What are the economic effects of extreme weather caused by climate change?

Human-made climate change is increasing the likelihood or intensity of extreme weather events. The costs of these disasters – and their effects on equality – are being underestimated, and greater urgency is required to implement mitigation policies.

Climate change is heating up the world. This year, the cherry trees in the main temple in Kyoto blossomed on 26 March. This is the earliest they have done so since the custodians of the temple started recording that date, more than 1,200 years ago. Sea levels are also rising and will continue to rise in the years to come.

These two phenomena – warming average temperatures and rising sea levels – are widely reported, but they are not the most important for assessing the economic costs of climate change. This is especially true for the costs that, collectively, we are already experiencing.

Figure 1: Full-flowering day of the cherry blossom in Kyoto, Japan

Source: NOAA, based on Aono, Kazui (2008), Aono, Saito (2010) and Aono (2012)

Figure 2: Cumulative ice mass loss of ice sheets

Source: IMBIE/EEA/ESA

Warming and sea-level rise may get a lot of attention, but that is because they are easier to measure. The more important climate change phenomena – those that are wreaking havoc on economies – are associated with the increasing frequency and intensity of extreme weather events. Events like the floods in Germany, India and China in July 2021, or the wildfires in Southern Europe in August are all relevant examples.

In some instances, these disasters would have been virtually impossible without climate change (for example, the Siberian heatwave in 2020 or the heatwave in North America in June 2021), but thankfully these instances are still rare. In most cases, however, anthropogenic (human-made) climate change is increasing the likelihood or intensity of the events that are already happening.

Figure 3: Number of relevant loss events by peril

Source: Met Office, based on Munich RE

Extreme weather attribution

Following a flood in his Oxford neighbourhood in 2003, and a fatal heatwave in France in the same year, Miles Allen (a professor of geosystem science at the University of Oxford) and his colleagues realised that the influence of anthropogenic climate change on individual extreme weather events can be quantified using climate models. This approach has become known as ‘Extreme Event Attribution’ (EEA).

Using EEA, we can now attribute the costs of an extreme weather event to the greenhouse gases emitted into the atmosphere in the past century. We can explain using the example of Hurricane Harvey, a tropical cyclone that hit the mega-city of Houston in Texas in 2017.

The Gulf of Mexico, the body of water through which the storm passed before hitting land in Texas, is becoming warmer because of climate change, and therefore the atmosphere above it contains more moisture. That led directly to Hurricane Harvey dumping a lot more rain on the city of Houston when it passed over it during the last few days of August 2017.

Running their climate models with and without the greenhouse gases that society has added to the atmosphere, climate scientists were able to estimate that the probability of the flooding event (the dramatic increase in rainfall from the storm) was more likely to happen by about two-thirds. Put differently, their best estimate was that about 38% of the rainfall during those few days fell because of climate change. In other words, without climate change, the hurricane would have been less likely or much ‘drier’.

What do we know about the economic costs of Hurricane Harvey?

We typically differentiate between damage (a stock measure – something that is measured at a specific time) and loss (a flow – measured over a period of time).

Damage from extreme weather events can occur during or immediately after a hazard event. The damage is usually measured in physical units (for example, units of housing, kilometres of roads, number of bridges or amount of crops) or the overall cost of these physical units (their market price, or the cost of reconstructing them, multiplied by the number of units that were damaged). Sometimes though, the damage is more difficult to measure (for example, destruction of cultural assets and heritage, and the environment). These all count stocks of assets.

The economic damage, in turn, can cause indirect economic loss – a reduction in the flow of economic activity after the event. These losses can include:

Micro-level reductions in activity (such as declines in firms’ revenue owing to business interruption or individuals’ loss of income).
Meso-economic impacts (for example, interruptions to transport networks, or stoppages in the flow of inputs through supply chains).
Macroeconomic impacts (including price and exchange rate changes, increases in government debt, negative effects on stock markets and declines in GDP)

Overall, the damage associated with Hurricane Harvey was estimated by the Centre for Research on the Epidemiology of Disasters in Belgium. It concluded that the overall value of the damage from Hurricane Harvey was about $95 billion, making it the second costliest hurricane in US history, after Hurricane Katrina, which flooded New Orleans in 2005 and killed around 1,800 people. Fortunately, the death toll associated with Hurricane Harvey was much smaller.

