The Uninsurable Coast

In September 2022, Hurricane Ian made landfall near Fort Myers, Florida, as a Category 4 storm. The damage was extensive: flooded neighborhoods, destroyed infrastructure, over 150 deaths. But the economic aftershock that followed was not simply a function of wind and water. It was a function of what happened to insurance in the months and years after the storm passed.

By 2024, home values in Lee County, which includes Fort Myers, had fallen more than 16 percent from their pre-storm trajectory.¹ The decline was not caused solely by physical damage. It was caused by the repricing of risk. Insurers who had been retreating from Florida before Ian accelerated their withdrawal afterward. Premiums for the policies that remained climbed to levels that many homeowners could not absorb. Prospective buyers, confronting insurance quotes that added thousands of dollars to the annual cost of ownership, adjusted their offers downward or looked elsewhere.

The mechanism is simple. When the cost of insuring a home rises faster than the local economy can support, property values fall. When property values fall, the wealth stored in those homes, for most American families their largest asset, erodes. The erosion is not distributed equally. It falls hardest on people who cannot relocate, cannot absorb the cost increase, and cannot afford to sell at a loss.

Fort Myers is one case. The pattern is national.

The map of withdrawal

Lake Charles, Louisiana, was hit by Hurricane Laura in August 2020 and Hurricane Delta six weeks later. The back-to-back storms caused billions in damage. Four years later, the city had lost approximately 5 percent of its population.² The departures were not all directly caused by the storms. Some were caused by what came after: the insurance cancellations, the premium increases, the difficulty of obtaining coverage at any price, and the slow realization that the financial infrastructure supporting homeownership in the region was contracting.

In Paradise, California, the 2018 Camp Fire destroyed nearly 19,000 structures and killed 85 people. Residents who rebuilt, many of them using fire-resistant materials and modern building codes, discovered that their insurance premiums had risen from approximately $1,200 per year before the fire to $9,750 per year afterward.³ The new homes were objectively safer than the ones they replaced. The premiums reflected not the individual building’s risk profile but the area’s risk classification. The zip code had been repriced, and every structure in it bore the cost.

These are not isolated stories. They are expressions of a systemic process. The insurance industry is withdrawing from climate-vulnerable regions across the United States, and the withdrawal is not random. It follows the geography of risk: coastlines exposed to hurricanes, wildland-urban interfaces exposed to wildfire, river corridors exposed to flooding, and increasingly, the broad swaths of the interior exposed to severe convective storms, hail, and tornado activity.

Since 2018, insurers have issued approximately 1.9 million policy non-renewals in climate-exposed areas across the country.⁴ The non-renewals are concentrated in the states where climate risk is highest, but they are not confined to those states. As reinsurance costs rise globally, the repricing pressure extends to markets that historically considered themselves low-risk.

The result is a new map of American viability, drawn not by planners or legislators but by underwriters.

The demographics of uninsurance

The insurance withdrawal does not affect all Americans equally. The disparities are stark, well-documented, and follow the contours of existing inequality with uncomfortable precision.

A 2024 study by the Federal Reserve Bank of Chicago found that Black homeowners are 8.3 percentage points less likely to have homeowner insurance than white homeowners, after controlling for income, home value, and other demographic factors.⁵ The gap is not explained by economic differences alone. It reflects a combination of factors: historical patterns of residential segregation that concentrated Black homeownership in higher-risk areas, lower rates of mortgage-held homes (mortgages typically require insurance), and the legacy effects of redlining on neighborhood infrastructure and housing stock quality.

The disparities extend across racial and ethnic groups. Approximately 22 percent of Native American homeowners, 14 percent of Hispanic homeowners, and 11 percent of Black homeowners lack homeowner insurance, compared to lower rates among white homeowners.⁶ These are not small differences. They represent millions of households whose primary asset is unprotected against the climate events that are becoming more frequent and more severe.

