Navigating the Storm: Natural Disasters and the Policy Process

FOCUSING EVENTS AND THE POLICY PROCESS

Thomas A. Birkland’s (1997, 1998) work on focusing events provides an important theoretical frame to understand how natural disasters impact the policy process. Focusing events are “sudden, relatively rare,” inflicting harms “… concentrated on a definable geographical area or community of interest, and that is known to policymakers and the public virtually simultaneously” (Birkland, 1997, 22). Focusing events “capture in a nutshell some sort of reality that people already sense in a vaguer, more diffuse way,” (Kingdon 1995, 103) bringing attention to a policy problem. Existing research defines natural disasters as focusing events to understand their impact on policy (Arceneaux and Stein 2006; Birkland 1997, 1998; Birkland and Warnement 2013; DeLeo et al. 2021; O’Donovan 2017; Zelin and Smith 2023). While scientists have noted that natural disasters are increasing as a result of climate change, natural disasters remain relatively rare. Harms from natural disasters generally accrue quickly to a specific geographic area or community, focusing public attention “suddenly and rapidly” (Birkland 1998, 54). Natural disasters are often described as “acts of God,” creating a “causal story” that makes it difficult to blame a single political entity (Stone 2007).

The term “focusing event” first appears in John W. Kingdon’s Agendas, Alternatives, and Public Policy (1984; 1995) introducing the multiple streams approach. The three streams that lead to policy change include problems, policy, and politics. First, policy actors must recognize that there is an existing problem. Next, policy alternatives must be generated to address the problem, including input from a wide range of actors such as scientists, bureaucratic agencies, and elected representatives (Kingdon 1984, 200). The politics stream includes partisan composition of government, preferences of interest groups, and public opinion. Problems, policy and politics streams converge to form a “policy window,” or opportunity for policy change. Focusing events may bring the problems, policy, and politics streams together through “a crisis or disaster that comes along to call attention to the problem” (Kingdon 1984, 100). Policymakers have the opportunity to seize the policy window, leveraging attention to the problem and offering a policy solution. However, policymakers must remain attentive to the parties’ positions and the “mood” of their constituents (Cairney and Zahariadis 2016).¹

Climate hazards may create the conditions for political change (Pelling and Dill 2006). Damages in the wake of natural disasters provide a powerful symbol of the impact of climate hazards for members of Congress and their constituents (Weber 2016). Alternatively, disasters can mobilize the public by pushing “event-relevant issues to the top of the public agenda and provok[ing] shifts in public opinion” (Reny and Newman 2021, 1499). Constituents may pressure legislators to reflect their new positions (Barberá et al. 2019; Unsworth and Fielding 2014). Voters may hold elected leaders accountable for the aftermath of natural disasters (Albrecht 2017), especially if they believe that government is responsible for disaster preparation (Arceneaux and Stein 2006). Finally, disasters may shift the policy agenda (Sabatier and Weible 2014), pushing representatives to join pro-environmental advocacy coalitions.

Focusing events act as triggers for agenda setting, policy change, and issue entrepreneurship (Birkland 1998; Jones et al. 1993; Reny and Newman 2021). They also widen the scope of conflict (Schattschneider 1975), bringing new perspectives to the debate through agenda setting. This may occur through institutional avenues, such as congressional testimony (Birkland 1998) or media coverage (Best 2010). Such events can mobilize policy change through executive or legislative action (Godwin and Schroedel 2000; O’Donovan 2017; Pralle 2009). They may also uncover problems with existing policies, providing an opportunity for policymakers to mitigate risk or reduce inequities (Albright and Crow 2021; Birkland 2006; Heikkila and Gerlak 2013; O’Donovan 2017). Last, focusing events may bring attention to an issue but fail to make substantive policy change. The size and scope of a focusing event shape the policy window and the opportunity for political change.

