- Split View
-
Views
-
Cite
Cite
Fengxiu Zhang, Eric W Welch, Explaining Public Organization Adaptation to Climate Change: Configurations of Macro- and Meso-Level Institutional Logics, Journal of Public Administration Research and Theory, Volume 33, Issue 2, April 2023, Pages 357–374, https://doi.org/10.1093/jopart/muac027
- Share Icon Share
Abstract
Climate change can bring about large-scale irreversible physical impacts and systemic changes in the operating environment of public organizations. Research on preconditions for organizational adaptation to climate change has produced two parallel lines of inquiry, one focusing on macro-level norms, rules, and expectations and the other on meso-level culture, design, and structure within the organization. Drawing on the meta-theory of institutional logics, this study proposes a configurational approach to link institutionally aware top managers with the combination and reconciliation of macro- and meso-level logics. We identify government authority, professionalism, and market as macro-level institutional logics, and risk-based logic and capacity-based logic as critical meso-level institutional logics. Our theory proposes that (1) the macro- and meso-level institutional logics co-exist in systematic ways as to produce identifiable configurations, (2) the configurations are differentially associated with climate adaptation, and (3) the effects of each logic differ across the configurations. Using a 2019 national survey on approximately 1000 top managers in the largest U.S. transit agencies, we apply latent profile analysis to identify three distinct clusters: forerunner, complacent, and market-oriented. Only the forerunner cluster is adaptive to climate change, whereas the two others are maladaptive. Findings from the multigroup structural equation model also demonstrate varied effects of each institutional logic on adaptation across the clusters, confirming institutional work at play to reconcile and integrate co-existing and potential contradictory logics.
Introduction
Climate change is the most significant challenge of our times. It has the potential to cause large-scale, far-reaching and irreversible disruptions to the operating environment of public organizations (IPCC 2021). Such disruptions include but are not limited to increased extreme weather events, crumbling critical infrastructure, ecological crises, social unrest, and surging demand for disaster management. In the United States alone, 308 billion-dollar weather and climate disasters have occurred since 1980, resulting in an aggregated cost of $2.085 trillion (Smith 2020). Situated at the crossroads of action and consequence, public organizations are responsible not only for responding to climate-induced disasters but also for adapting to an increasingly uncertain future through building resilience over time. Adaptation is the set of decisions and actions organizations undertake in response to experienced or anticipated perturbations (Nelson et al. 2007). Resilience is the ability to absorb disturbances and reorganize while going through adverse changes so as to retain the same function, structure and identity (Walker et al. 2004). Yet to public administration scholars, the promise of adaptation immediately brings to mind the institutional contexts of public bureaucracies, which are often characterized as highly politicized and structurally intractable. We are left wondering whether it is possible for public organizations to adapt to climate change. And if adaptation is possible, how does it happen and what does it look like?
Although climate adaptation is increasingly prescribed (Dovers and Hezri 2010; Zhang et al. 2018), systematic reviews show that adaptation actions are far fewer than needed (Berrang-Ford et al. 2011; Bierbaum et al. 2013). The adaptation deficit has triggered extensive efforts to investigate the mechanisms and preconditions for adaptation to occur. The institutional approach figures prominently, emphasizing macro-level norms, cultural expectations and beliefs that define and enforce the “rules of the game” by which organizations operate (Inderberg 2011; Miao et al. 2018b; Naess et al. 2005). Other research has examined how organizational culture, design, and structure, established over time based on history of events and experiences, condition organization-level expectations, understanding, and practices regarding climate risks (Berkhout 2012; Nowell and Stutler 2020; Page and Dilling 2020).
