MGIC Investment Corp (MTG) 2009 Q3 法說會逐字稿

(PLEASE STAND BY FOR STREAMING TEXT)

Good day, ladies and gentlemen, and welcome to the MGIC third-quarter earnings conference call. (Operator Instructions). As a reminder, this conference call is being recorded.

I would now like to introduce your host for today's presentation, Mr. Mike Zimmerman.

Good morning, and thank you for joining us this morning and for your interest in MGIC Investment Corporation. Joining me on the call today to discuss the results for the third quarter of 2009 are Chairman and CEO, Curt Culver; Executive Vice President and CFO, Mike Lauer; and Executive Vice President of Risk Management, Larry Pierzchalski.

I want to remind all participants that our earnings release for this morning, which may be accessed on MGIC's website, which is located at MCG.MGIC.com under Investor Information includes additional information about the Company's quarterly results that we will refer to during the call and includes certain non-GAAP financial measures. As we have indicated in this morning's press release, we have posted on our website supplemental information containing characteristics of our primary risk in force and flow new insurance written, as well as other information we think you will find valuable.

Also, this morning our filing -- SEC filing that contains the press release, we included a copy of the agreement with Fannie Mae.

During the course of this call we make make comments about our expectations of the future. Actual results could differ materially from those contained in these forward-looking statements. Additional information about those factors that could cause actual results to differ materially from those discussed in the call are contained in the quarterly earnings release. If the Company makes any forward-looking statements, we are not undertaking an obligation to update those statements in the future in light of subsequent developments. Further, no interested parties should rely on the fact that such guidance or forward-looking statements are current at any time other than the time of this call or the issuance of the press release.

With that, I would like to turn the call over to Curt.

Thanks, Mike, and good morning. The third quarter was clearly a disappointing one financially as we reported a net loss of $517.8 million, with a diluted loss per share of $4.17. The quarterly results reflect the impact that higher unemployment is having on the borrower's ability to staying current on their mortgage, which has caused our delinquencies and incurred losses to increase significantly.

However on the positive side, we made substantial progress on our capital plan that will allow us to continue to write business in the future, which I will discuss later. We also have retired over $300 million of debt throughout the year, and cash and investments increased to $8.7 billion at the end of the third quarter.

Relative to the quarter, we insured $4.6 billion of new business, down 22% from the first quarter. But in addition HARP refinances accounted for $450 million of insurance that is not reported as new insurance by MGIC, as we account for refinance transactions of existing insurance as a modification. I mention that in that some of our competitors count that as new insurance.

Our marketshare for the first two months of the third quarter was up slightly at 27.6% versus 26.3% in the second quarter. Persistency was flat at 85.2%. Insurance in force was down 1.5% from last quarter and now totals $216.8 billion. The average earned premium yield declined to 53.7 basis points versus 62.5 basis points last quarter. This decline is a result of refunding premiums that were paid to us on loans where coverage was rescinded during the quarter, as well as an increase in our estimate for premiums to be refunded on future rescissions.

Underwriting expenses totaled $59.1 million, which is down 4.2% from last quarter and through nine months is now down 12% from last year. Also, during the quarter we repurchased an additional $42 million of the 2011 senior notes, which leaves approximately $86 million still outstanding at quarter end.

Paid claims in the quarter were $417 million, which was up 9.7% from last quarter's $380 million, and up 26% from a year ago's $330 million. Year to date we had paid $1.15 billion versus $1.1 billion a year ago.

Losses incurred were $971 million compared to $770 million last quarter, with our loss reserve now totaling $6.3 billion. The increase in losses incurred was primarily driven by a 23,000 increase in our delinquency inventory. And the primary driver behind the net increase is the fact that the cure rate is still quite low, reflecting the impact of higher unemployment that I mentioned earlier, as well as the significant decline in property values and the delay in foreclosure activity due to past and present moratoriums. And finally, also, the servicers that are implementing the US Treasury's modification program over the last few months.

Loss mitigation was again active in the quarter as we modified loans with exposure totaling $122 million versus $109 million last quarter, which reflects the implementation of HAMP. Actually this number would be larger without HHAMP being implemented.

To date the number of delinquencies that have cured due to HAMP modification has been immaterial. However, based on the limited reporting from servicers and Fannie Mae, who serves as the Treasury's reporting arm, approximately 12,000 MGIC delinquent loans started the HAMP trial period in the third quarter, which is up from 2,500 in the second quarter. So we are starting to see significant activity, which should be beneficial to us in the long run.

Our investigation of submitted claims continues to find a great deal of fraud and misrepresentation, and as a result during the quarter we rescinded or denied $390 million of claim obligations that would have otherwise been paid or charged to a deductible. This compares to $286 million last quarter and $168 million a year ago.

