Skip to content

Doing more work on MGIC and Radian Group

Over the last couple of days I lightened up on my position in Radian Group and added to my position in MGIC. While I am nervous that this runs contrary to the claims of analysts (which have been getting on board the Radian train lately) I can’t find a hole in my work and cannot ignore the value I see at MGIC.

A few weeks ago I worked through a “blue sky” estimate for both Radian and MGIC. I was pretty surprised by the results. The following is not intended to be 2013 estimate or really an any-particular-time-period estimate. It is simply a look at what earnings might be once defaults “normalize” and each company’s reserve additions revert back to being those on new delinquencies only.


It is difficult for me to get past the fact that MGIC has very similar normalized earnings per share to Radian, and yet the share price is about a third of the price.

The difference in valuation is, I think, a result of a consensus that Radian has passed through to the other side, whereas MGIC has not. This is partly because Radian has been able to refinance itself and extend out most of its debt maturities, exchanging notes due 2015 for notes due 2017. The exchange offer was successful for about $200 million of a total outstanding principle amount of $250 million which leaves Radian with minimal obligations until 2017. MGIC meanwhile still has about $100 million of outstanding notes due 2015.

A second reason for the gap is because MGIC reserved more aggressively in the second and third quarter, so while Radian’s stock price was able to build up some momentum, MGIC lagged on worries of when and how much the next write-down might be.  This caught up to Radian in the fourth quarter, and they had to add to their reserves because they had been assuming too few denials would be resubmitted as claims.

While I am not sure I believe that the lower reserving by Radian reflects some sort of a shell game being played (as Oliver Davies has been suggesting in his Seeking Alpha articles) or if it’s just the fact that Radian has a larger post-2008 book than MGIC, what I do think is that MGIC is adequately reserved at this point (which I wrote about here). I would be really surprised if MGIC has to make any further additions to increase reserves on its existing delinquent book. Given the increase in cures over the past 4 months and the general downturn in delinquencies and foreclosures in Q4 (see this article from DS News) MGIC should be adequately reserved.

And while MGIC’s Q4 headline number is likely going to be ugly because of the Freddie Mac settlement, when I look ahead to Q1 I think it could be a pretty eye-popping result. If you assume the company is adequately reserved on its existing delinquent book, then they are going to be taking reserves on new delinquencies only. Running through the numbers on about 10,000 new delinquencies at $50K per delinquency and assuming a 25% delinquency to default – you get reserving of about $150mm for the quarter.

Quarterly premiums should be around $260 million, and there will be another $30 million from investment income.  On the cost side interest expense is $25 million and operating expense is $50 million. That is $290 million of revenue coming in versus total expenses of $225 million, which works out to  $65mm or 32c per share before tax, and about 23c after tax. For the quarter.

If I turn out to be right in this assessment, I don’t think the stock will continue to trade at $2.50.

The other concern that analysts have with MGIC is whether their risk to capital is simply getting too high. After the payment to Freddie the entire company’s risk-to-capital may be above 35 and at the MGIC subsidiary it will be somewhere between 30 and 35 (though it was recently pointed out to me that there could be a benefit from the aggregate loss limit cures that occurred this quarter). How high is too high?

I guess I have a few points. First, I believe that the regulators aren’t looking at risk to capital  – they are concerned with running cash flow scenarios on the book of business and looking at what the cash levels are. I have focused on this a bunch of times in the past. I think the risk to capital metric is over-emphasized, with its primary problem being that it doesn’t consider the future premiums coming in. And that is a key source of cash (it’s around $1 billion per year) that will help MGIC pay out its book.

Moreover, if I am right that the company is adequately reserved on the existing delinquent book, the risk to capital may very well have peaked.

This is how I think about MGIC’s risk to cap going forward:

First lets start with definitions: risk to capital is risk divided by capital using the following definitions:

Risk – this is quite simply the insurance in force (which is unpaid principle balance or UPB) multiplied by the percentage of that UPB that is covered by the insurance – usually around 25% in the aggregate, less any reinsurance that has been taken out.

Capital – While there are many individual items included in capital, for the benefit of both simplicity and clarity we can say that the essence of capital is the current cash and investments minus the current reserves.

Going forward the top line (risk) is going to stay flat at best. MGIC is not likely to be writing more risk than is peeling off the books due to natural causes. On the bottom line, if MGIC is adequately reserved on the existing delinquent book (again my big assumption here), new reserves are going to come exclusively from new delinquencies. We’ve already calculated that reserves for new delinquencies are going to run at $150 million per quarter and that quarterly premiums are going to more than cover that and expenses.  So just using my rough calculations from above, the capital line is actually going to start increasing by about $65mm each quarter.

