Flood Insurance Coverage of Federal Housing Administration Single-Family Homes
Extreme weather events such as storms and hurricanes have become more frequent in the United States, resulting in more than $1 trillion in damage over the past 40 years. Many insurance companies have removed flood insurance provisions from standard homeowner’s insurance policies, resulting in the rise of separate high-cost flood insurance policies. To protect families living in areas prone to flooding, the federal government established the National Flood Insurance Program (NFIP) in 1968, which designated Special Flood Hazard Areas (SFHAs), and passed the Flood Disaster Protection Act of 1973, which established the flood insurance purchase mandate for mortgage loans in those areas that originated from a federally backed or regulated lender. In March 2020, HUD initiated a study ― Flood Insurance Coverage of Federal Housing Administration Single-Family Homes ― to evaluate the flood risk exposure facing single-family homes insured by the Federal Housing Administration (FHA) and determine FHA homeowners’ compliance with the flood insurance purchase mandate from 2011 to 2019 in North Carolina and Florida. Beyond an initial check of a property’s flood risk and flood insurance status, HUD does not monitor long-run compliance, so this study fills a critical knowledge gap. This research also examines the relationship among flood insurance coverage and premiums, flood insurance claims, and loan performance, allowing HUD to examine the potential effects of increasing flood risks on its portfolio.
The first objective of this study was to determine the number of FHA-insured homes in SFHAs. With NFIP collaboration, HUD provided the researchers with data on 713,090 FHA-insured properties in Florida and 262,558 FHA-insured properties in North Carolina with mortgage insurance endorsement dates after January 1, 2011. Using ArcGIS software, researchers matched parcel and building information to each property based on the corresponding latitude and longitude coordinates. Researchers then used those coordinates to determine which parcels and buildings overlapped with SFHAs. Some properties had inaccurate coordinates, so three estimates accounted for these errors. Under the minimum or conservative estimate, the team designated properties with inaccurate coordinates as outside SFHAs, but under the maximum estimate, the team designated properties with inaccurate coordinates as inside SFHAs based on their ZIP Code. The moderate estimate deems properties inside SFHAs if their parcels or building footprints intersect with SFHAs. The research team also examined the number of single-family FHA properties that are outside but near SFHAs (within 600 meters). Properties in these areas are still likely to experience high flood risk even though they are exempt from the flood insurance mandate.
The researchers’ second objective was to examine whether the homes in SFHAs complied with the flood insurance mandate. To do so, this study merged three datasets for the first time: the FHA-insured single-family property data, the NFIP flood insurance policy data, and the NFIP flood insurance claims data. If a property fell within an SFHA in 2019, the researchers assumed that this was also the case for all prior years of the study. If a flood insurance policy was active for any amount of time in a given year, then researchers deemed it active for that year. The team assumed a claim was valid if the date of loss fell within the start and end date of the corresponding policy. The researchers also used the merged dataset to fulfill the third objective: evaluating the relationship between flood insurance coverage and premiums, claims against NFIP, and loan performance. A longitudinal data file and cross-tabulations showed these relationships for the years 2011 to 2019. Home loan default rates served as a proxy for loan performance. For each year, the study team indicated the claim rate for FHA-insured properties that had a flood insurance policy. Finally, a logistic regression model enabled the research team to determine the impact of a claim or change in policy premium on loan performance.
The study showed that in both states, the number of FHA-insured properties increased in areas inside an SFHA and near an SFHA. In North Carolina, the number of FHA-insured properties inside SFHAs increased from 1,277 to 8,673 between 2011 and 2019, while the number of FHA-insured properties near SFHAs increased from 14,202 to 95,840. Florida witnessed similar trends during the same period, with the number of FHA-insured mortgages inside SFHAs increasing from 12,692 to 101,128 and those near SFHAs increasing from 47,298 to 399,271. Under the moderate estimate classification, in 2019, about 80 and 54 percent of properties with FHA-insured mortgages in Florida and North Carolina, respectively, were near SFHAs. By contrast, approximately 20 and 5 percent of FHA-insured mortgages were located inside SFHAs in Florida and North Carolina, respectively, using the moderate estimate classification.