As we previously discussed, we continued to pursue capital strategies in the quarter and made significant progress to address our future writing capability. As we reported last quarter, the Wisconsin OCI issued an order allowing MGIC to contribute capital to a new mortgage insurance sub of MGIC to begin writing mortgage insurance, and that we would need to obtain the GSE's approval of this new sub, which we called MIC.

I am delighted to report that we have very recently completed those discussions with Fannie Mae, who has approved MIC as an eligible mortgage insurer to be used in states where MGIC is prohibited from writing new business due to minimum capital requirements. We are now working closely with Freddie Mac to approve MIC and hope to have their approval soon. We are also working with the OCI and are optimistic that they will issue a revised order in the very near future.

The bottom line is that by reactivating MIC we will accomplish our goal of being able to support the housing market and our customers in all states by being able to write business on an uninterrupted basis through a combination of MIC and MGIC.

In addition to the progress made on our internal excess capital plan, a great deal of progress was also made on seeking regulatory relief from those states that have a minimum capital standard. To date Arizona, North Carolina and California all have given their respective insurance regulators discretion to wave the minimum capital standard. Additionally, Wisconsin has informed us that they currently have the discretion to wave their NPP requirement. We are hopeful that the remaining states with such standards will follow suit over time, but if not, we will be able to use MIC to insure business in those states.

Finally, but yet very importantly to our future financials, the underwriting results on the business booked since the underwriting changes in early 2008 continues to remain positive. Even in the face of the current economy, we expect that business to perform at loss ratios of 50% or less.

With that, Howard, let's take questions.

(Operator Instructions). Donna Halverstadt, Goldman Sachs.

Good morning, gentlemen. I had a handful of things I wanted to ask you about. In your supp you have a page where you go through the delinquency rates by vintage bucket for various types of loan categories. Could you take us through the various vintage buckets and for each of those buckets give us an idea what the cure rate looks like across the vintages?

Donna, you are talking about the supplements of page 14?

Yes.

Let me take a shot at explaining the categories.

This is Larry talking.

The reduced documentation, those would be loans which are called stated income/stated assets, reduced stock, Alt-A, [ninos], all those types of loan [partnering] and regardless of FICO.

The other categories are all foldout categories and are broken dependent upon the FICO. Subprime would be below 575, whereas A- would be below 620 to 575, and prime would be above 620.

I guess relative to the cure rate I would say -- and new notice activity, Curt alluded to in the quarter we saw a higher level of new notice activity. That was primarily from the prime full doc sector. I think our belief is that employment impacts that sector more so than Alt-A and A- and subprime. Those other categories certainly react to employment issues, but they have other issues such as poor credit and maybe not having fully documented their income and the like. So the prime full doc is -- you see more response to them with employment changes, and that is what we saw here in the quarter.

Okay. So are you suggesting that the cure rate on -- let's just take a crosscut here -- prime 2004 is as bad as it is for non-prime 2007?

No, I think the major drivers of the cure rate would be geography and vintage. And behind that comment is more what has happened to the home prices in the economy in certain vintages and geographies. For instance, the older book of business still have, hopefully, some equity (inaudible) aren't as much in the hole as some of the newer vintages. So their cure rates are higher. The newer books of business, particularly Florida and California, those cure rates are much lower. And it is more of that than whether it is a prime, and A- or an Alt-A.

I also wanted to ask a question about rescissions and denials. There was an article I saw that said there was a lawsuit ruling on October 5, wherein a US District Court issued a ruling in favor of Countrywide to the potential detriment of United Guaranty, related to misreps of underwriting standards. And this article was saying that historically MIs have paid on claims where borrowers committed the fraud, but have not paid when a loan officer or the appraiser or the real estate agent committed the fraud.

So I was curious, one, is it true that the critical issue is who committed the fraud? And do you view this ruling as a dangerous precedent for the industry? And if not, why not?

This is Curt. No, relative to the impact on us -- I think you are referring to the United Guaranty/Countrywide case -- relative to the decisions that we are making on it, the loan by loan basis, and then in that case it was more on pool -- or the policy basis. So I don't think that has implications per se relative to how we are pursuing rescissions, which as I said, are on a loan by loan basis, case by case and not a general policy in total.

Okay, so has there been any change in your outlook for future rescission activity? And as we get more of the prime loans defaulting, how do you expect rescission rates on the prime paper will look relative to the non-prime?

Well, there is a -- in the supplement there is a updated page, page 15, that shows the accumulative rescission rates, both for bulk and for flow. The flow business, although it has some reduced stock in it and some A-, it is 80% plus or so full doc prime, and you can see the trends of the rescission rates here.