In this study, we apply the meta-theory of institutional logics to argue that organizations adapt when key institutional actors such as top managers integrate and reconcile macro- and meso-level institutional logics (Blomgren and Waks 2015; Hambrick and Mason 1984) to justify and support building resilience to climate change. Borrowing from Thornton and Ocasio (2008), we define institutional logic as the “socially constructed, historical patterns of material practices, assumptions, values, beliefs and rules by which individuals produce and reproduce their material subsistence, organize time and space, and provide meaning to their social reality.” At the macro-level, institutional logics often include market, democratic governance, and professionalism logics (Blomgren and Waks 2015; Friedland and Alford 1991). At the meso-level, institutional logics reflect localized structures and cultural orientations about risk, capacity, transparency, openness, and other dimensions operating within the organization (Gould 2021; Noordegraaf 2011). Combinations of macro- and meso-level logics are embedded in the roles, routines, and practices of all public organizations, providing deep justification for and acceptance of decisions and actions (Greenwood et al. 2011; Thornton and Ocasio 1999; 2008).
The institutional logic approach provides a means of linking exogenous climate change to organizational adaptation. It highlights the cognitive processes of top managers in detecting, interpreting, and enacting the external environment (Daft and Weick 1984; Maitlis and Sonenshein 2010) as well as the role of managerial agency in altering the combination and prioritization of new and existing institutions (Besharov and Smith 2014). When top managers recognize that institutional logics no longer provide valid guideposts or patterns of activity no longer fit experience (Lawrence et al. 2011), they mobilize a different combination of logics to account for decisions and actions that reduce uncertainty and boost resilience (Friedland and Alford 1991).
This study takes a configurational approach to integrate the macro- and meso-level institutional logics, asking: (1) What configurations emerge from the coexistence of macro- and meso-level institutional logics in public organizations? (2) How are the configurations differentially associated with climate adaptation? and (3) How do the relationships between the institutional logics and adaptation vary across the configurations? The empirical investigation focuses on extreme weather as the primary stimulus of adaptation to climate change (Berrang-Ford et al. 2011). Using a 2019 national survey of approximately 1000 top managers from the largest US transit agencies merged with institutional data from various sources, the analysis discovers three distinct combinations of the macro- and meso-level institutional logics, which we label as forerunner, complacent, and market-oriented. The forerunner combination shows a significantly higher level of adaptation, whereas the complacent and market-oriented are both maladaptive to climate change. The results further demonstrate that the effects of each institutional logic on adaptation significantly vary across the combinations. Of note are the macro-level professional logic and the meso-level risk-based logic that consistently promote adaptation across the three combinations.
The study contributes to the public administration, climate adaptation, and institutional logic literatures. First, it uniquely features physical impacts of climate change and applies a public administration lens to examine adaptation in public organizations. In doing so, it bridges the “marked degree of separation” between climate adaptation studies and public administration research (Dovers and Hezri 2010; Dupuis and Knoepfel 2013). Second, we add to the emerging scholarship on organizational adaptation to climate change by recognizing the role of institutionally aware top managers in navigating institutional logics in response to climate impacts. The findings offer evidence of institutional work in reconciling and integrating complex institutional logics in ways that promote adaptation and resilience. Third, we propose and test a configurational approach to integrate macro- and meso-level logics, providing both a holistic method and a means of examining cross-level institutional effects on adaptation. The results unveil new insights about facilitators and barriers to climate adaptation, confirming theoretical and analytical purchase provided by the configurational approach.
The next section presents the meta-theory of institutional logics that links macro- and meso-institutions, climate impacts, and the role of top public managers in reconciling tensions among logics within and across levels. We then introduce a configurational approach, positing that organizations will systematically cluster based on unique combinations of the logics, that the clusters are differentially associated with adaptation, and that the effects of each logic on adaptation will vary across the clusters. We test those propositions and end with discussion on this study’s theoretical and practical implications.
Meta-Theory of Institutional Logics: Linking Macro- and Meso-Institutional Environment
One way to understand the complexity of organizational adaptation to climate change is to examine the interplay of two systems of institutional logics, one operating at the societal or macro-level and the other at the organizational or meso-level. Top managers reconcile and integrate macro- and meso-level logics that results in, for example, governance practices in the organization (Hardy and Maguire 2016).
