Moving the Needle on Closing the Flood Insurance Gap

Flood insurance is a vital element of household and community resilience. It provides funding to pay the costs of repair and rebuilding without the need to draw down savings, divert consumption, take on debt, or rely on often insufficient and delayed assistance from the federal government. Yet, many households at risk of flooding remain uninsured. For example, less than 25% of the buildings inundated by Hurricanes Harvey, Sandy, and Irma had flood coverage.

In October 2018, the Wharton Risk Center’s Policy Incubator hosted a workshop designed to evaluate policy options for expanding the number of people with flood insurance in the United States, particularly those of low and moderate income. The workshop brought together roughly 50 experts across sectors and disciplines, including representatives from FEMA, the (re)insurance industry, lending institutions, state departments of insurance, non-profit organizations, and research institutions. Participants were divided into work groups, each focusing on different topics, and discussed policy options to help close the flood insurance gap.

The workshop generated active and wide-ranging discussion. Participants agreed that the impact of many policies considered to date may be marginal, perhaps only slightly increasing the number with flood coverage. However, participants also identified a number of ideas that would involve larger, structural changes that they believed could help substantially increase the number of households with flood insurance. In this brief, we present seven approaches that workshop participants felt had the potential to generate more substantial increases in take-up rates across the country.

FEMA has developed a moonshot goal of doubling the number structures with flood insurance in the United States by 2022. These policies have the possibility to contribute meaningfully to that goal. All, however, would require further refinement and development. We discuss the basic policy design of each proposal, as well as the challenges that would need to be addressed before implementation.

2.1. Flood Insurance Opt-Out

Previous research in behavioral economics has found that, in many circumstances, individuals tend to stick with default options. The “opt-out” proposal is predicated on the assumption that this tendency might apply in the case of flood insurance, as well. If so, making flood coverage a default part of homeowners policies (or coupling homeowners insurance to an NFIP policy), while still giving individuals the option to decline coverage if they chose (i.e. opt-out), might lead more people to purchase flood insurance. While the notion of increasing take-up rates through this feature of choice architecture is compelling, there are many open questions about the efficacy and design of an opt-out policy as it relates to flood insurance that would need to be addressed before it could be piloted or implemented.

Notably, most prior studies of opt-out designs have been for choices that are not financially costly for the individual. Flood insurance, particularly in higher risk areas, can be an expensive purchase. Given this, it is possible that many individuals would choose to opt-out and not stick with the default option, especially if they are not encouraged to do so by an insurance agent or realtor. If empirical data reveals that consumers, in fact, do have a tendency to stick with a default choice of flood insurance, then many questions regarding design of such a policy need attention.

Such questions include the following: Would the ability to “opt out” of coverage be mandated by state or federal regulators? Could regulators incentivize insurers to add an “opt-out” feature to their homeowners policies perhaps by providing a federal backstop or subsidized reinsurance to those that choose to do so? Would an “opt-out” feature replace the mandatory purchase requirement in Special Flood Hazard Areas (SFHA) or be in place outside this zone? How would the policy be structured? Would the default option be an NFIP policy attached to a homeowners policy? How would a flood insurance default work for insurance companies that are not Write-Your-Own (WYO) firms? Could homeowners make changes to the amount and types of coverage they purchase? Would any insurer choose to provide their own endorsement instead of an NFIP policy? Would an “opt out” work best for lower cost endorsements (or NFIP Preferred Risk Policies) outside high-risk areas? Would consumer reaction be different for the less expensive flood coverage in lower risk areas?

2.2. Mandatory Flood Insurance Offer

Another option discussed at the workshop is to require insurance companies to offer flood coverage whenever they are writing a homeowners policy. The reasoning behind this suggestion is that many consumers may be unaware that flood is not included in their homeowners policies or may not be paying attention to the possible need for flood coverage. Having it offered at the time a consumer is considering broader insurance coverages for their home may call residents’ attention to their flood risk and encourage more households to purchase flood insurance.

Whether a mandatory flood insurance offer would increase take-up rates is an empirical question and there may be good reason to be skeptical. A mandatory offer policy would likely need to be adopted on a state-by-state basis. The state of California has a mandatory offer for earthquake insurance and yet less than 15% of households purchase earthquake coverage. While the NFIP could back any flood policy, it is unclear how insurers would feel about a mandatory offer law. Non-WYO insurers would likely object to offering flood insurance sold by WYO firms with whom they compete on homeowners policies. It is possible that mandating a flood offer could incentivize more insurers to become WYOs or to develop and offer their own flood products. Indeed, a mandatory offer policy could be most effective if the federal government also offered a backstop or low-cost reinsurance (see below) to firms that develop their own flood products and make them a default part their standard homeowners policies (perhaps allowing consumers the option to opt-out as discussed above).

2.3. Federal Backstop for Including Flood in Homeowners Policies

Several workshop participants noted that the possibility of catastrophic flood loss years contributed to the difficulty in the private sector with writing flood coverage. They held up the Terrorism Risk Insurance Act as a possible model for overcoming this problem. More specifically, the federal government could provide either lower-cost reinsurance and/or a federal backstop to insurance companies when flood losses exceeded a high threshold. By essentially cutting off the tail of the loss distribution, many private insurers might then be willing to offer flood coverage. The workshop participants noted that issues of affordability for lower-income families would still have to be addressed.

