# Congress Renewed Flood Insurance but Left 13 Million High-Risk Properties Outside Its Purchase Map

Published: 2026-11-01

[Donald Trump](https://factrail.ai/people/donald-trump) is a relevant federal political actor, but congressional changes to the purchase mandate require legislation rather than presidential announcement alone.

# Congress Renewed Flood Insurance but Left 13 Million High-Risk Properties Outside Its Purchase Map

Factrail investigation | United States | 1 November 2026 | Documentary record current to 1 November 2026

The finding. Congress kept the National Flood Insurance Program operating through 11 December 2026, but the law it extended still makes compulsory flood insurance depend heavily on a federal flood-zone boundary that excludes many properties another risk model considers dangerous. The [Government Accountability Office's 1 October audit](https://files.gao.gov/reports/GAO-27-108012/index.html) estimates that approximately 13 million high-risk properties lie outside FEMA's designated special flood hazard areas or in unmapped locations. It also estimates that 86% of high-risk properties lacked federal NFIP coverage in April. That is not proof that 86% lacked every kind of flood insurance: private policies were unavailable at property level for this calculation. Congress, rather than FEMA alone, controls the statutory expansion GAO recommends. The immediate winners and losers are not a secret list of corrupt contractors. They are households facing uninsured losses, borrowers who avoid premiums today, lenders managing collateral, private insurers competing for policies and taxpayers exposed to disaster-recovery costs. No evidence examined establishes a political payoff or an unlawful decision.

## What the documents add beyond the headline

The new accountability question is not simply whether FEMA drew an inaccurate map. It is why federal law uses one kind of map to determine whether a borrower must buy insurance while the same federal program can price insurance using broader [flood-risk](/drivers/flood-risk) information. That difference appears in GAO's [full audit, GAO-27-108012](https://files.gao.gov/reports/GAO-27-108012/index.html), which explains that FEMA's Risk Rating 2.0 pricing methodology considers heavy-rain flooding even though the older special-flood-hazard-area designation generally does not. The public therefore encounters two related but legally distinct signals: a risk-based price and a regulatory line controlling a mortgage condition.

This investigation cross-checks three documentary routes: the GAO's property-level analysis; [Public Law 119-103, signed on 2 September](https://www.govinfo.gov/content/pkg/PLAW-119publ103/html/PLAW-119publ103.htm), which temporarily extended the program; and the [House's separate NFIP Extension Act bill](https://www.govinfo.gov/content/pkg/BILLS-119hr5577rh/html/BILLS-119hr5577rh.htm), which shows how lawmakers could debate continuation without rewriting the coverage trigger. The distinction matters because a temporary extension is a real enacted decision, while the four reforms GAO recommends remain proposals for congressional consideration.

There is no verified evidence here that an individual lawmaker deliberately sought to leave a particular home uninsured. The mechanism is structural and older than the current administration. The defensible criticism is that Congress has maintained a narrow statutory purchase test despite increasingly detailed information about rainfall-related risk. Whether any particular proposed replacement would be better depends on costs, model quality, affordability and the ability to challenge an erroneous classification.

## The documentary timeline: the insurance rule and the renewal are different decisions

Congress established the National Flood Insurance Program in 1968, combining insurance access with floodplain management. Subsequent law tied a mandatory-purchase requirement to designated special flood hazard areas for properties financed with certain federally backed mortgages. Those areas generally reflect a 1% annual probability of flooding, historically focused on river and coastal hazards. The requirement is conditional: not every property inside a zone must carry an NFIP policy, and private policies can satisfy applicable requirements. A cash buyer or an owner without a covered mortgage is not automatically subject to the same federal obligation.

The program's legal architecture became a rolling political decision. The [Congressional Research Service's March 2026 account](https://www.everycrsreport.com/files/2026-03-05_R44593_04ca583cd7d91f687e4bbdcf3c2122e6d94e6e11.html) documents federal borrowing and repeated authorization questions, including approximately $22.525 billion in outstanding Treasury debt at that reporting date. That is the program's accumulated borrowing balance, not a new bill generated by the October audit and not a measurement of the uninsured homes' losses.

On 26 September 2025, Representative Andrew Garbarino introduced H.R. 5577 to extend the NFIP. The [official House-reported text dated 15 January 2026](https://www.govinfo.gov/content/pkg/BILLS-119hr5577rh/html/BILLS-119hr5577rh.htm) identifies additional sponsors Buddy Carter, Jeff Van Drew, Michael Lawler, Jared Moskowitz, Nick LaLota and Josh Gottheimer. The text concerns authorization and financing dates, not a new rule for rainfall risk or a higher residential insurance limit. These are named legislators associated with a documented bill; sponsorship does not prove that they opposed any of GAO's later recommendations.

