MGIC (MTG) 2026 年第 3 季法說會：淨利 1.75 億美元與持續庫藏股買回

重點摘要

• MGIC Investment Corporation 2026 年第三季淨利達 1.75 億美元，或稀釋每股盈餘 0.85 美元，年化股東權益報酬率 (ROE) 為 14.2%。
• 新承保保額年增 3.6% 至 170 億美元。有效保額達 3,070 億美元，較去年同期成長 2.2%，而年度保單續存率維持在 83%。
• 損失率為 10%，其中包括 3,400 萬美元的有利準備金調整，主要受先前逾期通知案件的恢復正常還款情況優於預期所帶動。
• 逾期率上升至 2.52%，季增 15 個基點，年增 20 個基點。管理階層表示，該增幅反映出疫情後異常強勁的信用表現持續回歸正常化。
• MGIC 本季以 1.525 億美元買回 510 萬股，並發放每股 0.17 美元的股利。過去四個季度的資本回報總額相當於淨利的 122%。
• 管理階層目前預期全年營業費用將略低於先前預估的 1.9 億至 2 億美元區間下限。

關鍵財務數據

業務與營運表現

MGIC 表示，在底層信用資產品質與嚴格核保紀律的支持下，其承保投資組合表現持續良好。初期還款違約率維持在低檔，符合管理階層的近期信用展望。

該公司對新增逾期通知維持 7.5% 的初始理賠率假設。充分發展後的最終理賠率已逐漸移向約 3% 至 4%，雖高於異常偏低的水準，但仍低於初始假設。管理階層表示，若信用狀況保持強勁，此差距可能為進一步的有利準備金調整提供支撐。

投資組合產生了 6,100 萬美元的收益。固定收益資產的再投資利率持續高於投資組合約 4% 的帳面收益率，但資本回報限制了投資組合的成長及隨之帶來的投資收益增加。

MGIC 亦擴大了再保險保障範圍。該公司為符合條件的 2027 年新承保業務爭取到了最高 1.68 億美元的超額賠款保障，並修訂其 2023 年成數再保險協議，自 2027 年起將經常性成本降低約 35%；同時協議就符合條件的 2028 年業務承擔 40% 成數分保，以及 2029 年業務承擔 20% 成數分保。

管理階層估計遠期成數分保安排的稅後資金成本約為 3%。MGIC 計劃在成數分保、超額賠款及保險連結票據市場持續保持積極參與。

管理階層財測指引

管理階層預期全年營業費用將略低於先前公布的 1.9 億至 2 億美元區間下限。

該公司預期有效保費收益率將維持與近年類似的走勢，以零點幾個基點的幅度緩步下滑。管理階層指出，若四捨五入至整數基點，過去五個季度該收益率均維持在接近 38 個基點的水準。

MGIC 預期短期內保單續存率將保持穩定，而房貸保險新承保業務成長仍具挑戰性。管理階層亦預期逾期率將緩慢上升，因為投資組合逐漸集中於 2022 年及隨後年份的業務，這些業務的預期最終損失率略高於 2020 年與 2021 年的帳面資產。

風險與關注領域

信用表現正從疫情後異常強勁的水準回歸正常化。儘管近期年份業務的表現略優於管理階層最初的預期，但其最終損失率預計仍將高於 2020 年和 2021 年的業務。

管理階層表示，下次若調整 7.5% 的初始理賠率假設，調升的可能性高於調降。未來結果仍對失業率、房價以及借款人恢復正常還款的情況高度敏感。

房貸保險取消規則的潛在變動仍具不確定性。總統行政命令已指示聯邦住房金融局 (FHFA) 研議房貸保險取消的試辦計畫，但管理階層表示，具體設計、資格要求及潛在財務影響目前仍不得而知。

資本回報仍對 MGIC 的股價及其他資本用途高度敏感。管理階層表示，在股價出現不合理定價時可能會加快買回步伐，但若股價接近其預估的長期內在價值，買回節奏則可能放緩。