The exposure disparities compound the insurance disparities. A 2024 analysis by Zillow and First Street Foundation found that 81 percent of Black homeowners in the United States live in areas facing extreme heat risk, compared to 52 percent of white homeowners.⁷ The same pattern holds, with varying magnitudes, for flood risk, wildfire risk, and severe storm risk. The communities most exposed to climate hazards are disproportionately communities of color, and these are the same communities where insurance coverage is most likely to be absent.

This is the architecture of compounding vulnerability. Higher risk, less insurance, fewer resources for recovery, greater likelihood of financial devastation when a disaster occurs.

The credit score cascade

When a disaster strikes an uninsured or underinsured household, the financial consequences extend well beyond the immediate cost of repairs. A growing body of research has documented what happens to the credit profiles of disaster-affected households, and the findings are disturbing in their asymmetry.

Studies of post-disaster credit outcomes have found that communities of color experience an average credit score decline of 31 points in the years following a major disaster, compared to a 4-point decline for white communities in similarly affected areas.⁸ A 31-point credit score decline is not a minor fluctuation. It can push households below the thresholds that determine access to credit cards, auto loans, and mortgages. It increases borrowing costs for years. It compounds the wealth destruction caused by the disaster itself.

The mechanism is straightforward. Uninsured losses lead to unpaid bills, missed mortgage payments, and increased reliance on high-cost credit. These financial behaviors are recorded by credit bureaus and reflected in credit scores. The scores, in turn, determine the terms on which households can access the financial system. Lower scores mean higher interest rates, reduced credit limits, and in some cases, denial of credit altogether.

For households that were already financially precarious before the disaster, the credit score decline can trigger a spiral: uninsured loss leads to financial stress, which leads to credit deterioration, which leads to reduced access to affordable credit, which leads to greater financial stress. The spiral does not reverse itself automatically. It requires resources (savings, family support, access to affordable lending) that the most affected households are least likely to have.

Bluelining

In 2023, the Center for International Environmental Law published a report documenting a phenomenon it termed “bluelining,” a deliberate parallel to the historic practice of redlining.⁹

Redlining, as practiced by the Home Owners’ Loan Corporation and private lenders from the 1930s onward, involved the systematic denial of mortgage credit to neighborhoods based on their racial composition. The practice was made illegal by the Fair Housing Act of 1968, but its effects persisted for decades in the form of depressed property values, disinvested infrastructure, and concentrated poverty in formerly redlined areas.

Bluelining describes a structurally similar process operating through a different mechanism. Instead of race (at least not explicitly), the sorting variable is climate risk. Instead of mortgage lenders, the sorting agents are insurance companies. Instead of government-drawn maps, the boundaries are defined by actuarial models and catastrophe simulations.

The parallel is not exact. Redlining was a deliberate policy of racial exclusion. Bluelining is, at least in its formal logic, a response to physical risk. But the outcomes overlap in ways that are difficult to dismiss as coincidental. Because communities of color are disproportionately located in high-risk areas (a legacy of the same discriminatory housing policies that produced redlining), the climate-risk-based withdrawal of insurance falls disproportionately on the same populations that were targeted by the earlier practice.

The CIEL report documented specific cases in which the withdrawal of insurance coverage from flood-prone, hurricane-exposed, or wildfire-vulnerable areas produced demographic and economic effects strikingly similar to those produced by mid-20th-century redlining: declining property values, population loss, erosion of public services, and the concentration of poverty in the areas left behind.⁹

Whether bluelining is “discrimination” in the legal sense is a question that courts have not yet resolved. Insurers argue, with some justification, that they are pricing risk, not targeting populations. But the distinction between intent and impact is one that civil rights law has grappled with for decades, and the impact in this case falls with familiar precision along lines of race and income.

The migration signal

People respond to insurance signals, even when the signals are indirect. A Redfin analysis of 2024 migration data found a net out-migration of 29,027 people from counties classified as having high flood risk.¹⁰ The number is modest in absolute terms but significant as a trend indicator. It represents one year of a process that is projected to accelerate.

First Street Foundation’s modeling suggests that approximately 55 million Americans may need to relocate by 2055 due to climate-related risks, including flooding, wildfire, extreme heat, and the insurance and economic consequences of those risks.¹¹ The projection is subject to significant uncertainty, but even at half that scale, it would represent one of the largest internal migrations in American history.

