A Community-Based Flood Insurance Option (2015)

4

Community-Based Flood Insurance: Issues and Considerations

This report identified and described some of the major problems facing the National Flood Insurance Program (NFIP). These problems include the pricing of flood insurance premiums, takeup rates, efforts to support floodplain management, and socioeconomic factors. In theory at least, a community-based flood insurance (CBFI) option may have the potential to help solve some of these problems. For example, if a community purchased insurance for all dwellings within it, then the takeup problem would be solved.

Concerns exist about flood insurance purchases by individuals. For example, community members, perhaps motivated by altruistic feelings, paternalistic concerns, or by concerns that their lives would be affected by others’ uninsured losses, would prefer that their neighbors be insured. The flooding of uninsured properties in the community could have adverse impacts on property values and could generate other adverse physical, social, and economic consequences. Various government entities, most notably the federal government, also secure a positive externality from insurance purchase: covered individuals are less likely to require costly disaster relief.

A major objective of flood insurance is to ensure that both individuals and the community undertake appropriate policies to mitigate the risk of a loss, as well as the size of a loss should a flood occur. In theory, if insurance pricing was full risk and community leaders accounted their constituents’ costs (including insurance costs), then this objective would be achieved. Conceivably, if flood insurance premiums were paid at the community level, then leaders would become more strongly oriented toward mitigation efforts on behalf of their constituents. These are all conceptual arguments;

mere identification of their existence does not sufficiently support or oppose significant efforts to promote CBFI. A key requirement would be to evaluate their potential impact should communities purchase flood insurance.

Chapter 4 further discusses some of the flood insurance topics that merit consideration when evaluating the future prospects for CBFI. This chapter has two main sections: (1) a conception of CBFI, and (2) design considerations for CBFI.

A CONCEPTION OF COMMUNITY-BASED FLOOD INSURANCE

A basic premise underlying CBFI is that it may be less expensive and more effective to write single flood insurance policies at the community level than having insurance companies or the NFIP write multiple policies at the individual level. For this to be true, either the community must be more efficient than FEMA in writing policies covering individual entities, or the shift of responsibility from individual owners to the community will foster incentives to protect properties from floods.

The following section discusses the following topics: what is a community; the rationale for CBFI; spillovers across communities; how CBFI might provide solutions; and reasons why CBFI might not be successful.

What Is a Community?

There is no single definition of what constitutes a “community,” and many different definitions exist in an array of contexts. For example, definitions of “community” in Webster’s Ninth New Collegiate Dictionary include

FEMA defines a community as

A recent study (NRC, 2015) modified the FEMA definition and stated

The NFIP seems to focus on a community as a geographic entity that has powers (i.e., land use zoning authority [FEMA, 2005a]), at a minimum, to regulate behaviors of homeowners and other real estate interests to take actions that would reduce flood risk. It also may have the power to undertake projects that have the potential to limit flood risk.

As FEMA contemplates the future prospects of CBFI, a clear and explicit definition of the requirements placed on an entity that purchases such insurance would be most helpful. Although the previous definitions of a community are useful, it may be necessary to broaden the current FEMA definition. A town or city would clearly qualify. However, it is unclear whether a geographic area in a city, a gated community, or a business district would qualify. Minimum requirements may be the authority to enter into contractual arrangements and the ability to pay for insurance by assessment or cost levy. Another possible requirement may be that purchase at the entity level results in cost savings.

Rationale for Community-Based Flood Insurance

The implicit arguments for CBFI relate to at least two identified concerns: information and incentives. First, individuals and communities may not be sufficiently informed about levels of flood risk, and more importantly, efforts that would reduce flood risk. Second, even when they are well informed, individuals and communities may not have sufficient incentives to reduce flood risk, stemming from a variety of factors, including the expectation of disaster relief if insurance is not purchased. If insurance is purchased, then its price may not sufficiently recognize the risks associated with the covered entities. If none of these problems were present, then it would be appropriate to have individual property owners purchase insurance on a market, whether from private insurers or the government, just as they purchase toasters, clothes, and automobiles. Flood insurance would be no different.

One justification for devolving additional insurance purchase responsibilities to communities, either down from FEMA or up from the insured themselves, would be that neither communities nor individuals at present properly recognize some costs in the system. Perhaps relevant information is not available to the parties that must take actions to control flood risks, or perhaps incentives are insufficient. A second justification would be that placing insurance through communities would reduce or eliminate failures in providing and recognizing relevant information on flood risk. To take more appropriate actions the academic field and literature on “agency theory” can provide some guidance.

Agency theory looks at a broad array of situations when one individual (an agent) acts on behalf of another (a principal). The community,

or more appropriately the decision makers in a community, are agents for the stakeholders in the community (residents and other asset holders), who are the principals. The agents should faithfully represent the principals’ interests. If they do not, then this represents a situation of failed agency. The attractiveness of community insurance will always be based, in part, on the proposition that principals’ interests are faithfully represented by agents. Failed agency complicates enormously the relative attractiveness of community insurance (Pratt and Zeckhauser, 1985).

As an example of agency theory application, consider a hypothetical community, Riverville, whose losses from failing to implement appropriate policies take three forms: (1) expected losses that exceed the costs of their prevention, (2) reductions in insurance charges that are not being realized, and (3) expenditures to reduce risks that fail a cost/benefit test. The argument for getting Riverville to pay for insurance assumes that, for example, given current arrangements, its leaders would not vote to build a levee for $1 million that would save its residents an expected $1.1 million. However, if the community was directly paying a premium that would be reduced if it built the levee, then it would build the levee. Moreover, this would also be true if it paid the premium by collecting premium monies from its residents.

The Responsibility for Insurance is Irrelevant (RII) Proposition

The Coase Theorem from the discipline of economics addresses the common situation where there are two disparate parties—one imposes a negative externality on the other, say a polluter upstream on a river and a laundry downstream (Coase, 1960). The surprising result of the Theorem, is that it does not matter which party has property rights. Thus, if the laundry has rights, then it will charge the polluter for emissions until an additional $1 in cost to reduce emissions just avoids an additional $1 of damages. If the polluter has rights, then it will demand payment from the laundry to clean up to precisely this same point. A flood example might include damming of a river that negatively impacts downstream flooding.