Overlaying a parcel and building footprint with the SFHAs helped researchers identify whether such properties might require flood insurance. This technique, however, did not account for flood-mitigation improvements that homeowners may have undertaken that could exempt them from the insurance requirement, so the rates at which properties held flood insurance policies are only estimates of compliance. In both North Carolina and Florida, properties inside SFHAs using the conservative classification had higher insurance coverage rates than did properties near SFHAs and further away. During the study period, both states also saw an uptick in the total number of FHA-insured properties that carried flood insurance policies both inside and outside SFHAs. In Florida, the total number of properties with flood insurance policies rose from 14,391 in 2011 to 68,023 in 2019 using the conservative method, and in North Carolina, the total number increased from 586 to 3,978 during the same period using the conservative method. These findings show that the mandated purchase requirement likely influences a homeowner’s decision to purchase a policy. Consideration of risk also plays a role, because FHA-insured properties were slightly more likely to carry flood insurance even when they were within 600 feet of SFHAs compared with properties that were further away.
A logistic regression allowed the research team to examine the relationship between filing a claim and the likelihood of mortgage default in the next year. The next-year default rates are the proportion of FHA-insured single-family properties inside or outside SFHAs and with or without flood insurance in the current year that defaulted on mortgage in the next year. For both states, the likelihood of next-year default was larger for properties that had at least one claim in the current year than for properties without a claim. In North Carolina, on average, 2.61 percent of properties with flood insurance inside an SFHA defaulted on a mortgage in the next year. The proportion of North Carolina properties inside an SFHA with flood insurance coverage in 2017 that defaulted on mortgage in 2018 jumped to 4.40 percent compared with slightly more than 2 percent the previous year. The research team attributes this increase to either a delayed reaction to Hurricane Matthew in 2016 or an immediate reaction to Hurricane Florence in 2018. In Florida, on average, 3.21 percent of properties with flood insurance inside an SFHA defaulted on a mortgage in the following year. In Florida, the proportion of properties ― regardless of SFHA status or flood insurance coverage ― that defaulted on mortgage in the following year spiked in 2016, seemingly corresponding to Hurricane Irma. Unlike the fluctuations in North Carolina, the default rates in Florida were remarkably similar regardless of SFHA status or flood insurance coverage; yet, for some of the study period in both states, properties with flood insurance inside SFHAs actually had slightly higher default rates than did all other properties in the study. Further research is needed to explain these trends, especially before and after major storms.
High rates of flood insurance claims occurred in 2011, 2016, and 2018 in North Carolina and in 2017 in Florida, corresponding to hurricanes. North Carolina’s highest claim rate for the study years — 17.3 percent for properties inside an SFHA and 5.7 percent outside — occurred in 2018 because of Hurricane Florence. Florida’s claim rate was generally lower than North Carolina’s, peaking in 2017 at 2.5 percent for properties inside an SFHA and 1 percent for properties outside, likely because of Hurricane Irma. Flood insurance tends to be more expensive in North Carolina than in Florida, especially within SFHAs. The likelihood of default in the next year in North Carolina depended largely on both a high flood insurance premium and a claim. In Florida, however, having a flood insurance claim was a better predictor of default in the next year than the cost of the premium. The team indicates that more research is needed to determine whether flood insurance is adequate to cover the cost of damage and whether premiums are too costly for homeowners, causing them to default.
This study determined that the number of FHA-insured properties in high-risk areas increased in both states during the study period, as did the number of FHA-insured properties with flood insurance inside and outside SFHAs. Although more than half of the FHA-insured mortgages in both states are inside or near SFHAs, many homeowners still lack flood insurance coverage, even if it is mandated. The study’s results emphasized the need for HUD to undertake regular audits of compliance with the flood insurance mandate. Future research could expand the analysis to more states to determine wider compliance trends. Although this study is limited in determining the actual number of properties that are not in compliance, additional research could focus on property improvements, SFHA boundary changes, and other factors that influence whether a homeowner purchases flood insurance. The researchers also determined that flood insurance claims for FHA-insured single-family homes are associated with increased odds of mortgage default, with HUD liable for losses. This study represents a vital first step in understanding the risk to the FHA portfolio.