I saw that. I was just curious if your future outlook has changed at all or if it is the same as it was last quarter.

I would say it is the same as last quarter. A lot of the '06 and '07 book we are finding there was fraud and misrepresentation on. And as we get claims processed with us, we are seeing an increased percentage because more of those are coming to claim.

Okay, all right. Let me ask one more question and then I will get back in line for my others. It is a holdco liquidity question. You show that your holdco liquid assets are now $97 million, and at the end of last quarter it was $124 million, so that is down about $27 million. You took a gain of $6.4 million on the 42 that you repurchased, so that implies you spent about $35.9 million, yet your liquid assets are only down $27 million. So either I did my math wrong or you found an inflow of liquid assets of about $9 million.

Yes, we had a tax refund of about $6 million or $7 million, I believe.

Did you hear that, Donna?

Donna, did you hear that?

Yes, I did hear that. That is great to hear. What is the outlook for additional tax refunds the rest of this year and in 2010?

Nothing the balance of this year, and I don't know about anything for next year. Nothing significant on that line, other than what we currently have and the payment of interest.

And other than the payment of interest, other holdco cash expenses are still de minimis, correct?

That's correct.

Okay, great. Thanks. I will get back in line.

Mike Grondahl, Northland Securities.

Two questions on fraud rescissions. Can you talk about what percent you are in looking through the '06 and the 07 vintages? Are you 20% of the way through those books or 80% of the way through those books?

Mike, are you talking -- this is Mike Zimmerman -- are you talking about how far along we are of the delinquent loans we have in inventory for the '06 and '07 and where they stand at?

Yes, just -- do you think you are still at the front of the '07 that you're looking at or are you more near the middle or the end?

I guess I'm still trying to follow what you are saying. We have -- when we go back to page 14 for 2007 we have 67,000 delinquent loans in the flow category for 2007. I would say the majority of those have not reached claims status. That is where we begin our investigation typically.

So it depends how many go to claim, Mike, before you can say how far you are along in those vintages.

But from an incurred standpoint we take that hit when the new notice of delinquency is reported to us. So on an incurred basis that would -- we would be further into the lifecycle than on a paid basis.

Got you. Okay, so there is still a lot to look at there that could be rescinded.

Potentially, yes.

We are not laying anybody off (multiple speakers).

It depends on how many claims we have, Mike, but I think that is a fair assumption.

Just on the modifications, you talked about 12,000 of your delinquent loans started a modification (multiple speakers).

With the HAMP program. Yes, we have our own internal number, which I gave to you, the $122 million, and then you also have the programs that are being modified under the government programs, which we don't count in our numbers. But we received a preliminary estimate, although we have a lot of servicers that haven't reported in, that 12,000 of our MGIC insured loans are in some form of modification through the government program, which was up sixfold from last quarter. So that is a very, very positive aspect for us and should -- that should be very positive for MGIC and our industry.

How did it grow during the quarter? Was it backend loaded? Do you see a ramp in it on a monthly basis or do you just get a quarterly run?

Let me try to clarify that 12,000. That is new starts -- HAMP starts in the quarter. I think second quarter was 2,500, so we're a little over 14,000. 2,500 a quarter ago, 12,000 more recent. So it is ramping up.

I think it was just pressure on servers to get this done by the administration. And I think Fannie and Freddie are playing a key role in making this happen. And as a result, it is starting to happen. And again I think this will be very positive for all within the industry.

Do you have any reason to believe that MI companies in general are being negatively selected by servicers or the big banks?

I will say initially I think that was the case, and it wasn't -- it was the fact that our product was higher LTV. What we heard anecdotally from servicers was that they were going after the low hanging fruit, which was borrowers with a lot of equity and that were easy to get done.

I would say our loans, being on the other end of that, weren't easy to get done. So now the low hanging fruit is taken care of and they are getting on with the majority of the work, which is really higher LTV loans, whether they are insured by our industry or not, that now is the bulk of the work being done.

And keep in mind that these rules continue to be clarified throughout the year, so the more they get clarified, the more comfortable servicers get as to knowing what they could do with all different types of loans.

Got it. Okay.

And in particular one of them is the net present value calculation that was heavily debated throughout the last several months.

Okay, thanks guys.

Nat Otis, KBW.

Just a couple of -- actually quick follow-ups on some other stuff. On that 12,000 HAMP delinquencies, do you happen to have that broken out by prime and any other segments?

Not right off hand, no.

A quick shift to the captive benefit this quarter. Any -- it seems to be a little lower than several of the past quarters. Any commentary there or anything we should look into from that standpoint?

