Following the Path to Policy Change

Publication Date: 2022

Abstract

In September 2020, a series of wildfires in Oregon caused extensive property destruction and tragic loss of life, as well as widespread smoke and diminished air quality. This study uses qualitative data from 36 interviews and a process tracing approach to explore how potential focusing events—in this case, the Oregon 2020 wildfires—yielded opportunities for public policy change. The study offers empirical evidence about how extreme weather events shapes public policy, contributes to ongoing development of focusing events theory, and highlights the mechanisms by which and conditions under which public policy change occurs.

---

Introduction

The September 2020 Oregon wildfires were widely described as an unprecedented event with respect to both geographic scope and the number of communities affected by both smoke and wildfire (Schmidt et al., 2020¹). The event began abruptly on September 7, 2020, with the rapid expansion of several small wildfires throughout the state. It was unusual in provenance, occurring because of an abrupt shift of winds out from the east. The wildfires resulted in 9 fatalities, over 4,000 homes destroyed, and an estimated $1.15 billion in total response and recovery costs (Boudet et al., 2021²).

Literature Review

Disasters and extreme weather events have strong potential to act as “focusing events,” introducing windows of opportunity for policy actors to convene around defining the problem, identifying potential solutions, and promoting or suppressing policy change (Birkland, 1997³, 2006⁴; Kingdon, 1984⁵). Kingdon (1984) initially used the term focusing event to refer to any event, symbol, or experience with potential to focus attention on a policy problem and provide opportunities for agenda setting. Birkland (1997, 1998⁶, 2006) refined Kingdon’s concept, defining a potential focusing event as one that is sudden, rare, and harmful (or likely to cause future harm), and which becomes known to the public and policy makers at the same time. Such events tend to generate significant attention among the media, the public, and policy makers, providing opportunities for policy entrepreneurs—individuals who invest time and resources to advocate for a specific policy position—to emerge and influence policy discourse and outcomes (Kingdon, 1984).

Empirical evidence about policy change following extreme weather events has highlighted the role of both event characteristics and attention in the policy process, substantiating some of the key expectations of focusing events theory (e.g., Birkland, 2006; Giordono et al., 2020⁷; Rudel, 2019⁸; Yeo & Knox, 2019⁹). However, the dynamics of the mechanisms by which focusing events act to yield policy change or stasis is not as well developed (Birkland & Warnement, 2016¹⁰). Moreover, evidence about the how extreme weather events yield changes to climate adaptation and mitigation policy is mixed and would benefit from theoretical development (Giordono et al., 2021¹¹).

Finally, while the focusing events literature is most closely associated with the Multiple Streams Approach (Herweg et al., 2017¹²; Kingdon, 1984), it also draws from several other theories of the policy process. Such theories offer various lenses on the policy process, attending to “the interactions that occur over time between public policies and surrounding actors, events, contexts, and outcomes” (Weible & Sabatier, 2017¹³, p. 2). The Advocacy Coalition Framework (ACF), for example, examines how coalitions of policy actors coordinate and act strategically to influence public policy (Jenkins-Smith et al., 2017¹⁴; Sabatier & Jenkins-Smith, 1993¹⁵). Punctuated Equilibrium Theory turns to attention—especially media attention—to explain major policy changes that depart from the more typical experience of incremental policy change (Baumgartner & Jones, 1993¹⁶). The proposed study heeds several recent calls (Capano & Howlett, 2021¹⁷; Kay & Baker, 2015¹⁸; van der Heijden et al., 2021¹⁹) for more development of causal mechanisms in the broader policy process literature, including these frameworks. Figure 1 presents a simplified conceptual framework developed for the purpose of this project derived from the theories described above.

Figure 1. Conceptual Framework

Research Design

Research Questions

This study explores the lens on the causal mechanism(s) by which public attention surrounding a disaster event translates into policy entrepreneur and advocacy coalition activities to influence policy. Specifically, the study responds to Birkland and Warnement’s (2016) call for a closer examination of the role of post-event discourse in the agenda-setting process, as well as broader calls for policy scholars to attend more closely to mechanisms by which policy change occurs (Capano & Howlett, 2021; van der Heijden et al., 2021). Following these observations, we use process tracing methods (Beach & Pedersen, 2013²⁰) to address the following research questions:

1. What state-level policy changes, if any, were proposed and/or adopted in the wake of the 2020 Oregon wildfires?

2. What were the causal mechanisms by which the Oregon wildfires, as a potential “focusing event,” yielded policy change or stasis?