分析師問答環節重點

• 信用正常化：管理階層將逾期率上升主要歸因於 2022 年及隨後年份業務佔比提高。在當前條件下，預期將呈緩步上升而非急劇惡化。
• 資本回報：MGIC 表示，疲軟的內生增長、有利的再保險市場以及優於預期的業務表現，使資本回報得以達到或超過淨利水準。股票買回步伐仍將視股價水準而定。
• 準備金變動：借款人恢復正常還款的情況持續優於假設。管理階層認為在各種情境下準備金均屬充足，但若信用表現保持強勁，仍具進一步正向準備金調整的潛力。
• 房貸保險取消：過去 10 至 12 個季度中，約有一半的保單失效發生在底層貸款尚未結清的情況下。管理階層估計，每年約有 40,000 至 50,000 筆貸款由借款人主動申請取消，而根據《房屋所有權人保護法》條款取消的貸款數量也相當。
• 再保險策略：MGIC 認為遠期成數分保的定價具備吸引力，並表示再保險公司對美國房貸信用風險的長尾特性已越來越能適應。

法說會完整逐字稿

完整財報電話會議逐字稿

管理層陳述

Operator

Ladies and gentlemen, thank you for standing by, and welcome to the MGIC Investment Corporation Third Quarter 2026 Earnings Call. [Operator Instructions] I will now turn the conference over to Andrea Bolyard, Director, Corporate Communications. Please go ahead.

Andrea Bolyard

Thank you, Janelle. Good morning, and welcome, everyone. Thank you for your interest in MGIC. Joining me on the call today to discuss our results for the third quarter are Tim Mattke, Chief Executive Officer; and Nathan Colson, Chief Financial Officer.

Our press release, which contains MGIC's third quarter financial results, was issued yesterday and is available on our website at mtg.mgic.com under Newsroom. It includes additional information about our quarterly results that we will refer to during the call. It also includes a reconciliation of non-GAAP financial measures to their most comparable GAAP measures. In addition, we posted on our website a quarterly supplement that contains information pertaining to our primary risk in force and other information you may find valuable. As a reminder, from time to time, we may post information about our underwriting guidelines and other presentations or corrections to past presentations on our website.

Before getting started today, I want to remind everyone that during the course of this call, we may make comments about our expectations of the future. Actual results could differ materially from those contained in these forward-looking statements. Our 8-K and 10-Q filed yesterday include additional information about the factors that could cause actual results to differ materially from those discussed on the call today. If we make any forward-looking statements, we are not undertaking an obligation to update those statements in the future in light of subsequent developments. No one 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 our 8-K or 10-Q.

With that, I will now turn the call over to Tim.

Timothy Mattke

Thank you, Andrea, and good morning, everyone. Before I start my prepared remarks, I want to take a moment to recognize this is Andrea's first earnings call. While Andrea has been working behind the scenes since the beginning of the year in anticipation of the retirement of Dianna Higgins, this is the first time you are hearing her voice on the call. I want to say thank you to both Dianna and Andrea for working together on the transition and wish Dianna a long and prosperous retirement after 30 years with MGIC. Now back to my prepared remarks.

Our experience, financial strength and consistent focus on our customers remain foundational to our business. Our third quarter results reflect solid performance and disciplined execution across the company. In the third quarter, we generated net income of $175 million, delivering an annualized return on equity of 14.2%. We wrote $17 billion of new insurance in the third quarter, an increase of 3.6% from the third quarter of 2025. Insurance in force ended the quarter at $307 billion, up slightly in the quarter and up 2.2% from a year ago. At the end of the third quarter, annual persistency was 83%, flat compared to last quarter. Both insurance in force and the annual persistency were in line with the expectations we previously shared.

Our insured portfolio continues to perform well, reflecting the quality of the underlying credit and our underwriting discipline. Early payment defaults remain low, consistent with our near-term credit outlook. Our capital position remains strong with $5.5 billion of balance sheet capital complemented by a well-established reinsurance program. Reinsurance is an important part of how we manage risk and capital. It helps reduce risk volatility and loss volatility in stress scenarios while providing capital diversification and flexibility at attractive costs.