Yes, I think there was one treaty that was terminated. That is -- I think you see that in the supporting information. I think we have approximately $40 million returned. It was a treaty that canceled. So as you remember, that business is running down the percent of business and quota share is running down so that is part of the trend there.

But that $40 million, that certainly explains it. Last quick question. I have seen some data recently that implies that '08 booked -- the delinquency rate in '08 overall has kind of accelerated over the last several months. Any comments that you guys might be seeing from that standpoint or any commentary there?

I would say the '08 book is a tale of two cities kind of thing. The first quarter I think not only for MGIC, but for most of the industry, a little better was akin to '07 because I think certainly we, and most of the industry, introduced underwriting guidelines changes towards the end of '07, early '08, and really impacted the business at the end of the quarter. So a lot of the new insurance written in '08 Q1 had characteristics similar to '07.

But since that we have seen a lot of improvement such that I think, as we said in the past, business since the first quarter of '08 we think has a loss ratio of 50% or less.

So in essence we are not seeing that after the first quarter business insured.

Okay, so you're not seeing -- because the only thing is because I -- the data that I saw implied that so far that overall book, not certainly -- not just the first quarter, but that overall book of business is performing -- already performing even worse than '05 performed at that point in that stage in their seasoning.

I think, once again, that first quarter look of business was the larger piece of the book. It is not -- it was more than 25%, let's put it that way. And because it is further into the aging curve of delinquencies you're probably seeing more of that and that is why in total it is looking that way to you. But once again, you lose that first quarter and we are showing things to be -- that 50% loss ratio.

Again, I don't know what the delinquency number you're looking at, if that includes FHA and others. They continue to insure business in markets that were quite weak and with very high LTV's. So I am sure those are exhibiting the characteristics that we had in the first quarter relative to delinquencies they got throughout the year of 2008.

Keep in mind when you're comparing to '05, at least early on, the front end curve of '05, you're talking about a pretty good economy, housing market. So you are measuring the '08 book in this housing market with the employment versus the early development '05 with the economy and whatnot at that time.

Fair enough. That's very helpful. Thank you.

Matthew Howlett, Fox-Pitt Kelton.

Just on -- getting back to the 14,500 loans you have done so far through HAMP, is that --

Just to clarify, that's begun in HAMP. They are not completed.

Not completed, that is started. Okay. Is that -- include the latest release from the Treasury Department? They put out a report on October 8 saying that they up to 490,000. Is there any -- are we missing any months? Does that include up until the latest information we have from the Treasury?

That reflects through September reporting that we received.

Okay, September reporting. They also put out a -- they also say $3.1 million of overall 60 plus day delinquencies are eligible, meaning the servicers have a long way to go. In terms of your delinquencies, the 14,000 number, that is 6% of your current delinquencies. Is there any percent you can give us in terms of what total eligibility of that relative to that 3.1 million number they have put out?

This is Mike Zimmerman. I would tell you the only -- that we felt at the highest level is about two-thirds of our insurance in force and roughly two-thirds of our delinquencies are Freddie/Fannie loans. So whether they all are eligible is based upon individual borrower characteristics, but roughly two-thirds are Freddie/Fannie loans.

But you can't equate it to the 3.1 million they have already (multiple speakers)?

No, we don't have the current information, what their current debt to income ratio is -- if they have a job at all (inaudible).

Right. I was just thinking on your delinquent inventory there is no way to go and check (multiple speakers).

Servicer details -- the servicers touch the borrowers, we don't.

Okay, needless to say it is just getting ramped up.

It is just starting.

Okay, and it has had -- and essentially if those were to reperform over three months, presumably by 12/31 you will have the information on most of that 14,000 reperformance. And if that did (multiple speakers).

I don't think that is fair.

The information relative to our claims department on HAMP timeline, it takes 60 to 90 days or longer between the initial discussion with the borrower, collection of all their financial information and the beginning of the trial period. The trial period takes another 90 days. And there is a 30 day lag for servicers to report the cures to MGIC. So the total time until cure is about six months. So I would say that is going to be three months after that.

Okay. Have you baked in -- any of your reserving assumptions, have you baked in any potential benefit from?

No.

From (inaudible). Okay.

No, that is all upside.

Great, thanks. Then on the revenue side, I understand the concept of returning some premium to the trust for rescissions. Were there any one-time hits in that revenue number this quarter?

It wasn't a one-time hit, but it was an increase in our estimate of what rescissions will be going forward and we adjusted have that. As you saw, the premiums were down because we increased our estimate for future return premiums on the higher rescission levels. We will look at that every quarter.