This article is novel in its approach for several reasons. The dataset measures the impact of natural disasters on legislative voting records over a twenty-two year period. Studying long time periods ensures that broad trends in policymaking activity are considered (Sabatier and Jenkins-Smith 1999) and that sufficient numbers of events are analyzed to perform meaningful statistical analyses.² Second, I argue that local focusing events—natural disasters in a congressional district—impact legislative voting records on national environmental policy issues. Existing research focuses on the effects of natural disasters on local policy outcomes (Arceneaux and Stein 2006; Zelin and Smith 2023), or the effects of natural disasters on the volume of green legislation and partisan differences in legislative response (Jud and Nguyen 2024). This article focuses on shifts in behavior among legislators. In an increasingly polarized Congress, small changes in individual legislators’ policy positions may have reverberating effects for national policy. Focusing events can have deep impacts outside of their geographic area: local disasters shape legislative behavior, which in turn shapes federal policy. Environmental policy can mitigate or magnify structural inequalities in the United States.

Finally, qualitative data illustrates how legislators talk about natural disasters on the floor of the House. Natural disasters may spur legislators to engage in issue entrepreneurship and pursue policy change through congressional testimony (Birkland 1998). Alternatively, natural disasters may illustrate problems with existing policies (Albright and Crow 2021; Birkland 2006; Heikkila and Gerlak 2013; O’Donovan 2017): climate hazards may offer an opportunity for members of Congress to link environmental policy to other policy problems, such as segregation and income inequality.

DATA

The article uses a novel dataset that merges the FEMA declared disaster database with legislative voting records from the League of Conservation Voters National Environmental Scorecard from 2000–2022 (Federal Emergency Management Agency 2023; League of Conservation Voters Scorecard 2023). Table 1 presents the data sources for all variables included in the analysis. The outcome of interest is the change in a member of the House of Representatives’ legislative scorecard on environmental policy issues.⁵ The League of Conservation Voters and environmental policy experts from twenty organizations score roll-call votes on environmental issues in twelve categories, including air pollution, renewable energy, climate change, fossil fuels, drilling, private and public land conservation, ocean conservation, exposure to toxic chemicals, transportation, clean water and wildlife, and other. Members of Congress are awarded an annual legislative score calculated by dividing the number of “yay” votes for pro-environmental legislation by the total number of possible pro-environmental votes (excluding excused absences). Figure 1 shows the distribution of scores by party on a scale of 0 to 100, depicting wide variation in average legislative scores annually among Democratic and Republican members of the House of Representatives. The graph suggests a time trend across the sample: Democratic legislator scores increased slightly, while Republican legislator scores decreased. To address this time trend, the model uses year fixed effects.

• Download figure
• Open in new tab
• Download PowerPoint

Figure 1.

Distribution of Environmental Scorecard by Party, 2000–2022

Source: Data compiled from the Federal Emergency Management Agency Disaster Declarations, the League of Conservation Voters Congressional Scorecard, and The Cook Partisan Voting Index (Federal Emergency Management Agency 2023; The League of Conservation Voters 2023; The Cook Political Report 2023).

The independent variable measures whether a congressional district experienced a FEMA-declared disaster by year.⁶ The most common FEMA-declared disasters include fires, floods, hurricanes, severe storms, and ice and snowstorms.⁷ From 2000 to 2022, there were 35,894 county-declared disasters. Large natural disasters can yield numerous declarations: for example, Hurricane Sandy resulted in 559 county disaster declarations. The independent variable is a count of the total disaster declarations per congressional district year (ranging from 0–116, median = 0, μ = 1.69). Most congressional districts experienced at least one disaster declaration during the twenty-two-year time period, and the median number of disaster declarations is 12. There is wide variation in districts’ exposure to climate hazards: several districts in California, Illinois, Texas, and Washington had no disaster declarations, while 48 districts experienced over 100 disaster declarations from 2000 to 2022. There are limitations to using FEMA data: defining and measuring disasters are bureaucratic and political decisions.⁸ The executive branch is largely responsible for defining disasters to provide funding for protection, rescue, and recovery (Kousky et al. 2023). Existing research suggests that presidents use this discretionary power: a president facing reelection is more likely to grant a disaster declaration, and Democratic presidents award more disaster aid than Republican presidents (Husted and Nickerson 2014). To address this concern, the model includes year fixed-effects and dummy variables for presidential administrations during the time period.

