leonard shabman is a senior fellow at Resources for the Future, a Washington-based environmental research group.
Published April 27, 2018
Who pays for property damage caused by flooding? Just one in five flood-prone residences is insured by the National Flood Insurance Program, which is the only insurer in town. Now, in the wake of the spectacular flooding caused by Hurricanes Harvey, Irma and Maria (on the heels of Hurricane Katrina and Superstorm Sandy), the program must be reauthorized this year, giving Congress a chance to redesign the program.
There are four current reform bills — three in the Senate and one in the House. For the most part, they agree on two points. First, the changes should reduce, if not eliminate, the NFIP’s debt to the U.S. Treasury, which has been building since 2004. Second, the design should be structured to increase the number of flood-insured homeowners.
Here, I offer a perspective on why and how the federal government came to dominate flood insurance, and why the key to successful reform may be in creating public-private partnerships that sell such insurance, ideally in a competitive market. However, closing the coverage gap will require a federal funding commitment, whether in the form of rules for debt forgiveness, direct cash transfers to the NFIP, or standing ready to help pay claims caused by catastrophic storms.
Why Is Flood Insurance Different?
For insurance to be a profitable business, the insurer must set premiums based on valid estimates of expected claims, which in turn requires the insurer to have adequate models and data that predict exposure. That’s fairly straightforward for, say, auto insurance, where the characteristics of the vehicles and drivers can be pinned down — and, importantly, where claims are independent of one another. No one, for example, knows for sure which car will crash, or where, or how much damage will be done. But an insurer with a solid statistical model and tens of thousands of policyholders has a pretty good idea how many claims will be made next year and how much it will cost to settle them.
By contrast, flood claims are correlated because they are linked to a single weather event that may lead to hundreds or even thousands of claims in a locality in a single year. This means that the insurer needs to predict weather — a much tougher job than predicting traffic accidents — and would need massive reserves to be able to settle claims when a flood occurs. The costs to maintain these reserves make premiums unaffordable. Indeed, the unpredictability of claims and the possibility of widespread catastrophic loss explain why Washington has become the flood insurer of last resort.
The Origin Story
The great Mississippi River floods of 1927 and 1928 left private insurers bankrupt and claimants empty-handed. After World War II, the Truman and Eisenhower administrations both proposed that the U.S. Treasury enter a partnership with private insurers in hopes of luring them back into the flood insurance business. Congress responded positively, but never followed through with appropriations to fund a program. In the late 1960s, concerned that taxpayers’ cost of federal disaster aid after floods was escalating, Congress again took on the issue.
In the 1968 legislation that created the National Flood Insurance Program, Congress authorized both a public-private risk-sharing partnership and a wholly governmental program. The partnership approach was the preferred option. The private partner would consist of an industry pool, the National Flood Insurers Association. The pool, which ultimately included about 125 companies, spread liability among pool members even if the claims were concentrated in an area where a single company had sold most of the policies.
The government partner, the Federal Insurance Administration, would design the policies sold and serviced by the private insurers, with premiums set according to the FIA’s own rating system. Each member of the pool would collect premiums and deposit the receipts (net of the agreed-upon fee for administration and profit) in the National Flood Insurance Fund. After a flood, the pool members would assess damages and pay claims from the balance in the fund.
Premiums would not be the only source of funds to pay claims. NFIA members would pledge an agreed-to amount of capital that would be available to pay claims if the NFIF’s accumulated reserves were inadequate. If the NFIF reserve and pledged capital were still insufficient to pay claims, the Treasury would provide low-interest loans to be paid back from future premium revenues. Finally, Treasury would provide a catastrophic-loss backstop. If total claims in a given time period exceeded a specified level, the Treasury would provide the extra money, with no expectation those funds would be recouped from future premiums.
Congress recognized that if flood insurance were available, but few were able to pay for it, the coverage gap would not be reduced. This meant that the premiums had to be “reasonable.” Premiums would be kept affordable by assigning premium setting to the NFIA, tight management of administrative costs and the federal provision of low-interest loans and the loss backstop.