Right. In other words, would that imply you have already taken a hit on the revenue side and potentially on the paid loss line you could see a release of reserves, if you will?

As Larry talked about, the time lag is significant. But what happens practically is when that notice is actually rescinded then it comes out of reserves. And of course, we have built into our loss estimates a reduction in claims on giving notices anticipating some rescission activity, but we adjust that every month.

Great, okay. Then the last question on the new sub. Assuming you to get approval, how long would it take to get fully ramped up and deploy that $200 million of capital?

We think January 1, we will be able to write business in MIC.

Great, thank you.

Beth Malone, Wunderlich.

Just could you give us an idea -- looking historically you mentioned that it is pretty well known that unemployment rates and the reduction in values of houses is having an impact on delinquencies and foreclosures. Can you look at this experience that we've had in this environment and suggest whether -- is this comparable to an experience you have had in the past, or are we in totally new territory so it is hard to predict that if we start to see an improvement in unemployment, for example, should we anticipate a reduction in delinquencies?

Well, where the new territory is is that this is a national occurrence. Everything, again, in our Company's history, which is the industry's history, has been whereby the recessions happen regionally and the rest of the country is participating on a profitable basis. In this case it is happening -- both the decline in property values, as well as the employment, is happening nationally. So the impact is one we haven't seen before and hope to never see again.

But relative to the impact, as you say, clearly if unemployment starts improving that will be beneficial to us. I don't know how to generalize it any --.

Okay. Now on -- are you seeing any new competition for the new business, given that new business is probably a lot better profitable or better quality than maybe stuff that had been written before? Are there (multiple speakers)?

Yes. I think there's two new companies. (multiple speakers). One, we are for sure and another that is in talks to enter the business. So, yes, -- yes.

Neither are writing business at this point, but they are all -- it is (inaudible) and they have received licenses in 20 states and -- but they're not writing business at this time.

All right. Thank you.

[Shawn Faurot], Deutsche Bank.

Has anything changed with the Wisconsin regulator relative to the MIC sub? I know you made -- there were some comment in the earnings release about them having to reapprove, even though they obviously previously pre-approved the MIC sub back in July?

Yes, the ultimate approval that we were through relative to Fannie Mae differed from what the drop down was that was approved by the OCI. And so we have gone back to the OCI based on what we had structured with Fannie Mae, and would need a revised order on the basis of that new program.

But that is -- obviously the program should be less impactful to the regulator as it relates to the amount of capital and everything else. So it should be a relatively easy approval, is that how you guys are viewing it?

I would say we are optimistic.

That's it. Thanks guys.

Steve Stelmach, FBR Capital.

Of the 14,000 loans, is there any thoughts on what the [ultimate] cure rate is of those? Is it going to be the majority, you think? The minority? How should we think about the cure rate?

I think that we don't really know. I think we use a 50% redefault rate internally here on modifications that we -- that had been historicals that we have tracked over the many years that we have been in the business. So that has been our experience. Now these are somewhat different and -- but that is what we -- I would talk about as a rule.

See, the fundamental (multiple speakers) historic modifications that we have done typically did not lower the borrower's payments. These HAMP programs typically lower the borrower's payments, so one would think that would happen. We don't have any evidence of that yet, because very few have completed.

There's a couple of steps here. First off, the borrower has to complete this three-month trial period. If they do that successfully, as you maybe read in the papers, and then they have to complete some paperwork. And apparently there has been some problems getting all that paperwork completed, so the government is trying to simplify that. But then only after the trial period and the paperwork does the loan actually get modified and cured.

And then as Curt and Mike were saying, historically a high percentage of loan mods have redefaulted, but they haven't, I don't think, lower the payment like this plan. So there's a couple of steps that follow that. One, the borrower could fail the trial. Two, they could fail the paperwork. And three, they could redefault, but hopefully at a lower rate than what we have seen.

Yes, you would think -- that is a good point, that if they make it through to the final mod that should be at a lower redefault rate than what we have experienced historically.

Okay, that's helpful. Just to reiterate, you don't have any of this positive impact of HAMP in your reserving methodology as we --?

We do not.

You do not. So anything is just gravy at this point?

Everything is helpful.

Got you. Then can you talk about how you arrived at the $200 million for MIC?

It was looking at the states, and I think there are 17 jurisdictions that were impacted, either with risk to capital or MPP requirements. And looking at the volume of business that we would write in those states for a long period of time, so going out five years or longer, and saying how much capital would we need. And also considering that a number of those states along the way would get waivers as we discussed earlier, such as North Carolina, California and Arizona and Wisconsin could provide that would come out of MIC and then move over to MGIC.