The analogue of the Coase Theorem in the flood insurance context, where the interests of the two parties (community and residents) would be expected to coincide, is that which party bears responsibility for insurance does not matter. Call it the RII Proposition as an acronym for “Responsibility for Insurance is Irrelevant.” If the RII Proposition is satisfied, community insurance would neither help nor hurt. (The technical features of the RII Proposition are described in Appendix B. Appendix B identifies the parties to the insurance decision, and then explains how appropriate incentives can lead their decisions to produce an efficient outcome in the flood insurance case.) To understand why insurance arrangements

might fail, it is important to first recognize the properties of and conditions for an efficient outcome.

Appendix B shows that if both the community and residents must take actions to reduce the risks of and damages from floods, then an efficient outcome can result given reasonable behavior. Each party, that is each resident and the community, must spend until an additional $1 of expenditure just reduces expected losses by $1. To produce such an outcome, two properties must hold: (1) individuals must take the actions they would if they bore full responsibility for their own losses, and (2) the community, in choosing its actions (expenditures), must take full account of the losses to its residents.

Appropriate actions from the individuals (residents) can be achieved in two ways. First, they could be required to pay any incremental costs that their own actions impose on the costs of insuring their losses, for example, if they choose to live in a high-risk area. Second, the community could impose regulations that appropriately control residents’ actions, for example, by imposing building codes that efficiently control the risk and magnitude of flood losses to each resident.¹

For these conditions to hold, of course, the residents and the community must also be well informed about how their exposure to flood risks and losses is affected by the actions that they take. If the community employs the regulatory approach to secure appropriate actions from residents, then it must understand the costs to residents in adhering to such regulations, for example, how much they would pay to live in such areas.

Why the RII Proposition Might Fail to Hold

As an example, one could assume that given the current system where insurance is purchased at the individual level to an inefficient outcome in Riverville. Such a failure would refute the RII Proposition. The RII Proposition starts by observing that the community has no independent interests of its own, and is expected to act as a faithful agent and take full account of its residents’ interests, implying that Riverville and its residents would have fully congruent interests. In circumstances where the RII proposition may fail to hold, CBFI may be superior or it may be inferior depending upon the circumstances.

Why might the RII Proposition fail to hold? This is a more general and much more important question about public finance and public choice. Why might governmental entities not appropriately represent the interests

_______________________

¹ Prohibiting residents from living in high-risk areas might be inefficient—and would be, if the residents paid substantial amounts to live there. Further, FEMA has no legal authority to do this.

of its residents who were paying their insurance bills themselves,² or why might those residents not make optimal insurance purchases of insurance? And more broadly, would performance be improved or diminished if insurance were instead purchased at the community level?

Any argument for CBFI must start with an argument that the RII Proposition will fail to hold in practice. Failure of the RII proposition, however, is not sufficient to justify CBFI. It is required as well that insuring at the community level will work better than insuring at the individual level. Below are eight reasons why the RII proposition might fail to hold:

1. Free Riding. Residents might not purchase flood insurance because they expect disaster relief to bail them out; hence they would take a “free ride” and get coverage for free. To make CBFI preferred, the community would have to be better able than the current program to require its residents to buy insurance, to avoid living in certain areas, or to build to certain standards in flood-prone areas.
2. Alternative Financing. The community might choose a different mode of paying for the insurance than charging residents on the basis of their actuarial costs. For example, it could just use the general tax base. Such a shifting of responsibility might secure funding in a way that some would consider more equitable, but it would sacrifice efficiency, because residents would no longer have appropriate incentives to take their own damage-avoiding measures. And some, including many residents, would consider such a cross-subsidy program less equitable, quite apart from its efficiency implications.
3. Externalities. If owners or renters in the community would prefer that others purchase flood insurance, then such purchases would convey a positive externality. Such an externality could flow from altruistic or paternalistic feelings, or from real losses—for example, deterioration of the community—if the assets of others are lost and not replaced. Self-interested owners would not take account of this externality when deciding whether to purchase insurance. A community that accounts for all of the impacts on its constituents would account for all externalities.
4. Behavioral Mistakes by Residents or the Community. The residents may fail to purchase insurance because they underestimate flood

_______________________

² In economics, the agency problem is where there is a need for the community to take optimal actions on behalf of its residents. In this situation the community is the agent and the residents are the principals. Other prominent agency problems arise between a lawyer agent and client principal, or worker agent and owner principal. The principle explanation for losses from efficiency in those cases is that the agent and principal have divergent preferences. That explanation would not apply here, at least not if the community considered its interests to be the same as its residents’ interests.

5. risks for example because there has not been a recent flood. The whole panoply of heuristics and biases discovered in the past few decades by psychologists and economists, namely human tendencies that lead them to make decision-making errors, could lead to such mistakes (Tversky and Kahneman, 1974; Kahneman, 2011). So too, community officials might fall prey to these mistakes.

6. Community Pricing Is More Appropriate. Current pricing for flood insurance deviates from risk-based pricing. For example, it has grandfathering exceptions.³ It is also possible that the community, being closer to the problem and also utilizing federal information, can assess risk more accurately than can a distant federal agency. When assessing at the community level, to establish a community premium, the community decision makers need not be concerned if some residences are over assessed and others under assessed. Determining a reasonable overall level is important.
7. Political Distortion. Community political figures, such as mayors and city councilors, are elected with support from part of the community. Often that support comes differentially from different locations within the community. In making flood-control decisions, they may cater disproportionately to their supporters, implying that such decisions would be inefficient overall.
8. Out-of-District Owners Undervalued. In a special important case of reason 5, political figures might well feel that external (out-of-district) owners⁴ deserve lesser consideration than residents. Thus, when deciding what protective measures to undertake, they might give inadequate weight to such external owners’ damages or insurance payments. Community insurance, in theory, would account for damages to all dwellings in the community and would make under-weighting external owners less likely.
9. Greater Community Attention to Expenditures That Are Budgeted. The efficiency condition (see Appendix B) requires that the community value $1 to a resident equal to $1 of its own expenditure. It may not do so, for a variety of reasons. First, it may not notice or may under notice those dollars. Few mayors or city councilors may know what their residents are spending for flood insurance, for example. Second, officials are concerned with what the voters can easily monitor. If the town spends a great deal to rebuild a marsh, then that will be noted, perhaps much more than will insurance

_______________________

³ Such exceptions arise when, for example residents have mistakenly not been charged for something in the past, they are also not charged for it in the future.

