Simply put, uninsurable projects go unbuilt. Banks will neither finance a mortgage nor lend money to an industrial facility if the prospective homeowner or business cannot secure insurance. On the flip side, insurers also encourage risk mitigation, something seen in the industry’s role in the development of early building codes as a means for managing urban fire risks. Sprinkler systems, fire escapes, and inflammable building materials all grew from investments by insurers. These twin logics of loss speculation and loss prevention are integral to the sector’s model of capital accumulation.
Event attribution, a field of science that detects the influence of climate change on extreme weather, gets directly at this problem. Using counterfactual modeling techniques, scientists compare the likelihood that actual events would achieve their recorded intensities in both a climate-altered and non-climate altered world. This allows them to calculate the “fingerprint” of climate change on already transpired events. The American Meteorological Society has been publishing an annual series of studies in this rapidly maturing field since 2011. See “Explaining Extreme Events from a Climate Perspective,” American Meteorological Society, ➝.
On “non-calculability” see, for example, Ulrich Beck, Risk Society (New York: Sage Publications, 1992).
This database, called HURDAT, is maintained by the US National Oceanic and Atmospheric Administration (NOAA). Since 1900, there has been less than two landfalling hurricane per year on average in the continental United States—not really a time series that lends itself to robust statistical analysis. This absence of loss data is what the models solve.
Damian Carrington, “Climate Change Threatens Ability of Insurers to Manage Risk,” The Guardian, December 7, 2016. A growing body of science attributes worsening storms to the warming of the climate (cf. Adam H. Sobel et al., “Human Influence on Tropical Cyclone Intensity,” Science 353, no. 6296 (2016): 242–46), and the trend is expected to become more severe as anthropogenic change accelerates (cf. Hiroyuki Murakami et al., “Dominant Effect of Relative Tropical Atlantic Warming on Major Hurricane Occurrence,” Science 362, no. 6416 (2018): 794–799).
Paige St. John, “Creating an $82 billion threat: the formula—a hotel room, four hours and a dubious hurricane computer model,” Sarasota Herald-Tribune, November 14, 2010.
Cf. Marion Fourcade and Kieran Healy, “Classification Situations: Life-Chances in the Neoliberal Era,” Accounting, Organizations and Society 38, no. 8 (November 2013): 559–72; Eric Siegel, Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die (Hoboken: Wiley, 2016); Zeynep Tufekci, “Algorithmic Harms Beyond Facebook and Google: Emergent Challenges of Computational Agency,” Colorado Technology Law Journal 13 (2015): 203.
Shoshana Zuboff, The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power (New York: Public Affairs, 2019). According to Zubhoff’s critique, surveillance capital, as epitomized by companies such as Google or Facebook, is premised on the exploitation of surplus personal data which the companies collect by monitoring user behavior on their services. This surplus data (frequently called “data exhaust” in the industry), whose existence remains both unknown and completely invisible to the user, can be packaged and sold to other parties, thus creating new value streams for the surveillance capitalist.
There have traditionally been three major cat modeling firms globally: Risk Management Solutions (RMS), Applied Insurance Research (AIR) and EQECat (now called CoreLogic). RMS and AIR currently dominate the field, although a host of new weather and climate risk analytics firms such as Jupiter Intelligence, One Concern, and Four Twenty Seven are increasingly competing in this space. See also: Don G. Friedman, “Natural Hazard Risk Assessment for an Insurance Program,” The Geneva Papers on Risk and Insurance 9, no. 30 (1984): 57–128.
Patricia Grossi and Howard Kunreuther, Catastrophe Modeling: A New Approach to Managing Risk, Springer Science & Business Media, 2005.
Friedman, 1984, 55.
Barry Gilway, “Citizens Property Insurance Corporation,” CEO of Citizens Property presentation to the US Department of Treasury Advisory Committee on Risk-Sharing Mechanisms (ACRSM), July 28, 2017, ➝.
See Karen Clark, “Big Data Analytics: Mining Your Catastrophe Claims Data for Competitive Advantage,” Carrier Management, July 20, 2016. This is not “big data” on the order of daily tweets or Facebook hits, but big in terms of the amount of information encoded with each claim. In 2017, for the entire US, according to the National Bureau for Economic Research, there were over five million individual catastrophe claims with private insurers (catastrophes being events defined as causing over $25 million in losses for the insurance sector). See also: “Facts + Statistics: U.S. Catastrophes | III,” Insurance Information Institute, 2018, ➝.