A separate enacted instrument ultimately supplied the current temporary authority. Section 139 of [Public Law 119-103](https://www.govinfo.gov/content/pkg/PLAW-119publ103/html/PLAW-119publ103.htm) substitutes the expiration date in section 106(3), 11 December 2026, into the NFIP's authorization and financing provisions. The official legislative history records House passage in January, Senate action in August, House concurrence in September and presidential approval on 2 September. [President Donald Trump](/people/mig-donald-trump) therefore signed an extension, not the map-expansion legislation contemplated by the subsequent audit. It would be inaccurate to call the extension a program shutdown or to say that signing it caused the original map design.

On 1 October 2026, GAO issued its [new audit](https://www.gao.gov/products/gao-27-108012). Its congressional addressees include Senators Elizabeth Warren and John Kennedy, and Representatives French Hill and Maxine Waters, holding relevant banking or financial-services oversight roles in the report's addressee list. Receiving an audit establishes an institutional route for oversight; it is not evidence that any addressee voted on or rejected the four proposed reforms. By 1 November, the report's four matters for congressional consideration were shown as open on GAO's public page. We did not establish a later enacted reform in the sources reviewed.

## What 13 million actually measures

GAO compared FEMA's regulatory zones with property-level flood-risk information from First Street, a private modeling organization. First Street identified roughly 16 million properties with a risk corresponding approximately to a 1% annual flood probability. Around 13 million of those were outside FEMA's special flood hazard areas or in places lacking the relevant map. This is a comparison between two methods, not a finding that FEMA officially misclassified 13 million individual addresses under its current legal mapping rules.

GAO classed First Street flood scores of four through ten as high risk and matched approximately 79% of NFIP policies to property records. It adjusted its estimates for unmatched records. That leaves genuine model and matching uncertainty. Private flood insurance records were available as state-level totals rather than a full property-level match. Thus the striking 86% number means no identified NFIP coverage in the modeled high-risk group, not a verified nationwide uninsured rate across federal and private markets. [GAO's published methodology](https://files.gao.gov/reports/GAO-27-108012/index.html) is unusually explicit about these limitations.

The geographic consequences are unequal. In GAO's selected communities, it estimated NFIP coverage among high-risk properties at 52% in Orleans Parish, Louisiana, 39% in Lee County, Florida, 9% in Sacramento County, California, and only 5% in Buncombe County, North Carolina. These are modeled county estimates, not a census of every resident's insurance status. They illustrate why an inland rainfall event can surprise a market trained to associate flood insurance primarily with coastlines and riverbanks.

[Realtor.com's 7 October reporting](https://www.realtor.com/news/trends/13-million-high-risk-homes-fema-flood-zones-el-nino/) emphasizes that heavy rainfall is the predominant modeled threat for 62% of the high-risk properties outside the mapped zones. [NewsNation's 8 October report](https://www.aol.com/articles/nearly-13m-risk-properties-not-205550000.html) independently frames the same gap as a problem for household decisions during a wet-weather outlook. Those outlets corroborate the public significance and explain the hazard, but their repetition of GAO's count does not constitute an independent replication of First Street's property data.

## The decision and veto map

Congress controls the statutory test for mandatory purchase, the federal program's borrowing authority and the maximum NFIP coverage limits. It can amend the relevant law, require a risk disclosure or decide to fund means-tested assistance. Passing a bill requires the ordinary bicameral legislative process and presidential approval or a veto override. A committee chair can convene hearings and shape a bill but cannot unilaterally rewrite the statute.

FEMA, within the [Department of Homeland Security](/organizations/department-of-homeland-security), administers the NFIP, works with more than 22,000 communities on flood maps and uses modern risk factors in premium pricing. GAO says FEMA lacks explicit authority to replace the mandatory-purchase criteria without Congress. FEMA also faces public-review and appeal processes for map changes, a safeguard for property rights but one that can make updates slow. The audit cites a historical average of about seven years to complete a map as of February 2020, not a newly measured 2026 processing time.

Federal mortgage and housing actors implement the purchase rule in transactions they supervise or finance. Lenders must assess whether covered properties meet the statutory test. The Department of Agriculture's Rural Housing Service is especially important because it supplied the most concrete official opposition to GAO's proposals: it warned that expanding compulsory coverage could raise housing costs and that quote-and-declination paperwork could delay closings and create liability. These objections appear in GAO's agency-comment section, not as anonymous political speculation.

Private risk and insurance businesses have different powers. First Street produces a model used by the audit; it does not determine federal law. Insurers sell private flood policies or administer NFIP policies through federal arrangements. The [FEMA Write Your Own program explanation](https://www.fema.gov/pdf/nfip/manual200705/02ref.pdf) describes the mechanism: participating insurers service federally underwritten flood policies and receive expense allowances, while the federal government retains underwriting losses. That historical program document establishes a commercial channel, not a current payment to any specific carrier. Lenders, real-estate brokers and agents also influence whether buyers understand risk, but they do not possess Congress's legal veto.

