National Flood Insurance Program (United States, 1968 to present)

Congress created federal flood insurance in 1968 because private insurers would not offer it; the program has since borrowed $22.5 billion from the Treasury, kept some rates below risk, and tried to reform through laws in 2012 and a pricing change in 2022, while experts disagree on who benefits and how much to charge.

The record

The law and its purposes

• The National Flood Insurance Act of 1968 is Title XIII of Public Law 90-448 (Congressional Research Service). Congress found in the codified statute that "many factors have made it uneconomic for the private insurance industry alone" to offer flood coverage on reasonable terms. It called for "large-scale participation of the Federal Government," carried out as far as practicable by private insurers. It also said insurance should encourage "sound land use by minimizing exposure of property to flood losses" and tie into floodplain management, and it urged state and local land-use changes to limit building in flood-prone areas. The text we read does not mention subsidized premiums. (42 U.S.C. 4001, Cornell LII; opened. CRS, R45999; opened. The exact 1968 enactment date was not given in what we read.)
• CRS lists three purposes: offer primary flood insurance, reduce risk through floodplain management standards, and over the longer term lower federal disaster aid after floods. It reports more than 5 million policies in 22,500 communities and about $4.6 billion a year in revenue from premiums, fees and surcharges. (CRS R45999; opened. The report's own date was not confirmed.)

Debt

(CRS, NFIP Borrowing Authority and a March 10, 2026 version; both opened.) Two other figures circulate. Pew's 2016 infographic gave "$23 billion" and a 2016 thesis by Kevin Starbuck gave $24 billion from the 2004, 2005, 2008 and 2012 hurricane seasons. Those were earlier snapshots before the 2017 cancellation and are not the current balance.

Reform laws and pricing

• Biggert-Waters Flood Insurance Reform Act of 2012 (Title II of P.L. 112-141) began phasing out NFIP subsidies. The table of contents we opened lists Sec. 100205 "Reform of premium rate structure," Sec. 100212 "Reserve fund," and Sec. 100236 "Study of participation and affordability for certain policyholders." The Title II body text was truncated, so quoted provisions beyond those headings are from CRS. (P.L. 112-141, 126 Stat. 405, GovInfo; opened, headings only)
• The Homeowner Flood Insurance Affordability Act of 2014 (HFIAA) capped annual increases for primary residences at 18 percent per property, set surcharges of $25 for primary residences and $250 for others, and created a "newly mapped" subsidy for properties mapped into a high-risk zone on or after April 1, 2015. A reserve fund assessment was set at 15 percent of premium in April 2016 and raised to 18 percent in April 2020. (CRS R45999; opened)
• Risk Rating 2.0 "came fully into effect on April 1, 2022." Premiums now reflect "an individual property's specific flood risk," using factors such as distance from water, flood type, foundation type, lowest-floor height, prior claims and replacement cost. Flood zones no longer set premiums. Primary residences are capped at 18 percent a year; other properties, including non-primary homes, businesses and severe repetitive loss properties, must rise 25 percent a year to full risk-based rates. CRS notes there is no appeal procedure for Risk Rating 2.0 premium changes. Community Rating System discounts range from 5 to 45 percent. (CRS Insight IN11777, May 2024; opened)
• CRS R45999 reports that Risk Rating 2.0 is projected to raise premiums for "77% of NFIP policyholders." The CRS Insight above does not give the share of policyholders whose premiums go up or down. (Attributed projection; the FEMA original was not opened. FEMA's Risk Rating page could not be fetched.)

Timeline

Repetitive-loss properties: figures differ

| Source | Share of insured properties | Share of claims | Period and definition | |---|---|---| | GAO-04-401T (2004) | about 1 percent of 4.4 million | about 38 percent of claim costs | 1978 to Nov. 2003; two or more claims of $1,000 or more | | Pew infographic (Oct. 2016, FEMA data Jan. 31, 2016) | 1 percent | about 25 to 30 percent of flood claims; more than 150,000 properties; more than $12.5 billion | two claims above $1,000 in any 10 years | | Starbuck thesis summary (Dec. 2016) | 3.8 percent of policyholders | 35.5 percent of claims, 30.5 percent of payments | FEMA data, 1978 to 2015 | | GAO-26-109045 highlights (Mar. 2026) | about 2.5 percent of policies | 48 percent of claim dollars | properties with two or more losses |

Sources used different definitions, years and denominators, so the figures cannot be averaged. Pew also reported that about 1 in 10 repeatedly flooded properties received payments worth more than their value (FEMA, DHS Inspector General, GAO and CRS figures compiled by Pew). Starbuck reported that about 90 percent of repetitive-loss properties receive pre-FIRM or grandfathered subsidies (a FEMA estimate cited in his summary). (GAO-04-401T; Pew; Starbuck, Homeland Security Affairs; GAO-26-109045; all opened)

Subsidized rates: figures differ

(NAS Report 2; Insurance Business; opened. Neptune Flood is a private flood insurer; we opened the news report, not the study.)

