Home Insurance in Crisis


PENNY LIAO is an economist and fellow at Resources for the Future specializing in climate adaptation and natural disaster policy.

Published January 23, 2024


If you’re one of the millions of homeowners who was recently shocked by a fat increase in your insurance premium – or worse, were informed your insurance wouldn’t be renewed at any price – you don’t need to be told insurers (and thus their clients) are in trouble. A crisis is unfolding in the $125 billion U.S. market

as insured disaster losses have surged dramatically. Since 2017, the Western U.S. has been besieged by catastrophic wildfires, which has prompted an accelerating retreat of insurers from fire-prone areas. In 2019, insurers declined to renew coverage of 20-30 percent of policies in some of the riskiest zip codes in California. And in 2023, State Farm, Farmers and Allstate – which collectively write 40 percent of California homeowners’ coverage – have announced an indefinite pause in new policies within the whole state.

Meanwhile, the Gulf Coast states are also grappling with declining availability of insurance, in part due to escalating hurricane losses – this following on many major insurers’ exits from Florida and Louisiana in earlier years. Indeed, the exodus continues: after Hurricane Ian inflicted $1.5 billion in insured losses in 2022, Bankers and Farmers (among others) withdrew from Florida while AAA, which is based in the state, tightened its underwriting criteria. A similar, albeit more gradual, trend is discernable in Texas and the Carolinas – all states susceptible to hurricanes.

The issue of affordability has also grown more acute. Homeowners are discovering that even if they can renew existing policies, they face double-digit increases in premiums. The Consumer Federation of America estimates that between 2017 and 2022 home insurance premiums outpaced general price inflation by 40 percent. By no coincidence, the five states experiencing the steepest premium increases – Florida, New Mexico, Colorado, Idaho and Texas – are all vulnerable to hurricanes or wildfires. The crisis extends into other aspects of the insurance contract as well. Insurers are increasingly insisting on higher deductibles and limiting maximum coverage.

When homeowners find themselves excluded from the regular insurance market, they often turn to what’s called “the residual market” – programs backed or chartered by individual states that serve as insurers of last resort. For example, from 2015 to 2021 the number of policies in the California Fair Access to Insurance Requirement (FAIR) Plan nearly doubled, reaching 3 percent of all residential policies.

In Florida, the state-operated Citizens Property Insurance Corporation absorbed a staggering 150 percent surge between 2019 and 2022; it now underwrites some 1.3 million policies, which make up fully 18 percent of the market. For its part, Louisiana’s Citizens program grew by 270 percent in two years after the extensive hurricane damage of the 2020/2021 season. As these plans mainly cater to high-risk properties that the regular market would not cover, they come with stiff premiums in return for relatively limited coverage.

Liao Penny Home Insurance Crisis 2

While the blow of insurance non-renewals and premium increases is felt most acutely in high-risk areas, availability and affordability issues have a way of spreading. First, with major insurers halting new policy underwriting across a state, relatively safe areas have been affected. Second, recent research highlights a trend in which national insurers increase rates in states with lenient regulation to offset their inability to recover losses through rate increases in states with stringent rules.

Third, if a multistate insurer has extensive exposure to catastrophic events in one state, it may get out of the business entirely, affecting policyholders in other states. For instance, exposure to Hurricane Ian in Florida caused financial distress for two Louisiana-based insurers, Lighthouse and Americas Insurance, which went into receivership in 2022.

The tightening availability and affordability of homeowners’ insurance will reverberate in the broader housing and mortgage sector. Given that lenders universally require mortgage borrowers to maintain homeowners’ insurance, difficulty in securing such coverage at an affordable rate can lead to delays or even cancellations in mortgage applications and property sales, undermining the vitality of local housing markets.

And, of course, soaring premiums increase overall homeownership costs, raising the already- high barrier blocking low- and middleincome households from building wealth through home equity. Homeowners in risky areas may also find their properties harder to sell, or at very least, worth less in the market. This drag on the low and middle portion of the housing market could have major welfare implications since home equity stands as the largest asset category for the average American household. Moreover, the suspension of new underwriting is likely to slow construction, aggravating the affordable housing shortage plaguing many regions.