FIA, with no initial claims experience, had to set premiums using primitive watershed and claims estimation models. Within these constraints, rating began by creating flood insurance rate maps (FIRM) for each community. The map displayed the special flood hazard area for each community, defined as the area with at least a 1 percent chance of being inundated in any year. However, the mapping process was necessarily crude, lumping properties with varying risk of inundation into a limited number of zones within the SFHA and charging those properties the same premium. Meanwhile, some properties located in an SFHA were given a premium discount if they had been built before the first rate map was created. Of course, these discounts narrowed the revenue base available to pay claims. So Congress made “equalization payments” to the pool equivalent to the foregone revenue.
Congress’s objective of making flood insurance widely available had been achieved. Nonetheless, between 1969 and 1973, relatively few property owners chose to buy coverage. So, in 1973, Congress mandated purchase of insurance for properties in hazard areas that had been financed by federally insured mortgages. At the same time, Congress required the FIA to reduce premiums on three occasions in an effort to increase purchases. This time, however, the premium revenue foregone was not offset by equalization payments.
The design details of the partnership were conceptually coherent. There’s no doubt, though: it was difficult to manage. The FIA faced daunting technical challenges in fulfilling its premium-setting responsibility. The NFIA struggled to organize and motivate members to aggressively market policies. Adding to these difficulties, the FIA and NFIA squabbled over authority and responsibility.
The flash point in the conflict: disagreement over the administrative expenses and profit margins that private insurers could add to premiums. This set the stage in the waning days of 1978 for the termination of the partnership arrangement. Housing and Urban Development concluded the government would save on expenses and do a better job of marketing flood policies by foregoing the services of private insurers.
Congress still expects premiums be “affordable.” Yet premium setting still does not reflect the possibility of catastrophic loss.
Flood Insurance Now
Between 1979 and 1982, the FIA was fully responsible for program administration and claims payments. Then, in 1983, FEMA authorized private intermediaries, nicknamed “write-your-own” companies, to issue policies, charge premiums based on NFIP rating tables and required surcharges, and — after a deduction for operating costs and profit — deposit the premiums to the NFIF. This arrangement remains; claims are now paid by FEMA, and, if the reserve is depleted, FEMA has authority to borrow from the Treasury. Note that the write-your-own companies bear none of the risk, having no responsibility to pay claims from their own reserves. Today, many of the original partnership design features remain in place. The flood insurance rate maps define both the boundaries for mandatory purchase by property owners with federally insured mortgages and the zones for setting premiums. Discounts for properties built before the maps were created continue, and are only slowly being phased out. The Treasury stands ready to loan cash to the program if reserves built from accumulated premiums are depleted. Congress still expects premiums be “affordable.” Yet premium setting still does not reflect the possibility of catastrophic loss.
In addition, Congress no longer treats the Treasury as a catastrophic loss backstop, and the government no longer makes equalization payments to offset revenue shortfalls created by discounted premiums or other requirements Congress places on the program. Two consequences have followed.
First, even before the 2017 storms, the NFIP’s debt to the Treasury had reached nearly $25 billion. Some have attributed the debt to “subsidized” premiums, especially for “pre- FIRM” properties — those built before the flood insurance rate map was initially created for the area. Others attribute it to what the NFIP calls “severe repetitive loss” properties. These properties make repeated claims that the NFIP must pay, even if the cumulative claims paid exceed the market value of the property.
In fact, the predominant source of NFIP debt consists of claims paid after large storms, beginning with Hurricane Katrina in 2005. But for catastrophic events, premium revenue would generally cover claims. In September 2017, Congress forgave $16 billion of this debt without explaining why — presumably to avoid the creation of a precedent for subsidizing flood insurance down the road.
If Congress really wanted to create a competitive market, it would first give FEMA the funds and authority to support ongoing NFIP efforts to reduce reliance on crude zone-based ratings for setting premiums..
