Climate change and hailstorm damage: Empirical evidence and implications for agriculture and insurance
Introduction
Climate change may cause an increase in the frequency and intensity of weather extremes, which could have adverse consequences for the insurance industry and the economy at large. Global warming has already resulted in an increase in global surface temperatures of about 0.76 °C between 1906 and 2005, while global surface temperatures are projected to increase by between 1.8 °C and 4 °C by the end of the century, relative to the period 1980–1999, depending on greenhouse gas emission scenarios (IPCC, 2007). Emissions are expected to rise considerably in the absence of stringent climate policy agreements, particularly as a result of rapid economic development in emerging countries (Botzen et al., 2008). The Intergovernmental Panel on Climate Change (IPCC) concluded that the historical temperature rise is mainly due to anthropogenic greenhouse gas emissions, and, in the last few decades, has been responsible for the observed increased frequency and severity of certain weather extremes, such as droughts, extreme precipitation, and the increasing intensity of tropical cyclones. The current state of climate research cannot, however, yet make firm conclusions concerning the possible influence of global warming on small-scale severe weather phenomena like hailstorms (Trenberth et al., 2007: p. 300) and hailstorm damage (Vellinga et al., 2001: Table 8-1). The latter is the topic of this paper, in which the effect of global warming on hailstorm damage in the Netherlands is examined by extrapolating historical relations between insured hailstorm damage to the agricultural sector and indicators of temperature and precipitation.
Insight into the potential effects of climate change on damage caused by natural disasters is relevant for the effective design of climate policy. There is considerable uncertainty about the effects of climate change on extreme weather, which complicates the adequate inclusion of their costs in cost–benefit analyses (CBA) of climate policy (Nordhaus and Yang, 1996, Tol, 2002a, Tol, 2002b, Stern, 2007). Nevertheless, extreme weather events, such as storms, can have a major economic impact on modern societies. This highlights the importance of considering cost estimates of potential changes in the frequency and severity of extreme weather events due to climate change in the assessments of strategies for the reduction of greenhouse gas emissions. Stringent reductions in greenhouse gases may be warranted on the basis of the precautionary principle if research shows that large increases in damage from extreme weather events can be expected as a result of a changing climate (van den Bergh, 2004). This study contributes to the knowledge about the potential effects of climate change on extreme weather, notably hailstorms.
Research on the effects of climate conditions and change on insured natural disaster losses is especially relevant for the insurance sector. The latter is particularly vulnerable to weather extremes, such as storms (Vellinga et al., 2001, Mills, 2005). Economic losses and insured losses of natural catastrophes have increased rapidly over recent decades due to socio-economic developments and possibly climate change, as data collected by reinsurer Munich Re (2009) indicate (Kunreuther and Michel-Kerjan, 2007). Insurers, and especially reinsurers, may suffer from increased exposure to climate extremes if these risks are not adequately incorporated into premiums and risk management practices (Schiermeier, 2006, Botzen et al., forthcoming). The economic sectors that demand natural catastrophe insurance may also be affected by more weather related impacts. For example, hailstorms inflict considerable damage to agriculture, but not all of these losses are insured. Moreover, if climate change increases hailstorms, then it is likely that insurers will transfer part of the higher risk back to the agricultural sector by providing less coverage or increasing premiums. This paper outlines several responses of insurers and farmers to adapt to a potential rise in hailstorm risk.
Data on insured losses are especially useful in assessing possible climate change impacts (Changnon, 2003), as the quality of this data is relatively good. Insurance loss data contain information about both the frequency and the severity of hailstorm damage (Vinet, 2001, McMaster, 2001), and climate change may affect both. However, measuring the contribution of climate change in historical loss data is complicated due to the low probability nature of extreme events and problems with recording losses (Brooks, 2006). Very few long loss records exist that allow the determination of the impact of climate change on hailstorm damage. Partly due to these problems, great uncertainty remains about the role of climate change in the upward trend of natural disaster losses in the last decades. It is likely that socio-economic change is the major driver for the observed increase in these losses (Pielke et al., 2005). However, studies that account for socio-economic change (for an overview, see, e.g., Vellinga et al., 2001) still cannot attribute specific variations in losses to climate variations or climate change (Bouwer et al., 2007). Statistical analyses continue to be necessary to assess the dependence of natural disaster losses on variations in climate variables.
The main objective of this study is to examine the potential effect of climate change on future hailstorm damage. Few studies have examined the relations between climate indicators and insured hailstorm damage. Dessens (1995) and Willemse (1995) estimate simple correlations between temperature and hailstorm damage for France and Switzerland, respectively. Both studies find a positive and significant relation between temperature and hailstorm damage, which suggests that global warming may increase hailstorm losses in the future. The current study estimates the potential change in hailstorm damage for different climate change scenarios based on historical relations between damage and a range of climate indicators. The analysis covers minimum, average, and maximum temperatures, as well as precipitation. It uses data on insured hailstorm damage for the agricultural sector of the Netherlands. These data consist of aggregated monthly payouts to policyholders between January 1990 and June 2006. What are called ‘out of sample’ forecast tests are conducted in order to examine which indicators serve best as predictors of hailstorm damage.
The analysis performed here is richer than previous studies due to the explicit modelling of time dynamics and seasonal effects. The specifics of loss data are accounted for in this study by applying Tobit estimation methods, which result in more accurate estimations than OLS or simple correlation methods. Furthermore, separate models are estimated using observations for the whole year and only for months with high damage. This makes it possible to assess whether variation in hailstorm damage both throughout the year and in high damage periods can be largely explained by temperature and precipitation. Moreover, a comparison of results indicates whether relations are similar in these distinct sample periods. We estimate separate models for hailstorm damage to outdoor farming and to greenhouse horticulture, as changes in temperature and precipitation are expected to have different impacts on both these sectors. Finally, we present some possibilities for adaptation strategies to deal with the impacts of climate change on hailstorm damage.
The remainder of this paper is structured as follows. Section 2 discusses hailstorm indicators used in other studies, and presents the data on temperature and precipitation analysed in this study. Section 3 provides information about insurance against hailstorm damage in the Netherlands and explains the data on insured hailstorm damage used in our analysis. Section 4 analyses the estimation results for a range of models that relate hailstorm damage to climate indicators. The forecast performance of various indicators of hailstorms is tested, and the results are compared with other studies. Section 5 extrapolates hailstorm damage to the future, in accordance with a range of climate change scenarios. Section 6 concludes and discusses the economic implications of our results, as well as strategies for insurance and agricultural sectors to adapt to higher hailstorm risk.
Section snippets
Indicators of hailstorm damage used in other studies
It has been shown that temporal fluctuations in hailstorms coincide with fluctuations in weather, such as in temperature and precipitation, although this has been examined by only a few studies. In particular, high minimum temperatures are related to increased hailstorm damage (Dessens, 1995). The daily minimum temperature is usually recorded in the early morning and can be regarded as a rough estimate for the “wet bulb potential temperature” during the following day, which reflects the
Hailstorm damage insurance for the agricultural sector in the Netherlands
Insurance coverage for losses caused by hailstorms is generally available for the agricultural sector in the Netherlands. Hailstorm damage insurance faces fewer supply-side failures than other types of catastrophe insurance and, as a consequence, hailstorm insurance schemes have operated successfully in a number of countries (van Asseldonk et al., 2001, van der Meulen et al., 2006). This is because correlated risks and consequent high loss accumulation are smaller for hailstorm damage than for
Statistical model and estimation method
The simple analysis in the previous section suggests that a positive relation exists between minimum temperature and hailstorm damage. This is further examined by estimating alternative models for the different damage and temperature variables that also include precipitation. Time dynamics in hailstorm damage are explicitly modelled in order to account for autocorrelations that exist in the data and for seasonal dependence of damage. Autocorrelations may be present in hailstorm damage if damage
Extrapolations of hailstorm damage using climate change scenarios
The results in the previous section indicated that historical damage of hailstorms is positively related to temperature and precipitation. Over the past 50 years, average monthly mean, minimum and maximum temperatures increased by, respectively, 1.2 °C, 1.0 °C and 1.3 °C in the Netherlands.11 Climate change projections indicate that
Conclusions
This study has examined relations between normalized agricultural hailstorm damage and a range of indicators of temperature and precipitation for the Netherlands. Temporal dynamics are explicitly modelled and Tobit models account for the censoring of the dependent variable. A distinction is made between damage to outdoor farming and to greenhouse horticulture. The results of models that use observations for the entire year indicate that strong positive relations are observed between historical
Acknowledgements
The insurance companies Hagelunie and Interpolis provided data on hailstorm damage. Gijs Kloek and Roy Kluitman (Eureko Re) assisted us in interpreting and preparing the database. Frédéric Reynes made useful suggestions for the statistical analyses. Geert Groen from KNMI, our colleague Richard Tol, two anonymous referees, and Sjak Smulders gave helpful comments on earlier drafts. The project was carried out in the framework of the Dutch National Research Programme ‘Climate Changes Spatial
References (44)
Optimal climate policy is a utopia: from quantitative to qualitative cost–benefit analysis
Ecological Economics
(2004)Climatology of hail in France
Atmospheric Research
(2001)- et al.
Insurance against climate change and flooding in the Netherlands: present, future, and comparison with other countries
Risk Analysis
(2008) - et al.
Cumulative CO2 emissions: shifting international responsibilities for climate debt
Climate Policy
(2008) - Botzen, W.J.W., van den Bergh, J.C.J.M., Bouwer, L.M., forthcoming. Climate change and increased risk for the insurance...
- et al.
Confronting disaster losses
Science
(2007) Tornado and severe thunderstorm damage
- et al.
Regression model for generating time series of daily precipitation amounts for climate change impacts
Stochastic Hydrology and Hydraulics
(1996) - et al.
Temporal fluctuations in weather disasters: 1950–1989
Climatic Change
(1992) Measures of economic impacts of weather extremes: getting better but far from what is needed—a call for action
Bulletin of the American Meteorological Society
(2003)
Catastrophic winter storms: an escalating problem
Climatic Change
Severe convective weather in the context of a nighttime global warming
Geophysical Research Letters
Econometric Methods with Applications in Business and Economics
The behavior of maximum likelihood estimates under non-standard conditions
The dependence of precipitation on temperature
KNMI Climate Change Scenarios 2006 for the Netherlands
KNMI Scientific Report WR 2006-01
Climate change, insurability of large-scale disasters and the emerging liability challenge
University of Pennsylvania Law Review
The uses of Tobit analysis
The Review of Economics and Statistics
The potential impact of global warming on hail losses to winter cereal crops in New South Wales
Climatic Change
Hailstorm risk assessment in rural New South Wales
Natural Hazards
Insurance in a climate of change
Science
Cited by (75)
Advanced characterisation of photovoltaics for hail resistance
2024, Materials LettersHail detection from Meteosat satellite imagery using a deep learning neural network and a new remote sensing index
2023, Advances in Space ResearchEvolution of research on climate risk insurance: A bibliometric analysis from 1975 to 2022
2023, Advances in Climate Change ResearchQuantitative evaluation of drought risk related to vegetation productivity in China
2023, Journal of HydrologyClimate change adaptation for managing non-timber forest products in the Nepalese Himalaya
2021, Science of the Total EnvironmentCitation Excerpt :Hailstorms have impacted the provision of almost all NTFPs in the study area. Botzen et al. (2010) reported that hailstorm damage may increase in the future with climate change. An increase in pests and insect attack has impacted most of NTFPs.
Revisiting the diversification and insurance relationship: Differences between on– and off-farm strategies
2021, Climate Risk Management