⁴ This report uses of the term residents elsewhere in this section includes out-of-district owners.

These eight reasons, and some subcategories within the reasons, explain why the RII Proposition may not prevail. Other observers would add further reasons. Some of the above reasons, by themselves, would indicate that insurance at the community level is superior, others that it is inferior, and others that it depends on the underlying empirical situation. The most important lesson to emerge from this discussion is that any analysis of CBFI should begin with understanding of the RII Proposition and whether the evidence points in favor of, or against, insurance at the community level.

Spillovers Across Communities

Regardless of the flood insurance arrangements that are in place, a community has a responsibility to take action so that the sum of expected flood damages and/or insurance premiums avoided for the residents just equals, or exceeds, the cost of taking additional actions. This condition is appropriate if the community’s actions have no impact on other communities and their residents. However, many actions taken by community A may have a positive or negative consequence for nearby community B. For example, the construction of levees in upstream locations can raise flood stages downstream, and thereby contribute to increased downstream flood damages. The spillover could be positive, by contrast, for community D if nearby or upstream community C restored a wetland.

The formulation above assumes that each community optimized for itself. Moreover, this assumption would be satisfied whether insurance premiums were paid at the community or the resident level. With self-interested decisions by communities, which are to be expected, these spillovers will either be ignored or taken into account insufficiently. One potential solution would be to make a larger unit, a collection of communities that mainly impact each other, the unit of decision. Another potential solution would be to have the community receiving the spillover to pay (if positive) or get paid (if negative) so the spillover would be “internalized,” that is taken into account by the community creating the spillover. This latter system would work if the aforementioned Coase Theorem applied when

communities were the players, which seems like a heroic assumption in this context.

How CBFI Might Provide Solutions

A CBFI option, like any other policy option, has both strengths and weaknesses. Such an option might provide solutions—but not the sole solution—to recurring issues within the NFIP. The potential solutions identified and discussed in the following sections are whether CBFI can reduce administrative and transaction costs; increase takeup rates; and promote mitigation and floodplain management.

Reduce Administrative and Transaction Costs

Any insurance program entails administrative and transactions costs. Rates must be set, premiums collected, delinquent accounts pursued, and claims settled. Other NFIP costs include the federal policy fee and surcharges. Costs can vary depending upon the size of the institution that administers the insurance and the extent to which specialization may achieve savings. Many small employers contract with specialty firms to provide insurance-related programs because such outside firms may have lower administrative and transactions costs. In some circumstances, CBFI may entail administrative and transactions costs that are considerably lower than the aggregate of such costs to individuals who purchase insurance independently. In these instances, CBFI could be a less expensive option.

The extent to which a CBFI option has the potential to reduce transaction costs will depend upon the overall design of the option. Thus, for example, the NFIP currently pays 15 percent to write-your-own (WYO) agents, which reflects the cost of writing individual policies and provides agents an incentive to promote the policy option. A single policy covering an entire community may entail much lower costs because of economies of scale and the absence of the need to create incentives. Some costs will remain, however, including the costs of collecting data, setting rates, and settling claims.

The extent to which CBFI would reduce claim adjustment and processing costs is unclear. Adjusters presumably still need to evaluate damage and estimate repair costs structure by structure. There would still need to be some type of appeals process for policyholders who are dissatisfied with the initial insurance award. The administrative costs required to design and set up the program would also need to be considered. The return on this investment may not be large if few communities decide to enroll.

Increase Takeup Rates

One frequently expressed concern about flood insurance is that many high-risk properties go uninsured: takeup rates are low. Low takeup rates can occur for a variety of reasons:

• Owners without mortgages are not required to purchase insurance.
• Owners with mortgages, for whom insurance is mandatory, still might not purchase.
• Although owners or renters should buy insurance, they mistakenly do not.
• Mispricing makes it undesirable for some owners or renters to buy insurance.
• Owners or renters avoid insurance because they expect government bailouts should there be a flood.

Of course, some properties with very low risk—a house on a hill, a business far from any water—may not need to be insured. These are properties for which the administrative costs of issuing a policy outweigh any risk-spreading benefits.

This problem could be readily dealt with at the community level, assuming that the community has the power to require its constituents to purchase coverage. The simplest way, of course, would be to have the community just buy a policy that covers all properties (or all properties that merit insurance). This could be done even with minimal requirements on what constitutes a community, say a gated community or business district. If CBFI were a catalyst for increased takeup, then it could enhance resilience (when paired with floodplain management and mitigation activities) to flood events and, depending on how premiums were set, provide incentives to mitigate more properties.

The mandatory requirement to purchase flood insurance increased takeup rates, but compliance is not complete and takeup remains low among those not subject to the requirement. The issue of mandatory purchase is both complicated and controversial, especially in a world where mortgages are often securitized and resold. Ongoing attention has been paid to enforcing the mandatory purchase of flood insurance by homeowners (within the SFHA) that have a federally insured mortgage. Efforts to increase enforcement by focusing on the lending sector have been partially successfully (NRC, 2015). For example, the NFIP increased the responsibilities of the lenders by requiring them to notify property owners when their policies lapsed. The National Flood Insurance Reform Act of 1994 also established penalties for lenders who did not carry out specific requirements (NRC, 2013). A CBFI option may help with such an issue.

Promote Mitigation and Floodplain Management

One of the primary arguments in support of CBFI is that communities will take more actions to mitigate risks if they themselves must purchase insurance. This could be accomplished in two ways:

1. The community could undertake actions on its own, such as restoring a marsh that can mitigate storm flows and thereby reduce flood risks downstream.
2. The community could require owners to take actions that reduce risks, for example by not locating in a flood-prone area.