So we think ultimately there will probably be somewhere between five and ten states that would write business within MIC, and that the $200 million is more than sufficient to serve those states for a long time.

Great, guys. Thank you.

Mike Grasher, Piper Jaffray.

I wanted to follow up with a couple of questions here on these same topics. I guess first of all on the rescissions, you mentioned all the delinquent loans outstanding and how you look at those if they come to claim. Philosophically though wouldn't those delinquent loans, or those that are tied to fraud specifically, wouldn't those likely already have gone to claim, particularly the '06, '07 vintages?

No, unfortunately -- if we knew that, we would be having much lower claim expectations going forward.

I am just thinking about the trajectory of some of those delinquencies.

I think clearly on the bulk, Mike, you've got those hits first relative to the fraud side, but not on the prime side.

It is a long timeline, so let's say '07 and early '08 the housing market economy start falling apart, starting to shake out some of some these fraudulent loans, so then they go delinquent say sometime middle of '08. For us to get a claim from that initial delinquency to the time it comes into claim to us could be a year and a half, two years -- after they go through the delinquency foreclosure, they tried to get modified, the redemption periods in the various states. And then it comes in a year and we look at it, and if it has red flags then we investigate it, and then we give the servicer a chance to --

60 days to come back on what our findings are.

So that is all in all a good two years, if not more, from delinquency to time of recession.

Okay, fair enough.

There is no question that it will be front end loaded relative to fraud and rescissions over the next couple of years. But we've got -- there is a lot unfortunately that will play out yet.

Then in terms of the modifications, did you happen to mention, just in terms of what it looks like right now, those modifications that maybe have occurred over the past year, how many of those are maybe still good versus redefaulting?

Those would be the non-HAMP ones you're talking about, and that is where we have typically seen that 50% redefault rate.

Yes, we haven't I don't think seeing a material difference from that historic number.

There has been some studies I have seen in the papers that showed some of those old mods all they did was capitalize the amount past due to the backend of the loan, not really change the payment and just put it to the backend.

Just from off the top of my head, I think 70% or thereabouts of that type of modification, where they really didn't help the borrower and the situation that led to the delinquency redefaulted. But then there was a study that showed that if the payment reduced by 5%, 10%, 15%, that 70% redefault rate became 60% to 50%. Those aren't the exact numbers, but that was the thrust of the article.

So the HAMP program reduces monthly payment to a 31% debt to income ratio, so we are hopeful we get down below 50% redefault rate. And maybe 30%, 40% is a reasonable expectation. We will have to wait to find out.

Okay, that's helpful. Then as you look at your new insurance written this quarter and considering the $200 million capital contribution, what percent of your new insurance written this quarter, I guess, came from those subject states?

I don't know what the -- if you looked at the 17 in total, about 50% of our business came from those states. But as we said, probably about 10 of those are going to be writing business under MGIC, because we have received state waivers from California, Arizona, North Carolina, Wisconsin. (multiple speakers).

The ability to receive a waiver.

The ability to receive -- I am sorry -- the ability to receive a waiver in those states. So those will be written under MGIC. So I -- what seven states or so survive, I am not quite sure, but I can tell you the $200 million is, as I described earlier, is adequate for a long time for us.

Then should we be thinking about your appetite changing in the rest of the states at all? Do we just continue to watch the risk to capital ratio or how should we be thinking about that?

We haven't been limiting our business based on risk to capital. We have been insuring every loan that we think will repay. So I don't know if our appetite will change. As markets stabilize, you will see us writing more business. But the reality is we are insuring everything we can today that we think will repay.

Okay. Fair enough. Thanks very much.

Jordan Hymowitz, Philadelphia Financial.

Most of my questions have been answered. My only one is a follow-up to Grondahl. The 12,000, can you quantify it by month how many of each -- in each month please this quarter?

We don't have that information at the table here by month, but we can try and look that up and see if we can get that back, maybe for the duration of the call, but right now it is not necessarily available. But it has been ramping up.

Can you say approximately, if it is not exact, how many were in September?

Maybe we will try and look that up as we go through the balance of the questions and answer that.

The other problem with the data is that there is probably more started in September that didn't get reported to us by September, so there is some time lag. So I guess, whatever the September number is, it is short than actual.

I guess my other question then as a follow up, since there has been no impact of modifications, but now it is actually finally ramped up, do you think in the fourth quarter there will be some -- and we don't need to quantify some -- impact to modifications in the incurred number?

No, because the reality, as we mentioned earlier, it take six months before you see the benefit, so that will be an next year item.

I think the key in the fourth quarter will be just pure delinquency levels and what is happening and what markets are they in. Whether or not some of this mitigates delinquency levels and we actually experience it, it is too early to tell.