Shifts in voter partisanship may impact legislative behavior: members of the House of Representatives are responsive to the policy priorities of their constituents (Barberá et al. 2019). I use the Cook Partisan Voting Index to measure congressional district partisanship from 2000 to 2022 (Cook Political Report 2023). This comparative measure examines how strongly a congressional district leans toward the Democratic or Republican Party. One challenge in linking natural disasters to legislative behavior is a mismatch in the unit of analysis: FEMA declares natural disasters by county-year, and counties often overlap congressional district boundaries. To address this issue, I employ Andreas Ferrara and colleagues’ (2021) population-based geographic crosswalks for counties and congressional districts from 1990 to 2020.⁹

EMPIRICAL STRATEGY

This article utilizes a mixed-methods empirical strategy to understand legislative behavior after a natural disaster. The quantitative analysis illustrates the effect of a natural disaster on legislative behavior, or policy outcomes. Parties exert control over the legislative agenda, shaping whether or not members of the House have an opportunity to vote on environmental legislation (Cox et al. 2010; Aldrich and Rohde 2000); and the qualitative analysis examines whether legislators couple policy problems (natural disasters) with policy solutions.

Quantitative Analysis

I estimate the causal effect (β) of FEMA-declared disasters on members of the House of Representatives’ environmental policy scorecard during each legislative year y_it. I use linear panel data, where i indexes congressional districts and t indexes years. The outcome y_it is the annual environmental score assigned to each member of the House of Representatives by the League for Conservation Voters. The independent variable is either a count variable for the number of FEMA-declared disasters in a congressional district, or a 0/1 binary variable noting a major FEMA-declared disaster in a congressional district.¹⁰ The unit of analysis is within-legislator, or legislator-years: the analysis is restricted to members of the House of Representatives who were in office before, during, and after the natural disaster. The panel method helps to isolate the exogenous impact of FEMA disaster declaration, accounting for time-varying unobservable variables that may be correlated with both the independent and dependent variables.

In implementing an empirical strategy that examines differences before and after an event, the first assumption is that there are two discrete time periods in the sample (Angrist and Pischke 2009; Roth et al. 2023). In this case, the sample is divided into congressional districts before and after FEMA disaster declarations. However, there is heterogeneity in the occurrence of natural disasters—congressional districts experience natural disasters at different points in time. Many congressional districts do not experience any natural disasters during this time period. This article uses a heterogeneity-robust estimator for staggered event time to estimate causal impacts when the timing of natural disasters varies (Freyaldenhoven et al. 2019; Roth et al. 2023).

The second assumption is that the average outcome among the treatment and comparison groups follows a parallel trend in the absence of the event (Angrist and Pischke 2009; Roth et al. 2023). This assumption requires that members of the House of Representatives express ideological consistency in environmental policy positions, absent experiencing a natural disaster in their district. Empirical research in political science reports ideological consistency among members of Congress (Poole 2007): voting behavior of members of Congress is remarkably stable over time (Lott Jr. and Bronars 1993). Parties exert control over House of Representatives voting records, enforcing ideological consistency (Hager and Talbert 2000). Primary challengers present an exception: incumbents who face extreme ideological primary challengers may adopt more conservative or progressive policy positions (Jewitt and Treul 2019). Absent a natural disaster, existing research suggests that members of the House follow parallel trends in environmental legislative activity.

A third key assumption is that the event has no causal effect before it occurs. Pre-event trends provide evidence against the strict exogeneity of the event in a panel event study (Freyaldenhoven et al. 2019). Natural disasters are often described as “acts of God,” making it unlikely that members of the House of Representatives are anticipating a disaster (Stone 1989; Birkland 1998). The analysis will test for pre-trends to understand whether members of the House of Representatives are changing their voting behavior in anticipation of natural disasters in their congressional district.