Macro-Level Institutional Logics
Conventional understanding of macro-level institutional logics stems from the seminal paper by Friedland and Alford (1991) who recognized the contradictory logics embedded in major societal-level institutions, such as the capitalistic market, bureaucratic state, and democracy logics. Institutional logics operate at the societal level, but they permeate organizations because individuals who populate organizations are embedded in the broader social fabric and functioning. In organizations, institutional logics constitute the deeply held beliefs, values, and assumptions that justify and legitimize structures, routines, rules, decisions, and actions (Thornton and Ocasio 1999). The logics provide frames of reference that direct actors’ attention, condition their sensemaking, and impose structure on otherwise ambiguous signals from an organization’s environment. Thus, they define and prioritize problems organizations attend to, influence their interpretation, and delimit the range of options that are considered appropriate.
Over time, an enormous literature, much of which is found in organization sociology (Thornton and Ocasio 2008), has sought to understand how logics are observed in organizations, how they conflict with each other and how they are adapted, manipulated, prioritized, and reconciled by key individuals and organizations (see, for example, Greenwood et al. 2011). The institutional logic approach is useful in public administration as it helps frame the conflicts that all public organizations must reconcile, such as those between the value of efficiency that underpins the market logic and the value of responsiveness that embodies the democracy logic (Rosenbloom 1983; Zhang et al. 2020). Public administration scholarship has investigated combinations of private and public logics in public-private joint ventures (Saz-Carranza and Longo 2012), contradictions between policy and political logics in inter-departmental coordination (Hustedt and Danken 2017), and exploitation of public administration, new public management, and new public government logics in nonprofit public service innovation (Coule and Patmore 2013).
Meso-Level Institutional Logics
Meso-level institutional logics are the assumptions, values, beliefs, and rules predicated on locally-relevant and organization-specific histories, contexts, experiences, norms, and understandings. Meso-level logics are produced and reproduced in the organization (Thornton and Ocasio 2008), and legitimate decisions and actions that are strategically applied to highly localized contexts. Multiple meso-level logics—rule orientation, risk propensity, communication, transparency, etc.—exist simultaneously and can interact with each other in the organization (Noordegraaf et al. 2016). For example, the meso-logic of rule orientation might conflict with the meso-logic of risk propensity, such that the former supports stability, protection of public interest and political responsiveness, while the latter prioritizes innovation and may encourage decentralized decision making and “out-of-the-box” thinking.
We borrow from the management scholarship that posits organizational logics as a “meso-level construct that lies between institutional theory’s field-level logics and the sense-making activities of individual agents in organizations” (Spicer and Sewell 2010, p. 913). Spicer and others (Spicer 2006; Spicer and Sewell 2010) theorize organizational logics as sensemaking frames that provide “a coherent symbolic system articulating what constitutes legitimate, reasonable or effective conduct in, around and by organizations” (Guillén 2001, p. 14). This concept is similar to commonly shared group identity. Others have developed the concept of “logics of action” operating as consistently understood means-ends relationships for group behavior (Bacharach et al. 1996). Because meso-level logics are embedded in organizational structures and cultures, they can influence agency capacity to undertake sensemaking and adapt to climate challenges (Nowell and Stutler 2020).
Individual Agency and Climate Adaptation
The meta-theory of institutional logics further recognizes the dual role of the individual as both an institutionally constrained actor and a potential force for institutional change. On the one hand, the immutability of institutions means that individuals operate within stable and taken for granted norms, beliefs, and expectations (DiMaggio and Powell 1983). On the other hand, this approach also “brings the individual in” as an institutionally aware actor whose role is to reflect on, interpret, maintain, reconcile, destroy, or create institutions (Lawrence et al. 2011). That institutional work involves a sensemaking process during which individuals or groups of individuals attach meanings to external changes and influence subsequent action through strategic action or communication of their interpretations to others (Daft and Weick 1984; Maitlis and Sonenshein 2010). Because logics are “shared beliefs and values” among members of a community (Bévort and Suddaby 2016), the meta-theory makes room for dynamic cognitive processes that interpret, reinterpret, and eventually reify institutional complexity through creating diverse institutional hybrids (Battilana and D’aunno 2009; Blomgren and Waks 2015; Jay 2013). This process can be initiated or accelerated when exogenous shocks expose tensions among logics, calling into question existing and deeply held justifications for actions and behaviors (Maitlis and Sonenshein 2010). In the context of climate adaptation, the individuals would be the institutionally aware top public managers who are justifying decisions and actions in response to climate change.