Consider, too, that problems with insurance availability and affordability raise concerns about enlarging the already considerable gap between coverage and potential losses. Owing to the insurance requirement by mortgage lenders, about 95 percent of homeowners do carry insurance. However, several analyses have found that a majority of U.S. homes are underinsured – on average by over 20 percent. Coverage is further reduced when insurers impose higher “hurricane” or “wildfire” deductibles or simply set lower coverage limits. Should insurance premiums keep growing rapidly, one can expect some homeowners without mortgages to drop coverage altogether. Others might opt for the minimum coverage required by the lender rather than insuring the full replacement cost of their homes. Such decisions leave households vulnerable, potentially stalling post-disaster recovery and weakening a community’s overall resilience.

Below, I delve deeper into the complexities of insuring disaster risks, looking into how a confluence of factors – from the inherent nature of the hazard to the regulatory environment – contributes to the current crisis. It’s not all bad news, though: there are potential solutions.

The 1994 Northridge Earthquake in Los Angeles County caused $20-billion-plus damage and led insurers to pull back from California. The state ultimately filled part of the gap by establishing the California Earthquake Authority (CEA) in 1996 to provide standalone earthquake coverage.
Losses, Losses Everywhere

The decline in private disaster coverage following catastrophic events is not a new phenomenon. Following the great Mississippi River flood of 1927 that submerged 23,000 square miles and displaced hundreds of thousands, private insurers deemed flood risks to be uninsurable. Widespread flood coverage only became available after the creation of the federally underwritten National Flood Insurance Program (NFIP) in 1968. (Homeowners’ insurance covers fire and wind, not flood damage.)

Similarly, the 1994 Northridge Earthquake in Los Angeles County caused $20-billionplus damage and led insurers to pull back from California. The state ultimately filled part of the gap by establishing the California Earthquake Authority (CEA) in 1996 to provide standalone earthquake coverage.

The homeowners’ insurance market itself has seen similar contractions since Hurricane Katrina in 2005, when nearly three million households lost their coverage and half of them were unable to find a replacement. State Farm exited the Florida market in 2008, and several other insurers followed suit. Today, Florida’s market largely consists of small and local insurers.

Insuring disaster risks is inherently challenging due to the correlated nature of disaster losses. Catastrophic events such as floods, earthquakes, wildfires and hurricanes often cause a large number of concurrent claims – if the roof of my house blows off, chances are good my neighbors’ houses will meet the same fate. From an insurer’s perspective, this necessitates substantial capital reserves to honor commitments to policyholders while ensuring fiscal solvency. However, maintaining such liquidity, typically financed by a combination of equity, debt and reinsurance, is costly.

This financial challenge is further compounded by the extreme volatility of disaster losses. Billion-dollar loss events have become more frequent in recent years. The 2017 and 2018 wildfire seasons in California caused $20 billion in underwriting losses, wiping out double the accumulated profits of the industry in the state since 1991. In 2019, the Congressional Budget Office estimated that tropical cyclones, which include hurricanes and tropical storms, cause an average of $54 billion in damage annually in the U.S. For both types of disasters, the escalation in the frequency and severity of losses can be attributed to increases in the hazard – defined as the likelihood and intensity of a peril – as well as the exposure, which refers to the values of assets at risk.

Liao Penny Home Insurance Crisis 3

The increase in wildfire hazard in the Western U.S. is largely due to shifts in fire management practices combined with climate change. The policy of suppressing natural fires throughout the 20th century has culminated in a landscape that is densely vegetated and homogenous, thus more prone to high-intensity fires. Climate change exacerbates this hazard by extending the periods in which vegetation is dry as well as altering snowmelt patterns and modifying regional climates – thereby creating more opportunities for fires to ignite and spread.

Meanwhile, the increase in exposure stems largely from development patterns. In the wildland-urban interface (WUI) – a zone where low-density development intermingles with fire-prone wildlands – the number of new homes grew by 41 percent from 1990 to 2010, making WUI the fastest-growing landuse type in the U.S. and amplifying the potential for losses.