The challenge with this approach is that there is enormous path-dependency in the political system. We now have fifty years of the federal government writing first-dollar flood coverage. Transitioning from this system to an entirely different approach would be challenging. Before a move to do so, there would need to be more dialogue with the private sector to ensure they would indeed be willing to write significant amounts of residential flood coverage if the federal government provided financial backing, should private insurers experience catastrophic losses. Determining the nature of the federal backstop, whether based on firm-specific or industry-wide losses, whether a backstop or pre- or post- funded, and other design questions would have to be carefully assessed.

Though challenging, such an approach is not without precedent. The NFIP was originally structured as a public-private risk-sharing partnership in which claims were paid first from collected premiums, then by capital pledged by participating insurers, then by low-interest loans from the Treasury, and, as a last resort, the Treasury would provide a catastrophic-loss backstop, paying all claims beyond a certain threshold, with no expectation of repayment. This experience provides lessons that would be useful in developing and implementing a federal backstop for flood insurance today.

2.4. Community Policies for Defined Regions

One notion that has been proposed to help close the flood insurance gap is for communities to purchase flood insurance on behalf of their residents in the floodplain. Such a policy would be purchased by a local governmental or quasi-governmental body and would cover a group of designated properties. The local government could collect premiums for the policy through a property assessment and would be responsible for distributing the claims to victims. In this way, many households would be covered by flood insurance.

While such a policy would ensure coverage for more households, many communities are not interested in playing the role of policyholder or in assessing higher property fees. It may be feasible for certain entities, however, such as levee districts, which are already assessing property owners and taking on a flood risk management role. There may also be particularly active communities that are already engaged in flood risk management for which this would be an appealing option. Private reinsurers could provide such policies to interested communities or districts.

2.5. Insuretech to Ease Purchase

Numerous workshop participants noted that currently it is difficult and time consuming to acquire an NFIP policy. While some private insurers are trying to simplify the process and make it possible to purchase policies online or via a mobile phone, these options are not universally available. There was agreement among many participants that harnessing insuretech to speed and ease purchase of flood insurance could have a positive impact on demand. If consumers could use an app on their phone to apply for and purchase a flood policy in less than 10 minutes, many felt this could encourage more widespread purchase. Participants also pointed to being able to enroll in automatic monthly payments and being able to manage their policy—all aspects of it—either online or from their phone, as likely also having a positive impact on demand. Such technological advances could be pursued by FEMA and the private sector.

2.6. Escrowing Flood Insurance and Automatic Renewal

One concern voiced by participants was that policyholders may drop their flood policy over time. Some felt this was exacerbated by the fact that most NFIP policyholders are forced to actively renew each year and to make an annual payment. Some participants noted that findings from behavioral economics suggest that if policyholders could have their flood insurance escrowed with their homeowners insurance and their mortgage, paid in monthly installments automatically from their bank account to smooth the cost over the year, and have it automatically renew, this could potentially dramatically increase policy retention. While escrowing flood insurance is possible and in some cases required, many participants observed it is not widespread in practice. This is something FEMA could work with lenders on operationalizing and encouraging.

2.3. Low Income Voucher Program

Workshop participants agreed that policies to close the flood insurance gap should focus, at least in part, on low- and moderate-income (LMI) households. This population is particularly important because they are most in need of the financial protection offered by flood insurance, yet are the least likely to have it. In a 2018 report on flood insurance affordability, FEMA found that among households located in the SFHA, insured households have a median income ($77,000), nearly twice that of uninsured households ($40,000). FEMA also found that about 51% of uninsured SFHA residents meet the Department of Housing and Urban Development’s definition of “low-income.” Providing tools that enable these households to afford flood insurance could be a significant step toward expanding coverage.

One proposed solution in the FEMA report that had been examined in other studies, and was discussed at the workshop is the creation of a federally-funded voucher program to help LMI households afford the cost of flood insurance. Under this program, policyholders would be charged risk-based premiums that communicate the level of flood risk they face. Most households would pay the entire premium themselves, but qualifying LMI households would pay a portion (determined by a means-tested threshold) and then receive a voucher to cover the remaining cost. The program could be designed and funded in multiple ways, and even coupled with mitigation grants or loans to reduce flood risk and premium costs over time. However, an essential characteristic of an effective voucher program is that it be funded not by premium discounts, but rather by general taxpayer revenue, tax credits, or another source external to the NFIP.

While this solution is perhaps the most efficient way to expand flood insurance coverage among LMI households, participants also noted that it is fraught with political challenges and that policymakers would have little interest in creating a new entitlement program. However, the impact of this policy could be multiplied if it were implemented in concert with other potential solutions discussed here. For example, combining an opt-out or mandatory offer policy with a low-income voucher program would likely do much more to increase take-up than a voucher program alone.