By linking institutional awareness with the combination and reconciliation of macro- and meso-institutions, we can sharpen understanding and prediction of organizational adaptation to climate change. Our approach reinforces scholarship that underscores the role of individuals for creating heterogenous organizational responses under similar macro institutional environments, but it moves in new directions to embrace a multiple level approach to institutional logics. Figure 1 presents the theoretical framework. It conceptualizes climate adaptation as the joint outcome of an organization’s macro- and meso-institutions under the influence of climate change. The set of macro-institutions includes government authority, professional, and market logics. At the organization level, the study draws on and extends Grothmann and Patt’s (2005) socio-cognitive model of climate adaptation to investigate risk-based and capacity-based meso-logics. In the following sections, we characterize each of the logics, theorize their impact on adaptation, and set out propositions about systematic clustering of the five logics, the clusters’ relationships with adaptation, as well as the differential effects of each logic on adaptation across the clusters.
Characterizing the Specific Macro-Level Institutional Logics
Since our aim is to understand the role of institutional logics for public agency adaptation to climate change, it is first important to identify the macro-logics that are most relevant to the field of public organizations as a whole. Scott considers a field to constitute ‘a collection of diverse, interdependent organizations that participate in a common meaning system’ (Scott 2014, p. 106). Given the intellectual priorities and normative traditions in public administration, we believe that government authority, professionalism, and market efficiency are three prominent logics in the field, and potentially the three most relevant for understanding how institutional logics guide organizational actions and behaviors related to climate change.
Government authority logic emanates from politically constituted hierarchy in which government carries out programs and activities for the public interest (Perry and Rainey 1988). It establishes rules and laws, exercises oversight and sanctions, and controls resource flow to public organizations to ensure legitimacy and compliance (Frumkin and Galaskiewicz 2004). This logic is reified in formal plans, regulations, guidelines, incentive structures, and power relationships (Dovers and Hezri 2010; Naess et al. 2005; Zhang 2021).
Public organizations bear the responsibility for providing protection and security under a worsening climate. Failed or ineffective performance is usually met with political backlash and leads to unwanted consequences, including negative publicity, expanded oversight, diminished authority, reduced funding, and dismissal of key personnel. The Katrina fiasco provides a stark reminder of politicians’ inquiries and scrutiny over organizational mismanagement and ill preparation, culminating in the resignation of FEMA’s director (Boin et al. 2010). As a result, government authority logic can encourage public organizations to undertake strategies and actions that are adaptive to climate change.
On the other hand, government authority logic may preempt decisions and actions necessary for substantive climate adaptation for two main reasons: organizational inertia and constricted power or decision-making autonomy. Macro-institutions are highly resistant to perturbations, and public organizations may be unaware of the extent to which this logic constrains response, or they may be unwilling to challenge it to enable necessary changes without due process (Dovers and Hezri 2010). When subjected to intense authority logic, public organizations tend to react defensively by creating or altering rules or practices to accommodate expectations (Comfort et al. 2019; Zhang et al. 2020). As a prime example, organizations commonly develop “lessons learned” documents to demonstrate accountability after disaster, although the documents are regularly decoupled from policy learning or post-event implementation (Birkland 2009). Beyond inertia, the authority logic can also restrict an organization’s political power, thereby limiting its ability to act independently and commit resources to adaptive behavior. An ideal example comes from New Jersey where the revision of coastal building standards to accommodate growing weather severity required substantial commitment from the governor’s office (Tully 2020). When political influence generates strong pressure for “doing something,” the authority logic can be maladaptive, eliciting quick fixes that preclude proactive initiatives designed to reframe agency approach (Birkland 1997).