In addition, our capital allocation priorities remain consistent. We prioritize prudent growth in insurance in force when market opportunities support it while maintaining a disciplined approach to capital returns. Market conditions have constrained insurance in force growth in recent years. During that time, strong mortgage credit performance and our financial position have supported continued capital returns.

In the third quarter, we paid a common stock dividend of $0.17 per share. We also repurchased 5.1 million shares for $152.5 million. Over the prior 4 quarters, we returned $711 million through share repurchases and $135 million through shareholder dividends. Together, those capital returns represented a 122% payout of the net income earned over that period. Last week, the Board approved another quarterly common stock dividend of $0.17 per share. The strong financial position of both the holding company and the operating company was a key factor in that decision.

With that, let me turn it over to Nate to provide more details on our financial results for the quarter.

Nathaniel Colson

Thanks, Tim, and good morning. As Tim discussed, we had solid financial results for the third quarter. We earned net income of $0.85 per diluted share compared to $0.83 per diluted share last year. The loss ratio was 10% in the quarter, benefiting from $34 million of favorable loss reserve development in the quarter as we reestimated ultimate losses on prior delinquencies. The loss reserve development primarily reflects better-than-expected cure activity on delinquency notices received in 2025 and the first quarter of 2026. For new delinquency notices received in the third quarter, we continued to apply our initial claim rate assumption of 7.5%. Our account-based delinquency rate ended the quarter at 2.52%, up 15 basis points from the prior quarter and 20 basis points from a year ago.

The trends in the quarter were consistent with the credit normalization we have been experiencing for the past 3 years. The in-force premium yield was 38 basis points in the quarter, down less than 0.5 basis point compared with the third quarter last year. We expect the in-force premium yield to continue on a similar path to the past couple of years. This outlook reflects our near-term expectations for consistent persistency and continued challenges to MI origination growth. Investment income totaled $61 million in the third quarter and the book yield on our investment portfolio remains approximately 4%. During the quarter, reinvestment rates on our fixed income portfolio continue to exceed our book yield. However, our capital return activities have limited the growth in the investment portfolio and the resulting net income -- investment income.

Underwriting and other expenses in the quarter were $45 million, down from $50 million in the third quarter last year as we remain focused on disciplined expense management. We now expect operating expenses for the full year will be slightly below the low end of the $190 million to $200 million range we previously shared.

We have been very active in strengthening our reinsurance program while benefiting from the resulting financial flexibility. As previously announced, we executed a traditional excess of loss reinsurance transaction that provides up to $168 million of reinsurance coverage on eligible 2027 NIW. We also amended the terms of our quota share agreements covering 2023 NIW, which will reduce the ongoing cost of these agreements by approximately 35% beginning in 2027. More recently, we agreed to terms for a 40% quota share reinsurance transaction covering eligible 2028 NIW. We also agreed to terms on a separate 20% quota share covering eligible 2029 NIW, which is the furthest forward we have secured coverage in the 13-plus years of our reinsurance program.

With that, let me turn it back over to Tim.

Timothy Mattke

Thanks, Nathan. Turning from our financial results to the borrowers our product serves. We're proud of the role mortgage insurance plays in addressing affordability. For many people, the down payment remains a key barrier to homeownership, and our product helps borrowers overcome that hurdle and purchase a home sooner. Monthly affordability is an important consideration, and our product is designed to help borrowers achieve homeownership while keeping the ongoing cost of mortgage insurance low. As you likely saw, a recent executive order issued by the President includes provisions for the FHFA to explore a pilot on MI cancellation. We support efforts that focus on lowering costs for new homeowners. And as always, we look forward to working with FHFA and the GSEs on ways to make homeownership more attainable.

While our business adapts and evolves over time, our focus remains consistent. Our strong capital position gives us flexibility across market cycles and helps us manage and respond as conditions change. Just as importantly, it supports a consistent presence in the market and a daily focus on our customers. Together, our capital strength, experience and disciplined approach gives us confidence in our ability to continue delivering value to customers, homebuyers and our shareholders. With that, Janelle, let's take questions.

