Sep 5, 2018
Wildfires in California destroy thousands of structures each year, and in 2017 that number jumped to 10,800. In 2018, wildfires wrought even greater destruction, with more than 22,000 structures destroyed, according to the California Department of Forestry and Fire Protection. Those conflagrations can devastate homeowners and bring heavy costs for the insurance industry. To assess how fire risks will affect the California insurance market, and consequently homeowners, RAND researchers undertook a novel study in two fire-prone areas in Northern and Southern California. They found that, while the market in lower-risk ZIP codes within those two areas was working relatively well as of 2017, higher-risk ZIP codes faced challenges.
Those challenges may intensify as wildfire risks increase, potentially affecting the health of the insurance market, policy affordability, coverage adequacy, and insurer profitability. The RAND team found that wildfire risk is likely to increase markedly in some areas under business-as-usual greenhouse gas (GHG) emissions scenarios.
The study focused on two areas: a 1.9-million-acre region in the Sierra Nevada foothills spanning parts of Placer, Nevada, and El Dorado Counties in Northern California, and an 860,000-acre region in western San Bernardino County, east of Los Angeles in Southern California.
In the Northern California study area, the researchers found that the average number of acres burned each year may double by mid-century and double again by 2100 under a business-as-usual GHG emissions scenario. The researchers project that an aggressive and successful GHG emissions control strategy would stabilize the average number of acres burned in the second half of the century, keeping the average annual risk at mid-century levels. In the Southern California study area, the researchers project that the number of acres burned may not change substantially.
Policymakers, insurers, and homeowners all are exploring how to adjust to changing wildfire risk in California. This research is designed to help them consider policy options for the insurance market.
The RAND researchers focused on four main questions:
To answer these questions, the researchers used a combination of quantitative and qualitative methods: wildfire scenario analysis that accounted for multiple climate projections and population growth patterns, regression analysis to investigate the impact of the change in risk on indicators of insurance market health, and interviews with insurers, regulators, consumer advocates, and county staff on possible market responses to changes in risk.
Overall, the San Bernardino study area faced higher wildfire risk in 2017 than the Sierra Foothills area. However, by mid-century, wildfire risk in the Sierra Foothills study area is projected to roughly double in the riskiest one-third of ZIP codes under a business-as-usual GHG emissions scenario. By the end of the century, structure risk is projected to quadruple in the riskiest one-third of ZIP codes. The analysis showed little future change in this risk in the San Bernardino study area.
The market share of admitted insurers is high overall in the Sierra Foothills and San Bernardino study areas, but there is evidence that admitted insurers are facing challenges in higher-risk wildfire areas. The market share of admitted insurers is 7.8 percentage points lower in the highest-risk ZIP codes than areas with little wildfire risk when other factors are controlled for. There is also evidence that the challenges in the admitted market are accelerating: Market share of the admitted market declined much faster between 2007 and 2015 in the higher-risk ZIP codes.
Evidence suggests that insurer-initiated nonrenewal rates in the admitted market are somewhat higher in higher-risk ZIP codes, but major differences have not been observed so far.
Insurance rates per $1,000 of coverage in the admitted market increase substantially with wildfire risk, and have been growing more rapidly in recent years compared with lower-risk areas. Even so, many insurers believe that the difference between premiums for low- and high-risk properties does not fully reflect the actual difference in risk. At the time this research was published, the California Department of Insurance held that insurers haven't provided sufficient evidence to support a request for higher premium differentials.
Increasing structure risk does not appear to be decreasing homeowners' insurance policy take-up rate. However, homeowners in high-risk areas are purchasing less coverage relative to structure value. They also are selecting higher deductibles than homeowners in lower-risk ZIP codes. Further analysis of the adequacy of coverage is warranted.
For the homeowners multiple-peril line of insurance, admitted insurers broke even in terms of combined underwriting profits between 2001 and 2017. Policies written in the fire insurance line were profitable over the same period. Overall, the industry was profitable when investment returns are included. However, the pattern in annual underwriting profits shows that a large wildfire event can wipe out many years of underwriting profits.
The RAND team projects that by 2055 the insurance market in the Sierra Foothill study area for the highest-risk ZIP codes will experience several changes affecting insurers and homeowners:
Little change in these indicators is projected for the lower-risk areas of the Sierra Foothills study area because climate change is expected to have little impact on wildfire risk in the lower-risk areas. Similarly, the changes projected for the San Bernardino County study area were very small.
The RAND team identified regulatory and other factors that will influence how climate change will affect the residential insurance market. Insurance regulatory issues include the extent to which rates reflect the full difference in fire risk across structures, whether probabilistic models of wildfire risk are allowed in the rate-approval process, and whether the net reinsurance margin is allowed as an expense in rate filings. Although there is no indication that current FAIR Plan that rates are artificially low, the extent to which rates offered by the FAIR Plan keep up with the increase in risk will also be an important factor in how insurance markets respond to climate-induced changes in wildfire risk.