Using a panel event study, I compare changes in a member of the House of Representatives voting behavior after experiencing a natural disaster. I estimate the effect of natural disasters on members’ of the House of Representatives voting behavior during each legislative session, using the xtevent Stata package (Freyaldenhoven et al. 2022), where:

y_it represents legislators’ scorecard on environmental policy, at time t in congressional district i. Coefficients α_i and λ_t represent legislator fixed effects and time fixed effects, respectively. The covariates for k = −3, …, 3 represent event time dummies that are k years away from a disaster declaration in member’s congressional district i. and represent that the disaster declaration is at least three years in the future and the past, respectively. The coefficients of interest are β_k, which measures the deviation in the outcome variable k years before and after the natural disaster impacted their congressional district. The k refers to the window around the event to estimate a dynamic effect on the outcome. I use a window of two years, as recovery from natural disasters is often a long process, and congressional districts are likely still implementing recovery plans for years after a natural disaster (Cook et al. 2008).¹¹ Members of the House of Representatives also serve two-year terms. The visualization depicts the six coefficients in the window (−2, 0, 1, 2) and two endpoints (−3+, 3+). The year a disaster strikes is referred to as k = 0, and k = −1 is the reference period prior to the disaster declaration. The model controls for variation in constituent partisanship by congressional district-year (PVI), and includes dummy variables for presidential administration.

RESULTS

The results for the quantitative analysis are presented graphically. Each event study plots the legislative scorecard on environmental policy on the y-axis, and the time before and after a disaster declaration on the x-axis. The plotted coefficients depict the estimated effect of a single disaster declaration on environmental scorecards, relative to the year before the natural disaster occurs. The bands around each coefficient depict a 95 percent confidence interval. The figures report pre-trend and leveling-off p-values. The pre-trend p-value assesses parallel trends in legislative behavior before a disaster declaration, and the leveling-off p-value assesses whether the estimation window is sufficiently long to capture the dynamic effect of disasters on legislative behavior.

The analysis demonstrates wide variation in Representatives’ legislative behavior on environmental policy issues by party. Democrats exhibit legislative behavior that is more “pro”-environment after a natural disaster. In figure 2, panel A, the plotted coefficients can be interpreted as estimated effects for a one-unit change in the disaster declaration variable. For Democrats, time “zero”—the year the disaster strikes—represents a statistically significant increase in pro-environmental legislative behavior from the year prior. Specifically, every FEMA disaster declaration within the congressional district corresponds with a 0.27-point increase in the legislator’s environmental policy score. The effect grows over time, in the years after the event: in year two, members’ score increases by 0.49 points per disaster, and in the years following the disaster members’ score increases by 0.54 points per disaster relative to the year before the disaster. In sum, Democratic members of the House receive higher environmental scores in the period following the event. In contrast, Republican members of the House engage in anti-environmental legislative behavior after a natural disaster declaration, although the effect is not statistically significant.

• Download figure
• Open in new tab
• Download PowerPoint

Figure 2.

Results for Panel Event Study Testing Hypothesis 1: Partisanship

Source: Data compiled from the Federal Emergency Management Agency Disaster Declarations, the League of Conservation Voters Congressional Scorecard, and The Cook Partisan Voting Index (Federal Emergency Management Agency 2023; The League of Conservation Voters 2023; The Cook Political Report 2023). Author utilized Freyaldenhoven et al. (2022) xtevent Stata package for panel event study estimation.

Note: Top panel: Pretrends p-value = 0.00. Leveling off p-value = 0.43. Bottom panel: Pretrends p-value = 0.72. Leveling off p-value = 0.35.

Next, I test the hypothesis that major natural disasters impact legislative behavior on environmental issues. I categorize congressional districts as experiencing a major natural disaster if FEMA declared five or more disasters that year (occurring in only 10 percent of the sample). For example, New Orleans was gravely impacted by Hurricane Katrina and Hurricane Rita in 2005, resulting in many declared disasters in Louisiana’s Second District. I expect a major FEMA-declared disaster to have a greater impact on legislative behavior for Democratic legislators than a single FEMA-declared disaster. Figure 3 shows the plotted coefficients by party.

• Download figure
• Open in new tab
• Download PowerPoint

Figure 3.

Results for Panel Event Study Testing Hypothesis 2, Major Event

Source: Data compiled from the Federal Emergency Management Agency Disaster Declarations, the League of Conservation Voters Congressional Scorecard, and The Cook Partisan Voting Index (Federal Emergency Management Agency 2023; The League of Conservation Voters 2023; The Cook Political Report 2023). Author utilized Freyaldenhoven et al. (2022) xtevent Stata package for panel event study estimation.