Operator

[Operator Instructions] Our first question comes from the line of Terry Ma with Barclays.

分析師問答

Terry Ma

Maybe just start off with credit. The delinquency rate was up about 20 basis points year-over-year as you stated. But that rate -- the year-over-year increase accelerated each of the last 4 quarters. Any underlying drivers you can kind of call out as we think about our models going forward, like is that a good jumping off point for next quarter in 2027?

Nathaniel Colson

Terry, it's Nathan. Thanks for the question. I think we do still view it broadly as just the normalization of credit performance coming off of the peak kind of high-quality credit performance coming out of the COVID period. I think increasingly, too, the book is more concentrated in the 2022 and later vintages, which are performing, I'd say, a little bit better than we expected, but we did expect those to have ultimate loss ratios that were somewhat higher than we had in the 2020 and 2021 books. So as the overall book is more concentrated in those vintages, we do expect that the general trend is going to be a slow increase in the delinquency rate because the overall book has still benefited from the low delinquency rates from the 2020 and 2021 vintages.

Terry Ma

Got it. And then maybe just on capital returns. You continue to generate excess capital while the environment remains relatively muted in terms of growth. Just how are you thinking about the pace of buybacks from here? And is there room to become more aggressive with the buyback if NIW remains around current levels?

Nathaniel Colson

Terry, it's Nathan. I'll take that one, too. I think our approach really hasn't changed much over the last couple of years. Tim said this in the prepared remarks, but our preference is to grow the core business organically to reinvest capital back into that business. The market conditions really haven't supported that. So in that environment, along with a robust reinsurance program, attractive reinsurance markets, business performing better than we expected, we've been able to support capital returns at or even above the level of net income that we're generating. And we've been comfortable with that. And if those conditions persist, we could -- I think we can continue to do that.

The other variable is obviously the share price and what we think the value of share repurchases versus other capital deployment opportunities are, and throughout this period, we feel like the share price has supported long-term value creation via share repurchases. We still feel that way today. But if there became a price dislocation, we could get more aggressive or if prices got closer to our view of longer-term intrinsic value, then the pace might become a little bit more moderate. So we remain price sensitive. I think you see that a little bit in what we disclosed this quarter, the pace of third quarter buybacks was a little bit lower, but accelerated then in October as the share price got down closer to book value that period. So we remain price sensitive and, as always, are looking for ways to generate long-term value for shareholders.

Operator

Our next question comes from Bose George from KBW.

Bose George

Sticking to credit, can you talk about where the cure rates are relative to the -- your defaulted claim assumption? And the cure rates, I mean, is -- are they just going to remain elevated until something changes in terms of HPA or unemployment or other variables there?

Nathaniel Colson

Yes. Bose, this is Nathan. I'll take that question. We've kept our initial ultimate claim rate assumption of 7.5% for some time, but I think we have been operating somewhat better than that, resulting in consistent favorable reserve development. I think underlying that, we do see we're close to fully developed new notice quarters. We have seen an increase in the ultimate claim rate on those, but it's coming off of very low levels. The lowest level in our recent history -- actually in our company's history is the second quarter of 2020, the kind of peak COVID dislocation, where ultimate claim rates were approximately 1%. We've been rising from there but still have some cushion to that 7.5% ultimate claim rate that we're putting on.

So if credit performance remains good as it has been, we do expect that there's opportunity for favorable development in the quarter. We think we're well reserved for a wide range of outcomes. So if performance -- credit performance remains strong, that could certainly result. But we do see the underlying trends are that the ultimate loss rates are moving up into the, say, low single digits, call that 3%, 4% on a fully developed basis, which, again, is exceptional performance in the kind of longer-term broader picture. But as compared to the peak COVID periods, it is increasing towards our initial estimates.

Bose George

Okay. Great. And just on your expenses, the year-over-year number, is there some of that -- is there a ceding commission benefit as well in that?