A consequence of the failure to make equalization payments is an increase in “cross-subsidization” among NFIP policyholders. Legislation in 2012 that would have phased out pre-FIRM discounts was quickly modified in 2014 to slow the pace. At Congress’s insistence, the NFIP has even expanded the number of policyholders eligible for discounts.
That 2014 legislation codified “grandfathering,” a rating practice allowing owners to pay the premiums set at the time they built their properties, even if revised flood-risk maps call for higher premiums. Yet another premium discount is offered through FEMA’s Community Rating System. The system offers discounts to property owners if their localities adopt suggested community-wide flood-risk reduction programs, even if the individual properties in question are subject to the same risks as before adoption.
Faced with maintaining old discounts and adding new ones, the NFIP has tried to offset revenue losses with across-the-board premium increases, and in 2014 Congress required the NFIP to add across-the-board surcharges to increase revenues. The result is that some premiums plus surcharges have increased where claims exposure has not changed, while others pay discounted premiums that do not reflect the exposure they create for the NFIP.
There may be compelling policy justifications for such cross-subsidies. However, the original design called for equalization payments from the Treasury to make up for revenues lost due to discounting or other program requirements, rather than raising premiums for all.
The House reform bill would remove barriers the bill’s sponsors argue now limit the private sector’s ability to compete with the NFIP — competition they believe would lower premiums. For example, the legislation would affirm that property owners’ compliance with mandatory insurance purchase requirements can be secured with private policies as well as NFIP coverage.
Critics see the encouragement of competition as a slippery slope, worrying that cherrypicking by private insurers will deny the NFIP the revenue from cross-subsidies and leave the NFIP with only the high-risk properties the private sector refuses to insure. Lacking a plan for managing cherry-picking, the NFIP would have to increase premiums dramatically on its remaining policies or borrow from the Treasury to make up the losses.
Two features of the NFIP create the opportunity for cherry-picking. First, NFIP ratings still use a uniform water surface elevation, ignoring differences in exposure within a zone. Meanwhile, private insurers with up-to-date watershed modeling capability could set premiums based on more finely tuned estimates of flood risk — and in so doing, identify properties where they can offer lower premiums by taking advantage of in-zone cross subsidies. In addition, cross-subsidization as a revenue source has increased premiums relative to claims exposure across zones. There is some evidence of a broad cross-subsidy from inland to coastal regions, giving private sellers another opportunity to cherry-pick.
Note, too, that the NFIP is obliged to sell to everyone. By contrast, according to the proposed legislation, private insurers would be able to limit to whom and where they sell. This would allow them to diversify risk geographically and reduce catastrophic loss exposure, avoiding the need to maintain huge reserves and limit the premiums they might have to pay for reinsurance.
If Congress really wanted to create a competitive market, it would first give FEMA the funds and authority to support ongoing NFIP efforts to reduce reliance on relatively crude zone-based ratings for setting premiums. Second, it would authorize equalization payments that fully offset the revenue lost through discounts, reducing the need to cross-subsidize. These changes would allow the NFIP to reduce the cross-subsidies that make cherry-picking possible.
Equally important, Congress would acknowledge that a commitment to offering coverage to all comers at affordable premiums requires that the cost of catastrophic events be funded separately. Congress has instructed the NFIP to purchase private reinsurance for catastrophic loss. But there will be a limit on the coverage purchased — and, in any event, someone will have to pay for it.
The clean solution is to bring back the formal government commitment to limit the NFIP’s catastrophic loss exposure. Under this plan, if the total claims in any year exceed a predetermined level (after accounting for the mandated private reinsurance), Congress would forgive the debt.
