets||$3,687,687,122|
|minus claims it should have already paid (RG only)||$320,000,000|
|minus illiquid stock ownership||$1,360,022,597|
|RAA (financial guaranty)||$1,153,338,508|
|RMII (mortgage insurance)||$21,175,444|
|RIINC (mortgage insurance)||$169,209,893|
|RMAI (mortgage insurance)||$16,298,752|
|Total Radian Guaranty MTG equivalent liquid assets||$2,007,664,525|
|RDN liquid assets per delinquency||$28,854|
|PMI liquid assets per delinquency on seizure||$25,056|
|MTG liquid assets per delinquency||$38,964|
As you can see, Radian Guaranty's adjusted liquid asset position is only $2 billion and, if we add back the adjustment for delayed claims, that number is $2.3 billion, close to what Barron's stated. I'm guessing Barron's has made an adjustment for the cash burn in the most recent quarter. Note that Radian's liquid assets per delinquency is almost as bad as PMI's before seizure.
So, while Barron's wasn't completely correct in its assertion, the main point it made holds: Radian Guaranty has a very poor liquidity position, and investors should be careful.
Why does RDN trade at a premium to MTG?
The rest of the article focuses on Radian's poor holding company liquidity and loss reserve adequacy relative to MTG, as well as its inflated book value. MTG trades at a discount to its book value, while Radian trades in line, despite the fact that MTG has better holding and operating company liquidity, as well as a higher quality loss reserve. Furthermore, the Freddie Mac issue is receding for MTG, while becoming a potential concern for Radian.
As stated in my previous article, Radian is using a few tricks to make its reserve per delinquency appear stronger than it actually is. Radian tells investors it has one of the best reserve per delinquencies in the industry at $28.5k yet its true reserve per delinquency, on an MTG comparable basis is $23.2k.
There are three factors that lead to this lower true reserve per delinquency:
- Radian's IBNR ("incurred but not reported") reserve is inflated relative to MTG. Radian books IBNR to account for future denial reversals - on delinquencies not yet reported. Given Radian is denying at an elevated rate, their number is much higher than MTG's. But, since the IBNR is not associated with any delinquencies, it artificially boosts the reserve per delinquency number.
- Radian's delinquency inventory has a higher proportion of pending claims, which means more of its delinquencies will go to claim. In turn, that means Radian should have a higher reserve per delinquency than MTG.
- Even after adjusting for the first two factors, Radian's delinquencies are, on average, older than MTG's. In other words, more of Radian's delinquencies will go to claim. 56% of Radian's delinquencies are in the 12+ missed payments bucket vs. 52% for MTG. We need to adjust for this also.
The following table sums up these adjustments:
|denial reversal?||up to 10% worse|
|MTG per delinq||$25,890|
Notice also I have included two other risks to the downside: the possibility that denial reversals are worse than expected (up to $200 million addition according to management) and/or litigation lands on their desk (unknown risk). Radian is doing things other mortgage insurers aren't - there is the very real risk they get sued and lose. Note also that Radian's denials and rescissions this month were close to zero, meaning they are started to see a very high level of reversals of previous denials.
Radian's true book value
To calculate Radian's true book value, we need to deduct the necessary reserve addition, as well as remove the boost from the market value adjustment on their 2017 convertibles.
|Radian true book value calculation|
|Current book value||$915,253,000|
|minus true value of 2017 debt adj.||$125,000,000|
|minus minimum reserve adjustment||$235,658,257|
|plus estimated recoveries||$58,200,000|
|Radian true book value||$612,794,743|
Note that I also added back in Radian management estimated recovery on CDS transactions at its bond insurance business. If you don't trust Radian management, don't add this number.
Note also that this doesn't include adjustments for denial reversal and litigation risk.
MTG's true book value
To calculate MTG's true book value we need to adjust for the Freddie settlement, given it is likely that it will be signed.
|MTG true book value|
|Current book value||$667,976,000|
|minus Freddie Settlement||$267,000,000|
|minus convertible adjust||$20,000,000|
|MTG true book value||$380,976,000|
Radian trades at 1x its true book value while MTG at a 15% discount. So you would expect that MTG must have a liquidity problem to justify this difference. On the contrary, it is Radian who has the liquidity problem (and all the other problems).
Holding company liquidity significantly worse than MTG
Radian has about $350 million in liquid investments at its holding company, with $92 million of principal due in 2013 and $250 million in 2015. Radian won't have enough cash to last past 2015. It will need to raise capital or receive a dividend from its mortgage insurance subsidiary, which the regulator is unlikely to allow given Radian's operating company liquidity.
MTG, on the other hand, has $100 million of debt due 2015 and $350 million of debt due 2017. It has enough cash to last until 2017. Furthermore, the superior liquidity position of its subsidiary means it is more likely to be able to withdraw cash.
Yes both aren't perfect, but RDN is in a much riskier position. So it is baffling that RDN trades at a premium.
The return of Freddie Mac
As discussed on Radian's recent conference call, Freddie Mac is yet to extend approval for Radian Asset Assurance, Radian's MI company that will write business once Radian Guaranty breaches capital levels. For anyone that has followed MTG's troubles, this is no surprise - Freddie wants MGIC to be able to access the cash of MIC, just in case MGIC doesn't have enough cash to pay claims.
And, since MGIC has a vastly superior liquidity position to Radian Guaranty, it seems clear Freddie will want something similar from Radian Guaranty, which just won't be possible, given Radian Guaranty's liquidity position.
It thus becomes clear why Radian is so desperate to remain below the 25 to 1 risk to capital level, while Genworth and MGIC were happy to breeze right by.
The following table shows how Radian, over the last 2 quarters, has been incurring losses on new delinquencies at a lower rate than MTG, even after adjusting for size:
|MTG losses incurred on new delinquencies||$417,250,000||$393,420,000|
|RDN losses incurred on new delinquencies||$161,600,000||$153,300,000|
|adjusted for size difference||1.61||1.60|
Radian is underreporting incurred losses by 40% (0.61/1.61) compared to MTG - presumably to remain below the 25 to 1 risk limit. In other words, if Radian was using MTG reserving practices, Radian's incurred losses should have been 60% higher in the last two quarters.
The financial guaranty risk at Radian
It goes without saying that if you assign the same multiple to Radian as MTG, you are implicitly assuming that the risk at Radian's bond insurance unit is zero. Do you want to make that bet?
Book value is not intrinsic value
And remember, while book value is a useful tool for comparing two companies; it is not a measure of absolute valuation. Both these companies intrinsic value depends on the earnings they'll generate for shareholders long into the future. But I'll leave that calculation to you.