Nathaniel Colson

Yes, there is. We are -- the quota share program has gotten a little bit larger. So the amount of ceding commission is a little bit higher. I don't have that directly in front of me, but I think that's worth maybe a couple of million dollars of the $5 million overall change. But I know in the 10-Q, we do have that disclosed, the ceding commission broken out, so you can get back to the kind of before ceding commission expenses.

Operator

Our next question comes from Doug Harter with BTIG.

Douglas Harter

I'm hoping you could talk about the forward reinsurance you put on, kind of what type of cost of capital are you seeing on that and the ability to continue to kind of add kind of far out into the future.

Timothy Mattke

I'll maybe start and let Nathan chime in there, Doug. I think it's safe to say that we've been able to push out a little bit further in the future on the quota share and have done terms that we think are favorable. And I think on par with, if not better, with what we've seen historically. Otherwise, we wouldn't have the need to necessarily go out that far. I think there remains a strong desire for reinsurers to have access to U.S. mortgage credit. And so we like to work with our partners.

And I think the other part, too, is ultimately, they thought about this as a long-term line of business that they're underwriting. And so they are comfortable with the thought of going out multiple years, having seen our discipline in pricing through different environments, understanding our view on underwriting, pricing, all those different types of things. And again, I think it's a line of business that they've become much more comfortable with over time.

Again, we started -- our first deal was back in covering business in 2013. A lot of partners with us through that whole way and a lot of partners have been with us for a better part of a decade at this point. So again, I think we view it as favorable from a cost of capital standpoint, and it's great that they're willing to partner with us as we look out further. But Nate, I don't know if you want anything specific as far as exact cost or sort of view of how the transaction went itself?

Nathaniel Colson

Yes. Tim, I definitely agree with all that. I think the other thing to keep in mind, Doug, is that even the agreements that cover, say, 2026, they're not just 1-year agreements. They cover the next, say, 10 calendar years for that vintage. So I think all reinsurers kind of understand that this is a very long-tail business. So they have to get comfortable with being in this business for a long time, which makes it, I think, for some of them, makes them a little more comfortable going out even further on a new underwriting basis. But ultimately, the risk and pricing environment and return environment in 2028 and 2029 will ultimately drive the cost. But as we estimate it today, we would think about after-tax cost of capital for us around the 3% range.

Douglas Harter

Got it. And how does that compare to like the XOL that you executed in the third quarter? And just how you're thinking about the mix or the balance of your reinsurance -- sorry, yes, the mix of the reinsurance?

Nathaniel Colson

Yes, I would say -- right, the XOL, obviously, a little different structurally. We do place that on a forward basis, but not as far into the future because it is really just on the risk side, not as much on the corresponding premium side. So if there is a more stressed scenario, the quota share deals can benefit from MI pricing adjustments, whereas the excess of loss deals don't. So I think the cost of capital generally are slightly higher, but still quite attractive on the excess of loss deals. And for us, we have for some time, really tried to execute across all 3 major programs that we have related to reinsurance. That's the traditional markets with both excess of loss and quota share and also the insurance market. So it's varied over time, but I think generically about kind of 1/3, 1/3, 1/3. The one thing I would say, it ends up being over -- on a -- the 1/3 that I just mentioned is more on like a new underwriting year basis.

Over time, the program becomes a little bit more concentrated in the quota share deals because they cover first dollar loss. So they're not really subject to the delevering that both the ILN and excess of loss deals cover where the ceded layer runs away a little bit faster. So I think overall, we found, I think, great success in all 3 of those markets and want to be active in all the markets where capital for mortgage credit risk is being traded and risk. So really no plans at this point to change the approach.

Operator

Our next question comes from the line of Rowland Mayor with RBC Capital Markets.

Rowland Mayor

I understand the regulatory proposals are forward-looking. We've had some questions from investors on whether AI could accelerate voluntary detachment due to home price appreciation. I guess, could you maybe just walk me through how that process mechanically occurs, and how much capital was unlocked when older policies do voluntarily detach.