Note: Top panel: Pretrends p-value = 0.01. Leveling off p-value = 0.32. Bottom panel: Pretrends p-value = 0.82. Leveling off p-value = 0.08.

Democratic legislators’ scores shift after a natural disaster, and the effect grows substantively over time: in the year of the event, Democratic lawmakers’ environmental policy scores are 2.5 percentage points higher than their pre-disaster scores, holding all else constant. Major natural disasters substantially impact Republican lawmakers’ environmental policy record, as well. In the post-disaster period, Republican legislators receive lower environmental scores (see figure 3, panel B). Again, this effect increases over time: in the year after a disaster, Republican legislators’ scores drop by 2.7 percentage points relative to the year prior to the disaster. Three years after the event, Republican lawmakers’ environmental policy scores are 4 percentage points lower than their pre-disaster scores, all else equal. Republicans react to major natural disasters by adopting anti-environmental positions.

The third hypothesis examines the policy window created by natural disasters. In interpreting the causal effect of a panel event study, the plot helps to visualize the initial impact of the event, or a jump at time zero when disaster strikes. It can also help visualize the long-term impact of the event on the outcome of interest. I hypothesized that a natural disaster has an immediate, brief window of impact on legislative behavior on environmental issues. Figure 3 presents a different narrative: natural disasters have long-lasting impacts on legislative behavior for both Republican and Democratic lawmakers.

Each model controls for presidential era, to indicate the complicated process of FEMA disaster declarations and the role of the executive branch in agenda-setting. The results indicate polarization between the branches of government: during the Barack Obama administration, Republicans have lower environmental scorecards than during the George W. Bush administration, all else equal (table 2). During the Donald Trump administration, conversely, Republicans have higher environmental scorecards compared to the Bush administration, all else equal. This may reflect policy positions unique to the Trump administration: for example, Trump’s decision to withdraw from the Paris climate agreement was met with disapproval across the aisle (Tollefson 2017).

Voter partisanship is an important control variable that may impact legislative behavior. However, the Cook Partisan Voter Index variable is not statistically significant. Controlling for congressional district and year fixed effects likely accounts for time-varying changes in constituent partisanship.

The pre-trend p-value tests the joint significance of the coefficients in the pre-disaster period relative to the baseline year. For Democratic lawmakers, the pre-trend test yields a p-value of 0.00 (model 1) and 0.01 (model 3), which indicates statistically significant differences in trends prior to disaster declarations. We need to use caution when making a causal link between disasters and changes in environmental scorecards for Democratic lawmakers. For Republican lawmakers, the pre-trend test yields a p-value of 0.72 (model 2) and 0.82 (model 4), which indicates that the pre-treatment coefficients are jointly indistinguishable from zero. There is no evidence of differential trends among Republican lawmakers that experienced a disaster prior to the event, supporting the parallel trends assumption. This further suggests a causal link between disasters and legislative behavior for Republicans. For Democrats and Republicans, the leveling-off p-value is not statistically significant, suggesting that the effect of a disaster on voting behavior has stabilized over time.

In the online appendix, I present additional analyses that examine the relationship between ideology and partisanship in congressional behavior after a disaster: surprisingly, members of Congress whose ideological positions are more moderate (operationalized as in the middle of the distribution of DW-NOMINATE scores) have more partisan responses to natural disasters than Democrats and Republicans overall. I examine whether members of Congress’ response to climate disaster is tempered by partisanship at the state level. In Democratic states, Republicans respond to natural disaster declarations by starkly reducing support for environmental legislation (see online appendix for more details).

The most critical threat to validity is political polarization in environmental policy over the past twenty years. I address this by limiting my analysis to within-legislator behavior. I include year fixed effects to address the impact of frequent reelection cycles. I include presidential administration fixed effects to address the relationship between the executive and legislative branches, as well as the president’s authority over FEMA. The analysis is restricted to estimating changes in legislative behavior annually; more granular data might reveal different patterns in legislative behavior in the days, weeks, or months after a disaster, when lawmakers are facing increased scrutiny and pressure from their constituents. The mixed methods approach provides insight into the trends in the quantitative analysis.

