Description

Arch Capital Group is a Bermuda-based insurer offering P&C insurance, reinsurance, and private mortgage insurance (PMI). We believe Arch currently represents the best risk-reward opportunity in the insurance space.

Arch Capital has a stellar track record of returns to shareholders. The company’s book value per share has compounded at approximately 15% over the past two decades, a record unmatched by peers. Despite this, Arch trades at 1.1x P/BV and 10.4x P/E using trailing 2019 results. We believe Arch’s future is likely to be as bright as its past; below are a few early thoughts.

First, Arch has demonstrated an exceptional ability at investing into new lines at defining junctures for the insurance industry, including developing the P&C insurance business post-September 11, the reinsurance business post-Katrina, and the PMI business post-financial crisis. We suspect we are currently at a similar juncture, rendering an investment at current prices all the more timely for long-term investors. Arch is also well diversified within its individual businesses, with each of insurance and reinsurance covering a wide spectrum of lines, providing management with further optionality on capital deployment. In fact, just this past quarter, when many insurers aimed to de-risk their portfolios, Arch grew its gross premiums written within its insurance business by about 28% YoY and within its reinsurance business by about 52% YoY.

Second, we do not know what the insurance industry will look like after this pandemic. But with a secular rise in insurance demand accelerated by COVID-19, and an unequal pick up in supply, we expect the industry to undergo a long-overdue re-pricing of risk. Moreover, we do not know for how long interest rates will be near zero. Yet, differently from other insurers, Arch’s success has come less from its investment portfolio returns (the company’s portfolio is predominantly composed of fixed income exposure, with an average duration of just over three years) and more from excellent underwriting results.

Third, the sell-side wavers on its classification of Arch as primarily a private mortgage insurer, insurer, or reinsurer. We recognize that the private mortgage insurance business has led to nearly 90% of net income over the past three years, and we expect this business to continue performing well. However, we would note that the PMI segment holds about $5bn in statutory assets, which compares favorably to Arch’s $13bn of capital and $38bn of assets. And though Arch is the largest private mortgage insurer by most standards, Arch’s loss in market capitalization since the pandemic began in earnest is tantamount to the combined market capitalization losses of mortgage insurance peers Radian, MGIC, Essent, and NMIH.

In fact, we thought it would be worth exploring some thoughts on COVID-19 and its impact on Arch Capital. For a more in-depth overview of the business, we suggest readers to look into nassau799’s and gary9’s VIC posts from about eleven and sixteen years ago, respectively. For those with access to sell-side research, we also suggest reading Joshua Shanker’s (currently at Bank of America and previously at Deutsche Bank) reports on Arch Capital.

Mortgage Insurance

We understand that the words mortgage and recession are anathema to most investors. It was not long ago that inadequate mortgage underwriting nearly collapsed our financial system. But therein lies the opportunity: we believe COVID-19 will not be nearly as harmful to Arch Capital and the industry as recent share price declines suggest.

Arch Capital entered the private mortgage insurance business in 2014 following the acquisition of PMI Mortgage Insurance and then expanded the business following the 2017 acquisition of United Guaranty from AIG. Despite being the nation’s largest private mortgage insurer by most metrics, Arch Capital (and by extension, United Guaranty) has often acted as the private mortgage industry’s innovator. For instance, Arch introduced dynamic risk-based pricing models ahead of peers, moving policy underwriting away from anachronistic pricing cards. More pertinent to the issue at hand, Arch Capital’s also introduced a series of insurance-linked notes (ILNs) under Bellemeade Re.

We will discuss the ILNs in more depth below, but at this point, we think it would be helpful to dip our toes in the peculiar world of mortgage insurance accounting. Under GAAP standards, insurers reserve for losses upon a mortgage’s delinquency, which in turn occurs after two consecutive months of mortgage nonpayment. There are two main factors influencing an insurer’s estimates for losses per delinquency: the estimated size of the claim (that is, the actual payout if the mortgage forecloses), which is also known as the severity rate when calculated as a ratio of net risk in force; and the probability that the mortgage actually becomes a claim, called the claim rate.

The other major driver of any one period’s loss reserves are the number of newly reported delinquencies. This can be further boiled into a delinquency rate, which is simply the ratio of delinquent loans over policies in force. Loans that miss two payments while under CARES Act forbearance are, for better or worse, also considered delinquent.

Historically, the severity rate has been close to 100% of net risk in force. Risk in force (RIF) refers to the insurance in force (the principal balance remaining under the covered mortgages) times the insurance coverage percentages specified under the individual insurance policies. Net RIF excludes external reinsurance coverage. However, claim rates have wavered dramatically over the years: we believe claim rates were above 20% for the industry between 2007 and 2012, though rates thereafter began a lengthy decline to about 7% by 2019.