The migration is not centrally planned. No agency is coordinating it. No relocation assistance is available for the vast majority of households. The movement is driven by the accumulation of individual decisions: a family that decides not to rebuild after a flood, a couple that factors insurance costs into their choice of retirement location, a young household that crosses a community off its list because the all-in cost of homeownership (mortgage, taxes, insurance) exceeds what they can afford.

The communities receiving these migrants benefit from the influx of population and economic activity. The communities losing them face the opposite dynamic. And because the losses are gradual, they do not produce the visible crisis that triggers policy intervention. A city that loses 5 percent of its population over five years does not make national news. It makes local news once, briefly, when the Census data is released, and then the slow contraction continues without attention.

The municipal finance problem

Property taxes are the fiscal foundation of American local government. In most states, property tax revenue funds the majority of K-12 education, a significant share of public safety, and much of the infrastructure maintenance that keeps communities functional.

When property values decline due to climate risk and insurance repricing, the property tax base declines with them. The decline creates a fiscal dilemma: the community must either raise tax rates (imposing higher costs on remaining residents and businesses, accelerating further departure) or cut services (reducing the quality of life that retains residents and attracts new ones).

A 2025 study published in Nature Cities examined the exposure of the $4.2 trillion U.S. municipal bond market to climate risk.¹² The study found that a significant and growing share of outstanding municipal bonds were issued by jurisdictions with material exposure to climate-related property value declines, tax base erosion, and infrastructure costs. The bonds had not, at the time of the study, been repriced to reflect this exposure. The authors noted that when repricing occurs, it will increase borrowing costs for the most vulnerable communities at precisely the moment they can least afford it.

This is the fiscal dimension of the insurance withdrawal. It is not merely that individual homeowners face higher costs or reduced coverage. It is that the economic foundation of local government in climate-exposed areas is being undermined by a process over which local officials have almost no control. They did not cause climate change. They do not set insurance prices. They cannot prevent carriers from leaving their jurisdictions. But they bear the consequences, in the form of shrinking tax bases, deteriorating services, and the slow departure of the residents and businesses on which their fiscal viability depends.

The scale of uninsurance

Approximately 12 percent of American homeowners now lack any form of property insurance, up from roughly 5 percent in 2019.¹³ The increase has been rapid and shows no sign of leveling off. At current trends, the uninsured share of the housing stock will continue to grow as premiums outpace incomes and as carriers continue to exit high-risk markets.

Over six million homes are estimated to be completely uninsured. Millions more are underinsured, carrying policies with coverage limits or deductibles that would leave significant losses uncovered in the event of a major climate event. In many markets, insurance now represents 20 percent or more of the total monthly cost of homeownership (mortgage payment plus insurance plus taxes), up from single-digit percentages a decade ago.¹⁴

The projected scale of the problem is larger still. First Street Foundation estimates that climate-driven insurance repricing will erase approximately $1.47 trillion in residential property value across the United States over the next three decades.¹⁵ The losses will be concentrated in coastal areas, the wildfire-urban interface, and flood-prone river corridors. But because insurance costs feed into mortgage affordability calculations, which feed into property valuations, which feed into tax bases, the effects will ripple outward from the most exposed areas into regional and national economic systems.

The accountability gap

The insurance industry’s withdrawal from climate-vulnerable communities raises a question that the current policy framework does not answer: who is accountable for the consequences?

Insurers are private companies. They have no obligation to write policies in any particular market. Their fiduciary duty runs to their shareholders and existing policyholders, not to the communities they serve. When they conclude that a market is no longer profitable, they leave. This is how markets work.

But the consequences of their departure are borne publicly. When property values fall, the wealth destruction falls on homeowners. When tax bases erode, the service cuts fall on residents. When credit scores decline, the reduced financial access falls on households. When communities shrink, the social costs (closed schools, reduced public safety, deteriorating infrastructure) fall on the people who remain.