## Who receives money, and who carries the loss

Start with the policyholder. A homeowner outside a designated zone with a federally backed mortgage may avoid a mandatory premium even when a broader model assigns high flood risk. The short-term cash benefit is the avoided insurance expense. The longer-term liability is potential repair cost after a flood. That is not a subsidy payment from Washington to the homeowner; it is an exposure retained on a private balance sheet, potentially with downstream demands for disaster relief.

The magnitude of that exposure is not captured by a single average, but GAO offers a useful comparison. For flood-related events in 2015–2025, inflation-adjusted average residential NFIP claims were approximately $88,700, compared with average federal individual disaster assistance of about $3,500. The two programs have different eligibility rules and beneficiary populations, so the difference is not an individual's guaranteed shortfall. It shows why disaster aid cannot reasonably be treated as a substitute for an insurance policy. The audit also notes that most conventional homeowners policies exclude flooding.

For mortgage lenders, flood insurance protects collateral and can reduce losses from property damage. Extending a requirement to additional borrowers may protect some loans but also raise monthly ownership costs and potentially complicate originations. For insurance agents and companies, more insured properties can mean additional premiums or service fees, but also more claims exposure in privately underwritten business. It is not responsible to declare all insurers beneficiaries of reform without distinguishing federally backed policies from private risk-bearing contracts.

The NFIP itself combines premiums, federal borrowing and disaster-loss payments. The [CRS debt record](https://www.everycrsreport.com/files/2026-03-05_R44593_04ca583cd7d91f687e4bbdcf3c2122e6d94e6e11.html) illustrates why taxpayers have a financial stake: large claims have previously required Treasury borrowing. More premium-paying participants could spread some risk and improve household protection, but broader enrollment could also add high-risk properties whose claims exceed premiums. Without pricing and take-up estimates, neither fiscal outcome can be asserted.

The private market is not absent. GAO found that private policies accounted for 14% of flood insurance policies in 2025, with growth partially offsetting the NFIP policy decline from 5.5 million in 2010 to 4.5 million in 2026. The [American Bankers Association's 7 October account](https://bankingjournal.aba.com/2026/10/gao-urges-congress-to-expand-flood-insurance-coverage-for-at-risk-properties/) reports the audit's recommendation to expand coverage while noting means-based assistance and private-market barriers. Its publication is a stakeholder-industry source, not a vote or a legal endorsement by every lender.

There are also concrete coverage ceilings. Congress has left NFIP building limits at $250,000 for typical residential properties and $500,000 for commercial properties since 1994. GAO estimates inflation adjustment would put those figures near $492,000 and $984,000 in 2025 dollars. It found about 61% of residential NFIP policies were at the statutory building limit. Increasing a cap may allow higher coverage but can require higher premiums and greater exposure for the program. A higher home sale price is not the same as a higher insured replacement cost, and the two should not be casually equated.

## The strongest counter-case: a mandate is not free protection

The most substantial objection to GAO's first recommendation came from an agency inside government, not from a lobbyist. USDA's Rural Housing Service opposed expanding the criteria used for compulsory flood insurance, arguing that requiring more households to buy policies would significantly increase housing costs. It also opposed a new pre-closing quote and signed-declination requirement because of paperwork, closing delays and legal exposure. The agency supported better risk information and higher optional coverage limits. Its preference for removing barriers to private insurance is recorded in the same audit.

That objection is material. A household with little disposable income could be unable to absorb a new annual premium even when its flood risk is real. A broader federal requirement without affordability assistance might increase protection for some families while pricing others out of a home or causing mortgage friction. The GAO did not calculate the premium each newly included borrower would pay, and it did not propose that every one of the 13 million properties automatically be forced to buy an NFIP policy.

GAO's reply is that being uninsured can be even less affordable after a flood and that means-tested assistance could soften the burden. This is a defensible policy argument, not a completed cost-benefit proof for every community. The agency also the administrative burden of additional disclosure. It maintained that the information benefit would justify that burden, but lawmakers would still need to design the rule and determine who pays to implement it.

A second counter-case concerns models. First Street's risk scores include hazards FEMA's regulatory maps generally omit, but model divergence alone does not identify which address is legally misclassified. Regulatory maps are used for floodplain management and due-process decisions, not merely consumer prediction. Public property-level data could help buyers and insurers but may also raise privacy and fairness questions. The Rural Housing Service supported disclosure only with safeguards and separation from the legal flood-zone designation. This is a reason to design transparency carefully, not a reason to pretend the risk difference disappears.

