In July 2021, record-breaking rainfall brought severe floods to Europe, where 200,000 properties lost electrical power. In the same month, torrential rain with a maximum intensity of 201.9 millimetres in a single hour led to devastating floods (Fig. 1) in Henan province, China, forcing more than one million people to relocate. These flooding events each caused roughly US$12 billion in property damage. Such losses, incurred during or shortly after extreme events, represent a direct negative impact on the economy. But how does excessive precipitation affect macroeconomics indirectly in the longer term? Writing in Nature, Kotz et al.1 report a comprehensive assessment of changes in gross regional product (GRP) relating to excessive precipitation, and conclude that increases in the numbers of wet days and in extreme daily rainfall dramatically reduce worldwide macroeconomic growth rates.
The authors analysed global annual GRP estimates in the agricultural, industrial and services sectors from 1,554 subnational regions across 77 countries, combined with high-resolution records of global daily precipitation over the past 40 years. Most previous assessments of the macroeconomic impacts of excessive precipitation considered seasonal or annual average rainfall, calculated at the national level. Such coarse resolutions in time and space cannot capture the complexities of how local precipitation events affect regional economic activities. By using detailed statistics beyond simple averages, and by correlating these statistics with subnational economic output, Kotz and colleagues showed that incorporating the details of an extreme event — where and when it hits — can have a profound effect on the assessment of its macroeconomic impact.
Natural disasters often have direct negative economic effects when they occur, but they can also affect economic productivity or GRP indirectly over a longer term. Kotz et al. demonstrated that increases in the number of wet days and in rainfall extremes generally reduce economic growth rates. They found that high-income nations were harder hit by these increases than were low-income countries — a conclusion that overturns ideas held previously. The prevalent hypothesis is that, in large developed economies with well-funded recovery resources, the impact of natural disasters (including increases in wet days and rainfall extremes) should be small, and sometimes even positive, but that it is generally negative for low-income countries, because they are ill-equipped to respond to catastrophe2.
The study also suggests that the services and manufacturing sectors are worse off than the agricultural sector when subjected to increases in wet days and excessive precipitation. In fact, Kotz et al. found evidence that agricultural productivity is relatively insensitive to climate anomalies — a finding that seems at odds with the conventional wisdom that agriculture is affected by anomalous precipitation. In the United States, excessive rainfall can reduce maize (corn) yields by up to 34% , which is comparable to the loss incurred by extreme drought3. Worldwide, climate extremes during the growing season can explain 18–43% of the variance in yield anomalies for maize, soya beans, rice and wheat4.
The effect of climate on the US agricultural economy was previously examined using estimates of this sector’s total factor productivity, which is an economic measure describing the ratio of aggregated outputs to aggregated inputs5. The analysis showed that the climate dependence of the sector’s economic productivity increased considerably after 1980, and that regional climate anomalies can now explain around 70% of the variance of growth in total factor productivity. Should this trend continue, the productivity of the US agricultural economy could fall to pre-1980 levels by 2050. Another study6 found that ongoing climate change has slowed the worldwide total factor productivity for agriculture by around 21% since 1961.
Although crop yields are determined by the cumulative result of daily weather conditions over the course of a growing season, a few shocks such as floods or droughts at crucial growth stages can cause severe damage or total loss, affecting gross agricultural productivity. All of these factors suggest that the agricultural sector is vulnerable to extreme rainfall. The fact that Kotz and colleagues estimate the climate dependence of the agricultural sector to be lower than expected might indicate the need for new measures of climate. Improved metrics might capture, for example, an awareness of the importance of growth stages, or consideration of irrigation and other factors that partly mitigate the negative effects of deficient rainfall and high temperature extremes. Land drainage could be another factor that complicates the impact of excessive rainfall.
The causal mechanisms behind these statistical relationships are yet to be determined. Kotz et al. defined extreme daily rainfall as the annual sum of rainfall on days exceeding the 99.9th percentile of the distribution spanning 1979 to 2019. This amounts to counting only the rain that fell on the rainiest of 1,000 days. How do these rare and local events cause substantial economic shocks and cascade into long-term effects across all sectors at regional, national and global scales? Is it because they are linked in time and space and, together with other climatic anomalies, produce persistent and widespread impacts on economic activities?
The 2021 floods in Europe and China occurred at around the same time, and it is tempting to assume they were connected. But what about the snow, sleet and freezing rain that accompanied low-temperature extremes five months earlier in Texas? These compound effects resulted in a massive power failure, leaving more than 4.5 million homes and businesses without electricity. In fact, extreme rainfall events are linked through complex networks of global climatic patterns7, and might also be related to heatwaves, cold surges, droughts, storms and other weather extremes. The integrated result of these compounded factors could lead to substantial economic impacts worldwide. Rainfall extremes might also be related to worldwide economic changes through the globalization of trade — a natural disaster in one location can affect the economy of another if their economies are interdependent. These mechanisms could change local impacts of climate extremes into indirect positive or negative economic effects elsewhere.
The frequency and intensity of precipitation extremes have been increasing in recent decades, and this trend is projected to continue with global warming8. In attempting to forecast the impact of these increases, a major problem is that most climate models underestimate extreme precipitation, and so projections are associated with large uncertainties. Such models therefore provide unreliable estimates of the economic effect of events resulting from excessive rainfall. With the help of numerical modelling and machine-learning techniques, further research might uncover the physical mechanisms behind the increase in precipitation extremes, and offer ways to mitigate them9. Until then, improved understanding of the uncertainties associated with climate models will enable policymakers to estimate rainfall extremes more accurately and to manage the related economic risks more effectively.