For most legacy private mortgage insurers, the past few years have provided a bonanza of underwriting earnings. Delinquencies were low, claim rates were low, and mortgage insurers benefited from plenty of reserve releases. Yet, the music abruptly stopped with COVID-19: many a mortgage insurer lost well over 50% of their stock prices in the ensuing wave of portfolio de-risking. Arch Capital was caught in this wave, with its shares declining by nearly 57% from peak to trough.

We believe Arch Capital (and by extension the industry) is in a firmly stronger position now than ever before. Below we highlight a few technical comments from Arch’s management team from the first quarter earnings call, which helps paint the forthcoming analysis.

Francois Morin, CFO: (…) based on our current analysis, which tells us that the pandemic will represent an earnings event for our mortgage segment and not a capital event, our current expectation is that our pretax underwriting income for the entire mortgage segment will be minimal for the remainder of 2020, i.e., from the second through the fourth quarter of 2020.

Marc Grandisson, CEO: I think you were right on the forbearance comment. I don't think -- we don't think it's going to 5% or 6%. I think if you ask us, we're -- our modeling is pretty much like 15%. That's sort of what we would expect forbearance to peak at.

Grandisson: We're just assuming -- we apply a forbearance rate and then we apply conversion from forbearance to claims that we would do. And the severity is pretty close to 100% on most of those cases. So it's really more binary than you might think it is.

Grandisson: [incident assumption will be] higher, higher than normal. If normal is 7% of 1 in 14 right now, 1 in 13 on new delinquencies, I would expect it to be 1 in 7 or 1 in 6, 1 in 8. That's what sort of what the model implies. I'm not saying this is what's going to happen, but you're asking about the parameters of the deal, and this is what -- of the modeling.

We would first and foremost note that Arch is a conservatively managed insurer, which shines through not only the company’s underwriting judgment but also through its GAAP reporting decisions. Arch is also very transparent; most other private mortgage insurers omitted judgment on these metrics above.

Simultaneously, plugging in these assumptions (100% severity, mid-teen delinquency rate, mid-teen claim rate) through our bottoms-up model confirms the company’s announcement of little incremental underwriting income for the rest of the year. For reasons we describe below, we believe that Arch’s assumptions are unduly cautious for the short term, leading to results through 2020 that may be better than those originally noted. Moreover, in the long-term, we believe that private mortgage insurance will actually emerge out of this as a stronger (and vindicated) business, even if new home origination slows in the next few years due to the impending economic slowdown.

The latest release from the Mortgage Bankers Association, covering data through May 17, indicates that 8.36% of mortgages are currently in forbearance. This is a broad number and includes those mortgages securitized by Ginnie Mae, Fannie Mae, and Freddie Mac, as well as private-label securities and portfolio loans. Private mortgage insurers only provide coverage for loans securitized by Fannie Mae and Freddie Mac. These borrowers tend to have relatively high FICO scores, and as a result, are less impacted by the economic slowdown than individuals of lower credit scores.

Hence, if we focus on this subsect of loans, about 6.36% or policies in force are in forbearance. The rate of increase from week to week has also slowed: the week ended May 10 saw an increase in loans under forbearance of 0.17%, whereas the following week saw an increase of 0.11%. We believe that essentially all loans that become delinquent from now on will be represented under the CARES Act forbearance data (to not do so by any borrower would be uneconomical). However, not all loans under the forbearance plan are actually delinquent: according to the Federal Housing Finance Agency (FHFA) director Mark Calabria, about one-third of borrowers under forbearance are continuing to make their payments.

Historical data from Arch’s current PMI business is scarce, but we can turn to peer mortgage insurers for a broadly representative picture. According to Radian, peak delinquency during the crisis occurred in 2009, when delinquencies for prime mortgages (that is, those that were not alt-a or subprime) peaked at nearly 13%. The industry now only underwrites mortgage insurance for prime loans. Among these loans, the industry’s underwriting standards tightened substantially after the great financial crisis. Had today’s portfolio undergone a similar shock as 2009, we would expect the delinquency rate to be much lower than 13% for any one year. And we do not expect the pandemic to be remotely as punitive to homeowners as the financial crisis. Below are a few of our reasons.

First, we believe that the ultimate claim rate will be lower than the mid-teens percentage estimated by Arch. Comparing claim rates between the great financial crisis and now is an exercise of false equivalence: even though claim rates were over 20% then, Arch’s mortgage insurance portfolio is much safer by most standards. Peer private mortgage insurer MGIC had average FICO scores at origination of about 690 in 2007, versus about 750 in 2019.