3. What do the findings suggest about the conditions that shape post-event policy change?

Study Context

Similar to other Western states, Oregon experienced multiple wildfires in 2020, including 330 fires on protected lands (Oregon Department of Forestry, 2021²¹). However, Oregon’s September 2020 wildfires, including the Beachie Creek, Lionshead, Holiday Farm, Riverside, and Almeda Drive fires, were widely described as “unprecedented” with respect to their impacts on land and people. The September 2020 wildfires began in late August, but expanded rapidly after gusting winds out of the east began on September 7 (Urness & Whitney, 2020²²). Ultimately, the fires burned over 1.2 million acres (4,850 square kilometers) in Oregon and contributed to 64.7 million person-days of hazardous air quality across the Pacific Northwest (Boudet et al., 2021).

Several Oregon communities—including Detroit, Blue River, Vida, Talent, and Phoenix—experienced devastating losses, with 9 Oregonians losing their lives and over 4,000 homes destroyed (Oregon Office of Emergency Management, 2021²³). Many more Oregonians experienced evacuation warnings or orders, and large areas of the state were blanketed by hazardous air quality for days. The state estimated a total cost of $1.15 billion in wildfire/wind damage, response costs, and debris removal (Governor’s Wildfire Economic Recovery Council, 2021²⁴). Due to the unusual wind from the east, smoke from the wildfires impacted some of the most populous areas of the state along the Willamette Valley corridor even 10 days after the initial wind shift. Figure 2 shows an image of near-surface smoke on September 17, 2020.

Figure 2. Air Quality September 14, 2020.

Data, Methods, and Procedures

This study uses qualitative data and a process tracing approach to address the research questions. Process tracing is a method used to identify causal mechanisms that link causes with observed outcomes to test and refine or build theory. In this study, data were collected from multiple sources to inform findings surrounding causal mechanisms that translate public attention into action by policy actors and ultimately policy change or stasis. Process tracing requires in-depth information in the form of causal process observations (CPOs), defined as “an insight or piece of data about the context or mechanism and contributes a different kind of leverage in causal inference” (Brady et al., 2010²⁵, p. 158).

One-on-one interviews with policy actors comprised the primary data source used in the study, supplemented by policy documents and newspaper articles. In total, 36 interviews were conducted with 38 respondents (two interviews were conducted with two respondents at their request) during the four-month period of April 13, 2022 to August 9, 2022. Thirty-five semi-guided interviews were conducted via Zoom video, and two were conducted via Zoom audio-only or telephone. Interviews averaged 46 minutes in length, totaling 27.9 hours (1,672 minutes) of interview time. Interviews included policy actors with various roles, including advocates (12), elected policy makers (9), industry representatives (7), non-elected policy makers (5), educators (2), and scientists (1). Interviews were also conducted with policy actors from various jurisdictional interests, including state (24), county (9), and city (2).

During the interviews, respondents were asked about their personal and professional experiences with the 2020 wildfires; impressions of Oregon wildfire management strategies; impacts of the 2020 wildfires on Oregon communities; post-wildfire policy changes and related processes; and other impacts of the wildfires. Respondents were also asked to name other potential interview respondents with insights into the post-wildfire policy change process. Additional interviews were conducted until saturation was reached (Saunders et al., 2018²⁶), which for this study meant both the waning of new substantive information with respect to theoretical constructs, as well as completion of interviews with key policy actors named by multiple other respondents. The timing of the interviews allowed the pursuit of specific details about the entire nine-month time period during which problems were defined and new policies were developed, selected, debated, and voted upon, as well as insight into post-policy implementation and related discourse. All interviews were recorded and transcribed, with one exception in the case of technology malfunction. Written notes were also taken during all interviews to enable pre-transcription analysis and generation of preliminary findings. In addition to interviews, other policy-related documents (e.g., newspaper articles, legislation) were collected and catalogued for the purpose of both respondent sampling and qualitative analysis.

Sampling and Participants

The study followed the approach taken by McAdam and Boudet (2012²⁷) for identifying and sampling respondents. Relevant, publicly-available documents, including newspaper articles, active participation in the legislative process, and membership in a 2019 advisory committee assembled by Oregon’s governor were systematically gathered and catalogued. Prospective respondents were identified and ranked from those documents, as described in detail below.