Similar factors are at play with tropical storms. Warming ocean temperatures, a key driver of hurricane formation and intensity, foster more powerful and potentially destructive storms. The sixth assessment report of the Intergovernmental Panel on Climate Change found with high confidence that the proportion of intense tropical cyclones, their rain rates and peak wind speeds all increase with global warming. Concurrently, U.S. coastal counties are among the nation’s most densely populated regions, with some 49 million homes and at least $1.4 trillion in assets within an eighth of a mile of the coast. From 1970 to 2020, coastal population increased by approximately 40 percent, placing more people, homes and infrastructure in the potential path of storms.

Coping With the New Reality

Insurance operates by pooling risk across individual policyholders, geographic areas and time. And the viability of an insurance model hinges on the insurer’s ability to accurately measure and price the risk. Moreover, insurers need a solid financial strategy to handle the potential for massive payouts, especially for disasters with correlated losses. As the probable distribution of disaster damage becomes increasingly “fat-tailed” (i.e., with more occurrences of extreme losses), it complicates the core functions of the insurance industry.

The industry’s ability to adapt to the evolving risks also depends on the regulations they face. Insurance is regulated by the states. The regulations aim to protect consumers by preventing “excessive, inadequate, and unfairly discriminatory rates” – a standard adopted almost everywhere – and requiring insurers to have sufficient capital to remain financially stable. However, the specifics and strictness of these regulations differs widely, giving insurers varying degrees of flexibility in how they operate.

A decade ago, when wildfire damage was relatively modest, most insurers priced wildfire risk at the zip-code level using the retrospective approach. But the jump in losses since 2017 have fostered a broad consensus among insurers about the necessity of incorporating wildfire-related catastrophe exposure.
Measuring and Pricing Risk

Traditional insurance takes a look-back approach to setting prices, drawing on historical claims data to calculate expected losses. Yet, relying solely on two or three decades of past claims can overlook the financial implications of infrequent yet catastrophic events. This results in premiums that are inadequate to cover “tail” events.

That shortcoming became all too evident following Hurricane Andrew in 1992, which caused $30 billion in losses and bankrupted at least 11 insurers, as well as the aforementioned Northridge Earthquake in California in 1994. These events precipitated the widespread recognition of the necessity to integrate catastrophe exposure into property and casualty insurance pricing. Consequently, in 2000 the Actuarial Standards Board formalized this approach.

A core actuarial technique used to estimate catastrophic losses is called catastrophe (“cat”) modeling. Originated in the 1980s, cat models employ a forward-looking and probabilistic approach to account for tail risk. Such models simulate myriad hazard events, subsequently evaluating the potential damage to individual properties based on their distinct locations and attributes, such as roof type and building materials. These models can also integrate projections of future scenarios linked to climate change and other evolving conditions. Compared to the traditional methods, cat models allow for a better characterization of the possible range and distribution of catastrophic losses.

Today, cat models for wind damage are quite sophisticated and widely used by insurers, having been augmented through recalibrations following significant events like Hurricanes Katrina, Ike and Sandy. In 1995, the Florida Commission on Hurricane Loss Projection Methodology was created to evaluate hurricane models. The commission certifies models deemed suitable for rate-setting. Many other states also follow FCHLPM’s certification. And over the years, catastrophe modeling has been largely successful in aiding insurers to make better risk management decisions and avert insolvency in the face of catastrophic loss events.

By contrast, few insurers have adopted cat modeling for wildfires. A decade ago, when wildfire damage was relatively modest, most insurers priced wildfire risk at the zip-code level using the retrospective approach. But the jump in losses since 2017 have fostered a broad consensus among insurers about the necessity of incorporating wildfire-related catastrophe exposure.

Liao Penny Home Insurance Crisis 4

Several obstacles, though, inhibit the widespread adoption of cat models for wildfires. For one, these models are still in an early stage of development. Some variables within the models are imperfectly calibrated because the data are just not available. Moreover, large inconsistencies exist across models, leading to insurers’ reservations about fully relying on any particular model. These complex proprietary models can also be quite costly to adopt.