Second, the housing environment is much healthier today than through the past crisis. Though home prices may decline due to an impending economic recession, we do not believe that the contraction will be material and long-lasting. This is important since lower home prices both decrease the recovery value on foreclosed homes and submerge some borrowers into negative equity on their high LTV mortgages, incenting these borrowers to default.

Third, the U.S. Government is likely to bend over backwards to support borrowers and prevent widespread foreclosures. With the CARES Act, mortgage borrowers can already enter into a total of 360 days of forbearance – a period in which the servicer cannot charge or collect any fees, penalties, or interest. Several states have instituted extensions to the foreclosure and eviction moratorium first introduced by the CARES Act. FHFA announced on May 13 that borrowers will be able to make up for deferred payments at the time the home is sold, refinanced, or at maturity. We believe more measures are to come to ensure borrowers do not default and face foreclosures, particularly in a presidential election year.

Fourth, pricing for PMI has increased materially since the pandemic began. According to industry commentary, pricing has increased well into the double digits, with peers such as NMIH noting rates were up between 10% and 70%. All other things the same, we expect the hard pricing market to support industry profitability into the foreseeable future.

Fifth, and if our assumptions are correct, Arch will likely experience very material reserve releases over the upcoming years. It generally takes two to three years for delinquent loans to make their way through foreclosure. While companies reserve for losses upfront, these reserves are consequently adjusted as the mortgage insurer refines its estimates of ultimate claim payments. Due to Arch’s conservativeness in booking losses, Arch experienced a favorable impact from reserve releases between 2017 and 2019 worth about 9% of its net premiums earned. We expect that the COVID-19 related delinquencies will mostly occur through 2020, and that the world will trend toward normalcy by late 2021. As the loans marked as delinquent are cured through the next ~18 months, we expect Arch to experience very material reserve releases. In fact, we would not be surprised if the reported loss ratio for the mortgage insurance segment dropped well below 20% by 2022.

Sixth, even if our assumptions are incorrect and either the delinquency or the claim rates are higher than we expect, we still do not expect the downside risk to be very material. As previously noted, Arch holds substantial reinsurance coverage through its Bellemeade Re ILNs. In order for the ILNs to attach and provide nearly $3.1bn of coverage, Arch would need to face $1.7bn of qualified losses. As a percentage of Arch’s risk in force net of quota share reinsurance (but not of excess-of-loss reinsurance), Arch’s first layer retention equals about 2.6% of net RIF, and the ILN coverage equals an additional 4.8% of net RIF of downside protection. In other words, Arch would only face an unlimited downside event in case booked losses amounted to over 7.4% of net RIF (and even then, claim payments would likely still be lower than reserved losses due to the conservative upfront accounting of losses). Such downside risk is most definitely not expected.

Moreover, there was some confusion early in the pandemic related to the capital coverage necessary for the delinquencies that occur under the forbearance plans. A regulatory capital standard set by the GSEs, called the Private Mortgage Insurer Eligibility Requirements (PMIERs), requires mortgage insurers to risk charge about 55% of the risk in force upon two to three missed mortgage payments under any one policy (the ratio goes up following subsequent missed payments). However, the PMIERS prescribes a 70% haircut to the required asset coverage for delinquencies in FEMA-declared disaster zones. Note that all 50 states in the U.S. were declared disaster zones, allowing Arch and the industry to substantially reduce the necessary capital coverage. According to Arch, it would require twelve-month delinquencies in excess of 20% just to exhaust its excess capital coverage.

We believe that Arch is well positioned to remain a dominant player, even as GAAP underwriting income suffers over the near term. We also note that Arch is a very conservatively structured mortgage insurer and that private mortgage insurance offers a critical function in the GSE securitization market, one which the government would be hard pressed not to meddle with. If push comes to shove, we would expect Arch to either be a net winner in the industry due to its superior reinsurance and excess capital coverages or to direct excess capital into more accretive opportunities within its insurance and reinsurance segments.

P&C Insurance and Reinsurance

There are a few insurance lines such as event cancellation, travel, and trade credit that are directly impacted by the pandemic. The potential payouts to these policies can often be calculated upfront, and we believe Arch has accounted for a substantial portion of these direct COVID-19 losses through its first quarter charges. However, there are two main insurance lines whose ultimate losses are less clear at this point: business interruption and workers’ compensation. We will try not to belabor much on these, and invite readers to inspect the multitude of superior legal and industry commentary available elsewhere.

Business interruption insurance covers the loss of income that a business suffers after a disaster. Most of the time, these disasters (fire, hurricane, earthquake and others) are expected to inflict direct physical property damage to qualify for policy coverage. In the U.S., most business interruption policies follow the Insurance Services Office (ISO) recommended language. Following the SARS epidemic in the early 2000’s, the ISO issued a circular that endorsed the following exclusion: “no coverage for loss or damage caused by or resulting from any virus, bacterium or other microorganism that induces or is capable of inducing physical distress, illness or disease.”