The market is making decisions with public consequences, and there is no public process governing those decisions. Insurers are not required to provide advance notice of market withdrawal in most states. They are not required to consider the community-level impacts of their pricing decisions. They are not subject to the kind of democratic accountability that would apply if a government agency were making the same choices.

This is the accountability gap at the center of the insurance crisis. The decisions are private. The costs are public. The populations bearing the greatest costs have the least influence over the decisions.

What the Senate found

In 2024, the Senate Budget Committee conducted an investigation into the scope and consequences of the insurance withdrawal from climate-vulnerable communities. The investigation documented approximately 1.9 million policy non-renewals since 2018 in climate-exposed areas, representing one of the largest reductions in insurance coverage in American history.⁴

The committee’s findings underscored the disproportionate impact on low-income communities and communities of color. They documented cases in which entire neighborhoods had been effectively de-insured, not through a single dramatic decision but through the accumulation of individual non-renewals that, taken together, left the community without viable insurance options.

The investigation also revealed the opacity of the process. Insurers are not required to report non-renewal data at the zip-code level in most states. There is no centralized database tracking the geographic pattern of coverage withdrawal. Researchers and policymakers attempting to understand the scope of the problem must piece together information from state insurance departments, industry reports, and surveys, none of which provides a complete picture.

The opacity is itself a problem. A process that is reshaping the economic geography of the United States is proceeding without the basic data infrastructure that would allow policymakers, communities, or the public to understand what is happening, where it is happening, and to whom.

The inadequacy of current responses

The policy responses to the insurance crisis have been, for the most part, reactive, incremental, and insufficient.

State-level responses have focused on two approaches: rate suppression and market stabilization. Rate suppression (limiting how much insurers can charge) reduces costs for current policyholders but accelerates carrier withdrawal, because companies that cannot charge risk-adequate rates will eventually stop writing business. Market stabilization (expanding state-backed insurers of last resort, creating reinsurance funds, providing subsidies) treats the symptoms without addressing the underlying cause.

Federal responses have been similarly limited. The National Flood Insurance Program continues to operate despite its $22.5 billion debt and a rate structure that is only beginning to reflect actual risk. Disaster relief, while essential for immediate recovery, does nothing to prevent the next event or to address the long-term affordability and availability of insurance. Tax policy, building code standards, and land use regulation, the tools that could most directly reduce the underlying risk, remain largely unchanged.

The absence of a comprehensive federal strategy is itself a policy choice. It means that the insurance market continues to function as the de facto climate adaptation mechanism, sorting communities by risk and viability without democratic input and without any system for mitigating the consequences.

A direction

The insurance withdrawal from climate-vulnerable communities is not a problem that will resolve itself. The underlying climate risks are increasing. The insurance industry’s capacity and willingness to absorb those risks is decreasing. The gap between the two will widen until something structural changes.

Structural change requires action on multiple fronts simultaneously. It requires reducing the underlying physical risk through building code modernization, infrastructure hardening, natural barrier restoration, and in some cases, managed retreat from the most exposed locations. It requires reforming insurance regulation to balance affordability with actuarial sustainability, drawing on models like the UK’s Flood Re program that link subsidized premiums to risk reduction timelines. It requires creating the data infrastructure to track insurance availability and affordability at the community level, so that the process of withdrawal is visible and accountable. And it requires confronting the equity dimensions directly, ensuring that the costs of the transition do not fall disproportionately on the communities least responsible for the climate change driving it.

None of this will happen quickly. All of it is technically feasible. The obstacle is not knowledge. The obstacle is the political will to intervene in a process that is generating losses for diffuse, low-income, disproportionately nonwhite populations while generating profits for concentrated, well-organized corporate interests.

The insurance industry did not create climate change. It is not responsible for the historical patterns of segregation and disinvestment that concentrated vulnerable populations in high-risk areas. But it is the mechanism through which the consequences of both are now being distributed. And the distribution, in the absence of deliberate intervention, will follow the path of least resistance: toward the people with the least power to resist it.

The map of insurable America is being redrawn. The question is whether anyone other than actuaries will have a say in where the lines fall.