Policy Entrepreneurship: Content Analysis of the Congressional Record

Speeches on the floor of the 117th Congress illustrate how Representatives engage in policy entrepreneurship, focusing on problem definitions and policy solutions, or how the Representative connects climate hazards to policy. The search parameters yield 1,245 congressional records mentioning a natural disaster or a specific weather event. We exclude entries that mention natural disasters as rhetorical flourishes or reference a specific candidate (one example that came up frequently is the phrase “flood of immigrants,” or references to Representative Mike Flood). After coding for context, 25 percent of the sample remained, or 318 individual speeches.

Climate hazards are frequently identified and defined as policy problems. Legislators who spoke about natural disasters on the floor had, on average, a higher number of disaster declarations in their district (mean 3.21) compared to those who did not (mean 1.83). Most legislators mentioned natural disasters only one time on the floor (n = 117). However, other legislators spoke more frequently about a natural disaster in their district: for example, Doug LaMalfa (CA-01, R) and Sheila Jackson Lee (TX-18, D) made twenty-six and seventeen speeches, respectively. Over 70 percent of legislators representing congressional districts impacted by a disaster declaration mentioned the disasters’ effect on their congressional district in a speech. There was no significant or substantive difference in the likelihood that Democrats and Republicans identified a natural disaster in their district as a policy problem.

Partisan differences emerged when defining and framing climate hazards as a policy problem. While the scientific community linked increases in global surface temperatures with extreme weather events and natural disasters, legislators were more reticent to connect disasters to climate change. Among legislators who mentioned natural disasters, only 53 percent connected the disaster to climate change in a speech. There is wide variation in problem definition by party: 21 percent of Republicans mentioned natural disasters and climate change, and about 70 percent of Democrats mentioned natural disasters and climate change in speeches on the floor.

Many legislators took the opportunity to define the problem by assigning blame (n = 96, or 30.2 percent of the sample). Figure 4, panel A presents differences in blame attribution by party. Republican legislators were more likely to define the problem by assigning blame (n = 24 legislators, or 39 percent of the sample). Republicans attributed blame to a wide variety of actors, ranging from inefficient bureaucracy to the Joe Biden administration to environmentalists. Democrats and Republicans united in blaming their fellow legislators for failing to mitigate the harms from natural disasters. Democrats overwhelmingly blamed climate change for the harms accrued to their district.

• Download figure
• Open in new tab
• Download PowerPoint

Figure 4.

Results for Qualitative Coding of Speeches on the Floor of the House of Representatives, Blame Attribution and Policy Solutions

Source: Data compiled from the Congressional Record (2023).

To illustrate how Democratic and Republican lawmakers vary in defining the policy problem of natural disasters, I share two speeches on the floor of the House of Representatives. The lawmakers represent districts that were devastated by Hurricane Ida in 2021. Representative Sheila Jackson Lee (TX-18, D) earned an environmental scorecard of 100, and mentioned natural disasters and climate change in seven of her speeches on the floor. Representative Jackson Lee argued that “natural disasters intensify existing social and racial and ethnic inequities. After a natural disaster, people in disadvantaged communities have worse outcomes due to stresses or exposures to other risks associated with a disaster. Low-income people and people of color are more likely to live in disaster-prone areas, live in lower-quality housing that is less safe when disasters occur, and have chronic illnesses that make them more vulnerable” (“House proceedings” 2022c).

Representative Jackson Lee connects natural disasters to climate change, using an intersectional lens to understand how climate hazards disproportionately burden low-income and minority communities. In congressional district LA-06, Republican Representative Garret Graves offers a different framing: “Just to explain what happens here, in the aftermath of a disaster … you have an alphabet soup of agencies that come in and offer assistance. It could be FEMA, it could be the Small Business Administration, it could be the Department of Agriculture, it could be the Department of Housing and Urban Development, or others. Madam Speaker, the last thing in the world the victim of a disaster needs is to be revictimized by their own government bureaucracy” (“House proceedings” 2022c).

