In this first article of a two-part series, we examine how insurance risk, consumer costs, and resilience incentives respond to the stress of a changing climate in Broward and Miami-Dade counties.
Across the United States and elsewhere, homeowners make long-term commitments to property ownership in order to raise families, and to build wealth and social roots.
The 30-year mortgage is the backbone of the housing market, secured by hazard insurance. Thanks to the U.S. government’s National Flood Insurance Program (NFIP) and a growing private flood insurance market, many homeowners are increasingly aware of flood risk and enjoy an expanding set of options to protect what is typically their largest investment. But homeowners are also increasingly aware of the Earth’s changing climate and its effect on several drivers of insurable risk, such as floods, hurricanes, and wildfires. Climate-related volatility in these perils operates both globally and locally, manifesting in physical changes such as sea level rise and increased average precipitation from tropical storms. Those phenomena, in turn, impact storm surge and inland flooding risk.
The visibility of climate’s impact on property hazard is increasingly leading individuals and their chosen leaders to ask: how might an increase in hazard affect the desirability of living in various communities, and how do we manage the socioeconomic impacts? Recent news stories have highlighted the concerns of “climate gentrification,” or potential migration from low-lying but relatively well-off areas to areas of higher elevation but sometimes higher poverty.1 At Milliman, we view the problem through the lens of managing shifts in insurable property value at risk and economic incentives associated with insurance costs.
For this article, we worked with Jupiter Intelligence, a climate risk analytics provider of forward-looking and probabilistic hazard data for future conditions, to develop a framework for analysis that may spark insights for community leaders in the public and private sectors who are charged with managing change and planning for a resilient future.
Defining the Risk: A Workable Classification of At-Risk Communities
For this pilot study, we chose to investigate a region of Southeast Florida, spanning parts of Miami-Dade and Broward counties. This area represents a “climate crucible,” associated with the following characteristics:
- A mostly urban, highly populated area
- An unusually varied range of income levels
- Experiencing documented sea level rise2
- Subject to a known extreme level of quantified, mapped flood hazards associated with Atlantic hurricanes
Our plan for studying climate gentrification risk starts by defining four quadrants for mapping the region:
- Stable areas of relatively high ground,3 encompassing mostly communities with relatively high-incomes.
- Emigrating areas of relatively low ground, encompassing mostly high-income communities.
- Destination areas of relatively high ground, currently populated mostly by low-income communities.
- Crisis areas that face high flood risk in extreme events, yet are mostly populated by low-income communities.
A visual representation is provided in the chart in Figure 1.
Figure 1: Quadrants For Gentrification Study
How were these regions quantified? First, Milliman used the U.S. Census estimates of poverty status by household at the geographic resolution of Census Block Group (CBG)4 to determine relative income distribution across the region—likely a better measure of the potential for migration than absolute income. Using the statewide CBG data, we defined the bottom quartile of CBGs—the 25% of CBGs having the highest proportion of households living below the poverty line—as “low- income.” Remaining CBGs were defined as “high-income.”
Next, we defined a geographic measure of hazard risk within the region. Jupiter provided a geospatial data set indicating modeled flood inundation extents for areas experiencing greater than one-foot water depth for a flood with a 1% annual chance of occurrence (1-in-100-year flood event) under a “high” sea level rise scenario,5 estimated for the year 2050. Jupiter’s model output incorporated the combined probability of flooding caused by storm surge from tropical cyclones, seasonal high-tide flooding, precipitation, and the resulting overland and riverine flooding. Jupiter’s technology and implementation of current climate science is described further in the sidebar.
The final step in mapping the region into quadrants was to track the intersection of income levels by CBG and future flood hazard. Any given land area in the region thus fell into one of four classifications: high-income/low-hazard, low-income/low-hazard, high-income/high-hazard, or low-income/high-hazard. We use the Stable, Destination, Emigrating, and Crisis keywords, respectively, as descriptors for this article.
Notably, the goal of this study is to define a methodology and framework for extracting insights that are not overly dependent upon the sources for either income or hazard data. Jupiter is among a number of private and public sector organizations that continue to advance our understanding of physical climate risk at a local level. In addition, while the U.S. Census is an excellent source for U.S. income data, other data could be used to define relative socioeconomic status. Our methodology was designed so that such model inputs could “plug and play” with our approach.
Actionable Insights from Climate Science
Jupiter’s role in the study is to bring together the latest climate modeling methodologies and scientific research to provide actionable insights for climate resilience. Beginning with the consensus Intergovernmental Panel on Climate Change (IPCC) representative concentration pathways (RCPs) for future greenhouse gas (GHG) emissions, Jupiter provides probabilistic modeling of an expansive set of planetary responses, such as storm surge, precipitation, and ocean/tidal dynamics as a result of each emissions scenario that captures extreme events in the future. Modeling is based on state-of-the-art dynamics augmented with machine learning, resulting in probabilistic geospatial data that capture the non-stationarity (changing extremes in addition to changing mean states) of our climatic system while fully accounting for uncertainty to better inform risk management. Those data are then combined into peril-based reporting on, for example, flood, heat, wind, etc., to enable analysis of physical risks associated with extreme events in a changing climate.
- 1 Harris, A. (December 28, 2018). Climate gentrification: Is sea rise turning Miami high ground into a hot commodity? Miami Herald. Retrieved September 11, 2019, from https://www.miamiherald.com/news/local/environment/article222547640.html.
- 2 The Unified Sea Level Rise Projection report, published by the Southeast Florida Regional Climate Change Compact Sea Level Rise Work Group, states in part: “In the short term, sea level rise is projected to be 6 to 10 inches by 2030 and 14 to 26 inches b y 2060 (above the 1992 mean sea level). In the long term, sea level rise is projected to be 31 to 61 inches by 2100.” The report is available at http://southeastfloridaclimatecompact.org/wp-content/uploads/2015/10/2015-Compact-Unified-Sea-Level-Rise-Projection.pdf.
- 3 More precisely, low risk of inundation beyond a given threshold depth in extreme event scenarios, as defined by Jupiter and Milliman using Jupiter’s technology.
- 4 U.S. Census Bureau (2018). 2013-2017 American Community Survey 5-Year Estimates, Table B17017: Poverty Status in the Past 12 Months by Household Type by Age of Householder. Accessed via ACS Summary File.
- 5 This regionally specific sea level rise scenario corresponds to a scenario resulting in 2.0 meters of global average sea level rise by the year 2100. It is roughly consistent with the Intergovernmental Panel on Climate Change (IPCC) RCP8.5 emissions scenario, which is expected to result in global temperature increases of 2.6 to 4.8 degrees Celsius relative to current global temperature. It represents persistent global emission growth and no additional efforts to reduce emissions beyond those already in place. This high sea level rise scenario includes physically plausible and higher-end estimates of sea level rise from melting polar ice sheets and warming oceans.