The Federal Insurance Office (FIO) recently proposed to solicit information from insurance companies about how climate risks may impact the affordability and availability of homeowners insurance policies. We submitted a comment supporting the call in which we emphasized the importance of nationwide data collection efforts and encouraged FIO to release its data publically. We also conducted analyses to support some potential modifications to the data solicitation, which we describe below.
The FIO proposal would involve collecting information nationwide from companies writing more than $100 million in policies. Additionally, in 10 states identified as the most “climate vulnerable,” FIO will solicit additional information in order to cover at least 80% of the state insurance market. FIO identified these states using FEMA’s National Risk Index (NRI) product. With enough resources, soliciting information to meet that 80% threshold nation-wide would of course be preferable, but as an initial step, we were pleased to see the targeted requests in regions that have been most heavily impacted by climate disasters to date.
Assessing the climate vulnerability of states, however, depends on the choice of metrics included in the analysis. The NRI captures expected annual losses from agriculture, buildings, and population, as well as the summed total of all three sectors. These losses are determined separately for 18 hazards. In its proposed approach, FIO ranked states according to the summed total expected annual loss for a subset of 15 climate hazards. As Figure 1 demonstrates through an interactive tool, changing the type of loss, or choosing a different subset of hazards, can change the ranking of most vulnerable states and result in a different set of top 10 most vulnerable states than the ones FIO identified. See our GitHub repository for more details on the analysis.
In our comments to FIO, we suggest they focus their initial selection of 10 states to reflect only losses to buildings, and to only include hazards that are likely to worsen with climate change. We make this recommendation because the data collection proposal specifically targets homeowners’ insurance losses from physical climate risks, as opposed to population or agricultural losses or losses from natural disasters more broadly. We hope that the interactive figure above, combined with our recommendations, can help FIO and others interested in FIO’s data request see how historical damages vary across states and use that information to better understand how analytical choices affect the selection of the most climate-vulnerable states.