Anything else, Jordan?

No, thank you very much.

Donna Halverstadt, Goldman Sachs.

A couple of quick ones. Are you ready to give paid claims guidance for 2010?

I don't know. We would give it a 9 or an 8, because we didn't do a very good job of it.

Okay. Another thing I was curious about, are there any assets tucked away anywhere, the sale of which could result in some cash proceeds to the holdco?

To the holding company, no, everything is done underneath MGIC. So there isn't any significant assets that are under the holding company other than MGIC. Obviously, we have some other assets that are under MGIC, but they would have to be dividended up through MGIC, the holding company.

Okay. The tax benefit of $100 million that you recorded, I thought from previous tax related disclosure there was an expectation that you wouldn't be able to take any sort of tax benefits related to the [channel vault].

Yes, we had that, but remember, we had that $290 million unrealized gain we had to book through equity, so we had to book that tax provision. So normally -- normally speaking on an operating basis we would be at a zero tax provision. So except for unrealized gains activity, they have to be recorded through equity. So you are correct.

Our guidance was correct on an operating basis, but we had that unrealized gain that we had to book through equity this quarter.

Okay. Then the other thing I wanted to ask about, in your release you made the comment that Obama & Co. have announced that they will announce their plans regarding the future of the GSEs in early 2010. For those of you who live and breathe residential mortgages 24/7 what do you expect them to come out with?

I have no idea.

Okay.

Hopefully the healthcare takes them a long time.

All right. Thank you very much.

Mike Grondahl, Northland Securities.

Just two things guys, so I can understand a little further. Are you taking any benefit in your current reserves for future fraud rescissions? Could you describe how you're doing that or how are you thinking about it?

And then could you talk a little bit -- I think you said you bought back $42 million of debt in the quarter -- how you're thinking about buying back debt in future quarters?

I wanted to give a -- the second part -- this is Mike Lauer -- the second part of the question, just obviously has to do with cash availability at the holding company. So we don't have any planned dividends out of the writing company, as I just talked -- mentioned earlier. So the only source of cash at the holding company is what is there for the most part. So any cash that we would use, or any debt repurchase that we would have, would be subject to that availability and also the market price, etc. We have been buying that at a discount obviously, diffusing that liability at a discount. That has been advantageous for us.

Right. I think you have $97 million of cash at the holding company.

Correct.

How much of that do you view as in the bucket to buy back debt, if you so choose?

Well, if you think about the cash flows for next year, we would have approximately $20 million some plus in interest payments, and so we would have some availability. But then the next question would be, what about the 11.

So needless to say, we don't have sufficient cash to buy back all the debt today and pay the interest, so there is a difference there. So we would need to find some sort -- type of dividend out of MGIC to further that. So that all would be subject to increasing the cash at the holding company to continue to repurchasing before 2011, October 2011. All right?

Then on the reserve and the benefit you may or may not be taking for future fraud rescissions?

Well, the reserve calculation takes into effect an estimate for total rescissions, albeit fraud or whatever. So in other words, the given calculation that we have today estimates how many notices will go to claim of the existing delinquency notices we have. And that factors in an assumption for future rescissions, and then obviously future severity. So it has factored in an estimate for future recessions.

Can you talk about what you factor in that conversion rate to be?

Not really, no. I wouldn't break that out. It is an estimate that we have and we reestimate that every quarter.

Okay.

(Operator Instructions). Robert Alpert, Atlas Capital.

Following up on Mike's question, while not asking for the exact estimate, can you talk about your view of the way you have changed -- what the rate of change of expectations, is it increasing?

Well, I think if you look back, one of the things that you have obviously, the availability, is we do give you the cumulative recession rates and what has been happening to those. So you can see the recession rates are going up.

And then secondly, we have experienced a significant amount of increase in delinquencies this year. And if you look at the average case basis, if you will per reserve, it hasn't gone up significantly. Obviously there has been some adjustments with respect to mitigation assumptions on rates.

Severity hasn't changed much. It has gone up a little bit, but rates have been mitigated for the recession assumptions. So it would be one way I would say that you could look at it. If you look at it historically, you would see that not withstanding the fact that we have had a significant increase in delinquencies, the average case basis hasn't changed materially.

Donna Halverstadt, Goldman Sachs.

I figured I would ask one more. This is more of an industry question. You didn't talk about it today, but at least through the second quarter 10-Qs the industry kept listing TARP as a potential source of help. Personally I have always assigned that a very, very low probability. But is that topic actually still alive? And if it is not alive, why does it keep appearing in people's 10-Q?