The sampling process began with a search of local Oregon newspapers for wildfire-relevant articles and letters to the editor dated September 1, 2020 to June 30, 2021, using the keyword “wildfire” to search the Oregon News Sources database via NewsBank. This effort yielded a database with a total of 3,082 articles. I then contracted with a machine learning expert to extract person names from the articles, yielding a database with a total of 12,310 person names, including duplicates. Finally, I used RStudio to perform several data cleaning operations, including dropping duplicate articles; merging the names and articles databases; and constructing pivot tables with articles counts by unique person name. The vast majority of individuals (82%) were named in just one or two articles, although one individual, Governor Kate Brown, was named in 270 distinct articles.

I then gathered information about individuals who participated in the legislative policy process and added it to the contact database. I began by searching the Oregon State Legislature website for all bills introduced during the 2021 regular legislative session, using the keyword “wildfire,” which yielded a total of 35 unique bills. Information about bill—including number, title, chamber (House or Senate), fiscal impact, revenue impacts, and status/location (enrolled and/or current committee), chief sponsor(s), and regular sponsor(s)—was collected and entered into a database. I also collected detailed information about testimony provided for all bills, including the submitter’s name and organization. Bill sponsorship and testimony information was merged into the existing contacts database. Finally, names and information for individuals who participated in the 2019 Governor’s Council on Wildfire Response were merged into the database.

As a result of these initial data collection processes (newspaper articles, legislation, council membership), 3,765 unique individuals were identified; these prospective respondents served as the pool from which respondents were selected for initial contact. Potential respondents were selected on the basis of participation in the policy process, with priority given to those who participated in all three modes of the policy process: article citations; legislative testimony; and sponsorship of or membership on the 2019 Council on Wildfire Response. Detailed respondent information, including organization, title, and contact information, was then collected for individuals with evidence of a high level of participation. Additional referrals from initial interview respondents were added to the database, and selected prospective respondents were targeted based on those referrals. Prospective respondents were invited to participate in the interview via e-mail, which included the initial invitation, an overview of the study, links to IRB clearance, and a calendar link for scheduling the interview. E-mails were sent to a total of 113 respondents: 67 in April/May, and 42 in June/July.

Institutional Review Board Approval and Other Ethical Considerations

After review of study documents, including interview guides, the study was approved under Oregon State University’s Institutional Review Board study IRB-2019-0137. In the interest of acknowledging the intensely personal experience of the 2020 fires for many Oregonians, respondents were provided with an opportunity to describe their own experiences, both personal and professional, with the 2020 wildfires. Respondents were also asked if they would like to be added to a distribution list for study-related products; all respondents replied yes and will be sent a link to final publications.

Data Analysis

The transcripts were recently completed and therefore have not yet been fully analyzed. Preliminary analyses and findings are based on an initial scan of the interview notes. This section describes both the preliminary analysis process and intentions for final analysis.

Preliminary Data Analysis Process

Preliminary data analysis was conducted using the initial notes taken during interviews and broad elements derived from the conceptual framework, namely (1) evidence of policy change; (2) causal mechanisms; and (3) conditions. A review of the notes and related documents yielded initial observations that will be tested further in subsequent analysis. The results from the preliminary analysis were presented at the Natural Hazards Researchers Meeting in July 2022.

Final Data Analysis Process

Full data analysis will proceed as follows. First, transcription documents will be uploaded to Dedoose 9, a cross-platform software application developed for qualitative and mixed methods research. Next, a system of codes will be used to identify and categorize policy change, policy actors, and key policy process elements (attention; leadership and coalitions; agenda-setting and problem/solution definitions; and policy-oriented learning) based on the conceptual framework identified above. Codes will be applied to the existing interview transcripts. This coding approach identifies positive evidence of the concept in question and allows for qualitative assessment of element presence, counts, and code co-occurrence.

Text analysis will also be used on newspaper articles to isolate topical trends with respect to media coverage during the nine months following the 2020 wildfire events. Latent Dirichlet Allocation (LDA) is a common method in the field of text analysis that uses a form of unsupervised machine learning to identify clusters of words based on their co-occurrence (Blei et al., 2003²⁸). LDA has been used in previous publications to measure policy-oriented attention and problem definition (e.g., Giordono et al., 2020). While it is a computer-driven method, it allows for analyst interpretation of the findings. Findings from the analysis will be used in the final data analysis as evidence about media coverage and topical focus.

