Holding the Housing Market Bag, Part II

In Part I of this article, I described the evolution of the private mortgage insurance industry. This industry has evolved to the point where it now depends on continued home price appreciation to solve most of its challenges. In Part II, I will provide more detail about the inner workings of the largest player in the industry, MGIC Investment Corp. (MTG).

The August issue of Strategic Investment profiled MGIC as an attractive long-term short selling opportunity. MGIC is the leading provider of private mortgage insurance to the home mortgage lending industry in the U.S., insuring $173.4 billion worth of mortgages as of September 2006. The company competes with six other major industry players including PMI Group, Radian, Triad Guaranty, Genworth Financial, United Guaranty Residential (an affiliate of AIG), and Republic Mortgage (an affiliate of Old Republic).

Not only must MGIC price its premiums to compete with six other major private players and a reinvigorated FHA, but also "piggyback loans," or "80-10-10s," have grown to replace much of the demand for PMI in recent years. The standard example used to illustrate this is financing 80% of a home purchase with a traditional 30-year fixed mortgage, 10% with a second ARM tied to the prime rate, and 10% with a cash down payment. Since financial engineering has become a comparative advantage of the U.S., there will likely be more products that develop into significant competition for PMI in the market for down payment financing.

Threats to MGIC: Asset Impairment, Underreserving, and Low Persistency

MGIC’s assets consist primarily of a $5.2 billion bond portfolio. The portfolio’s duration is 4.9 years. Duration is a measure of the approximate change in the portfolio’s value in response to future fluctuation in interest rates (recall the inverse price/yield relationship in bond investing). So if interest rates across the yield curve increase by 1 percentage point, the portfolio’s value will decline about 4.9%; and conversely, if rates decrease by 1 percentage point, the portfolio’s value will increase about 4.9%. The company faces the risk of having to take major asset impairments if rates continue spiraling on future inflation fears.

The bond portfolio is the repository of the "float" generated by the constant inflow of insurance premiums. It is used to pay out claims on defaults, and excess investment income after paying claims can be reinvested in the portfolio or returned to shareholders via stock buybacks or dividends. Either way, the stockholders of MGIC have a residual claim on this portfolio and the income it generates.

The difficult part in calculating the size of the priority claim (held by policyholders) is that it is entirely dependent upon mortgage default rates. Under a worst-case scenario for MGIC, a huge wave of defaults would quickly consume its premium inflow, forcing the company to tap into its bond portfolio to pay all claims. The company’s insurance rating would be downgraded to a level that would effectively put it out of business. After all, an insurance company has no perceived value if it cannot pay the claims on policies it has written.

Because of the uncertainty of future mortgage default rates (and the PMI claims that ensue), MGIC management must estimate "loss reserves," a huge determinant of earnings. This makes the risk of underreserving a threat to future earnings. If default rates pick up far more than management is currently anticipating in its reserve estimates, future EPS estimates will be slashed and the stock will likely decline by at least as much as estimates.

As previously mentioned, the mortgage industry only insures against the "top" 25-30% of a default. In the case of MGIC, its coverage ratio is 26%. "Insurance in force" is the total dollar amount of mortgages on which MGIC has written policies ($173.4 billion). "Risk in force" reflects total potential exposure to claims, and is calculated by multiplying insurance in force by the coverage ratio ($44.8 billion).

This is where the importance of reserves comes into play: The year-end 2005 level of $1.1 billion in reserves against risk in force of $44.8 billion tells you that management expects a maximum of 2.5% of policyholders to default on mortgages over their insured lives. But look at the subtotal of "losses paid" from the "Loss Reserves" table; payouts for claims have increased from $434 million in 2003 to $612 million in 2005:

Here is a table showing trends in key MGIC financial metrics over the past five years. In this table, "direct primary risk" refers to the risk in force. The increasing trend in "percentage of loans in default" combined with the bearish macro housing environment is an indication management is likely underreserving:

Relatively few defaults occur in the first 12-18 months of a mortgage, so it may not be until the end of 2007 before MGIC discovers just how bad the defaults in its book of business will ultimately be. The pie chart showing insurance in force by policy year reveals that nearly 80% of the company’s current book of insured mortgages has been written since January 2003, with a large portion of it (33%) having been written around the top of the market in 2005:

The final threat to MGIC that I will address is low persistency. This is measured as the percentage of insured mortgages that remain on the books over the course of the year. Low persistency has plagued the industry over the past three years as the refinancing boom led to mass cancellations of PMI. PMI is usually no longer required once the loan-to-value ratio on a mortgage drops below 80%. The pie chart also illustrates this effect; very few insured mortgages written before 2003 remain on the books. Nearly everyone with a mortgage refinanced during 2003-2005, and in the process, many either raised their equity stake above the conforming level of 20% or entered into a more aggressive financing arrangement (i.e., extracting equity and replacing it with a "piggyback" home equity line of credit).

At its most basic level, a company’s intrinsic value is equal to the cash that shareholders can extract out of the operating business in the future. Of course, the size of the cash flow will depend mostly on how large its obligations turn out to be (in the case of MGIC, covering defaults and overhead). These future cash flows, which last as long as the business remains in operation, must be discounted back to the present at a rate high enough to compensate for the uncertainty surrounding their size and timing.

In the case of MGIC stock, its valuation support (or lack thereof) is impossible to calculate with much precision. As the market revises its expectation of highly positive future cash flows to an expectation of highly negative cash flows, selling pressure will mount. The extreme uncertainty of future cash flows is probably why PMI stocks never get high EPS multiples. This is what I expect over the next 18 months: The odds of MTG running upward like a growth stock are tiny and the odds of the market fully acknowledging the risk behind this business are high.

