Radian Group RDN
December 31, 2007 - 4:02pm EST by
max318
2007 2008
Price: 12.00 EPS
Shares Out. (in M): 0 P/E
Market Cap (in $M): 970 P/FCF
Net Debt (in $M): 0 EBIT 0 0
TEV ($): 0 TEV/EBIT

Sign up for free guest access to view investment idea with a 45 days delay.

  • Mortgage Insurance

Description

RDN   Long   $12
 
3 to 6 month upside target of $20 (based on peer group valuation).  Longer term upside to $38 (worst case FY08 exit book value, with forward visibility into 10%+ RoE).
 

Why Radian is an Absolute Value

Immediate value could be realized in RDN if an activist shareholder argued for a spin out of their Financial Guaranty division.  Because of its minimal RMBS and CDO exposure, Radian Assurance (the FG division) should be worth at least $16 per share.  At $12 per share, RDN’s stock is trading well below the value of Radian Assurance.  Negative value is ascribed to RDN’s Mortgage Insurance division (MI) and remaining ownership in Sherman (a company that buys and works out distressed consumer debt).  Positive value is clearly ascribed to competing MI businesses such as MTG and TGIC that have worse Alt-A exposures and trade a 0.3x to 0.5x book value.
 

Why Radian is a Relative Value

If you are still strongly negative on the mortgage insurers and financial guarantors, you can short out a comparable sector basket (MTG & ABK for the MI and FG businesses, respectively).
 
For the most part, Wall Street thinks of RDN as the worst mortgage insurer because they wrote some non-traditional MI.  In reality, they wrote a manageable amount of non-traditional exposure in the form of coverage of NIMs (net interest margin securities) and 2nd liens.  After adjusting for write-offs of non-traditional products, RDN has more favorable exposure on the MI and FG sides relative to its sector.  RDN has significantly less CDO exposure in its FG biz and less Alt-A exposure in its MI biz.  Yet RDN has half the valuation of its peers and lower agency ratings that its peers.  RDN has little downgrade risk and no need to raise capital because of a recent sale of part of their remaining stake in Sherman Financial Services.
 
Sum of the Parts Analysis:
Radian is roughly half bond insurer (Radian Asset Assurance, the “FG”) and half mortgage insurer (Radian Guaranty, the “MI”).  Their remaining ownership of Sherman Financial, a workout buyer of distressed consumer debt, is 22%.
 
Numbers in $M unless per share
 
 
FG Q307 Equity
1,429
FG Q307 BVPS
$17.83
 
 
MI Q307 Equity
1,925
MI Q307 BVPS
$24.01
 
 
Total Q307 Equity
3,447
Total Q307 BVPS
$43.01
 
The small amount of extra total book value is for the remaining Sherman interest.  Ar $12, RDN is currently trading at ~28% of overall stated book value.
 
Radian FG is worth at least $16 per share
 
Radian FG simply does not have the billions of multi-sector CDO exposure that has dogged SCA, MBIA, ABK and others.  As long as pricing is rational, the bond insurance model for municipalities and certain structured transactions makes sense (see Buffett’s recent entry into the business).  Reinsuring municipal risk from the large primaries trying to free up capital (MBI and ABK) could be a very profitable business right now.
 
On Dec 19th (and under a huge amount of heat), S&P gave its updated stress case review of the subprime exposure of the bond insurers.  It is clear that Radian AND is covering rating agencies are sources of controversy right now.  But S&P had a lot of time, and a lot of incentive, to make a fair assessment on Dec 19thS&P stated that Radian’s worst case losses are $68M against a capital cushion of $550 to $600M:
 
http://www2.standardandpoors.com/portal/site/sp/en/us/page.article_print/3,1,1,0,1148450120384.html#ID208
 
What if we want to estimate worst case losses ourselves?  To be extremely conservative, we can take the recent marking of A) multi-sector CDO exposure and B) subprime exposure by Swiss Re (0 cents on the dollar for CDOs with subprime and 62 cents for outright subprime RMBS coverage).  It is important to note that Swiss Re’s marks (reported on Nov 19th) were taken on credit default swaps and were mark to market losses, not yet realized losses.
 
For A):
Radian’s recent presentations on its FG division detail its multi-sector CDO exposure at $765M.  2/3’s of that is backed by RMBS, with only a 1/3 of the $765M being subprime RMBS.  A full after-tax writedown of the RMBS CDO exposure would be $330M.
 