Well, for other (inaudible) I think you have to talk to them. But -- go ahead.

I think from our standpoint it is not necessary. So -- or relative to our Company it is not a topic. What other companies may be doing is truly their work. I know on an industry we had pursued that, not successfully. And I think the talk had died down on that as funds were used elsewhere. But relative to our Company, we are through what we have discussed. And with the hopeful approvals of Freddie Mac and the OCI on the new plan, our capital needs are taken care of.

Okay, great. Thanks again.

David Dusenbury, Dalton Greiner.

I am going to -- I want to ask a question, kind of a follow-up to this discussion on recession rates. I will probably end up butchery it, but let me see if I can get throught it a little bit here. So if I understand, you receive a notice of delinquency and the notice develops into a claim over an 18 to 24 month period.

Yes, probably 12 to 18, more likely.

So you -- as far as rescissions, you can't actively go after the claim until you see the claim at the end of that period.

Correct.

In terms of assumptions that go into your reserving process, you are now serving today -- let's say the claims you got this past quarter, or excuse me the notice of delinquencies you got this past quarter, will incorporate a higher rescission rate than let's say the reserves you put away a year ago.

In theory, yes, absolutely.

So at some point this process of seeing what I'm going to call reserve releases coming down from higher rescission rates will start to even out, because you're taking some of that benefit on the front end?

We should see some leveling off in theory as you get farther into the time cycle that Larry talked about earlier, yes.

But by and large the reserve assumptions would -- they are not forward-looking, they are based upon recent actual historic experience. So to the degree the cure rate is higher down the road or rescission rates are higher down the road, that is not incorporated. They are based on historic.

Right. That will be incremental. Got it. All right, well, thanks very much.

Shawn Faurot, Deutsche Bank.

I just wanted to follow-up on Donna's second set of questions and your response. So given that you don't think TARP is relevant for you guys at this point, and you clearly have capital needs at the holdco, am I to take your response that you guys don't see TARP as being necessary as you guys seem very confident that MGIC is going to be able to dividend cash to the holdco before that maturity -- the 2011 maturity?

Yes.

Okay, that's helpful. Thank you.

(Operator Instructions). Bruce Harting, Barclays.

Curt, in the way the auditors, or the way the formula works for reserving, if we continue with the rate we had in the quarter, and book drops to two quarters forward another $6 to $8 a share, you're just -- they are just making you shift capital, right, from equity into reserves? You could be the company that has $8 billion of reserves and $0.5 billion of capital. What is the sense behind --? I mean, at some point when do you -- when does the reserving match the paids and [loose] the capital?

Let me help you. Of course, that was one of the issues that we outlined the last 18 months was the fact that these delinquencies were rising rapidly. We were punishing statutory surplus, and as a result eroding surplus. So therefore we discussed the impact on risk to capital, and the result would be our inability to write business, so that was all those disclosures we had all year long. That is exactly what is happening.

So we had to find a way, if you will, to continue writing business. That was the idea of MIC, that we would put excess capital down in another company, continue to write business.

So you are right. That is exactly what is happening. And until we get some further clarity in delinquency development and loss development to support the case that you mentioned, we are in this particular situation. Albeit, some type of improvement in -- one of the things I remember that is happening mathematically is the '07 book was such a large book. And it is -- I think -- I don't have the numbers in front of me, but you have them in the disclosure, but the '07 delinquencies have been increasing significantly. That will turn mathematically in time. And so we would anticipate that all being -- all other things being equal that mathematically the delinquencies ought to start to [fail off] their slowdown sometime next year on the flow book. We think the bulk book is pretty close to peak.

So that could be a benefit also, notwithstanding what happens to the economy. And then secondarily as the rescission factors get farther into the reserve development, that may mitigate your incurred issue. But we are not there yet and we have to suffer along the way here in the interim.

I think we are being impacted also by delinquencies that really aren't delinquencies, but they are people that are going delinquent to get the attention of servicers, or if you will, the HAMP program or something to reduce either their payments or they are waiting for another program.

And I think when there comes to the finalization that this is it, this is the best -- the HAMP program is it, that would also lead to some cure activity as people realize, well, that it longer pays to be delinquent, let's cure the mortgage, because there is either no help or this is the best program being made available. So I think all these things come to some finalization early next year.

Ladies and gentlemen, this concludes our Q&A session. I will turn the conference back over to you, Mr. Zimmerman.

Actually it will be Mr. Culver, although Mr. Zimmerman could say it also. Thanks again for your questions and your intention -- your questions and your interest in our Company. Have a good day.

Ladies and gentlemen, thank you for participating in today's conference. This concludes the program. You may now disconnect. Everyone have a wonderful day.