That is why, in my opinion, the regulators would be silly to prevent MGIC from writing new business even if they are at a risk to capital of 30 or 35 to 1. We are getting over the hump and the slope back down should be fairly steep.  The only way I can envision that risk to capitat doesn’t begin to come back down is if MGIC is A. under-reserved on their current book (which I don’t think they are), or B. delinquencies start picking up again (which I don’t think they will).

Finally, one last consideration: When trying to predict what the regulators are going to do we need to look specifically at the individual subsidiaries, in particular the MGIC sub.  The regulator is evaluating each sub independently. The fact that the holding company has other subs in run-off isn’t going to affect their decision on the MGIC sub (which is the one, along with MIC, that is writing new business). I haven’t really dug down enough to see how the MGIC sub is doing in relation to the holding co but I’m pretty certain its somewhat better because its risk to capital is 31 vs. 34 for the company as a whole.


I feel like a bit of a lone wolf here, and that really makes me wonder if there is something I am missing. If there is though, I don’t see it, and I would love for someone to point it out for me. Because from what I can tell, MGIC is likely to post some very positive results over the coming year. And those results are in no way priced into the stock.

32 Comments Post a comment
  1. Dan #

    The folks at NMI say that in the current pricing environment they should earn a high teens ROE once they have reached scale… presumably in a few years’ time pricing will contract as capital is coming into the business (note Arch acquisition of PMI)… so say that the “normalized’ ROE for a mortgage insurer that has worked through its legacy business is 13%… what is that worth as a multiple of book? I would say 1.5x… you hardly ever see an insurance company trading at a higher multiple… So what will the book value per share be at MTG when it becomes profitable? The Macquarie guy says $1 at the end of 2013. Say it earns $1 in 2014 and again in 2015 then you’re at $3 by the end of 2015… say the pricing environment has become more competitive by then and ROEs are coming in… apply a 1.5 multiple to $3 and you’re at $4.50 in three years… now MTG’s ROE will be higher than others’ because it is so highly leveraged but there is a risk that they would feel obliged to raise equity capital which would dilute the ROE…

    February 23, 2013
    • IMO basing an analysis on the assumption that book value will be $1 per share, or $100mm at year end 2013 is pretty bearish. I would like to see how Macquarie arrives at that number. That would suggest $660mm in losses over the next 5 quarters wouldn’t it? They must feel that MGIC is quite under-reserved.

      February 23, 2013
      • Dan #

        All I’ve read are comments on the report… Macquarie says they lose $1.69 in the 4th quarter of 2102 which would take bvps from $3.31 to $1.62… then they lose another $0.62 per share over the first few quarters of this year taking bvps down to $1, which would be about $202m total book value. Losing another $125m in the first half of this year doesn’t seem too crazy…

        February 23, 2013
      • Thanks that helps. Yes I can see how they get there, I wasn’t thinking about the size of the Freddie settlement, just wrote my response off the cuff. Still, might not turn out the way Macquaries is projecting – 2013 could actually be positive to book value if cures continue this way. I also question whether basing future return estimates off of book is the right way to think about it. I mean for example MGIC has a $600mm deferred tax asset that is entirely written off right now against a valuation allowance. If things improve and the company begins to put that tax asset back on the books, has anything really changed for the company’s prospects? Yet book value could go up significantly.

        I tend to think the right way to look at it is what is the capacity of the company to write new business? And that’s where the earnings number I gave is helpful to me (albeit simplistic). If in 2016 they are writing business for lower premiums and earnings drop, it will take years before the higher return business is run off and replaced by the lower and you see that full effect.

        Thanks for the comments so far and getting me to think differently about the company.

        February 23, 2013
      • Dan #

        fair enough about the deferred taxes… and during the period before they capitalize the DTA they will probably not be reporting an income tax provision on the income statement…

        February 23, 2013
  2. Taymere #

    Thanks for your article and for sharing all your hard work for free on other stocks too. I’m very into dividends, could you point me towards which of your holdings that you expect to start paying substantial dividends?

    Credit Suisse calls for MTG BV=$1.60 at the end of 2013, if they are right then they are properly valued today. Some day they may start paying a dime a year in dividends again, but a 3.7% yield later isn’t worth the risk today IMO.