Preliminary Findings

State Policy Changes

During the regular legislative session in Oregon, which began on January 21, 2021, 98 bills related to wildfire were introduced in the Oregon House and Senate chambers, compared to 12 bills in 2020 and 25 bills in 2019 (Oregon State Legislature, n.d.²⁹). Due to COVID-19’s impact on activities in 2020, 2019 may be considered a more typical legislative year. By the end of the 2021 session, 35 out of the 98 introduced bills had been enrolled. By the end of the 2020 session, 35 bills (out of the original 98 that were introduced) had been passed and enrolled by both the House and Senate.

Of special note, Senate Bill 762 (SB 762)—described as a comprehensive attempt to address wildfire risks—was passed during the last week of the session (Stites, 2021³⁰). Most respondents highlighted SB 762 as the most important policy change and focused on the process leading to the bill’s passage. SB 762 was omnibus and bipartisan legislation that implicated 14 state-level agencies (see Figure 3) and allocated a historically high level of funding (approximately $220 million over two years) to efforts to address future wildfire risks. Several stakeholders described the bill as being unique to Oregon and indicated that the bill has set a new standard for other states.

Figure 3. Government agencies implicated in Senate Bill 762 text.

Moreover, 15 distinct key objectives were incorporated in SB 762 (Relating to Wildfire; and Declaring an Emergency., 2021³¹). Among these, five objectives, marked by an asterisk (*) below, received the most attention from respondents with respect to innovation, impact, and contention.

• Develop electric system plans
• Develop statewide wildfire risk map *
• Establish local defensible space requirements *
• Identify land use updates
• Adopt building code standards
• Develop program to mitigate risks of wildfire smoke
• Coordinate emergency response activities
• Develop landscape resiliency and fuel reduction program *
• Establish Small Forestland Grant Program *
• Adopt rules clarifying prescribed fire regulations
• Adopt rules establishing wildfire protection for lands outside of forest protection districts
• Establish expanded system of automated smoke detection
• Define and establish classes of wildland-urban interface *
• Establish State Wildfire Programs Director
• Establish Wildfire Programs Advisory Council

Importantly, while respondents noted the comprehensive and innovative nature of the legislation, they also remarked that most of these objectives were derived directly from previous work by the 2019 Governor’s Wildfire Council. Also, despite the apparent primacy of SB 762, it should also be noted that multiple other bills were passed early in the session that explicitly addressed community recovery needs in the wake of the event. Though these bills were smaller in scope and less comprehensive, they were highlighted by some respondents as evidence of legislator responsiveness and innovation, especially those with direct connections to affected communities.

Causal Mechanisms

Preliminary evidence points to several causal mechanisms of change, including attention; leadership and coalition-building; agenda-setting and problem/solution definition; and policy-oriented learning.

The 2020 Oregon wildfires resulted in considerable attention in the media, as well as from the public and policy makers. Media attention during and after the wildfires was prolific, with 1,202 newspaper articles published in the month after the wildfires in Oregon newspapers alone. Over the course of the nine-month period following the wildfires, 2,960 relevant articles, including letters to the editor, were published in Oregon newspapers. Even though media attention declined sharply after September, 195 articles referencing wildfire were published per month on average from October to June, as noted in Figure 4.

Figure 4. Oregon Newspaper "Wildfire" articles.

Despite wide assertion among respondents that attention to the wildfires was a key mechanism leading to the passage of Senate Bill 762—specifically with respect to its breadth of influence and funds invested—respondents also pointed to several other change mechanisms. Specifically, they highlighted the important roles played by (1) leadership and coalition-building; (2) agenda-setting and problem/solution definition; and (3) policy-oriented learning.

Respondents called attention to the strong and sustained leadership played by state representatives from locally-affected areas, as well the influence of at least three major organizational alliances with varying core beliefs about the best direction for policy change. Respondents also highlighted the role of the leadership surrounding the 2019 Governor’s Wildfire Council. Specifically, they spoke of the inclusive nature of that process and successful attempts to address a wide variety of viewpoints, which helped to minimize contention during the 2021 decision-making process. They pointed to close alignment between the recommendations from the Governor’s Council and the objectives ultimately adopted in SB 762, and attributed that alignment to effective leadership at both stages.