We therefore expect government authority logic to either stimulate or hinder adaptation contingent on specific contexts. In some organizations, top managers may act on this logic through adjusting its value and priority to legitimize and enable adaptation, while in others this will not be the case.
Professional logic is predicated on professional values, norms and standards, and advances a forward-thinking mindset toward climate action (Henstra 2012; Hovik et al. 2015). In public organizations, it is fundamentally anchored in expertise, rationality, ethical conduct, and impartiality (Skelcher and Smith 2015). This logic is generated and maintained by national associations, certified through colleges and universities, and learned through training and on the job experience (DiMaggio and Powell 1983).
Some professional associations recognize the systematic threats from climate change and subscribe to resilience-based thinking by incorporating long-term consequences of climate hazards in planning and design. Professional societies such as the American Society of Civil Engineers (ASCE), American Society for Adaptation Professionals, and American Association of State Highway and Transportation Officials are instrumental in forming and diffusing knowledge, practices, and standards about climate adaptation and resilience. Among others, the ASCE created the Committee on Adaptation to a Changing Climate in 2011 to “evaluate the technical requirements and civil engineering challenges for adapting to a changing climate” and institute appropriate standards and criteria for critical infrastructure protection (ASCE n.d.).
Professional logic transfers to public organizations as they employ individuals in multiple professions—management, engineering, urban planning, trades, etc.—who have received certified training and are members of professional associations (Blomgren and Waks 2015). It is further transmitted through overlapping sequences of workshops, conferences, and seminars that create a common frame of reference and promote professional familiarity with and acceptance of adaptation as an emerging paradigm (Noordegraaf 2011). Communication through written media (e.g., annual report, newsletters, and journals) or unwritten media (e.g., interpersonal exchange) additionally contributes a continuous flow to the knowledge base (DiMaggio and Powell 1983; Suddaby and Viale 2011). Empirical evidence shows that professional associations are conducive to problem awareness (Henstra 2017; Hovik et al. 2015) and pivotal in diffusing climate adaptation practices and facilitating adaptation implementation of member agencies (Zhang and Maroulis 2021). As a result, organizations that integrate and prioritize professional logic can create favorable institutional conditions for adaptation.
Market logic, which permeates the private sector, took hold in the public sector in the 1990s as part of the general movement legitimizing private sector efficiencies and customer orientation (Osborne and Gaebler 1992). Newer than the other two macro logics in the public sector, market logic values competition and efficiency, and stresses short-term and tangible outcomes. Legitimacy is derived from implementation of market-based instruments, practices, and structures, such as contracting to the private sector, partnering with private firms, expanding user charges, and prioritizing individual preferences.
This macro-level logic can influence strategic response to climate change. For example, the practice of service contracting can legitimize private sector mechanisms for incorporating risks into costs and decision making that are counterproductive to organizational resilience. These mechanisms are visible through insurance premiums (Otto-Banaszak et al. 2011) or contractual provisions exempting contractors from performance under extreme events due to “impossibility, impracticality and frustration of purposes” (Egan 2010, p. 48). Failure to honor contractual agreements can occur during emergency response (Egan 2010), as evidenced by FEMA contractor’s failure to deliver millions of meals to Puerto Rico amid Hurricane Maria (Heavey 2018). By reducing the liability of either the focal public organization or its contractors, market logic can disincentivize both parties from undertaking adaptation.