However, this potential for action is not a sufficient argument for CBFI. It must also be the case that the community would not take this action if insurance was purchased given current arrangements. The reason why it might and might not take such actions was the subject of the rationale for CBFI above.

Floodplain management and flood hazard mitigation are long-term activities, which may not garner political backing during the short tenures of elected officials. Local communities receive the property tax benefits from development but do not pay insurance premiums for structures developed in flood hazard areas. Often, local officials make land use decisions based on short-term economic gain that in the long term increases flood hazard exposure. Public support for flood mitigation peaks and fades with flood events and tends to be sustained only after a community has repeated flooding. A CBFI option has the potential for creating a monetary measure to enhance community floodplain management and mitigation action over the long term without waiting for flooding to become chronic. Flood mitigation has increasingly become the responsibility of local decision makers (Brody et al., 2010). Community-level interventions and particularly nonstructural mitigation can reduce flood losses. For example, there is empirical evidence to suggest the importance of strong organizational capacity—at the local level—to implement flood mitigation strategies (Brody et al., 2010). Therefore, it may be helpful for a community-level option to consider the contextual conditions and capabilities of local jurisdictions (Brody et al., 2009).

CBFI also has the potential for making local officials more aware of their land use decisions. The structure of payment for insurance by citizens could be devised to encourage them to take action to reduce their risk. Some local government units, such as drainage and levee districts, are organized to provide flood protection. The residual flood risk in areas protected by levees is acknowledged in NFIP reform legislation and the U.S. Code pertaining to the National Flood Mapping Program. A CBFI policy made available to local units of government could provide coverage for residual

risk. A strategy suggested for increasing private-sector involvement included the option of the NFIP providing residual insurance (GAO, 2014c). Within this strategy, GAO suggested that the federal government could encourage private sector involvement by providing coverage for the highest-risk properties that the private sector will not insure. Providing residual insurance could increase the program’s exposure, but NFIP would be insuring fewer properties.

Key Challenges for a CBFI Option

It is easy to envision a community that purchases insurance optimally on behalf of owners, collects the required monies inexpensively either through premiums or taxes or some combination, undertakes efficient mitigation actions on its own, and requires its constituents to take efficient actions. This, however, may be unrealistic. Taking responsibility for a program of insurance and mitigation is a significant responsibility, and many communities would not be up to the task. The following section identifies and discusses some reasons why CBFI may not be successful, which are lack of community interest, limited capabilities to implement CBFI, variations in the size of the population and geography involved, ability to regulate land use and authority to collect revenue, and goals not achieved, despite adopting CBFI.

Lack of Community Interest

It is unclear how many communities would participate if FEMA put a CBFI option into practice. From a practical perspective, community leaders might be reluctant to participate in a CBFI option, because it would require adjustments from the status quo. Even if CBFI were financially and economically attractive, there would be some losers. For example, if the community simply imposed the required premiums on owners, then those who had considered it economically desirable to go uninsured (e.g., free riders) would bear costs that they had previously escaped. If some or all of the premiums were paid through taxes, then individuals whose tax burdens increased above the level of their current insurance costs would likely object. And if the premiums were paid simply from tax revenues, then leaders would be confronted with constituent complaints, which characteristically occurs when taxes are increased. In addition, current leaders might be reluctant to purchase for others: the vast majority of the time it does not pay off. Thus, the community would incur a cost that would have little chance of being paid off during the leaders’ time in office. Political leaders tend to operate with very short time scales. Thus, the prospect of requiring people to purchase insurance against an event that might occur in the distant future

might be unattractive. The same argument would apply to undertaking mitigation measures. However, if a marsh is restored or a river bank raised, then at least the constituents can see what they paid for. The promise of protection is visible, and retains its value even if there is no flood.

Capabilities to Implement CBFI

Even in the absence of political concerns about undertaking CBFI, community leaders may have to assume responsibilities to successfully implement CBFI. Some of these responsibilities may require capabilities that many communities do not currently possess. Consider, for example, the simple case where the community pays the premiums but passes the charges to individual property owners. The community would require expertise in pricing that involves risk assessment and property valuation. It would also require the capability to oversee compliance with mandatory purchase requirements. Responsibility for mandatory purchase and maintenance of flood insurance could remain the responsibility of the lender. Additionally, the community would need to perform more traditional roles of establishing and enforcing building and zoning standards for areas susceptible to floods. These latter tasks would require engineering expertise and processes for managing and resolving appeals.

To help address issues related to community capability to carry out a CBFI option, multi-jurisdictional programs based on experience with hazard mitigation planning under the Disaster Mitigation Act of 2000 (DMA) (Box 4-1) were considered. Communities (e.g., counties, municipalities) can chose to consolidate or create separate mitigation plans. Communities sometimes cover both county and municipal jurisdictions, often called multi-jurisdictional plans under DMA. Small municipalities, for example, with limited capability to prepare and implement plans often choose to combine their planning with counties. These joint actions may help in coordinating spillover effects of different jurisdictional actions, setting insurance rates, and reducing transaction costs.

Variation in the Size and Homogeneity of the Populations Served

Some NFIP-participating communities have many policies-in-force, while others have few policies-in-force; therefore, there is little evidence for an average NFIP community (see section on “Participating Communities,” Chapter 3). This set of observations suggests that small communities might not have the executive/professional capability to oversee a CBFI option, or if they did, it would be at too small a scale to be efficient. If a community turned program management over to a third party, then it would just substitute that party for the federal government. However, this does not mean

that large communities would be ideal for implementing CBFI. Such communities might do well in passing through insurance premiums, but there could be difficulties when cross subsidies are involved (e.g., from high-lying to low-lying areas). If the cross subsidies are between different ethnic or income groups, then the difficulties could be greater still. In a small community, the residents of the two different areas might have more connection to each other or have more concern for each other than they would in a large metropolis that comprises a city, a single political unit. Indeed, it might be optimal to even have protective expenditures financed on a smaller scale than a large city. For example, the lack of concern of most residents for those living in a high-threat area might prevent the undertaking of efficient protective expenditures. (Of course, that phenomenon is already at work, given that such expenditures would come out of the budget of the city as a whole).