Despite the bipartisan nature of the final bill, respondents also acknowledged a moderate level of contention during the policy development process. Most elements of the discourse revolved around a largely undisputed community protection frame; in other words, the bill and related discussion were centered on how to protect communities in the event of future wildfire. However, there was considerable debate about the underlying problem and related solutions (e.g., excessive versus insufficient use of fire suppression; excessive versus insufficient fuels reduction; insufficiently resilient communities; workforce availability; etc.). There was also some debate about governance strategies, especially the role of local versus state government. Climate change was acknowledged by most respondents but was not generally described as the main source of policy discourse, and climate change mitigation was not prioritized as a solution during the policy process.

Finally, there was some evidence of learning within the policy process. Respondents described a direct connection between 2019 recommendations from the Governor’s Wildfire Response Council and the key elements of SB 762, but also acknowledged opportunities to learn from the 2020 wildfires. For example, the wildfires primarily burned on the West slope of the Cascades, which was unusual due to the high degree of moisture typically experienced in that area (Flavelle & Fountain, 2020³²); respondents suggested that additional research is required to understand how to prevent and respond to wildfires in such areas. Moreover, SB 762 prioritized the production and dissemination of information in its call for the development of wildfire risk maps for use in subsequent decision-making, suggesting a data-driven approach to policymaking ^{Endnote 1}.

Conditions

Both event attributes and the policy context stood out as key conditions for change. Respondents pointed to the geographically widespread impacts of the wildfires, including smoke and evacuation orders, as conditions that yielded both media and public attention. They also underscored the role played by the tragic nature of more localized consequences, including property destruction and loss of life.

Finally, several respondents noted that political conditions also contributed to the adoption of SB 762, namely fiscal health conditions (i.e., budget surplus). While some respondents also pointed to the Democratic “trifecta,” (i.e., Democratic Governor, plus Democratic majority in both House and Senate), this condition has been in place since 2013, so it is unlikely to have been a direct contributor to passage of the legislation.

Conclusions

Summary of Preliminary Findings

Preliminary findings suggest that this case is demonstrative of the major elements in the conceptual framework. In particular, attention (e.g., media, public, policy maker) appears to be a key causal mechanism that, in combination with other elements (e.g., leadership and coalition activity, policy discourse, and policy-oriented learning), yielded considerable policy change in Oregon. That said, the connections between these elements have not been fully explored. A more thorough analysis is likely to provide important insights about the direction and degree of influence.

Limitations, Strengths, and Future Research

This study offers an opportunity to take an in-depth look at how state policy makers, advocates and the public responded to an extreme event that generated widespread attention. However, it only represents one state and a small group of committed policy actors and does not offer the commonly pursued generalizability to the broader population. However, this case and the respondent sample offer considerable insight into the policy process, yielding analytic and theoretical generalizability (Yin, 2013³³). Moreover, the study provides a model for conducting similar research in other states, potentially yielding a selection of cases conducive to a larger cross-case comparison.

Implications for Theory, Policy, and Practice

Once full analysis is complete, this study will make important contributions to the empirical and theoretical literature. First, the research yields detailed information about the degree to which the Oregon 2020 Wildfires yielded state-level policy change. We are learning about key strategies and mechanisms—for example, narrative, framing, and learning—by which policy actors have attempted to influence the policy process, as well as conditions that led to their success or failure. Additionally, the study contributes to focusing events theory, which is currently underdeveloped, especially in the context of extreme weather events (Giordono et al., 2021). As extreme events increase in frequency and severity, we are likely to see more efforts at the state level to address future risks through public policy, and this study contributes insight into conditions that are likely to be conducive to policy change.

It is always challenging to take insights about the policy process and apply them to the policy process. However, in this case, there are at least two preliminary observations that may be valuable to policy makers and advocates alike. First, the alignment between the 2019 advisory council recommendations and the ultimate 2021 legislation is promising. A well-designed and inclusive planning process that yields actionable items can be an effective use of time and resources, even if the recommendations are not acted upon until a later date. Second, it highlights the post-event period as being ripe for policy change. While extreme events can yield destruction, chaos, and tragedy in affected areas, they also offer a window of opportunity for reconsidering how we choose to support our citizens and protect our communities against future risks.

---

Endnotes

Endnote 1: The wildfire risk mapping process was outside the timeframe of the study, which focused on the initial policy development. However, it is worth noting that implementation has not proven as smooth as the legislative process. Considerable backlash against initial wildfire mapping results has resulted in the subsequent withdrawal of the initial maps (Profita, 2022³⁴).↩

References