Rate regulations might have also lowered insurers’ incentives to employ cat modeling and cover high-risk properties. Adopting more refined risk assessments often results in marked premium increases, evident from the high premiums of FAIR Plan policies and wind-damage coverage in hurricane-prone areas. If insurers anticipate that such adjustments will hit regulatory constraints on rate increases, they will be less motivated to invest in risk measurement – or, for that matter, to continue to operate in high-risk markets. For example, until recently, California regulations imposed two major restrictions on rate increases. First, filings for annual increases exceeding 7 percent triggered public hearings, which substantially prolong the approval process. Second, an overall rate increase could not rely solely on cat model outputs for justification.

There’s evidence that these constraints prevented a smoother adjustment to escalating wildfire risk. While the regulations may well have moderated premium growth for homeowners in high-risk territory, it has likely exacerbated the access problem for others.

Capital Management

As disaster losses are often correlated, prudent risk management by insurers requires them to be prepared to pay claims not only for commonplace events but also the more extreme ones. For instance, the California Earthquake Authority ruled that insurers would need $19 billion to pay claims in a 1-in-350-year earthquake event. Using the 1-in-350-year rule, catastrophe modeling by PartnerRe (a big reinsurance company) estimated that $33 billion was needed for wildfire compensation in California. In Florida, a 1-in-250-year hurricane event was estimated to cause $87 billion in losses.

Insurers can finance these expenditures by issuing stock or borrowing or purchasing reinsurance. Reinsurance tends to be the cheaptest alternative and is extensively used to manage insurers’ capital requirements. Essentially, reinsurance is insurance for insurance companies and operates by pooling risk at a global scale. Reinsurance is particularly important for smaller insurers, given their lack of geographic diversification and limited capacity to absorb large losses. In markets dominated by small insurers (such as Florida), reinsurance is key to keeping insurers open for business.

Liao Penny Home Insurance Crisis 5

Yet reinsurers must also price-in underlying changes in loss distributions. And because the reinsurance market is global and unregulated, reinsurance costs can change rapidly. According to the Guy Carpenter Rate-on-Line Index, the cost of U.S. property catastrophe reinsurance has nearly doubled since 2017. In just the first half of 2023 alone, Florida experienced a jump of 30-40 percent in reinsurance costs.

Higher reinsurance costs directly impact insurers, making it costlier for them to provide the same level of aggregate coverage. And this, of course, heightens the tension between affordability and availability for homeowners. In states where insurers can pass on the extra costs to consumers, the rapid increase in premiums can financially strain many households. Conversely, in states with stringent rate regulations, insurers can respond by limiting their exposure or dropping customers. A case in point is California, where, for a long time, regulations barred insurers from justifying rate increases using reinsurance costs. Notably, the three insurers that halted sales of new policies in the state all cited reinsurance costs as a factor.

Construction and Litigation Costs

Other cost components also feed premium inflation. For instance, since 2020 construction costs have surged due to a striking 19.2 percent annual increase in building materials prices and a chronic labor shortage driving up wages. These higher costs directly scale up the replacement value of insured structures and, especially in high-hazard areas, drive significant premium increases.

Then there is litigation. Florida, in particular, is drowning in litigation. While the state accounts for 9 percent of all U.S. homeowners’ insurance claims, it represents a startling 79 percent of insurance-related lawsuits.

Actually, it’s even worse than it looks. Over the past decade, 71 percent of the insurers’ $51 billion payouts in Florida went to attorney fees and public adjusters. This results, in part, from several of the state’s rules governing insurance litigation. Since 2021, there have been multiple legislative changes intended to reform the system, but the jury is out on their success.

toward solutions

A fundamental fix is needed to address the challenges of sustaining a viable insurance market, and it lies in risk mitigation. There are a host of cost-effective interventions at both the household and community level that can significantly reduce risk. Both experimental and field evidence illustrates that using advanced building materials and adopting specific structural and landscaping strategies can bolster buildings’ resilience against wildfire, flood and wind damage. Moreover, local government can assert its authority in curbing development in highrisk areas through zoning, promoting disaster- resistant housing via tougher building codes and investing in protective infrastructure or nature-based solutions like underbrush removal.

A larger residual market collects the highest-risk properties into a single state-wide pool. From a risk diversification perspective, this is less desirable than dispersing these properties among national insurers capable of spreading risk across larger geographic areas.