For B):
For Radian FG’s subprime exposure (outside of CDOs), there is a total of $561M of exposure.  If this were written down to 60 cents on the dollar, there would be a $226M after tax effect.
 
Radian could absorb this total hit today and still be within its capital cushion.  There is $1.4B of statutory capital at the FG on an insured portfolio of $110B.  Of the $45B which is CDO exposure, 90%+ is corporate exposure, which simply is not having the negative credit experience that residential mortgage backed CDOs are having right now.  In total, relative to the rest of the industry, RDN has about 1/5th of the proportional non-CDO RMBS exposure that a comparable FG would have.  They have even less of a proportional amount of multi-sector CDO exposure.
 
There is no need for capital at Radian FG and no downgrade risk.  There is no collateral posting for its transactions.  The FG is rated AA, outlook stable, and meets S&P’s target of a 1.25x excess capital ratio for a AAA rating.  That is not a misprint:  they roughly meet the AAA requirements.
 
There is $850M of unearned premium at the FG, and although 1/3 of the recent premium earned (in the form of direct public and direct structured) may be much lower in the next few quarters, the reinsurance business is extremely strong.  Net operating income in the FG could be $200M+ in 2008 from strong reinsurance opportunities.
 
Reinsurers that have not been tainted with the subprime exposure, such as AGO, are able to capitalize on this environment and are trading at premiums to stated book value.  RDN arguably deserves a discount because of its AA rating, but a quick comparison of the FG’s RMBS and CDO exposure in Table 5 and 6 of S&P’s recent FG sector review shows a significant lack of the exposures that have been so toxic for RDN’s comps:
 
http://www2.standardandpoors.com/portal/site/sp/en/us/page.article_print/3,1,1,0,1148450120384.html#ID208
 
RDN’s FG has significantly less problematic exposure as a multiple of its equity than ABK and MBI.  Also, RDN is not in a situation of having to raise capital under the threat of downgrade.  The reinsurance environment is extremely strong right now with ABK and MBI looking to raise capital by shedding profitable exposure.  RDN should be able to capitalize on this environment along with the other reinsurers and deserves a more reasonable multiple to book value.
 
Radian MI is worth at least the 0.5x book multiple given to its peers
RDN’s MI and FG are separate subsidiaries with separate capital pools and ratings.  It is fair to say that RDN’s MI is currently being given negative value by the market.  Even if you argued that RDN’s FG deserved the same book multiple of ABK and MBI (both are trying to raise capital and face downgrade risk), that would imply a multiple smaller than 0.1x book for its MI.  This is less than 1/3 of the multiple given to RDN’s worse positioned peers such as TGIC and MTG. 
 
Why is RDN’s MI given such a discount?
MTG famously broke off its merger with RDN in the fall after the deal had been in doubt for months.  RDN has traded at a discount since the merger was brought into question and has never recovered from that discount.  The market seems to be signaling that RDN is the weakest of the mortgage insurers.
 
The chief reason for this could be because RDN insured $800M of NIMs and $1B of second liens.  They experienced an accelerated loss on these exposures when reporting Q307 (which coincided with the deal breaking).  Unlike traditional MI losses, which are recognized as the policies go bad, NIMs losses are pulled forward and are subject to mark to market losses for reporting purposes.  The NIMs mark to market and acceleration of second lien losses accounted for half of RDN’s horrific Q307 loss of almost $9 per share.  Besides the mark to market for the FG credit exposure, the rest of the loss was the writedown of C-Bass (RDN’s subprime joint venture with MTG).
 
A little over of 1/3 of RDN’s NIM and second lien exposure has been written off and writeoffs have significantly exceeded loss experience so far.  Its equity in C-Bass has been entirely written off.  Though RDN is no longer writing non-traditional exposure such as NIMs and second liens,  RDN could face another $200M of losses on its non-traditional risk in 2008.  These losses will be experienced over time and offset by premium collected in the traditional MI business.  The NIM writedowns and pulling forward of losses in Q307 suggest that further losses in non-traditional risk will be manageable.
 
Why is RDN’s go-forward MI book more favorable than its peers?
Although this fact has gone unnoticed in the current mortgage insurance debacle, RDN significantly cutback on its Alt-A and subprime books relative to its peers in recent years.  This can be seen on page 29 of Radian’s September 5th presentation on its website.  In the Alt-A category, risk in force (RIF) / insurance in force (IIF) – roughly the % of each loan that Radian insures – has fallen from 26% (a few years ago) to 12% in the more recent vintages.  This pullback has not occurred at comps such as MTG and TGIC who have continued to offer close to 25% coverage on the recent vintages.   
 
