Credit Risk Transfer bond ratings may fail to capture insurance-driven credit risk in agency mortgage pools

Over the past few years, many homeowners in the United States have seen their home insurance rates increase dramatically. The insurance crises in California and Florida have dominated headlines over the past few years, but the affordability challenge extends well beyond these two states. Homeowners in states from Iowa to North Carolina to Massachusetts to Louisiana have lost coverage, been forced into policies that include less coverage, or seen their rates increase for the same coverage.

This crisis is not going away. The number and cost of billion dollar disaster events in the United States has increased markedly since 1980, setting a record in 2023 with 28 events and a cost of over $90 billion. A recent New York Times investigation also found that insurance companies lost money in more than one third of states in 2023. In still more stark terms, DeltaTerra Capital, a leading climate and mortgage analytics firm and ICE collaborator, estimates that the ‘insurance gap’ — the gap between what insurance companies are currently charging in premiums, and the amount they would need to charge to cover their risk — is about $30 billion per year.

Credit ratings may often fail to fully capture insurance risks, even for mortgage securitizations made up of loans against high value homes with affluent borrowers. Models developed by DeltaTerra Capital suggest that despite the high credit quality of borrowers in these deals, falling home values due to rising insurance prices and insurability problems may still drive some loans to default, especially for borrowers whose mortgage amounts begin to exceed home values.

Many homeowners in the town of Orinda, California for example are currently feeling the impacts of rising insurance costs. Orinda is an affluent community, with a median household income upwards of $250,000 according to the U.S. Census’ American Community Survey in 2022 , and yet this recent article discusses the plight of some residents who face both high exposure to wildfire risk and dwindling options for home insurance. For these homeowners, the state’s insurer of last resort is often the only option, offering coverage in exchange for premiums that can often be multiples of previous rates.

Insurance risks translate to investors. Take for instance the following comparison between two securities (both are actual outstanding bonds) described in Table 1 below: Bond A is a diversified Credit Risk Transfer security issued by Fannie Mae, while Bond B, an MSRM bond, is backed by prime jumbo loans and is highly exposed to mortgages in Orinda. The securites are of similar vintage and rating — but have vastly different climate and insurance risk exposure.

The Klima Discount Factor is a modelled scenario value haircut that considers important aspects of credit risk assessment including asset appreciation response to rising premiums, loan structure and borrower metrics, and the structural enhancements and other features of each bond.

Bond B has an estimated Klima Discount Factor of -12.4% under a ‘base’ (RCP 4.5) climate scenario, compared to only -1.2% for Bond B—a difference of over an order of magnitude in the possible value decline if the market begins to expect the modelled income and asset value shocks to play out in the real world economy. The difference is even more extreme under a ‘bear’ (RCP 8.5) scenario, with a potential value decline of over 40% for Bond B and only 4.6% for Bond A.

Of course, Bond A and Bond B are just one example. The scale of potential value declines is similarly large for many other securities (and often unexpected, given their ratings). The Klima Discount Factors provided by ICE Climate Credit Analytics and DeltaTerra Capital fill this information gap, and allow investors to quickly assess potential discounts due to underappreciated value risks for exposed tranches.