Additionally, placing a high value on efficiency has led to institutionalization of tightly coupled systems and removal of system slack such that extreme weather events can more easily cause cascading risks and catastrophic failures (Perrow 1999; Stark 2014). As a costly undertaking that can hurt short-term efficiency (Zhang 2021), adaptation is usually shunned in market-oriented organizations due to their short-termism and uncertainty avoidance (Slawinski et al. 2017). For instance, the new public management reform in the Norwegian electricity sector fundamentally undermined its climate adaptation through privileging economic efficiency, creating responsibility gaps, causing goal replacement, and eroding safety culture (Inderberg 2011). Extensive involvement of private firms in the production and delivery of public services also runs the risks of hollowing out public organizations’ capacity to make adaptive change, as woefully evident at the outset of the COVID-19 pandemic (Balz 2020). Consequently, organizations that prioritize market logic as a basis for decisions and actions would build in structures and processes that deter or delay climate adaptation.
Characterizing the Specific Meso-Level Institutional Logics
Risk-based logic is defined as an organization-wide understanding, held by top managers, about the probability of a hazard and its consequences if the hazard materializes (Gould 2021). A strong risk-based logic entails deep thinking or reflection on the underlying assumptions, values, and beliefs about an organization’s risk environment, usually serving as the necessary trigger for adaptive changes deviating from routine understanding and institutional expectations.
Public organizations typically exhibit low risk-based logic regarding climate risks. They have long operated in a relatively stable atmospheric climate, with a bureaucratic structure and design better-suited for day-to-day routines (Nowell and Stutler 2020). The climatic stability not only fosters complacency and risk denial toward climate change (O’Brien et al. 2006), but also results in scarcity of a priori experience with climate hazards, especially extreme ones, to facilitate learning. Recognition of potential climatic risks is simultaneously undercut by the lack of monitoring systems and inadequate organizational slack to support effective analysis, reflection, and sensemaking. Meanwhile, bureaucratic hierarchies and rigid chain of command confine members to their “in-the-box” thinking and discourage open exchanges that can collectively reveal risks and organizational vulnerabilities (Nowell and Stutler 2020). The habitual treatment of extreme weather events as one-off isolated episodes further blinds organizations to the systematic nature of risk generation, made worse by the structural designation of emergency management units to tackle climate risks that denies opportunities for organization-wide sensemaking and problem perception (Hughes 2015; Zhang 2020). The resulting organizational complacency underlines the role of risk-based logic in channeling climate signals to adaptive behavior.
Greater prioritization and integration of risk-based logic can be activated by environmental impacts that disrupt ongoing activities and cause a sufficiently large “discrepancy between what one expects and what one experiences” (Maitlis and Christianson 2014, p. 70). When existing logics no longer provide sufficient guidance, top managers mobilize new logics or recombine existing ones to meet emerging demands from deteriorating climate conditions (Battilana and D’aunno 2009; Blomgren and Waks 2015). Accordingly, when top managers prioritize risk-based logic to legitimize decisions and actions in response to climate change, organizations will be more likely to demonstrate adaptive behavior.
Capacity-based logic is defined as the beliefs and expectations in the attributes and ability of the organization to undertake adaptive measures that meet climate challenges. Building on confidence and learned self-assurance, the logic supports and justifies top managers’ commitment to non-incremental and non-routine adaptation. It overcomes structural and budgetary inertia, and counteracts the tendency of organizations to stay satisfied with the status quo (Grothmann and Patt 2005).
Conditions common to public organizations often do not support the development, inclusion, and prioritization of meso-level capacity-logic. Public organizations receive resource allocations that are based on their business-as-usual routines and proportional to the scale of their services and goods provisions. They are typically financially strapped yet operate under mounting pressure to deliver more with less. Over the past few decades, bureaucratic values toward rationality and efficiency have removed much of the slack in public organization systems (Stark 2014), depleting organizations of the necessary capacity reserve for responding to or planning for emerging climate risks. Even though organizations can receive additional resources in the aftermath of a weather disaster, the funding resources are usually tightly restricted to repair and recovery functions while granting little discretion for resilience-enhancing projects (Zhang and Maroulis 2021).