    February 23, 2013
  3. J #

    I recently upped stakes in both RDN and MTG on a down day. RDN’s risk to capital looks better but MTG is just so beaten down with the beyond-freddie-settlement quarters looking bright if delinquencies continue at present clip.

    RDN has been cutting prices and will probably gain market share..hurting profitability of the entire industry eventually. But considering that so many MI’s are out of business the overall market share these two should have (before the new goldman-backed and hayman-backed newcomers start making a dent) should be considerably larger than they previously held.

    They both have their merits at this point so an even split between the two is the position i’m holding.

    February 23, 2013
  4. I think we might be focusing on the wrong number. While $MTG’s book value has decreased dramatically to ~$668M. Most of that was due to increasing the Loan Loss Reserves, which now stands at $4B. LL reserving is a very judgement call dependent on macroeconomic factors. Most important being unemployment rate (key driver of claim rate) and housing prices (key driver of loss severity). Both of these factors have been improving lately.

    The relative valuation argument between RDN and MTG also rests on how conservative/aggressive they are reserving for losses.

    Looking purely at the current book value is missing the key dynamic here. As long as new underwriting revenue can adequately cover losses, sooner or later, the LLR will slowly be released and added back to BV. That is the driver of this potential multi-bagger.

    February 24, 2013
    • Thanks for the comment. It seems to me that the flaw with looking at book is that while book accounts for defaults there is no accounting for future premiums. To put it another way, I have heard estimates that if you took MTGs current book of business and ran it out (without writing any new business) the present value of that book would be around $10, or much higher than book, because it includes the cash coming in from premiums.

      February 25, 2013
  5. q #

    The main reason MTG is priced so low, is they are priced for bankruptcy/run-off, which seems like a possibility; regardless of what the book value is or could be. The biggest concern with MTG is the risk to capital ratio. This cannot be overstated. States can deny waivers like they did to PMI group and they simply can’t write new business. At this point, MTG is like playing with fire. Even though the gains could potentially be exponential- the risk is too great with MTG. MTG is under-reserved and have been bleeding losses based on the last couple of quarterly reports. I think this trend continues. There risk to captial ratio seems like it is only going to get worse and if that is the case, it is end game. You can’t stay afloat if you can’t write new business. Radian is the safe play and most likely be over 20 per share within 24 months. Simply look at the market share trends and the expriing legacy book. In addition, Why did you think Radian group just announced issuing more shares and convertibles? The main strategy of Radian has been to write as much new business as possible and keep the risk to captial ratio under 25:1. It seems like they have accomplished this for the near future. The key to to MI industry is the risk to captial ratio. This ensures Radian will be in business for the near future and will not be forced to waviers and shutdown. Radian will be the last man standing in this industry and will domniate the market share; that is the way I see it. Based, on the MTG report this week, MTG may be under $1 approaching zero or over $4. Why would you want to take that risk when the probability of doubling your return with Radian over the next 24 months is substially greater and less riskier. Something to consider.

    February 25, 2013
    • As I stated in the post I don’t think that the risk of capital is what the regulators are looking at and why I think MGIC’s risk to capital has peaked. My opinion is that the importance of risk to capital is actually overstated.

      I’m not going to repeat why i think that concerns about MGIC going into run-off are ill-founded. If you are interested you can read through my past posts to see my reasoning. I believe you are right that this is the disconnect between the share price of MGIC and that of Radian, but my take is that it is a disconnect to take advantage of by trading in Radian for MGIC, which I have done with some of my Radian position.

      I don’t believe the comparison to PMI is apt. PMI was well under-reserved on their current book and the regulator knew it. MGIC is not under-reserved on their current book. The housing market is also on firmer footing now then it was then.

      The point that MGIC is limited in how much new business they can write is true. MGIC would have difficulty growing their insurance in force much further without raising concerns (though I think these would probably be unfounded).

      In summary I agree that MGIC ifs priced for bankruptcy/runoff and I believe this is a mistake because when you look at the book of business and the cash it will generate it significantly exceeds the reserves of even a stressed scenario. Thus I believe the current price of MGIC presents an opportunity.