In short, given the administrative capabilities required for running an insurance program, communities that are too small would likely be ill

equipped. However, as communities grow they tend to become more diverse in their risk levels. In addition, there tends to be more social variability and a consequent wider range of preferences in such situation than in smaller communities. This suggests that very large communities might also have difficulties administering insurance, given political divisions. Should the NFIP offer a CBFI option, the optimum scale may be midsized communities, which may be better suited to balance needs for administrative capability against the dangers of insufficient concerns of some citizens for the well-being of others (Tiebout, 1956).⁵

Concerns about aggregate risks, and the need for a large population to spread them, are not part of this consideration. CBFI would be paid out of a federal pool, and millions of insureds in hundreds of communities would be covered. Even in hypothetical Riverville, with its two residents, aggregate risk would not be a concern; its risks would be laid off at the higher federal level. No supporter of CBFI has been proposing that communities also self-insure. Were they to do so, aggregate risk would certainly be an overwhelming concern.

Land Use and Revenue Collection Authority

The discussion above implicitly and explicitly assumes that communities would be able to regulate land use to achieve flood protection at the lowest possible cost, accounting for all cost elements. Absent this authority, individuals might choose to live where risks are high, particularly if they could expect—perhaps incorrectly—either subsidized insurance rates or post-flood disaster relief. If the community lacks the authority to restrict land use, whether insurance is at the individual or community level, then considerable inefficiencies will result.

A second implicit assumption is that communities contemplating CBFI would have the authority to raise revenues to cover the premium. If the community were merely a pass through, then it would collect insurance premiums from its residents. If the premium were to be financed in part by taxes, then the community would require the ability to raise taxes to cover the expense. Collecting premiums could be challenging administratively. Raising taxes, particularly if the beneficiaries were a relatively small group, might be difficult politically.

_______________________

⁵ The Tiebout model suggests that forces sort individuals so that those within a community will have fairly similar preferences. Unfortunately, geography often implies that different locations within the same community will be at very different risk levels. Thus, even after sorting, although residents may be similar in terms of preferences on taxes and school spending, they may have quite different preferences on whether a wetland should be created or flood insurance subsidized by the municipality.

Political Distortions

As mentioned above, political processes can never fully reflect the views and preferences of constituents. The problem stems in part because, politics aside, methods for aggregating individual preferences fail to capture relevant information, for example how strongly they care about an issue. Even if all preference information were known, however, the electoral process assures that elected officials will feel more beholden to some people than to others; hence the former will have their values over-weighted. There is considerable evidence in addition that those who participate in particular decisions are those who care the most. In effect, the intense interests of a minority may outweigh the diffuse interests of the majority. The exact balance of these forces in any situation is unclear, and it may lead the high risk to exploit the low risk, the rich the poor, or vice versa in both cases. What is clear is that politics can never assure an efficient outcome. The critical question in this context is how well a community-based process, however designed, performs relative to the current imperfect situation.

Additional Challenges

The implementation of CBFI as a policy option presents additional challenges. First, many communities that should adopt it will not do so, perhaps because leaders are influenced by politics or do not want to assume new responsibilities and tasks. Second, some communities that lack adequate capacity and resources to effectively implement and administer CBFI may nevertheless choose to do so. They may overestimate their capabilities to administer the program, or their ability to build those capabilities swiftly. The result would be a poorly administered program in these communities. Possibly, there would be swift reversion to the current program. If CBFI were implemented, then there may be potential for delay in disbursements to individual property owners or renters. An additional administrative layer may extend this period even longer, particularly for those who are un- or underinsured, which may lengthen the recovery period and make it more variable for the most socially vulnerable households and neighborhoods within a community.

Risk and Uncertainty

Insurance companies, and other insuring entities, deal with risk in a variety of ways. The first line of defense is to properly underwrite risk, which begins with estimating the probable frequency and severity of losses associated with risk exposure. Generally, a new risk exposure is added to an existing pool of exposures if it meets the insurer’s underwriting criteria

and does not cause the insurer’s overall risk to exceed its appetite for risk. Pooling a larger number of independent exposures enables the insurer to more accurately estimate total losses to be paid and to establish appropriate premiums for coverage. Diversifying across geographic areas or lines of business can also help reduce risk, if the areas or lines of business are independent. An insurer’s risk can be defined by a distribution of expected losses which, more fundamentally, captures unknown outcomes whose likelihood of occurrence can be measured or calculated.

In addition to bearing calculable, objective risk, insuring entities face uncertainty in the form of events that cannot or at least have not yet been described. Various actions can be taken to address uncertainty. Through time, of course, uncertainty regarding some types of losses can be reduced by learning more about them, that is, collecting and analyzing event frequencies and severities, and better understanding correlations between losses. Various manifestations of climate change, described in Box 4-2, are examples of uncertainties that could be faced by flood insurers. These are detailed in general in the following paragraphs. Specific examples drawn from the climate change example are identified and detailed in Box 4-2.

To deal with uncertainty, insurers may contractually limit coverage to known events (e.g., through a named perils policy) or limit the amount of losses to be paid under the contract. The NFIP currently employs both of these approaches.

Yet another approach for dealing with both risk and uncertainty is to transfer a portion of risk to a reinsurer. In such an arrangement, an insurer transfers its catastrophic risks to a reinsurer or reinsurance facility in exchange for a known premium amount. Because the insurer is primarily concerned that losses might exceed a target level for maintaining financial solvency, a reinsurance contract could be purchased to cover losses exceeding this target. In addition to transferring the known risk, such a contract would, consequently, also transfer to the reinsurer the uncertainty associated with unknown catastrophic events.

The methods of managing risk and uncertainty are more problematic in the context of floods because of the fundamental nature of the risk. Diversification across geographic areas can help to reduce risk, but pooling a large number of exposures that are correlated does not help the insurer reduce its risk because additional exposures do not reduce the variation of expected losses. This largely explains why individual private insurers do not write flood risk exposures: their ability to diversify geographically is often limited. Uncertainty, in the flood risk context, stems primarily from the unknown nature of future flood events (e.g., changes in event frequency and severity due to climate change), but also from uncertainty about the effectiveness of mitigation activities and neglected contributors to damages such as failing structures or storm drainage systems.