Many Republican representatives voiced frustration with the inefficiencies of the federal government and FEMA, including Troy Nehls (TX-22) (“House proceedings” 2022a), and Randy Webber (TX-14) (“House proceedings” 2022a).

Legislators also differ in their specific policy solutions. Speeches that mentioned a policy alternative or solution connect the disaster to an action, such as sponsoring new policy or revising existing legislation. Republican representatives focused on disaster response, resource management, and infrastructure development rather than broader climate change policies. Consistent with Republican core ideology favoring small government, many speeches called for revising or reauthorizing federal legislation rather than ambitious new policies. Republican lawmakers had a distinct focus on policy related to agriculture, farmers and ranchers (n = 26), and food supply (n = 14), tailoring their policy solutions to suit their constituents in rural areas. Democrats, in contrast, focused on sponsorship or advocacy of new legislation (n = 156 or 85 percent) aimed at climate change policy or expanding federal capacity to respond to climate hazards. During this congressional session, Democrats had many opportunities to connect disasters in their district to climate change policy during debates on the Inflation Reduction Act. The online appendix includes a detailed list of legislation discussed by representatives.

To illustrate partisan differences, I offer two examples where Democratic and Republican lawmakers connected the policy problem of drought and wildfire to their preferred policy solutions. Democratic Representative Joe Neguse (CO-2), founder and cochair of the Bipartisan Wildfire Caucus, advocated for new legislation focused on climate hazard mitigation strategies: “As climate-change-induced drought means long-term reductions in rain, snow, and necessary water supplies, this legislation invests in droughtproof water projects like water recycling, modern desalination, advancing water data management and technology, protecting and restoring ecosystems” (“House proceedings” 2022b).

Representative Neguse sponsored the Wildfire Response and Drought Resiliency Act, which passed narrowly with one House Republican voting in favor. In contrast, Republican Representative Bruce Westerman (AR-04), chairman of the Natural Resources Committee, used concerns over droughts and forest fires to argue against conserving almost three million acres of public land across Colorado, California, Arizona, and Washington state: “Do you care about the health of our forests? Then don’t vote for this bill. The most basic types of preventive maintenance are illegal under wilderness designation. You can’t take a chain saw and cut away fire traps, and you can’t trim branches or overgrown timber stands and underbrush. (…) Should this become law, with its restrictions on fire prevention, we are going to see bigger and hotter fires in Colorado than ever before (“Colorado Wilderness Act of 2021” 2021).”

Representative Westerman uses the problem of droughts and fires to argue against federal wilderness designations. While Westerman does not represent an area routinely affected by drought and wildfire, his position as chairman on the Natural Resource Committee carries weight in this policy debate.

Democrats and Republicans occasionally work across the aisle. Representatives of California and Colorado used disasters in their districts to mobilize for new legislation. In California, Representative Zoe Lofgren (CA-18) thanked the bipartisan House Committee on Natural Resources for their work on the Wildfire Response and Drought Resiliency Act, reminding legislators that “fires don’t rage only in Democratic districts. This calamity knows no political divide” (“House proceedings” 2022c).

CONCLUSION

This article makes a theoretical contribution to the policy process literature, helping scholars understand the dynamics of policymaking in a polarized era. Partisanship is a critical frame for understanding legislative behavior in response to climate hazards. A natural disaster serves as a focusing event if the political context is right: disasters may change the legislative behavior of Democrats, increasing their votes in favor of environmental policy issues. Content analysis of speeches on the floor of the House of Representatives demonstrates how Democrats link natural disasters to climate change, using their policy window to advocate for pro-environmental policy during the 117th Congress. Democratic Representatives also focus on environmental inequalities, demonstrating how climate hazards disproportionately impact low-income, minority residents.