Not only do rules for building design change, but so do institutional responses to natural disasters. When estimating the post-disaster replacement cost value of a home, modelers now anticipate non-linear social factors such as community investment in disaster preparedness (i.e. pump stations; rapid response teams), demand surge on construction materials, tightening labor markets under times of crisis, and even looting, among other factors. They build these factors into their risk assessments (thereby adding justifications for higher premium rates to meet these modeled contingencies).
The BCEGS sends teams to communities to assess numerous factors of “enforcement,” including how many people a city’s inspection department employs, whether the city planning office has specific rules governing constructing for natural hazards, who licenses building contractors within a community’s jurisdiction, etc.
See: Martha Lampland, “False Numbers as Formalizing Practices,” Social Studies of Science 40, no. 3 (2010): 377–404; Paige St. John, “Hurricane models: garbage in, gospel out: In secret calculations, a stew of flawed data,” Sarasota Herald-Tribune, November 15, 2010.
Cf. Big Data for Insurance Companies, eds. Marine Corlosquet-Habart and Jacques Janssen (Hoboken: Wiley-ISTE, 2018); Tony Boobier, Analytics for Insurance: The Real Business of Big Data (Hoboken: Wiley, 2016).
Cf. Evan Mills, “The Greening of Insurance,” Science 338, no. 6113 (2012): 1424–25; Dimitrios Tselentis, George Yannis, and Eleni Vlahogianni, “Innovative Insurance Schemes: Pay As/How You Drive,” Transportation Research Procedia 14 (2016): 362–71.
Cf. Lacie Glover, “3 Ways Car Insurers Use Technology to Monitor Driving and Offer Discounts,” Chicago Tribune, July 31, 2018; Dimitris Karapiperis et al., Usage-Based Insurance and Vehicle Telematics: Insurance Market and Regulatory Implications (Washington, D.C.: National Association of Insurance Commissioners, 2015), ➝.
See: Suzanne Barlyn, “Strap on the Fitbit: John Hancock to Sell Only Interactive Life...” Reuters, September 19, 2018; Bernard Marr, “How Big Data Is Changing Insurance Forever,” Forbes, December 16, 2015.
“CoreLogic: Taking Big Data Analytics to New Customers and Consumers,” Insurance Journal, December 2, 2014.
Large private property owners will hire building inspectors to gather secondary characteristics in order to reduce their insurance premiums. In a big portfolio of properties, perhaps only a handful of buildings will account for a large percentage the modeled risk estimates, and a company seeking insurance can sometimes drive down the price of their insurance by gathering more bespoke data showing that the one or two large buildings (say warehouses with inventory, or crucial manufacturing structures) that account for most of their loss estimates are less risky than suggested by the models.
Joseph Emison and Holly Tachovsky, “A National Loss Study: The Billion Dollar Impact of Underestimated Roof Age,” Claims Journal, October 3, 2014. This estimate does not account for lack of knowledge about roof structures and building materials.
While the market in the US is further advanced than anywhere else, Verisk recently acquired UK-based aerial imagery company Geoinformation Group in 2016 to build up equivalent property imagery capacity in Europe.
Author interview with senior cat modeling executive, June 29, 2018.
As one modeling executive told me, “Because it's a lot of investment on an insurance company’s part to collect this (secondary) information. Do you just believe the homeowner when he gives you this information? Because he might lie to get his premium reduced, so you might have to send out someone to inspect it right? But this is money.” Interview, December 14, 2018.
“About Geomni,” Geomni, ➝.
“Property Imagery for Insurance—CoreLogic,” CoreLogic, ➝.
Ibid.
Bradley Hope and Nicole Friedman, “A Hotter Planet Reprices Risk Around the World,” Wall Street Journal, October 3, 2018.
A wealth of social science research makes this point. Cf. Kai Erickson, Everything in its Path: Destruction of Community in the Buffalo Creek Flood (New York: Simon and Schuster, 1978); Eric Klinenberg, Heat Wave: A Social Autopsy of Disaster in Chicago (Chicago: University of Chicago Press, 2015).
Denny Jacob, “Managing claims in a changing climate”, Property and Casualty 360, National Underwriters, January 7, 2020.
See interview with Todd Stennett at the 2018 SPAR3D Expo and Conference in Anaheim, CA. Point of Beginning (POB), “Get to Know Geomni,” YouTube, September 7, 2018.
The crucial questions Shoshana Zuboff says should be asked of companies that surreptitiously collect and exploit user data for profit are: “Who decides how the data is used? And who decides who gets to decide how it is used?” Zuboff, The Age of Surveillance Capitalism, see particularly Chapter 6: Hijacked: The Division of Learning in Society, 176–198.
Karthik Ramanathan, “Modeling Fundamentals—Anatomy of a Damage Function,” AIR Currents, 2017, ➝.