Monitoring hazard events

Amid growing evidence that climate events can impact financial markets, how can participants assess risk?

Over the past decade, communities across the world have experienced disruptions due to extreme weather and natural disasters, including floods, droughts, tornados, wildfires, and hurricanes. These kinds of events have the potential to affect every sector of the economy. Storms often do significant damage to property and infrastructure, which in turn impacts the insurance industry; healthcare systems may experience heavy patient loads under extreme heat conditions; tourism may decrease in the wake of hurricanes; and entire harvests may be lost due to extreme flooding and drought events. In some cases, entire communities have been forced to permanently move.

The impacts of extreme weather and natural disasters extend to financial markets. There is evidence, for example, that municipal bond yields in the United States tend to be higher after major floods in unexpected locations1. In Florida, hurricanes have also been associated with long-lasting and disparate effects on housing markets and patterns of home ownership.2 These effects are also observed globally: major storms and natural disasters can increase sovereign default premiums3 and even the likelihood of sovereign default.4

Under realistic emissions scenarios, the frequency and severity of extreme weather events are projected to worsen, and communities around the world will likely face more weather-related disruptions. These disruptions will in turn likely have significant implications for the interconnected financial markets — residential housing, commercial real estate, corporate, sovereign, and municipal debt, among others — that tie us all together.

To inform market stakeholders about extreme weather and other hazards in real time, ICE Sustainable Finance monitors forecasted, observed, and historical hazard events data from the National Weather Service and other sources.

Case study: Flood and hurricane maps from the National Weather Service over a six-month period

Between July 1, 2023, and December 31, 2023, there were over 600 observed alerts — based on observations, rather than forecasts — for floods and hurricanes across the 48 contiguous states.

New England experienced some of the most frequent and widespread flood alerts during this six-month period. In Vermont during July, rainfall amounts broke local records and resulting floods killed two people and destroyed infrastructure.5 Significant portions of the state capital were underwater; months later in January 2023, local businesses remained closed as recovery efforts continued. In mid-December, another storm swept the Northeast, cutting power for over 800,000 customers along the eastern seaboard.6 In late August, Hurricane Idalia caused significant flooding as it cut across Northern Florida, Georgia, and the Carolinas.

On a state-by-state basis, Georgia had the largest total area under observed hurricane or flood alerts due to Hurricane Idalia.

However, to understand the scale of potential damage that may be caused by any event, it is important to quantify more than just the affected area. While Georgia had the largest total area affected over this period, the total number of people affected was highest in New Jersey, Connecticut, and Pennsylvania. (Figure 4).7

These exposure differences arise from differing population densities of the affected locations. In places like Connecticut, the National Weather Service issued flood alerts for the entire Connecticut River corridor, an area that included Greater Hartford and its mid-sized international airport. Alerts were also issued for the mid-sized city of Danbury in the western half of the state. In Georgia, by contrast, a significant number of alerts with large geographical footprints were issued in less populous areas.

In terms of infrastructure, over this same period, Pennsylvania had the largest number of built structures in areas under flood and hurricane alerts. But in terms of the estimated value of infrastructure in these areas, Connecticut ranked highest.

These state-by-state patterns of exposure (Figure 3-6) are complex: over the same six-month period, Georgia has the largest total area under flood and hurricane alerts, New Jersey the largest affected population, Pennsylvania the largest number of affected buildings, and Connecticut the largest total value of affected buildings. This complexity underscores the importance of integrating information about where these alerts are happening with information about what is exposed in those locations.

Of course, floods and hurricanes are not the only hazards that affect communities and markets around the world. ICE Sustainable Finance also incorporates alerts for other hazard types within the United States from the National Weather Service (like tornados; Figure 7) as well as global disaster alerts form the Global Disaster Alert and Coordination System. At the global scale, active alerts from GDACS include wildfire, flood, volcano, drought, earthquake, and hurricane alerts. For a historical perspective, ICE Sustainable Finance also integrates historical GDACS alerts (2002-present) and the National Oceanographic and Atmospheric Administration HURDAT2 historical hurricane database into the geospatial platform (1851-2002).

The tools that power this analysis

ICE Sustainable Finance links the geographic footprints of obligors — whether they are municipalities, corporates, or sovereigns — to securities. This geospatial framework allows us to identify obligors (and securities) that may be affected by hazard events and link those securities to the demographic, socioeconomic, and physical climate risk profiles of the obligated communities and assets. ICE Sustainable Finance already incorporates extreme weather maps for wildfires and hurricanes into monitoring of municipal bond securities obligated to affected communities. Going forward, global maps will be incorporated into monitoring of other securities, providing market stakeholders with information about the exposure of their entire portfolios to observed and forecasted hazards.