For Republican legislators, major natural disasters increase conservatism on environmental policy issues. Republican legislators are less likely to vote in favor of environmental legislation after a major disaster, and their anti-environmental policy stance is substantively and statistically significant. There are myriad mechanisms for increased conservatism: a disaster may spur a “rally around the flag” effect, boosting support for the legislator and their ideological position (Groeling and Baum 2008). Alternatively, fear and insecurity stemming from major natural disasters may renew support for anti-environmental ideological positions among conservatives (Bar-Tal 2020; Dörries 2010).¹⁴ While members of Congress who support pro-environmental legislation after a disaster have no change in their total campaign contributions (Jud and Nguyen 2024), it is possible that fear of donor reprisal impacts legislative behavior on the right. Content analysis suggests that Republicans define natural disasters as policy problems that do not require pro-environmental legislation. By connecting the problem of natural disasters to inefficiencies in the legislative and executive branch or the federal bureaucracy, the clear policy solution is less government oversight, not more.

Natural disasters have a long-lasting impact on legislative behavior. Major disaster recovery takes time, and communities must grapple with housing and neighborhood reconstruction. Legislators are especially responsive to issues related to their constituents’ daily life (Canes-Wrone and Shotts 2004). Lags in emergency assistance from FEMA and long-term aid from federal agencies such as Housing and Urban Development, inconsistencies in federal funding, and state and local capacity all impact disaster recovery (Teles and Martin 2021). The realities of disaster recovery help explain the long-lasting impact of natural disasters on legislative behavior: congressional districts feel the effects of disasters for years. Rather than inspiring a brief change in behavior after the event, natural disasters alter legislative behavior on environmental policy issues.

Congressional testimony illustrates how policymakers seize policy windows to champion policy and distribute blame. Representatives define natural disasters as policy problems that lead to their preferred policy solution. For Democrats in the 117th Congress, natural disasters offered proof of the existential threat of climate change for their districts, allowing them to advocate for pro-environmental policy. Representatives connected climate hazards with structural inequalities in access to housing, employment, and health, in vulnerable communities, the economic burden of climate hazards causes disproportionate financial strain, building on existing inequalities. Many Republicans critiqued the Biden administration’s response to disasters in their districts, and some argued pro-environmental policies, such as forest preservation, added fuel to forest fires.

Natural disasters occur locally, but have the power to shape national policy. While numerous studies have examined the effect of natural disasters on local policy outcomes (Arceneaux and Stein 2006; O’Donovan 2017; Zelin and Smith 2023), this article suggests that natural disasters have deep impacts outside of their geographic area. For example, the landmark Inflation Reduction Act (IRA) of 2022 is the first domestic legislative effort to combat climate change in recent history, and will help the United States meet its emission reduction commitments. The IRA passed the House of Representatives on party lines with a vote of 220–207. Legislation passes by slim margins: a small change in an individual legislator’s environmental policy position may have a large impact on national policy.

What do asymmetric findings along partisan lines mean for environmental policy in the United States? In the recent 2024 election, Republicans gained control of the executive and legislative branches, suggesting an uncertain future for climate policy. I find that partisanship conditions response to natural disasters, but there are important nuances to this partisan divide. For example, investment in green energy and job growth may impact perceptions of environmental legislation: about 85 percent of the announced investments from the IRA benefit Republican districts and encourage job growth in rural areas, which may shape perceptions of environmental legislation for conservative voters (Gaffney 2024). Younger elected representatives may also depart from President Trump’s views on climate change: in interviews, young conservatives identify climate change, conservation, and energy as salient policy issues (Bustillo 2024). Younger Republicans are much more likely to say that human activity contributes to climate change (PEW Research Center 2024). Finally, the legislature may pursue environmental policies that resonate more with Republican voters: conservatives support expanding both fossil fuel and renewable energy sources (PEW Research Center 2024).

In lieu of concerted federal efforts, state legislatures have accomplished important policy victories in environmental protection. Many states have established commissions to address environmental injustice and climate hazard mitigation policy. However, uneven state investment in climate adaptation policy may exacerbate inequality: regions that are more likely to be impacted by climate hazards are also less likely to invest in state climate adaptation policy (Avtar et al. 2023). As we grapple with an increase in climate-change-related natural disasters and a divided electorate, scholars and activists must find ways to frame environmental policies to appeal to voters and elected representatives. Future research should examine how scientists, academics, and government officials can galvanize conservative voters and elected representatives to support environmental policy after natural disasters. Without concerted federal efforts in the legislature, inequality in climate hazard burden and adaptation will continue to grow.