In CBFI, the task of managing risk and uncertainty would fall on the risk-bearing entity, the federal government, as it does in the current NFIP. Moving to a CBFI option, however, might yield new opportunities for managing risk and uncertainty. For example, the NFIP could require that communities seeking coverage provide a comprehensive analysis of the flood risk.

The above section discussed topic areas and questions that require further evaluation when assessing the strengths and the weaknesses of a CBFI

option. It also considered aspects of risk and uncertainty that pertain to a new insurance option. The next section discusses design considerations of a CBFI option, if it is pursued.

DESIGN CONSIDERATIONS

Fundamental considerations in the design of CBFI include the following and are discussed in detail below:

• Who bears the cost of risk and how is it shared
• Who writes the policy and adjusts for losses
• Coverage limits, standards, and compliance
• Underwriting, pricing, and allocation of premium costs
• Administrative capabilities
• Confirming compliance within mandatory purchase requirement
• Pricing expertise, including valuation of mitigation measures

Who Bears the Cost of Risk and How Is It Shared?

When feasible, risk of any nature, catastrophic or not, is ideally transferred to the entity that is in the best position to bear it. This is done largely for reasons of solvency, to ensure that the entity has sufficient capital, or access to capital, in the event that larger-than-expected losses are incurred.

The fundamental nature of floods complicates private insurers’ ability to underwrite flood risk. Specifically, floods constitute a catastrophic risk which, for noncommercial residential properties in the United States, is typically not insured by private companies except to the extent that they underwrite excess flood coverage beyond the NFIP’s policy limits. Adding full coverage of flood risk to their portfolios would present several challenges, because it would substantially increase their overall exposure to catastrophic losses. One such challenge results from the compounding of flood risk associated with tropical storms and hurricanes with the wind damage from these events.⁶

The federal government (including the U.S. Treasury) is the risk-bearing entity for all flood insurance policies sold through the NFIP. Insurers that participate in the WYO program do not bear risk. Individual homeowners bear risk only in the form of deductibles and losses that exceed policy limits.⁷ The movement to a community-based option would conceivably shift risk-bearing to communities, private insurers, or individuals depending upon how it is structured. Although, as a federal entity, the NFIP may be best positioned to bear the risk, movement to a CBFI option allows for a reexamination of how some risk might be transferred to and/or shared by other stakeholders. The manner in which flood risk is shared among different parties will affect their respective incentives for mitigating losses.

A community-based option could allow for the transfer of some flood risk to the community (e.g., through a high deductible paid by the community), but consider the community’s risk-bearing capacity. A larger com-

_______________________

⁶ Wind insurance is typically underwritten by private insurers or state-run wind pools.

⁷ Policyholders with pre-FIRM subsidized rates may purchase a minimum deductible of $1,500 if the building coverage is less than or equal to $100,000; if the building coverage exceeds $100,000, then minimum deductible is $2,000.

munity, for example, does not necessarily have a greater ability to bear risk. In addition, not every type of community has the ability to raise capital (e.g., a flood district vs. a county). Shifting some risk to the community may incentivize the community to mitigate flood risk. Deductibles and risk-based premiums that reflect mitigation behavior in a timely manner would provide incentives for a community to undertake mitigation.

In terms of the risk borne by individuals, those individuals with flood insurance bear risk in the form of deductibles and policy limits. Deductibles, policy limits, and risk-based premiums provide incentives to mitigate; but many potential mitigation measures may be out of the control of the homeowner (e.g., reinforcing a federal levee). Property owners can undertake some measure to make their homes/buildings more flood resistant (e.g., elevating structures), but these measures can be costly.

Who Writes the Policy and Adjusts for Losses?

The transfer of flood risk from the individual homeowner (or community) to an entity better suited to bear the risk (NFIP and community) requires a contractual agreement which is typically an insurance contract. Fundamental considerations in the design of this contract are the coverage limits, standards for coverage, cost-sharing mechanisms (i.e., deductibles), and the insuring agreement itself, which outlines the coverage provided.

A CBFI option could capitalize on the already-existing expertise and capability of NFIP and WYO insurers for writing policies for individual homeowners. That is, CBFI could involve the NFIP and the WYO insurers in writing policies and collecting premiums, but at the community level rather than the individual level. The processing of applications and claims from communities versus individuals would likely require additional administrative effort. In theory, however, the claims process may not change dramatically. For example, the insured incurs a loss and notifies the policy issuer (i.e., the community), who will notify whoever they got the policy from (either NFIP directly or through a third party), who will contact the adjuster, who will report on damages to the party responsible for sending claim payment to the property owner.

Coverage Limits, Standards, and Compliance

Another design consideration for CBFI is the extent of coverage that will be provided under the community policy. Currently, individuals are provided a limited set of coverage options. Homeowners may elect coverage for building property up to $250,000 and personal property up to $100,000. Within these coverages, they also have a choice of deductible amounts. Under CBFI, new limit guidelines would have to be established

for policies that are written for communities, with such limits and choices of deductibles potentially depending on the size of the community, the nature of the risk (e.g., type of flooding), existing infrastructure, and other community characteristics, for example, floodplain district versus a county.

Movement to a CBFI policy option may provide an opportunity to reconsider flood exposures; that is, defining the community’s exposures as extending beyond individual homes to public infrastructure (e.g., bridges). To the extent that the community policy can cover a variety of exposures, additional consideration is needed to establish coverage standards and limits that reflect this expanded group of exposures.

If designed to include all individual homes in a defined flood-prone area, then a community-based policy has the potential to achieve full compliance with lenders’ mandatory purchase requirements. This is because all properties included in the community policy would have the required coverage by design and, furthermore, the ability of individual homeowners would not be able to opt out of coverage. This constraint would depend on how the community establishes the rules of participation. For example, if the community charges homeowners a monthly premium or collects payments in the form of taxes. These considerations are addressed further below.

