Do Natural Disasters Lead to More Mortgage Defaults?
Almost a year after Hurricane Sandy ravaged the New York/New Jersey coastline, one question that is difficult for the mortgage industry to answer is what would the default rate be if a hurricane with the same type of impact made landfall?
In the present day, mortgage risk can be demonstrated through either prepayment risk, credit default risk, fraud risk or collateral risk. These types of risks have been measured by the industry for a long time.
However, one type of risk that traditionally has not been directly managed but rather insured against by the borrower is natural hazard risk. Right now, we know very little systematically about the level to which mortgage portfolios are exposed to natural hazard risk.
When a property resides in a flood zone, a borrower is required to have some sort of standard homeowner’s hazard insurance which covers a variety of additional natural hazard insurance risks.
But in a recent CoreLogic marketplace report titled “The Nature of Risk,” economists Mark Fleming and Katie Dobbyn tested the hypothesis that natural hazard risks, after controlling for traditional mortgage risk characteristics, increase the propensity of mortgage default. Then, the economists translated the level of risk of natural hazard default into loan-to-value (LTV) risk space.
“Utilizing an inherent understanding of, all else equal, the different risk profiles of an 80 LTV loan versus a 95 LTV or 125 LTV loan, the propensity to default caused by natural hazards can be represented in terms of the amount one would need to adjust the LTV to account for this risk,” the economists said in the report.
In order to understand the likelihood of default due to natural hazards, the Irvine, Calif.-based firm created a natural hazard single risk score based on all types of storm hazards, such as tornado, hurricane and straight-line winds, hail, wild fires, earthquakes, storm-surge flooding, inland flooding and sink holes for a particular geographic location.
CoreLogic said each natural hazard risk is modeled individually based on the geospatial characteristics which are predictive for that specific risk. Then, the risk score is created as a weighted combination of the individual natural hazard risks accounting for the severity-weighted likelihood of the individual natural hazards at that particular location, therefore providing a national view of which regions have the greatest risk severity.
To test the hypothesis that the propensity for natural hazard risk at the property level increases mortgage default risk, the two economists built an illustrative model to predict the probability of default for a traditional borrower based on their loan and property characteristics. Within this model, borrower credit worthiness, ability to pay, equity level and loan purpose are all included to predict whether a loan will become seriously delinquent, in foreclosure or REO.
To estimate the model, 1.5 million samples of first lien loans that were active between January 1995 and June 2013 were randomly selected from a universe of prime, subprime and Alt-A loans with monthly payment history. Then, the probability of default model was estimated on this data using a logistic regression technique, CoreLogic said.
Overall, the CoreLogic economists said that when a natural disaster occurs, a mortgage holder is almost twice as likely to default in high risk areas as compared to low risk areas. Generally, the risk adjusted LTV is highest in the western coastal markets, along the Gulf Coast, as well as the southern Atlantic coasts.
For example, the greatest exposure to mortgage default risk due to natural hazards is in Miami, followed by Riverside, Calif., Palm Bay, Fla., Vallejo, Calif., and Cape Coral, Fla. Almost all of the top 20 markets that have the highest natural hazard adjusted LTV are in either Florida or California, with Charleston, S.C., Reno and Houston the three exceptions.
Another way to determine default risk implied by the natural hazard single risk score is to determine how much an LTV would have to be adjusted in order to account for a possible default. The higher the natural hazard single risk score, CoreLogic stated, the more the LTV has to be adjusted to account for the risk.
CoreLogic said the propensity to default because of natural disaster of a high natural disaster risk loan is almost double that of the propensity for a low risk loan. By comparison, the propensity to default because of lack of equity, as measured by origination LTV, of a high LTV loan is a little over double that of the propensity for a low risk loan.
“Our research demonstrates that borrowers, after controlling for their propensity to default based on traditional mortgage credit characteristics, default at a higher rate the higher the propensity of natural disaster is at the property level,” the report concluded.
“This may be because they are either under or uninsured against the natural hazards to which the property is exposed,” the economists added. “The nature of risk is that work needed to successfully manage it is never finished. Natural hazard risk is another new frontier of risk management requiring ongoing attention.”