      February 25, 2013
      • q #

        Your points are valid and I wish you success in your investment with MTG. Is there an opportunity with MTG? Yes, absolutely. You could have a 5 bagger return within less than one year. But at the same time, you could lose your entire investment on this company. I simply would not like to make assumptions on how regulators will react to their numbers. I’d rather have sure singles and get on base, than swing for the fences on this one. As noted, I like to think of investing as probabilities. At this point, I think there is a high probability you make money with Radian over time with low risk. Also, if you look at Radian over the last 1-2 years- one of the reasons it was being held back is it was a victim of its own success by writing additional new business, it had to keep higher reserves. Hence, the true value of the company was not being realized in its stock price until the last couple of months. Once, Radian feels they are over-reserved within the next 24 months, and starts releasing reserves- the EPS and share price could truly be reflected towards a fair value of 20 per share incorporating future profits. I think one of the concerns analysts have with MTG is if they do write new business and gain market share, can there reserves support their new business and potential defaults? Just something else to consider. Finally, hypothetically- if MTG did go into runoff, consider how profitable Radian would become. Keep up the good analysis and opinions, best of luck to your investments. Based on your blog, it sounds like you like looking at turnaround companies that are simply undervalued. Sony and Ford might deserve a look. Cheers.

        February 25, 2013
      • Thanks for you comments

        February 25, 2013
  6. J #

    Is my math correct here? RDN increasing it’s shares outstanding by approx 22%..
    Up to 30m shares.. Currently there are roughly 134m..

    February 26, 2013
    • Sounds about right. I like RDN more after this deal. They should have plenty of capital to grow business now. If I can find the cash from somewhere I will look at adding if it falls far enough.

      February 26, 2013
      • J #

        Yeah, good move really. Downside on this dip could be 22% though… Good buying point but those holding already could see a solid loss in next day or so..

        February 26, 2013
  7. J #

    Increase float by 22% and stock is up 3%.. I’ll take it..but confused.

    February 26, 2013
    • I think its a combination of investors that are looking ahead to the new business Radian can right now that it is well-capitalized and covering shorts who realize the game is up and the worst case they hoped for is off the table. this morning I added back what I sold last week. I like that RDN will be able to take advantage of the dearth of competition

      February 26, 2013
  8. www3 #


    I’m a little confused on how you get your 10,000 new delinquencies a quarter. When you go to the MGIC website they say that new notices are 11k per month. Multiply that by 3 and you get 33k new notices a quarter. I’m kind of new to MI’s (thanks to you actually for getting me to look into them), but why do you have it droping from 33k to 10k? Thanks

    February 27, 2013
    • I got the 10,000 from last years Q1 adds to the delinquent book. So basically new delinquencies less cures plus paids (because i don’t want to adjust for reserves that came off the books). It worked out to about 10,000. I decided not to reduce that yoy even though we have seen a 20% reduction in delinquencies yoy.

      February 27, 2013
      • I don’t know how other people do it but I think you are significantly under reserving. Again I don’t really know what I’m doing but the logic makes sense to me so correct me if I’m wrong. First off I can’t tell if your formula is either new delinquencies – (cures + paids) or new delinquencies – cures + paids. New delinquencies – cures + paids gives a higher number of delinquencies so I’ll just try to address why that formula is under reserving. Let me just give an example: Lets say 20% of your loans go bad and require you to pay and the other 80% are cured. Lets say you have 10 loans and over the period 2 of them go bad and 8 of them are cured. However at the same time 10 more loans become delinquent. Using your formula new delinquencies – cures + paids you get 10 – 8 +2 = 4. Your formula says that only 4 more loans are at risk for default and on average 4*.2=.8 will default, however we clearly have 10 loans at risk for default and on average 10*.2=2 will default.

        It seems to me that you are double counting the number of cures which is mitigated a little by adding the paid loans but all and all the formula is a little funky. It seems logical to me that when a company sets a loss reserve they predict ahead of time the number of cures and defaults lets keep going with 80% and 20%. So they only put in reserves equal to 20% of the delinquent amount. Lets say a couple years down the road the cure and default rate is indeed 80% and 20%. Those rates, insomuch as they are accurate, don’t effect how much the company has to reserve on new delinquencies coming in because they already have been accounted for. Maybe MI’s reserve differently, but that is how companies reserve for bad debt in accounts receivable.

        In Q1 2012 delinquent adds were indeed around 10,000 however losses incurred were 337 million (280 million for the current year (although they constantly increase reserves for past years suggesting that they understate incurred losses often)). Although my estimate delinquency to default percentage times new delinquent loans times average amount is .25*34000*50000 is about 425 million and way overestimates the loss it is a better estimate of losses incurred and Q2 and Q3. I don’t know maybe I’m wrong, but I’m here to learn (and make money), but either way I really enjoy your blog and it’s super informative.