Underwriting, Pricing, and Allocation of Premium Costs

How the CBFI is priced will have important implications for the entity that bears the risk, the community, and the individuals in a community. If an insurer were to underwrite and bear all of the risk of a CBFI policy, then it would need to charge an actuarially sound premium that reflects all expected costs, including an appropriate catastrophe loading, which could be used to purchase reinsurance in the private market. If the NFIP were to provide reinsurance to the insurer, then the premiums charged by the insurer would only need to account for whatever the NFIP would charge for this reinsurance. This assumes that the NFIP reinsurance would cover any losses that would be considered catastrophic. There is also the issue of how the premium of a community policy would be allocated among the properties covered under the policy. If a private insurer were to bear the risk assumed under the policy, then its primary concern would be that the premiums would be sufficient to cover all of the costs associated with underwriting the policy. This assumes that insurers would not be exposed to higher than expected costs due to moral hazard or adverse selection. If the NFIP were to underwrite the risk assumed under the policy, then this issue would likely not be a matter of concern to the insurer if its role is confined to servicing the policy.

Other questions arise with respect to how the premium of a community

policy would be allocated among the properties covered under the policy. A detailed discussion of alternative cost allocation and financing schemes is beyond the scope of this report, but it is useful to consider two alternative approaches that would represent the opposite ends of a spectrum of cost allocation/financing schemes. With the first approach, the community charges all property owners an amount that depends on property value and flood risk. This could be accomplished through special assessments, a property tax surcharge, or other financing methods. However, this approach would likely be viewed as highly unfair by low-risk property owners and would exacerbate the moral hazard created by the NFIP’s current pricing structure. With the second approach, the community allocated the premium in such a way as to require each property owner to pay an amount commensurate with his or her risk of loss. This approach would likely be viewed as fair by low-risk property owners, but it would constitute a departure from the NFIP’s current pricing structure and would be opposed by property owners who currently pay discounted/subsidized premiums under that structure.

Ultimately, CBFI requires consideration of how underwriting connects risk-bearing to incentives for mitigation. Pertinent concerns include whether entities and/or individuals are allowed to opt out of participation, or whether their ability to participate is restricted, for example, by requirements for certain mitigation activities. Rules governing participation have important implications for the pricing. The ability of individuals to opt out of a CBFI policy option could result in greater adverse selection.

Administrative Capabilities

If insurance contracts remain the vehicle for transferring risk, then private insurers would remain the most efficient entities for handling administration of contracts. Communities (perhaps regardless of definition) would likely not have adequate expertise to administer policies. However, some definitions of community include entities that could effectively and efficiently collect the revenue needed to pay for a community policy through means such as special assessments, property taxes, and other means.

The majority of NFIP flood insurance policies are written and serviced by WYO private insurers. Typically, several WYO companies may operate in the same community. The bundling of all of the flood risk in a community into a single policy begs the question of who would write and service the policy. Servicing a flood insurance policy after a flood disaster requires knowledgeable professionals to visit sites to assess damage. A single provider may not have the staff capacity to address, in a timely manner, all the needs of a community after a flood event.

Alternatively, community purchase of a base policy (presumably directly from the NFIP) would provide a minimal level of community-wide

coverage, lowering the premiums for individual policyholders and leaving in place the current system of involving WYO companies. This design consideration requires no changes to the existing administration system.

Confirming Compliance with Mandatory Purchase Requirement

Under the current NFIP, policies are written and administered and claims adjusted by WYO insurance companies and, to a lesser extent, directly through FEMA. When flood insurance is required because the structure is collateral for a federally backed mortgage (mandatory purchase), the required coverage is for the amount of the outstanding loan (with maximum coverage limits). Lenders have the responsibility to enforce the mandatory purchase of flood insurance for structures in SFHAs; the level of coverage varies with the amount of the mortgage. Additional coverage may be purchased at the discretion of the individual. The level of any non-mandatory purchase of flood insurance, such as replacement cost, renters insurance, and contents insurance is the choice of the individual.

Complying with the mandatory purchase requirement is currently a loan-by-loan, structure-by-structure process. The administrative aspects of monitoring compliance with a CBFI policy option in place may depend on coverage provided through the policy. The mandatory purchase is only for the amount of a federally backed mortgage: a bundled community-based policy providing the minimum required coverage would likely need to maintain some aspect of individual coverage and monitoring, which would be administratively burdensome. Alternatively, if the CBFI option covered all structures to a defined limit, such as replacement costs, then the mandatory purchase requirement would be met. However, using this benchmark for coverage, the aggregate premium for the properties subject to mandatory purchase would likely be higher than the sum of individually required policies for those properties. This leaves unresolved the issue of individuals wanting higher coverage, such as for contents. The important point here is that property owners may be reluctant to pay for CBFI that covers replacement costs on structures. Another, perhaps better, alternative would be a CBFI option that provides a set base coverage amount, for example $30,000. Some homeowners would need to purchase additional coverage to comply with the current mandatory purchase requirement. Individual coverage written by WYO companies would be for mandatory or desired coverage above the $30,000. The administration burden for WYO companies would be comparable to that for tracking CRS-participating communities where premiums are discounted.

Pricing Expertise, Including Valuation of Mitigation Measures

A community might allocate the cost of CBFI in multiple ways. Each method presents unique considerations for pricing and valuing individual mitigation efforts. If private insurers were to underwrite the risk associated with community-based policies, then they would price them on an actuarially sound basis so that premiums would be sufficient to cover the cost of the risk being underwritten and so that incentives for loss mitigation would exist at the individual and community levels. To perform this function, private insurers would have to possess or acquire the information and expertise to price community-based policies to reflect the risk underwritten and the expected savings from mitigation measures. If the NFIP were to assume the risk of CBFI, then presumably it would also price policies to account for the expected savings from mitigation measures. FEMA has the expertise in setting premiums based on flood risk and clearly would have to work with communities to communicate the costs of bundling individual properties into a community policy.

There is a lack of data about which policies are subsidized and grandfathered. A policy cannot be both, although a policy can be eligible for both types (a prospective policyholder has to choose one). One could advance the generalization that structures built before the initial identification date of the flood hazard would tend to be grouped geographically. Similarly, structures built in compliance with the floodplain mapping effective at the time of construction and at a later time identified as being in a higher risk area would also tend to be grouped geographically. Specific conditions must be met for structures to retain a low-risk policy pricing.⁸ Given this generalization that structures with grandfathered or subsidized rates would be geographically clustered, high-risk areas with structures dating to a time when flood risk was unknown or considered low could be addressed at the community level for mitigation action consistent with community values. The protection of areas in communities with historic or cultural value is the preferred course of action. At the other extreme, large-scale buyout of structures in other areas would be advantageous if desired by the current residents. Sorting out which structures have subsidized or grandfathered rates might be best undertaken at the community level.