Finally, if the NFIP cannot deny coverage or raise premiums on high-exposure properties, Congress should explicitly create a “high-risk pool” within the program. Each year FEMA, the administrative home of the NFIP, with creation of a separate high-risk pool, would allow the NFIP to compete with the private sector on equal footing. But in this regime, reality would still bite: competition couldn’t possibly lower premiums sufficiently to significantly reduce the coverage gap. One solution, I think, may lie in a new regime of public-private partnerships.
Back to the Future
The administration and one of the Senate bills support public-private partnerships as one way to reduce the coverage gap. The administration proposal makes reference to the original partnership system and then calls for legislation to “encourage private insurance companies to risk private capital in selling federal flood insurance in SFHAs” by “authorizing FEMA to enter into arrangements with pools or individual insurance companies.” The Senate bill would authorize layered coverage as a way to cap private exposure. It would instruct FEMA to set up risk-sharing pilot programs where private insurers assumed a first-loss position for claims below $50,000, with the NFIP operating in a secondary loss position.
No matter how you slice and dice the problem, though, flooding remains a hazard difficult to insure at premiums property owners are willing and able to pay. Insurance industry interest in expanding its offerings would increase if Congress formally accepted the role of backstopping catastrophic losses of any and all sellers of private flood insurance. This, by the way, is how the Terrorism Risk Insurance Act, passed in the wake of 9/11, allocates the risk of terrorist attacks between the public and private sectors.
There are other routes to similar ends. A national pool of insurance companies served as a private partner to the government in 1969, when most insurers were smaller than they are now. Today, as a result of consolidation, several larger insurers have no need for (or interest in) a national pool. However, the insurers most interested in offering flood coverage are small. To avoid higher administration costs, FEMA might encourage smaller companies to form a pool that would serve as the private partner offering flood insurance at the state or regional level.
Following the approach envisioned in 1968, the private partner, either a large company or a pool, would sell and service policies and be compensated for costs and allowed a profit. Premiums would be deposited in a reserve fund. But unlike the old system, the partners would bear some financial risk before they could borrow from the Treasury. The partnership would rely on some form of federal catastrophic loss backstop. Unlike the original design, moreover, responsibility for setting premiums would be shared between FEMA and private partners.
A partnership, with the private sector as the face of the program, could stimulate demand by offering a variety of coverage rather than one-size-fits-all NFIP policy. For example, some homeowners may prefer a low deductible with a cap on total payments. Others may prefer a high deductible with a high coverage limit.
The most promising way to increase purchases is to package flood coverage with private- sector homeowners’ and renters’ policies, which most have. But providing flood coverage to supplement a homeowner’s policy may need state regulatory approval. And private insurers may fear that state regulators will resist innovation — not to mention resist the size of the premiums required to make the sale of flood coverage profitable. A federal partnership might assure regulators that the private product is financially sound and that premiums are reasonable.
Finally, a partnership would fit into an initiative to offer means-tested assistance to meet Congress’s affordability goal — the most daunting hurdle. One possibility: FEMA-issued vouchers that could be redeemed when insurance is purchased — though where FEMA would get the funds to cover the costs is not clear. Another suggestion: means-tested income tax credits. That approach, however, would surely raise the issue of why this specific line of insurance warrants a voucher program.
An administratively simpler program would make use of the old approach of equalization payments. FEMA would develop a means test to determine who would qualify for premium discounts. The private partner would sell the policies at the specified discount, and FEMA would request an equalization payment in its annual appropriations request to Congress. Congress could choose to appropriate the funds or simply forgive that amount of debt to the Treasury.
* * *
Readers who have gotten this far may wonder if I’ve put the cart before the horse. Is there a compelling reason for the federal government to play an outsized role in making flood insurance both affordable and available? I think so. While it has become widely expected that Washington will (and should) come to the aid of disaster victims, the reality is that aid is both uncertain and limited. Absent insurance, people must depend on personal savings or commercial-bank borrowing for their recovery — and these financial resources are especially limited for low- and moderate-income households. The fact is that insured property owners are better able to recover, and recover more quickly, than those without insurance. And when households recover more quickly, so too will their communities and local economies.