        February 27, 2013
      • I don’t really want to get into too much of a discussion about this because like I said I may be (not sure yet) missing the effect of aging delinquents but I think your missing the release of reserves. You have to account for the release of reserves that had been previously taken on the loans that have cured. The way you described it you would always be adding reserves and never releasing them even when loans cure.

        February 27, 2013
    • I was thinking some more about this and I think I may have done the calc incorrectly. In my calc I basically took into account new additions to the delinquent inventory and estimated that they would be reserved for at about 25% of the fully paid cost of $50,000 (which is what new 1-3 month delinquents get reserved at on average). I’m pretty sure this is right. But I now think I should have also accounted for movement of the existing delinquent book becoming older and I didn’t do that. There would be additional reserves associated with existing 1-3 month delinquencies becoming 4-11 months (and thus having greater probability going to paid claims) and the existing 4-11 month bucket going to 12+ (and again having a greater probability of becoming paid claims). I am going to have to do a bit more work on this and will post if I think I need to re-estimate the number.

      Thanks for getting me thinking about this some more.

      February 27, 2013
      • Ya I didn’t see your comment when I wrote mine. Ok I get what you are doing now. You’re using the “balance sheet” approach and I’m using the “income statement” approach (I’m taking accounting right now so it is fresh in my mind). If you are going to use a balance sheet approach to estimate Q1 loan losses it will probably take a lot of work but you have to estimate the trends in all the buckets, that might not even be possible if there are no trends. Then all you have to do is subtract the projected amount of reserves needed in Q1 from the total amount reserved in Q4 and then you have to add back in the amount paid and that should give you the loan losses for that quarter. You don’t need to worry about the number of cures or the number of new delinquencies b/c that is already factored into your balance sheet approach.

        It’ll be a b%&#& to do and it will probably be more accurate than an income statement approach, but you can always use the income statement approach too, both are GAAP accepted for setting bad debt reserves at least for AR.

        February 27, 2013
      • I just responded to your previous comment but you’ve got it – that is exactly the way I was looking at it. I didn’t think of it with the term you used but yes its basically a balance sheet approach.

        February 27, 2013
  9. J #

    4Q EPS -1.91

    Q4 Rev – 371.4

    February 28, 2013
  10. Earnings out. Q4 is bad but I guess that was expected. I wasn’t expecting the $100mm reserves addition they made for Countrywide. Risk to cap is quite a bit higher because of the extra reserves. I have to admit that while the comments from MTG and RDN in past have suggested regulators aren’t directly looking at risk to cap, the number (44.7) is getting pretty high (higher than I thought it would be b/c of countrywide). So I hope that it turns out they aren’t. I may reduce my position a little today because of this uncertainty.

    February 28, 2013
    • J #

      Seems other thinking the same. Down 15% in in premarket..

      February 28, 2013
    • J #

      MTG noticeably absent from Hayman Capital’s latest 13F after they held a 5% stake of the company.

      February 28, 2013
      • Justin #

        If I remember correctly, Hyman got out of MTG a while ago.

        February 28, 2013
  11. As per some comments left I started thinking that I needed to account for aging of the delinquent book but the more I think about I don’t think that’s true if, and this is the big if, if the existing delinquent book is adequately reserved for. If it is then you shouldn’t have to add reserves because the current reserves should reflect how that delinquent book develops over time.

    With respect to the stock price I sold about 1/5 of my position in MGIC yesterday in the morning but as I tweeted about bought back most of what I sold during the CC when the company said they have access to capital. This is a game changer IMO and you are seeing that in the stock price

    March 1, 2013
  12. Azam #

    I find these comments thought provoking. With regards to the who’s better MGIC or Radian, I hesitantly believe that Radian is the better choice. For the following reasons:

    * Earnings Power is much higher per share. Expected 2013 Risk in Force per share (after capital increase) is $250,000 for Radian vs. $140,000 for MGIC per share.

    * Legacy Insurance Book structure for Radian is better than MGIC. They will reach blue sky sooner. E.g. 2007 RIF book represents 17.5% vs. 23.1%. The 2012 RIF book represents amazing 25.1% of quality insurance vs. 13.4% for MGIC.

    * I feel that Capacity and further dilution risk through capital increases is higher for MGIC than Radian going forward.

    Please correct me if I am wrong.

    April 14, 2013

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: