Elsevier

Energy Economics

Volume 114, October 2022, 106316
Energy Economics

Deadly tornadoes and racial disparities in energy consumption: Implications for energy poverty

https://doi.org/10.1016/j.eneco.2022.106316Get rights and content

Highlights

  • Tornadoes affect household energy expenditures in the US.

  • Tornadoes induce declines in expenditures for electricity and home heating fuel.

  • Infrastructural damage from tornadoes in the energy sector may play a key role.

  • Tornadoes exacerbate disparities in energy consumption across different groups.

Abstract

Although tornadoes are considered to be one of the deadliest natural disasters in the United States, there exists limited evidence on how such economic shocks may induce disparities in energy-related expenditures across socioeconomic groups and geographical locations. This article shows that a 10 percent increase in tornado-induced fatalities in the same county where a household resides causes a 1.5 percent decrease in annual expenditures allocated to electricity and a 1.1 percent decline in annual expenditures allocated to home heating fuel. Findings indicate that the low income-high cost (LIHC) measure of energy poverty induced by tornadoes is strong, negative and statistically significant among US households. Results further demonstrate that the severity of tornadoes exacerbates disparities in energy consumption between (i) whites and non-whites, and (ii) English speaking and non-English speaking households. Among 22 states where tornadoes caused at least one fatality, white households in Wisconsin, Pennsylvania, North Carolina, Mississippi, Louisiana, Georgia, Arkansas and Iowa experience decreases in home heating fuel expenditures. Declines in electricity expenditures induced by tornadoes are pronounced among non-white households in Pennsylvania, Indiana, Missouri, Louisiana, Kentucky, Oklahoma and New Hampshire. These findings can help policymakers determine which socioeconomic groups to target, and design policies to address energy-related needs in areas affected by natural disasters.

Introduction

In the United States (US), tornadoes are considered to be one of the deadliest natural disasters, resulting in property damages of $50 million and an average federal disaster aid of $13.3 million per fatality between 2004 and 2017 (Deryugina and Marx, 2021). While the direct linkage between climate change and the frequent incidence of tornadoes in the US is unclear, Brooks et al. (2014) report that the variability in the occurrence of tornadoes has increased over the last fifty years. Although the relationship between natural hazards and subsequent demographic changes is well-documented (Raker, 2020), there exists limited empirical evidence on how such disasters may influence energy-related outcomes (Lee et al., 2020, Paudel, 2021a). Understanding household-level behavioral changes in energy consumption in response to severe tornadoes contributes to evaluating the economic impact of natural disasters. From a policy perspective, it is important to delve into micro-level responses because the incidence of deadly tornadoes may further induce disparities in energy poverty across different socioeconomic groups and geographical locations.

This article estimates the economic impact of deadly tornadoes on household expenditures for energy consumption in the US. The focus on severe tornadoes is important for two reasons. First, a plausibly exogenous distribution of tornado-related fatalities varies over both space and time and offers a unique natural experiment setting. Second, while tornadoes are unpredictable, improvements in tornado detection technology have resulted in higher number of tornado reports in recent years (Lee et al., 2013). This study combines nationally-representative samples of households from American Community Surveys (ACS) with geospatial information on tornado-related fatalities available from the National Oceanic and Atmospheric Administration (NOAA). The empirical sample comprises of 8,085,105 households from 22 US states where tornadoes caused at least one fatality in the years 2005–2017. The identification hinges on the assumption that, conditional on county and year fixed effects, tornado fatalities are not correlated with any unobservable determinants of household-level indicators related to energy consumption. These indicators include (i) annual income, (ii) annual electricity expenditure, (iii) annual expenditure for utility, bottled, tank, or liquid petroleum gas and (iv) annual expenditure for home heating fuel, including liquid and solid fuels such as oil, charcoal, kerosene and wood. This article further explores the heterogeneous effect of tornadoes on these four indicators across race and language and offers policy implications on energy poverty in the US.

Results indicate that the impact of severe tornadoes on energy-related indicators of household well-being is mixed. First, regression estimates show that a 10 percent increase in tornado-induced fatalities in the same county where a household resides causes a 1.5 percent decrease in annual expenditures allocated to electricity and a 1.1 percent decline in annual expenditures allocated to home heating fuel. However, annual income and annual cost for utility, bottled tanks, or liquid petroleum gas do not change significantly in response to fatal tornadoes among US households. Second, results illustrate that tornadoes exacerbate racial disparities in energy consumption. For example, a 10 percent increase in tornado-induced fatalities in the same county causes a 1.6 percent decrease in annual expenditures allocated to electricity among US non-white households, with a non-significant effect among US white households. Similarly, a 10 percent increase in tornado-induced fatalities causes a 1.3 percent decline in annual expenditures allocated to home heating fuel among US white households, while it does not have a statistically significant effect among US non-white households. Results further demonstrate that tornado-induced fatalities widen disparities in electricity-related annual expenditures between English speaking and non-English speaking US households. Finally, the empirical analysis suggests that energy poverty in response to tornadoes is more severe among white US households. More specifically, the effect of tornadoes on a (i) binary indicator of energy poverty, and (ii) low income-high cost (LIHC) measure of energy poverty is strong, negative and statistically significant among white households in the US.

To the best of the author’s knowledge, a study by Lee et al. (2020) is the first one to evaluate the causal impact of natural disasters on energy consumption across the globe. Specifically, Lee et al. (2020) focus on the incidence of flood, wildfires and pandemics in 123 countries over a sixteen year-long period to conclude that the severity of natural disasters leads to a significant decrease in energy consumption. Recent empirical studies on the relationship between environmental disasters and energy outcomes have delved into the incidence of the ongoing pandemic and the severity of human-induced forest fires. For instance, Cicala et al. (2021) show that the COVID-19 pandemic has resulted in a 6% decline in hourly electricity consumption across different US electricity regions. In a developing country setting, the severity of fire radiative power from Nepal’s forest fire events reduces household-level energy expenditures and subsequently causes a 0.36% decrease in energy poverty (Paudel, 2021a). Other studies have focused on trade and production disruptions (Evgenidis et al., 2021), natural gas prices (Huang and Etienne, 2021) and oil price risk (Wen et al., 2021).

This article also contributes to the broader literature on how economic agents adapt in response to economic shocks, ranging from earthquakes and wildfires to pandemics (Paudel, 2021b). For example, Fujimi and Chang (2014) show that private business firms decreased their use of air conditioning after the incidence of the 2011 Great East Japan earthquake. In a different study, Herington and Malakar (2016) show that households in Nepal switched from modern fuels to biomass-based traditional cooking practices in response to seismic shocks from the 2015 earthquake, implying that disaster-induced adaptive mechanisms are common during periods of energy insecurity.

More recently, Deryugina and Marx (2021) show that charitable giving in the US increases by about $ 2 million per tornado fatality in areas that lie 20 miles away from a tornado’s path. In addition, Boustan et al. (2020) show that severe natural disasters increase out-migration rates at the county level by 1.5 percentage points and lower housing prices by 2.5–5.0 percent in the US. Majority of quasi-experimental studies in the developing world have explored the repercussions of natural disasters on economic development and labor migration (Ishizawa and Miranda, 2019, Paudel, 2022, Shakya et al., 2022).

This article further contributes to a large literature on mechanisms behind energy poverty across the globe. Some of these mechanisms include demographic characteristics, employment status, living conditions and home features (Legendre and Ricci, 2015). Recent micro-level studies from Australia and Nepal show that ethnic fractionalization induces a significant rise in energy poverty across different dimensions (Awaworyi Churchill and Smyth, 2020, Paudel, 2021c). Research in the developing world indicates that income poverty and social backwardness are positively associated with different measures of energy poverty (Sadath and Acharya, 2017). Although the effect of energy poverty on health status and economic well-being is well-documented in the literature  (Awaworyi Churchill and Smyth, 2021, Kahouli, 2020, Awaworyi Churchill et al., 2020), the extent to which the severity of environmental shocks may influence indicators of household energy poverty remains understudied in the context of the US. This article fills in the gap and provides the first empirical evidence on the impact of severe tornadoes on energy poverty among households in the US. These findings on energy poverty contribute to ongoing debates over decarbonization, which has direct implications on energy transition and socioeconomic inequality (Lowans et al., 2021).

Finally, this article sheds light on state-level effect of severe tornadoes on annual expenditures allocated to electricity and home heating fuel among white and non-white households in the US. The heterogeneity across different states has direct consequences on understanding the distributional impact of policies aimed at mitigating the adverse effect of tornadoes on household welfare. For example, the study shows that non-white households residing in Pennsylvania, Indiana, Missouri, Louisiana, Kentucky, Oklahoma and New Hampshire experience a significant reduction in electricity expenditures, ranging from −21.25% to −4.95%, in response to the occurrence of tornadoes. In addition, the adverse effect on home heating fuel expenditures is prominent among white households in Wisconsin, Pennsylvania, North Carolina, Mississippi, Louisiana, Georgia, Arkansas and Iowa. These point estimates range from 9.00% in Wisconsin to −2.49% in Arkansas. Policymakers can apply these estimates to address energy-related needs among areas with severe tornadoes and determine which socioeconomic groups to target for optimal policy design.

The remainder of the paper is structured as follows. Section 2 presents a brief background on tornadoes in the US and discusses the data. Section 3 presents the empirical strategy. Section 4 describes the empirical findings. Section 5 discusses implications for energy poverty and findings from previous studies. Section 6 concludes.

Section snippets

Background

Tornadoes can be among the most violent phenomena of all atmospheric storms we experience, with winds that can exceed 120 m/s (Elsner et al., 2019). Tornadoes are spatially and temporally unpredictable, allowing researchers for a useful natural experiment setting. Deryugina and Marx (2021) report that a fatal tornado in the US, on average, kills 4 people and results in property damage of US$ 50 million. While tornadoes vary over both space and time, they are short-lived, unpredictable and

Empirical strategy

This section estimates how deadly tornadoes affect energy-related consumption expenditures among US households. The identifying assumption of the estimation approach is that the number of fatalities caused by tornadoes is not correlated with other determinants of energy consumption, conditional on county and year fixed effects. To evaluate the impact of tornadoes on energy-related consumption expenditures, this article estimates: Yict=β1fatalitiesct+ηc+δt+Xict+ϵictwhere Yit is the inverse

Impact of tornadoes on energy expenditures

Regression results in Table 2 indicate that the relationship between the severity of tornadoes and energy consumption expenditures is negative. Moving from left to right in the table, the analysis includes four dependent variables: annual household income and expenditures allocated for electricity, gas and heating fuel. Each column estimates Eq. (1) with county fixed effects and year fixed effects while including demographic controls and state-specific linear time trends.

Table 2 indicates that

Implications for energy poverty

The previous section showed that the impact of severe tornadoes on energy-related indicators of household well-being is mixed. A more policy-relevant question remains: what is the effect of tornadoes on energy poverty? A detailed exploration on energy poverty is important because the energy expenditure–income approach may underestimate or overestimate true energy poverty rates (Awaworyi Churchill and Smyth, 2020). To address this problem, this article makes use of different indicators of energy

Concluding remarks

This article estimates the economic impact of deadly tornadoes on energy-related consumption expenditures among US households. To account for the severity of natural disasters, it exploits a plausibly exogenous distribution of tornado-related fatalities over both space and time. Combining nationally-representative samples of US households with geospatial information on tornado-related fatalities, the study focuses on changes in energy-related expenditures among households from 22 US states

CRediT authorship contribution statement

Jayash Paudel: Conceptualization, Methodology, Software, Validation, Writing – original draft, Writing – reviewing and editing, Data cleaning, Visualization, Investigation.

References (36)

Cited by (0)

1

Jayash Paudel is an Assistant Professor in the Department of Economics at the University of Oklahoma. The author thanks editor Richard S.J. Tol and two anonymous referees for helpful comments in the review process. This study was supported by a summer research grant from the College of Business and Economics at Boise State University, United States of America .

View full text