Federal Flood Insurance

Word today that President Obama has indicated his support for a Senate bill to reform and extend the National Flood Insurance Program (NFIP). Here is WSJ coverage of the proposed changes, which would reduce subsidy for second homes, commercial property, and homes with a history of flooding, reducing the cost of the program by $4.7 Billion over the next 10 years.

I wrote the words below just as Hurricane Irene was bearing down on the North Carolina coast last August, that provide an overview of how the U.S. has done flood insurance for the past 45 years or so. Many fascinating issues about public v. private insurance, aggregate cost benefit analysis and distribution of the costs and the benefits, and the role of government in society. I also did a 5 part series comparing principles embedded in the NFIP to health care. I will write more about that later.

***

Flood insurance is provided in the United States by the federal government via the National Flood Insurance Program (NFIP), in two ways. First, the government directly provides coverage for some properties. Second, the government works in concert with around 90 private insurers who function as servicing contractors. In the second case, the profits from such flood insurance are private, but the losses are socialized as private insurance companies bear none of the underwriting risk associated with this insurance. How did this come to be the case?

As this American Academy of Actuaries monograph (July 2011) on the NFIP notes (p. 30):

The National Flood Insurance Act of 1968 filled this void and provided flood insurance for properties that were otherwise uninsurable. The creation of the NFIP has meant that homes and businesses could be built in places where they almost certainly would not have been absent this insurance. That was the overriding public policy goal of the NFIP. This outcome of course has costs and benefits in both a net, as well as a distributional sense. Further, the NFIB has had a sunset date for its entire existence, meaning it must be reauthorized to continue to operate, at intervals set by Congress. The next sunset date is September 30, 2011, and neither the House nor the Senate has moved to reauthorize the program, which will cease to operate even for those who have paid premiums on that day absent congressional action.

Since its inception, three principles have guided this program:

• identification of risk and the development of maps that delineate flood risk (roughly 5 risk bands, with elevation serving as a risk adjuster within bands)
• flood plain management, designed to mitigate risk of flood
• the provision of flood insurance for uninsurable properties

A related goal has been to reduce the reliance on federal flood relief after catastrophic events.  Why could private insurance not cover such flood prone properties? Two main reasons. First, the goal of private insurance companies is to earn a profit, and they at least must earn excess premiums to cover their cost of capital (they have to pre-fund losses), whereas the U.S. government can easily borrow money to cover catastrophic flood events after they occur. Second, private companies could not compel the purchase of their product absent legislation, so would face tremendous adverse selection problems and/or no one buying their insurance. Into this situation stepped the federal government in 1968, and with a variety of modifications, it has remained the only flood insurance provider in the United States for the past four-plus decades.

Several notable aspects of the NFIB:

• can compel the purchase flood insurance if the property resides in Special Flood Hazard Area (SFHA) (since 1974)
• around 80% of covered properties are assessed full-risk premiums based on Army Corps of Engineers models; the other 20% have subsidized premiums because the covered dwelling was built prior to the identification of SFHAs in 1974
• because of subsidized premiums, in some years the NFIP will run a loss, in others a “profit”
• there are a variety of reasons that full risk premium policies may not in fact represent the expected value of future damages (the goal for such premiums)
• has a statutory limit on the aggregate amount of money that can be borrowed to pay out damage claims; this limit was $1.5 billion from 1996-2005; after Hurricanes Dennis, Katrina, Rita and Wilma the cap is now $20.725 billion; once the cap is exceeded, no claims may be paid
• Hurricane Katrina alone cost the NFIB $17.75 billion
• Individual policies are limited to $250,000 for dwelling/$100,000 for contents for a home; $500,000/$500,000 for a business
• Annual premium increases are capped at 10% annually
• risk of loss is highly concentrated; 1% of the insured properties resulted in 38% of the losses incurred by the NFIP from 1978-2004
• 60% of the single family dwellings covered by the NFIP are in the South; 38% in Florida alone
• Perceptions of regional bias are incorrect; the true risk of flood differs by region of the country (a similar risk band property on the Florida coast would have the same premium as a home by a mountain river in Colorado; there are just more such similarly-risky properties in Florida)
• Over time, more insured property has been serviced by private insurance companies, with the federal government holding all the underwriting risk
• One large private insurer ceased participation in the program in 2010 because Congress 3 times in 2010 allowed the sunset date for NFIP to lapse but then reauthorized the program retroactively. All told, Congress has done this 11 times since 2002. Even though private insurers are not liable for losses in such cases, they fear bad publicity.
• Very active Hurricane seasons in 2004, 2005 and 2008 have lead to a call for a comprehensive review of the NFIP