Barron’s Weighs in on MGIC

The stock price of MGIC has fluctuated between the high-$50s and the low-$60s since the August issue of Strategic Investment went to press. Since that time, MGIC has delivered earnings results pretty much in line with expectations. But cracks are beginning to appear in the foundation of the mortgage insurance business outlook, and mainstream media outlets are starting to report on it:

In its Nov. 20 edition, Barron’s published an article entitled "No MGIC Solution" by Jonathan Laing. Using many of the same reasons I laid out in the August issue, Laing concludes that the company is far more exposed to credit risk than Wall Street is letting on. From the article:

"There is hardly a discordant note on Wall Street these days from the chorus of analysts covering MGIC Investment, the largest U.S. mortgage insurer. Buy and Outperform ratings abound. At $61.50 a share, or 8.5 times prospective earnings, the company’s stock is unquestionably cheap, so long as the Street’s consensus profit forecasts of $6.70 a share this year and $7.20 in 2007 come through.

"Don’t bet on it. With home prices up nearly 60% in five years, and speculative construction widespread, the current downturn in the U.S. housing market promises to be no ordinary correction, least of all one that ‘Magic,’ with its insured loan portfolio of $173.4 billion, will escape unscathed. To the contrary, as sales of new and existing homes fall, dragging prices lower, foreclosures and loan default claims are bound to rise. By some estimates, MGIC could see its profits cut in half, leading to similarly steep losses in its shares."

Laing goes on to explain how the consensus opinion fails to fully appreciate the importance of a steadily rising housing market to the health of mortgage insurers:

"Until this year, property price appreciation covered a multitude of sins. A homeowner running into payment problems generally could sell his home at a high enough price to more than pay off his mortgage debt. That’s changing, however, as loans made last year or in the first half of 2006 ‘season,’ or age, as collateral values decline. As real estate speculators and subprime borrowers hit the wall, there may be no embedded property gains to bail them out.

"Compounding the problem, owners of some $800 billion of subprime mortgages will face huge increases in their monthly payments next year, as two-year teaser loans with low fixed rates reset to floating-rate loans with interest charges several percentage points above London interbank offered rate or short-term Treasury rates. These wildly popular 2-and-28s (two-year fixed and 28-year floaters) were the financing weapon of choice as the bubble expanded.

"Many borrowers could see increases of 50% or more in their monthly payments, particularly if their mortgages also change at reset from interest only to principal repayment. Without price appreciation, these borrowers no longer will be able to refinance, either at 2-and-28 or a 30-year fixed rate. Those with negative equity just might mail their keys back to their mortgage servicers…

"Adjustable-rate mortgages, or ARMs, have grown to more than 25% of MGIC’s insured risk, up from 8.6% in 1999 [emphasis added]. Of today’s total, 5.2% consists of ‘pay option’ ARMs, recent confections of the mortgage-lending industry that permit borrowers to make smaller monthly payments than required, with the unpaid portion added to principal. Option ARMs made up just 0.8% of MGIC’s insured risk at the end of 2004. Insured interest-only (I/O) debt also has grown mightily…

"There are other indicators, as well, of the growing risks residing in MGIC’s insured portfolio. At the end of the third quarter, so-called Alt.-A mortgages accounted for 17.2% of the company’s insured risk, roughly double this category’s contribution at the end of 1999. These loans are made to borrowers just below the prime-risk category, who are deemed solid enough credits to forgo verification of their claimed income and net worth…

"Then there’s 100% loan-to-value mortgage debt, which made up more than 16% of MGIC’s insured portfolio last year, compared with just 4.5% at the end of 1999. These loans provide no homeowner-equity protection to buffer MGIC from claims losses if a loan goes bad."

Laing concludes the Barron’s article by noting a respected financial services analyst’s outlook for MGIC:

"Bill Ryan, a managing director at the New York research boutique Portales Partners, sees intimations of trouble in MGIC’s third-quarter report, however. He points to a jump in the average size of paid claims in the period, to $29,600 from $26,700 in the year-ago quarter. In insurance lingo, the size of claims measures ‘severity.’…

"Assuming a 5% decline in home prices and a modest rise in MGIC’s frequency rate, to 10% from the 7-8% rates that have obtained in recent years, Ryan estimates the company’s 2007 earnings could fall 50% from analysts’ current forecasts, to around $3.50 a share. Even moderate declines in home prices historically have had an outsized impact on the severity of loan losses " [emphasis added].

Mortgage insurers must constantly add new customers from the yearly crop of new homeowners. This new premium income, added to the premium income from the existing book of insured mortgages, must be more than enough to offset the inevitable losses due to defaults in its existing book of business. While recent favorable tax legislation will widen the field of potential mortgage insurance customers, the potential time bombs in the existing book of business should remain the focus of investors in these stocks.

There’s every reason to expect both continued slowing in the crop of new homebuyers (MGIC’s source of new customers), and growth in loss frequency and loss severity. Losses in the key Midwest (Michigan, Indiana, Ohio) region will only worsen with the deteriorating auto industry employment picture. This region accounts for only 12% of MGIC’s risk in force, but produced a whopping 35% of recent default claims paid.

A rapidly cooling housing market has taken away mortgage insurers’ ability to "mitigate" the losses it must pay to the banks upon foreclosure. Loss "mitigation" usually entails working with the bank to sell the repossessed house, or working out a restructured payment plan with the homeowner in default. But these options will be effectively cut off in the worst local housing markets.

MGIC turns the concept "risk-free return" on its head. This term is often used to describe short-term Treasury bills, since there’s zero chance of the federal government ever defaulting on its debt (ignoring the principal erosion of inflation, of course). At its current price of $63, MGIC stock represents a case of "return-free risk."

Good investing,
Dan Amoss, CFA

January 9, 2007