Nathaniel Colson

Rowland, it's Nathan. I mean I think generically, the way it would work today, and maybe be specific to a GSE loan. I mean, we also insure a lot of non-GSE loans and investors may have their own separate requirements. But all loans that are at least borrower-paid primary residence, like the core of what we do, they're subject to the Homeowners Protection Act. So those loans are really automatically canceled once they reach an original 78 amortization based on the original loan amount.

There is an opportunity, I think, for borrowers to request earlier cancellation due to making additional payments or at 80 LTV. I don't think we see a ton of evidence that, that feature is used very much. But on the GSE space, where borrowers are requesting cancellation for current value, I think the process is that they would reach out to their servicer. The servicer would let them know what the requirements are, which may involve getting an appraisal or a broker price opinion. The GSEs currently have LTV requirements based on current value and seasoning requirements. So not every loan would be eligible.

Those are published in their servicing guides. The GSEs today also have pay history requirements. So there needs to be good pay history, which I think is defined based on the number of, say, 30-day delinquencies in a time period or 60-day delinquencies in a longer time period. So there's an eligibility aspect, and then if all of those requirements are met, then the servicer would initiate cancellation of the MI coverage.

So I think mechanically, that's maybe generically, that's how it works today. I'm sure there are other ways that it could work, but they're really just following the published GSE guidelines for borrower-requested cancellation based on current value, whereas the Homeowners Protection Act is really focused on kind of original value.

Rowland Mayor

And then I guess as my follow-up, would you ever lower that initial claims rate assumption? I know the ultimate is moving up, but it still seems like it's well below the initial pick.

Nathaniel Colson

Yes. We've talked about it quite a bit internally. The questions come up on prior calls. I think for us, I think we understand we're setting an initial assumption today and really looking out into the future at a very wide range of possible outcomes. I think it is true that over the last 3 or 4 years, we've been down a particular path that's been quite constructive for mortgage credit performance. But that wasn't a given at any point along the way. So I think that -- I think what we've said previously and what I continue to believe is that our next change to the new notice claim rate assumption is likely to be to increase it, not to decrease it.

I think we feel very comfortable with where it is today and that it represents the uncertainty that we face going forward. And if we continue to operate in a relatively low unemployment environment, a relatively stable home price environment that it's entirely possible that, that could continue to come out as too conservative. But I think that's certainly where we're comfortable and where we expect to stay.

Operator

Our next question comes from the line of Mihir Bhatia with Bank of America.

Mihir Bhatia

I had 3 questions. Maybe I'll start with the net premium yield. If you will, as you think about 2027, some of the quota share changes you made, the 2023 quota share falls out, I guess, but other quota share becomes a larger part. Can you just -- and then you also have the new XOL, just help us with that net premium yield. Are we looking at it being higher or lower in 2027? Any kind of quantitative metrics you could give us on that, too.

Nathaniel Colson

Yes. Mihir, it's Nathan. I'd say we've typically talked about the in-force premium yield. One of the challenges with the net yield is based on the way the quota share agreements work, we get a profit commission that in low loss environments like we've been operating in, it varies dollar for dollar with the amount of losses that we're ceding. So any guidance on the net yield is really kind of guidance on near-term losses. And that's not something that we've ever really done and wouldn't feel comfortable doing just because we don't know what the future holds relative to losses. That's why we're holding so much capital. That's why we're trying to reserve prudently and conservatively.

So I think on the in-force yield, the top line of that, we do expect it's been, I'd say, grinding lower over time. It's not -- it's been coming down maybe tenths of basis point. We think that, that is likely to continue, but no real change to the path or pace of the in-force yield. And I would say the net yield, it will be much more impacted by loss expectations in 2027 or 2028, than it will -- the underlying cost of the reinsurance that we're buying.

Mihir Bhatia

Okay. I guess maybe then on the gross premium yield, I mean, do you expect that to just stay stable?