Who benefits: named studies

• Neptune Flood Research Group (as reported): the poorest fifth of census tracts receive the smallest share of subsidy, 15.8 percent (about $419 million a year), with an effective discount of about 27 percent versus 33 to 36 percent for the upper three quintiles. Non-primary properties hold 42 percent of subsidy dollars on 28 percent of policies. Properties with replacement cost above $1 million are 4 percent of the book and at least 17 percent of subsidy. Florida holds about $1.25 billion, roughly 48 percent. (Insurance Business, Sept. 10, 2026; opened)
• National Academies Report 2 (2016): defining affordability needs a policy choice on cost burden (income, housing cost, or premium relative to property value). Means-tested simulation is not possible with NFIP data, which lacks income, wealth and housing cost fields. It rejected capping rates that transfer with property regardless of owner ability to pay as contrary to the 2012 act's goals, and described grandfathering as perpetuating cross-subsidies. (NAS; opened)
• Landry, Turner and Petrolia (2021): coastal Gulf households expecting disaster aid were 25 to 42 percent less likely to hold flood insurance; the authors estimate "charity hazard" may explain about 817,000 uninsured homes and about $526 million a year of forgone NFIP revenue, with caveats about a small sample and extrapolation. (NSF PAR record; opened)

What this case shows

1. The program's written purpose was risk sharing plus land-use change, and the second half is hard to see. The statute says insurance should encourage sound land use. Debt, claims and rates are tracked; whether local land-use decisions changed is not shown in the sources we opened.
2. Prices that sit below risk send a weaker signal to buyers. The National Academies report notes that grandfathering and non-risk-based classes weaken the signal for mitigation. The signal is hidden in a rating table most owners never read.
3. Costs are shared widely and decided at a distance. The program borrowed $22.5 billion from the Treasury and had $16 billion cancelled by Congress in 2017. The people who bought policies, taxpayers who did not, and homeowners facing higher bills did not vote on those particular choices.
4. Both reforms ran into the same affordability problem. Biggert-Waters ordered subsidies phased out and also ordered an affordability study (Sec. 100236). HFIAA 2014 added caps. NAS says current data cannot simulate means-tested help because income is not collected. A fix is hard to design when the data needed to design it is not held.
5. Appeal rights are thin. CRS reports no appeal procedure for Risk Rating 2.0 premium changes.

The strongest objections, and our answers

1. Objection: Subsidies protect families who cannot afford true risk prices. Mitigation: We state this. NAS recommends targeted assistance rather than blanket subsidy, and notes the affordability question is a policy choice about who counts as burdened.
2. Objection: Risk-based prices are the only fair system. Mitigation: We state this too. Risk Rating 2.0 aims at property-specific risk. CRS says 77 percent of policyholders were projected to see increases and no appeal exists. Both facts are in the record, and we draw no conclusion about fairness.
3. Objection: Moral hazard is overstated. Mitigation: The evidence is mixed. Starbuck's thesis argues that subsidies and repeat-loss coverage reduce incentives to mitigate. Landry et al. find a different effect, that expected disaster aid lowers insurance purchase ("charity hazard"). NAS says takeup responses to price changes are poorly documented. We do not claim moral hazard is proven.
4. Objection: The debt is not a program failure but a result of catastrophic years. Mitigation: Starbuck reports that significant events account for most losses. We also note that interest of $6.17 billion has been paid since 2005, and $2.82 billion of principal repaid. We do not assign blame.
5. Objection: Repetitive-loss statistics are inconsistent. Mitigation: Agreed. We tabulate four versions with definitions and do not choose one.
6. Objection: The Neptune distribution figures come from an insurer. Mitigation: True. We label it as an industry source reported by a trade outlet and did not open the study itself.

Open questions: what would change our minds

• The enactment date of the 1968 act and the text of P.L. 90-448 Title XIII were not opened; CRS gave the title and public law number only.
• Biggert-Waters body text was truncated in our fetch. The 25 percent annual increase and subsidy phase-out descriptions come from CRS.
• The FEMA Risk Rating 2.0 page failed to load, so FEMA's own description and the 77 percent figure's original were not opened.
• The date of CRS R45999 was not confirmed; its 2018 subsidy shares may be dated.
• Neptune's study and methodology were not opened.
• We did not open the GAO 2014 report on forgone premiums ("cannot be measured") named in the NAS appendix, so any total subsidy cost estimate is not given as established.
• Evidence that the program changed local land use (the 1968 statute's stated aim) was not located.
• How many policies lapsed after Risk Rating 2.0 is not known from what we read.
• Whether Congress's current reauthorization status or any 2026 change affects the debt figures was not checked beyond the March 2026 CRS update.

Sources

Evidence status and correction history

Key claims

Correction history

Found an error or a better source? We keep every correction on the record with its date and reason. Nothing here is quietly rewritten.