## Four choices Congress has not resolved

The [GAO recommendations](https://www.gao.gov/products/gao-27-108012) identify four congressional levers. First, lawmakers could direct purchase criteria to consider heavy rainfall and other flood sources rather than relying only on the current regulatory-zone test. Second, they could require accessible property-level risk information, with safeguards for sensitive claims history. Third, they could make a flood-insurance quote and signed refusal part of closing a federally backed mortgage, even where purchase remains voluntary. Fourth, they could increase NFIP coverage limits and create periodic adjustments.

These proposals differ sharply in coercion and cost. An information rule might increase awareness without imposing insurance. A quote requirement adds transaction work but preserves the option to decline where coverage is voluntary. A broader mandate changes household cash flow directly. A higher coverage ceiling increases the amount a willing buyer can insure but does not automatically raise the number of buyers. Presenting all four as a single expansion plan would hide the legislative choices.

The [Association of State Floodplain Managers' 2 October assessment](https://www.floods.org/news-views/research-and-reports/most-at-risk-properties-remain-uninsured-gao-finds-urging-congress-to-act/) emphasizes the need for congressional action. Its members have professional experience in floodplain management, but it is an interested stakeholder voice, not an independent fiscal scoring agency. Conversely, the [Associated Press's earlier account of authorization lapses](https://apnews.com/article/bc5fb2154d39fd80217bc7a7e60454aa) documents the market disruption that short-term renewals can avert. Avoiding a lapse is a legitimate achievement even if it does not address the deeper mapping question.

## The public consequences and the test for accountability

Three observable indicators would distinguish a successful reform from a press release. First is take-up among modeled high-risk properties, measured with both NFIP and private policy records rather than federal policies alone. Second is affordability, especially premium burden and mortgage completion among lower-income owners newly affected by a rule. Third is loss absorption after flooding: the distribution of insured claims, uninsured repairs, disaster assistance and NFIP borrowing over time. A reform that raises policy counts but excludes poorer households would not be an unqualified success.

There are also clear institutional checkpoints. Congress can hold hearings on the four open GAO matters, publish a bill with a precise risk threshold and affordability mechanism, and disclose a fiscal estimate. FEMA can explain what existing Risk Rating 2.0 data can lawfully support, what is withheld for privacy and how an owner could challenge a property-level assessment. Housing agencies and lenders can quantify closing costs and propose a limited disclosure test. Those steps would create an evidence trail that allows voters to judge actual decisions rather than rhetoric.

The documentary verdict is narrow but consequential. The 2026 renewal retained an operating federal insurance program, while the older statutory purchase boundary remained intact. GAO's new comparison shows a large gap between that boundary and one credible modern risk model; the gap is associated with much lower federal insurance uptake outside designated zones. The record does not prove 13 million uninsured households, intentional concealment, a corrupt transfer to insurers, or a single politician's responsibility for every flood loss. It does show where legal power resides and what evidence Congress would need to confront before claiming that America's flood insurance system protects the properties most exposed to loss.

## Sources and verification boundaries

The central primary records are the [GAO's October 2026 audit and methodology](https://files.gao.gov/reports/GAO-27-108012/index.html), [Public Law 119-103, sections 106 and 139](https://www.govinfo.gov/content/pkg/PLAW-119publ103/html/PLAW-119publ103.htm), the [House-reported H.R. 5577](https://www.govinfo.gov/content/pkg/BILLS-119hr5577rh/html/BILLS-119hr5577rh.htm), the [Congressional Research Service's March 2026 NFIP analysis](https://www.everycrsreport.com/files/2026-03-05_R44593_04ca583cd7d91f687e4bbdcf3c2122e6d94e6e11.html), and [FEMA's explanation of private servicing under federal underwriting](https://www.fema.gov/pdf/nfip/manual200705/02ref.pdf). Independent and interested-source perspectives include [Realtor.com](https://www.realtor.com/news/trends/13-million-high-risk-homes-fema-flood-zones-el-nino/), [NewsNation](https://www.aol.com/articles/nearly-13m-risk-properties-not-205550000.html), the [American Bankers Association](https://bankingjournal.aba.com/2026/10/gao-urges-congress-to-expand-flood-insurance-coverage-for-at-risk-properties/), the [Association of State Floodplain Managers](https://www.floods.org/news-views/research-and-reports/most-at-risk-properties-remain-uninsured-gao-finds-urging-congress-to-act/) and [Associated Press](https://apnews.com/article/bc5fb2154d39fd80217bc7a7e60454aa). These do not all independently recalculate the GAO numbers. No personal donation trail, carrier-level current reimbursement, unverified Factrail internal URL or image licence is asserted.

## Methodology

https://factrail.ai/methodology

Canonical: https://factrail.ai/articles/us-flood-insurance-congress-13-million-risk-map-gap-2026