Nathaniel Colson

Yes, I think we've said follow a similar path to the last couple of years, which has been -- it's not coming down even 0.1 quarter. But I think if you look over the last couple of years, it's down maybe a basis point in the last 2 years. Maybe a little bit more than a basis point over the last 3 years. So just kind of grinding a little bit lower. But again, we're disclosing this in 0.1 of basis point. So if we were disclosing this in full basis points, each of the last 5 quarters would be 38. So again it's quite stable.

Mihir Bhatia

Okay. Maybe just turning then to the housing EO and just the noise around potential consumer-driven cancellations. I'll ask a 2-part question there. One is just -- the first part is like how much consumer-driven cancellations are there happening currently? And when I say consumer-driven, I mean, not like the contractual HPA, Homeowners Protection Act requirement or anything like that? I mean, like just like the consumer calls their servicer and says, "Hey, I need to cancel MI because the home price has increased in value." How much of that is happening currently? And to the extent this pilot is implemented or starts being put in place where you have AVMs and automatic usage. What would MTG's response be? Like do you have to adjust your pricing engine to reflect cancellation risk by loans or accounting for that a little bit more? Just trying to understand how shareholders and they are going to stay -- are going to continue to earn this mid-teens return if we start seeing a lot more cancellations or shorter policy life.

Timothy Mattke

Maybe I'll start and let Nathan maybe sort of fill in with maybe the first part of the question. I think high level, it's too early to tell, right? We don't know any designs of a pilot. We're obviously highly supportive of the administration's focus on sort of future homeowners and being able to think about affordability. But we talked through a little bit of the different levers that you can think about as far as the cancellation process that Nathan talked about earlier.

It's tough to know like where they might plug anything from a current LTV, seasoning requirements, pay history, all those things. So it's really difficult to sort of work in that abstract to say what exactly the response would be. I think it's safe to say that as FHFA looks at this and determines if they should do a pilot, and presumably working with the GSEs that we'll hopefully share some of our thoughts on what some of those triggers should be, what some of those levers that they could pull and certainly resulting ramifications. But it's really difficult, I'd say, to sit here and know what the exact response would be because there are so many different things that could sort of be done or not done, quite frankly. But Nathan, I don't know if you want to sort of chime in a little bit on some of the history we've seen.

Nathaniel Colson

Yes, happy to. I think -- if you think about MI persistency, there's kind of 2 main categories. There are situations where the loan itself is going away. That could be due to refinance or being paid in full for some other reason, someone's moving or something like that. Over the last, say, 10, 12 quarters, that's been about half of the reason for loans policies lapsing.

The other half is situations where the loan is persisting, but the MI coverage is canceling. And so that represents a ballpark of about 8 percentage points for our 1.1 million insured loans. You're talking about maybe 80,000 or 90,000 loans a year that where the policy or the -- I'm sorry, the loan is continuing, but the MI policy is canceling. And then within that group, it's about a 50-50 split as we can kind of best estimate between what would be kind of HPA-related cancellations and what we infer as borrower requested. And the way we think about that is what is happening more than 180 days prior to the HPA cancellation date. We're kind of inferring that, that is borrower requested activity.

So we're looking at on our book, maybe 40,000 or 50,000 loans a year where that's happening. And then another maybe 40,000 loans where the HPA cancellation is taking place. So absent the ability to already the ability to cancel MI and we'd see persistency rates that look closer to prepayment rates in the broader market, north of -- or persistency rates in the broader loan market north of 90%. So there's already quite a few ways to cancel loans, and that's already built into all of our models. It's not -- we don't just model prepay activity. We're also modeling both Homeowners Protection Act cancellations and the borrower behavior associated with borrower-requested cancellations. And that -- yes, it's been that way for decades.

Operator

There are no further questions. I will now turn the call back over to management for closing remarks.

Timothy Mattke

Thank you, Janelle. I just want to thank everyone for your interest in MGIC. Have a great rest of your week.

Operator

Ladies and gentlemen, this concludes today's conference call. Thank you for participating. You may now disconnect.