Nevertheless, the hurricane caused a lot of infrastructure damage, as well as destruction to commercial property along the gulf. Therefore, the follow-up losses to businesses were inevitable, and were especially severe for small and medium-sized enterprises (SMEs). But research shows that these indirect losses did not persist for very long, and largely dissipated about two years after the hurricane. This is not unusual in a high-income country, where recoveries are usually well-funded through insurance or assistance from the state. In contrast, the recovery process in lower-income countries can last much longer.

Joining attribution science with economic damage and loss assessments

By combining the EEA calculations that were done for Hurricane Harvey – identifying the impact of anthropogenic climate change on the likelihood and intensity of the storm – with all the accounting that is available of its economic and social impacts, we can reach two very important conclusions.

The first has to do with the overall assessments that economists undertake to evaluate the costs of climate change. Typically, economists use Integrated Assessment Models (IAM) to forecast or predict the cost of climate change. But these models are using damage functions that are based on the average temperature experienced (in a country) and not on the extremes, where most of the costs of climate change typically are.

We can compare the anthropogenic climate change costs associated with Hurricane Harvey, with the climate change as these are calculated in a typical IAM. When we examine the IAM constructed by William Nordhaus (an economics Nobel laureate), we find that the anthropogenic costs associated with this one single hurricane are much larger (by a factor of about three) than what Nordhaus predicts with his Dynamic Integrated Climate Change (DICE) model for the whole of the United States for the whole of that year. Note too that this was a year in which there were two other destructive hurricanes, many devastating wildfires, tornadoes and droughts, and many other more local adverse weather events.

This is not a perfect comparison, but the inevitable conclusion seems to be that the current quantifications of the economic costs of climate change, obtained with IAM techniques, vastly underestimate the cost of climate change as it is experienced right now (and therefore also what it predicts about the future). If the current costs of climate change are much higher than what most economists derive from IAMs, the profession needs to reassess its lukewarm support for more aggressive greenhouse gas emission reduction policies (such as much higher carbon taxes).

The second conclusion we can reach when combining attribution science with the economics of hurricanes has to do with equity and justice. Using the attribution conclusion about the amount of rainfall that was generated by anthropogenic climate change during Hurricane Harvey, and hydrological modelling of the flooding that occurred in Houston, we can analyse exactly where flooding happened directly because of climate change.

This shows that about half of the homes that were flooded during the storm were flooded only because of climate change (about 105,000 homes flooded overall). In other words, without the impact of climate change, about 50,000 homes would not have flooded at all.

Using detailed maps of property locations and census information about the owners of these properties, we can identify four observations. First, Latinx-owned homes were much more likely to flood because of climate change than their relative share of homeowners in Houston. Two, high-income white and non-Latinx homeowners were more likely to get flooded than lower-income households from these groups. Three, this relationship between income and ethnicity was reversed for Latinx homeowners: for them, it was lower-income Latinx homes that were most likely to get flooded. Four, the ‘bias’ that led to the storm hitting more low-income Latinx homes was especially pronounced in flooded homes that were outside the officially designated flood-zone.

This distinction is important, as for the most part it is only the homes that are located within the flood-zone that have flood insurance. Thus, the lower-income Latinx homes that were much more likely to get flooded, were also likely to be uninsured for floods. Without insurance, their recovery trajectory will inevitably be longer and more difficult, thus exacerbating their disadvantage.

Conclusion

Combining what we now know from the scientific discipline of extreme weather attribution with assessment of the economic costs of extreme events shows that economists typically underestimate the cost of climate change and are therefore likely to underestimate the urgency required for cost-effective mitigation policies.

Further, climate change is not an equal opportunity menace, and its impact on minority and low-income uninsured families is potentially much higher. This is certainly the case in the US example, but partial evidence suggests that this is applies elsewhere as well, and is likely to be much more prevalent in low-income countries.

These findings, taken together, suggest that the economics profession can contribute a lot to societies’ attempts to deal with the climate crisis, but that this contribution is still only in its infancy.