The mechanism used to distribute the cost of flood insurance for the community could be structured to encourage citizens to reduce exposure through individual actions. Stormwater management, for example, offers some interesting examples of community efforts to reduce damages by raising awareness of individual contributors. Some communities have stormwa-

_______________________

⁸ A policy must be in place prior to the effective change of risk zone and the policy must be continued to maintain the lower risk zone rating.

ter management fees, assessed on the basis of the impervious surface area on the property. Many communities work to reduce stormwater damages by offering programs such as vouchers for backflow valves and rain barrels. These strategies could be applied to flood insurance assessments.

SUMMARY

CBFI may help to address some of the current challenges within the NFIP. A basic premise of CBFI is that it may be less expensive and more effective to write insurance policies at the community level. There are many definitions of what constitutes a “community.” The NFIP seems to view a community as a political entity that has land use authority. An explicit definition of the requirements placed on an entity that purchases a community-level policy would be helpful.

Understanding of the rationale for CBFI requires some understanding of agency theory, specifically the Coase Theorem from the discipline of economics. The Coase Theorem holds that where parties (both individuals and groups) account for all costs and benefits, markets are functional, information is freely and widely available, transactions costs are zero, and economically efficient outcomes are reached irrespective of with whom the property rights are vested. If the collective economic interests of communities and residents fully coincide and are fully accounted for, then the outcomes of a flood insurance purchase decision do not rely upon which party—communities or residents—bears responsibility for insurance.

In considering practical applications of the Coase Theorem to flood insurance, at least eight reasons explain a Proposition of Responsibility of Insurance is Irrelevant may fail to hold. Two reasons are free riding—when some or all residents do not buy insurance because they expect disaster relief to provide adequate post-flood aid—and externalities—when self-interested parties fail to account for all of the impacts, most of which are not monetized, of buying or not buying insurance. These reasons help to guide the identification of circumstances when CBFI may be superior, inferior, or either one depending on underlying empirical circumstances. Choosing the CBFI option requires confidence that insuring at the community level will work better than at the individual level.

There are circumstances where CBFI may provide partial solutions to NFIP challenges. Solutions to challenges include reducing administrative and transaction costs, increasing takeup rates, and promoting flood mitigation and floodplain management. Even if CBFI does not effectively address these challenges, it could help in certain areas. For example, moving insurance to the community level would likely enhance attention to risk-reduction activities at that level. Under certain circumstances, however, CBFI may not help to solve NFIP challenges, such as when a community is

not interested or lacks the capability to implement CBFI. Successful implementation would likely require communities to restrict land use, adopt complementary flood risk management measures, and raise additional revenue to pay CBFI premiums.

Central to the concept of insurance is protection from losses incurred from uncertain events such as fire, automobile accidents, and floods. Regarding floods and flood insurance, past patterns of climate and hydrology are limited predictors of future patterns. In addition, changes in land use and population sizes influence flood risk and flood damages in uncertain ways. Scientific evidence shows that flood losses are mostly explained by what is or is not done to the landscape; therefore, efforts to improve landscape management are important. For insurance purposes, uncertainties would need to be evaluated on a case-by-case basis.

The chapter also identified design considerations that require further assessment when planning for and designing a CBFI option. These were as follows:

• Risk Bearing and Sharing
  A CBFI option could conceivably shift risk-bearing to communities, private insurers, or individuals depending on how it is structured. Although as a federal entity, the NFIP may be well positioned to bear the risk, movement to a community-based option allows for reexamination of how some risk might be transferred to and/or shared by other stakeholders.
• Responsibilities for Writing Policies and Loss Adjustments
  Write-your-own insurance agents write policies and collect premiums under the NFIP, but CBFI policies would have to be written at the community level. A range of administrative duties would need to expand to process applications from communities.
• Coverage Limits, Standards, and Compliance
  Under CBFI, deductible choices would depend on the community’s size, the nature of the risk (e.g., type of flooding), existing infrastructure, and other community characteristics. A CBFI option may provide an opportunity to reconsider flood exposure.
• Underwriting, Pricing, and Allocation of Premium Costs
  Several complex issues fall under this topic: the extent of actuarial principles to be used in setting premiums (NFIP premiums are legislatively and administratively constrained; see NRC, 2015); the extent to which catastrophic losses would be reflected in premiums for a given community; and the allocation of premium costs among property owners (and renters) in a given community. The third issue could involve deriving some portion of funding from owners of properties that are not in areas subject to flooding.

• Administrative Capabilities
  If insurance contracts remain the vehicle for transferring risk, the private insurers would likely remain as efficient entities for handling their administration. Communities (probably regardless of definition) would likely not have adequate expertise for undertaking policy administration. However, some definitions of community include entities that could effectively and efficiently collect the revenue needed to pay for a community policy through special assessments, property taxes, and other means.
• Confirming Compliance with Mandatory Purchase Requirements
  Currently, the mandatory purchase is only for the amount of a federally backed mortgage. A bundled community-based policy that provides a minimum required coverage would need to maintain some aspect of individual coverage and monitoring, which could be administratively burdensome. CBFI could cover all structures to a defined limit. Another alternative could be a CBFI that provides a set base coverage amount.
• Pricing Expertise, Including Valuation of Mitigation Measures
  If private insurers were to underwrite risks associated with a CBFI policy, then they would want to price them according to actuarial principles. Private insurers thus would have to possess or acquire the information and expertise to price community-based policies to reflect the risk underwritten and the savings expected from mitigation measures. If the NFIP were to assume the risk of community-based policies, then presumably it would also assume the function of pricing these policies to account for the savings expected from mitigation measures. FEMA has expertise in setting premium costs based on flood risk and would have to work with communities to communicate individual property coverage costs bundled into a community policy.