The viral exclusion should be generally sufficient to deny claim coverage for the vast majority of business interruption policyholders in the U.S. For those policies without such a viral exclusion, these policyholders would still have to prove that the virus causes direct physical property damage. This is particularly hard to prove since the SARS-CoV-2 virus is fortuitously easy to kill. We have heard of studies suggesting that two minutes of direct sunlight exposure is enough to destroy the organism; this is also why washing our hands with soap for 20 seconds is sufficient to destroy the virus. We are not lawyers or doctors, but we believe that it would be difficult for policyholders to argue that the virus remains on surfaces for more than just a few moments to qualify for physical property damage.

The primary risk with business interruption coverage in the U.S. comes from state legislators. In Connecticut, New Jersey, Ohio, New York, and others, legislators have proposed bills that would alter insurance policies retroactively to cover for business losses related to COVID-19. We believe there are constitutional problems with an ex post facto modification of existing contracts; moreover, such a heavy-handed contractual alteration could lead to an erosion of trust in the rule of law. And of course, insurers did not originally underwrite for a material and widespread risk as a pandemic. We do not believe legislators would willingly risk sinking an industry that is paramount to the functioning of our capitalist system in order to provide what essentially constitutes as corporate-sponsored fiscal policy.

So far, business interruption bills have not moved on to legislation. Insurance companies are well integrated into our existing political system and have held strong relationships with legislators for years. Insurers also benefit from being particularly active at the local level since insurers are mostly regulated by the states. But if bills do move on to becoming law, we believe that (i) the ensuing lawsuits would be fodder for our highest federal courts, (ii) the federal government would step in to bring order back to this system, and (iii) by the time this legislation makes its way through the courts, the disruption from the pandemic will have hopefully subsided meaningfully.

With all that said, Arch Capital is well protected against business interruption risk. CEO Marc Grandisson recently noted at a Wells Fargo conference that “98% of our policies have a physical damage-specific -- physical, direct physical damage coverage; and 94% plus of our policies have a virus exclusion.” According to Goldman Sachs research, Arch’s market share for statutory commercial multi-peril non-liability is 0.3% nationally – and this line represents only about 1.6% of Arch’s total U.S. direct premium written. Therefore, even though the company will likely experience some losses from business interruption, particularly on foreign policies that have already been set as IBNR, we do not expect these to be material.

The second risk we would touch on concerns workers’ compensation. This form of insurance provides wage replacement and medical benefits to employees injured in the course of employment. Even for frontline and healthcare workers, it is difficult to prove that an individual has contracted the virus from their line of work given the highly contagious nature of SARS-CoV-2.

To ease the burden of proof, over 20 states have either passed or proposed such presumption rules, often by executive decree. For instance, California governor Gavin Newson signed an order that “creates a time-limited rebuttable presumption for accessing workers’ compensation benefits.” Note that the order applies to essentially all Californians deemed essential workers.

The impact of workers’ compensation orders could range from very minimal to severe, weighing the industry with upwards of $90bn of losses according to a Willis Tower Watson study. However, the Workers’ Compensation Insurance Rating Bureau of California recently updated its estimates for the cost of COVID-19 workers’ compensation claims in the state, settling at a midpoint of $1.2bn – and that is for a region that represents 20% of the country’s workers’ compensation market. In addition, according to the National Association of Insurance Commissioners, Arch Capital only has a 1.1% share of the workers’ compensation market in the U.S., ranking the Company 21st among all insurers. As such, we believe losses will be manageable as this plays out in the near term.

Summary

The insurance industry is in for an interesting year or two. The slowdown in economic growth will likely slow new policy origination, even as insurance becomes increasingly sought and prices for policies rise. Due to some of the points mentioned above, the industry is expected to face increasing costs from litigation – all the while the U.S. president makes dubious comments as “I would like to see the insurance companies pay if they need to pay, if it’s fair. And they know what’s fair. And I know what’s fair.” And despite our optimistic take on the near and long terms, we do expect for some amount of value leakage, particularly as a number of lawsuits are sided with plaintiffs and insurers decide that the cost of litigating is higher than that of settling.

Overall, however, we expect Arch’s future to be as bright as its past. To this day, Arch’s share price is down 39% from its 52-week high, which we attribute to the market’s oft manic-depressive behavior. For those readers searching for mispriced, well-managed, and well-positioned long-term compounders, we encourage a more thorough look into Arch Capital. We believe that through a combination of high sustained ROE, meaningful growth in book value per share, and a re-rating of Arch Capital’s multiple, returns are likely to compound in the high teens to low 20’s annually with low risk over a five-year period.

Catalyst

Dissipation of COVID-19 fears; continued business growth.