However, past experience makes it pretty clear that, left to their own devices, developers, households and communities typically underinvest in damage protection. To get the job done, a multifaceted approach is needed – one combining information/communication initiatives such as hazard mapping and public disclosure, regulatory actions like the implementation of statewide building codes and zoning for high-risk areas, and financial incentives ranging from insurance premium discounts for resilient properties to grants for hardening property at heightened risk.

As we learn more about the distribution of losses, it is imperative that rating methods evolve in tandem. Insurers have significant opportunities to improve the measurement and pricing of catastrophic loss exposure, particularly with wildfires. As cat modeling for wildfires advances, state regulators could help by providing more guidance on how various models stack up in quality, emulating the Florida Hurricane Commission’s approach with wind-damage models.

Regulators should also consider relaxing some of the more stringent rating restrictions and administrative procedures to allow insurers to adapt more promptly. Indeed, California has recently put in motion regulatory reforms in this vein. However, a potential repercussion could be faster rate increases in high-risk zones. Targeted, means-tested assistance subsidies could help reduce the pain for low-and middle-income households – not to mention, blunt the political pushback.

Alternative ways of shuffling the burden of risk might help address the consequences of escalating reinsurance costs. One option is to create a public reinsurance program that offers lower and more stable pricing than private reinsurance.

The Florida Hurricane Catastrophe Fund serves as an example. It operates as a mandatory reinsurance scheme pooling risk across insurers in the state. However, it is not without limitations. The insolvencies of several insurers despite the FHCF’s backing during Hurricane Ian underscored that public reinsurance alone is not a panacea. It needs to be complemented by a robust regulatory oversight on insurers’ capital management. Moreover, Hurricane Ian has weakened FHCF’s financial position. Currently, it faces a shortfall of $9.5 billion if a massive event were to trigger the statutory maximum loss for the fund.

Going beyond individual states, a national public reinsurance program could prove more efficient in risk pooling across diverse geographies and disaster types. In addition, the past two decades have witnessed the emergence of innovative risk-transfer mechanisms to the broader capital market – think catastrophe bonds, collateralized reinsurance and reinsurance “sidecars.” While the jury is still out on the efficacy of these instruments, their rapid deployment suggests they could be integral to shaping the future solution.

Liao Penny Home Insurance Crisis 6
Expanding the Residual Market

One way to offset the withdrawal of private insurers is to expand the aforementioned residual market to cover consumers left out in the cold. There are two models for these state insurers of last resort. The first is the FAIR Plan, a state-mandated yet privately run insurance pool backed by all regulated insurers in the state. This model is currently implemented in 34 states and Washington, DC, with Colorado recently passing legislation to establish its own. The other model is the state-managed and underwritten Citizens Insurance Program. Presently, only Florida and Louisiana employ this model.

The residual market does not have inherent advantages over the private market in handling catastrophic losses. Both residual market models described above are intended to be financially self-sufficient. Given that these programs predominantly attract the highest-risk properties, the premiums they set are correspondingly high. Consequently, while they might ensure the availability of coverage, they do not make it more affordable. Moreover, a larger residual market collects the highest-risk properties – whose losses are highly correlated – into a single statewide pool. From a risk diversification perspective, this is less desirable than dispersing these properties among national insurers capable of spreading risk across larger geographic areas.

Perhaps a clear benefit of residual market programs is their reliability. However, for better or worse, a residual market approach may lead to rob-Paul-to-pay-Peter cross-subsidization if the state assesses lower-risk policyholders or taxpayers to sustain these programs following extreme loss events.

the way forward

The bad news is that rapidly rising claims are exacerbating the existing problems in the state-regulated housing insurance markets as they cope with political pressures, interest group demands and limited technical and financial capacities. The good news is that this state of play offers opportunities for a free lunch – well, a somewhat cheaper lunch than might be expected.

Reinforcing private incentives to reduce the risk of damage and rearranging risk to minimize the total societal cost could go a long way toward offsetting the cold fact that enormous amounts of property are already in harm’s way and that climate change is making the places where people want to live ever more vulnerable. Much depends on the good sense, flexibility and innovation of the public and private institutions that set the course of the insurance industry.

main topic: Finance: Insurance