For the MI industry, the most brutal reserve addition in the current environment is from California and Florida Alt-A exposures.  These are from high-balance loans that have spiking defaults and generally much larger average $ losses per loan.  These loans were performing well just two quarters ago and accounted for very little of the MI’s reserving activity – they have quickly grown to be the largest part of the MI industry’s reserve additions.
 
As mentioned, reserve additions in the MI business are taken as defaults happen – they are a function of average claim size, cure rate, and the default rate.  MTG’s average claim paid for Alt-A jumped from the low $40k’s per claim in H107 to $57,000 in Q307.  RDN’s average claim jumped as well, but from the $40k range to $47,000.  This corresponds with management’s claim that they pulled back on the risky Alt-A business relative to the industry.  These significantly lower claim sizes will make a difference in reserving activity going forward.
 
It is also important to note that RDN’s default rate for Alt-A was 7.5% in the most recent quarter vs above 12% for MTG.  RDN wrote less $ coverage per loan and less ARM coverage than the industry in recent years in the Alt A segment.  This is the exposure that is wreaking havoc on the MI business as live-in speculators are abandoning their FL and CA homes in the face of falling home prices and rate resets.
 
On the subprime side of MI, Radian was the only MI player that raised funds in the capital markets and ceded exposure (through Smart Home).  This will also have the effect of mitigating RDN’s non-prime losses.
 
RDN MI Losses Stress Test
Slide 29 of RDN’s September 5th presentation gives management’s negative case loss experience (after full policy seasoning).  One could argue that Radian’s worst case claim rates are too low on Alt-A (and perhaps subprime), but their claim rate assumptions are very negative for recent vintage prime (and well above historical loss experience on prime mortgages).
 
Although slide 29 implies a cumulative loss of $2.1B, this scenario will take many years to play out and RDN MI will continue to generate ~$1.2B per year of earned premium and investment income in its MI subsidiary.  MI book value may not grow for a while, and may go lower in the next few quarters, but there is plenty of earnings power to offset projected losses.
 
If slide 29 isn’t draconian enough for you, you can look at Goldman Sachs’ Sept 25th piece on the MI industry.  They take the model of RDN’s slide 29 and apply it to the rest of the MI players.  On page 29 of the GSCO report, they give their version of RDN’s slide 29 in 3 scenarios – scenario 1 is the best case and is essentially lifted right from RDN’s presentation.
 
It should be noted that GS’s estimate essentially puts RDN’s MI into runoff mode and greatly reduces RDN’s MI premium volume (despite the GSE’s repeated statements that they will continue to do business with any downgraded MI).  Goldman’s scenario 2 is extremely negative and elevates the claim rate on RDN’s 2005-2007 insured book to 24% from management’s current estimate of a 9% claim rate.  After all that damage, RDN still emerges in 5 years with a book value of $45 per share (with MI losses partially offset by FG earnings).  RDN’s sale of 20% of its interest in Sherman financial services in the 3rd quarter allowed it to bring in cash to prepare for such a negative scenario.
 
Something between Goldman’s Scenario 1 and Scenario 2 seems like a reasonable worst case estimate.  Overall claim rates of 24% in Scenario 2 might imply 50% subprime and 25% Alt-A delinquency rates.  This scenario is certainly possible, but it has to be politically unfeasible at a certain level.  We would likely see lower interest rates and greater government assistance to homeowners with such destruction in the home ownership rate.  Any effort to limit foreclosures greatly helps the MI business (which pays out on foreclosure).
 
Conclusion
There are no liquidity needs at RDN and there is no downgrade risk that threatens its current rate of premium collection.  Because of mark to market credit exposure at the FG, RDN will report another very negative Q407 EPS number.  However, after mark to market losses from spread widening subside in 2008 and Radian continues to collect loss offsetting premiums, investors will look ahead to Radian’s substantial remaining book value ($40 + per share) and its ability to participate in the strong pricing environment for MI and bond insurance.  Radian has a profitable bond insurance business (with extremely limited CDO and RMBS exposure) that will enable it to offset further significant losses in its MI business and participate in the strong pricing cycle ahead of us.

Catalyst

Q108 results and beyond as mark to market losses stop in the FG
Lack of downgrades, visibility of continued underwriting
Outside chance that an activist demands that MI and FG businesses be separated to reveal value
    sort by    
      Back to top