Resource constraints, in addition to pressure for performance, have led many public organizations to prioritize goals and tasks on which they can deliver immediate and visible outcomes, leaving many public organizations underequipped for a worsening climate. Barriers related to inadequate leadership support, competing agendas, turf battles, short-term priorities, under-trained staff, as well as mismatch between climate science and organizational needs all exacerbate the capacity deficit (Biesbroek et al. 2013; Moser et al. 2019; Naess et al. 2005). For instance, financial constraints have been persistently identified as a major limiting factor in public organization adaptation to climate change (Biesbroek et al. 2013; Miao et al. 2018a). Moser and Luers (2008) note that coastal management agencies in California are restricted by analytical and technical capacity, funding shortages and political power struggles in their preparation for climate adaptation.
Nevertheless, drawing on their unique position and prior experience deploying resources or creativity to work around resource constraints, some organizations are able to legitimate adaptive decisions and actions based on a confident and positive understanding of organizational capacity (Battilana and D’aunno 2009). The “can do” institutional logic endows the organization’s top management with a sense of empowerment (Zhang and Welch 2021) and motivates them to exploit diverse existing institutions and resources to create propitious conditions for adaptation.
Configurations of Macro- and Meso-Level Institutional Logics
The study takes a configurational approach to investigate the multiplicity of logics in organizations. A configuration denotes “any multidimensional constellation of conceptually distinct characteristics that commonly occur together” (Meyer et al. 1993, p. 1175). A configurational inquiry allows researchers to understand and explain how orders emerge from the co-existence and complex interactions of those parts as a whole. The number of possible combinations is bounded (i.e., limited diversity) by the interdependence of the attributes that often can change only discretely or intermittently (Meyer et al. 1993; Ragin 2009). Scholars suggest that many forces cause the attributes to fall into coherent patterns, such as the macro-level homogenizing institutions and meso-level organization processes and outcomes (DiMaggio and Powell 1983; Spicer and Sewell 2010).
We apply the configuration approach to understand organizational adaptation to climate change. The three macro-level logics—government authority, professionalism, and market—structure the operating environment of an organization, shaping its resource base, scope of authorities, strategic priorities, and range of options in direct or indirect ways (Scott 2014). Meso-level logics—risk-based and capacity-based—are produced and reproduced within an organization. On the one hand, macro-level institutions can influence meso-level logics. For instance, strong professional logic can attune an organization to climate anomalies and raise meso-level risk-based logic. On the other hand, however, meso-level logics also emanate from localized values, beliefs, rules, and practices that may or may not align with the macro realities the organization faces. The relationship between macro- and meso-level logics suggests the potential for a diversity of co-existing patterns. Drawing on the configurational approach (Meyer et al. 1993), institutional logics from both the macro- and meso-levels can systematically converge yielding distinct patterns of co-existence.
Proposition 1: Public organizations will cluster into groups based on recognizable combinations of macro- and meso-institutional logics.
As noted above, the five institutional logics influence adaptation in distinct ways. At the macro level, government authority logic can stimulate adaptation by elevating organizational sensitivity to climate change impacts and heightening the urgency for adaptive response. Alternatively, restrained autonomy and associated institutional inertia might unconsciously incentivize maladaptive fixes to demonstrate political responsiveness. Professional logic primes public organizations to climate risks, while institutionalizing standards and practices to guide legitimate behavior (DiMaggio and Powell 1983). Market logic gravitates a public organization toward efficiency and short-term tangible benefits, creating structural bias against purposeful adaptation due to its upfront costs, interruption to organizational routines, and ambiguous payoffs (Slawinski et al. 2017; Wilson et al. 2010; Zhang 2021). At the meso-level, risk-based logic facilitates motivation for and commitment to adaptation, while capacity-based logic fosters the necessary efficacy for engaging in non-incremental adaptation activities.
The institutional logic meta-theory tells us that public managers, as institutionally aware actors, are able to perceive and proactively manipulate, prioritize, reconcile, and integrate multiple logics in response to the external environment. In cases when co-existing logics offer contradictory rationales for action, the ways in which public managers interpret the institutional arrangements and influence the logics configuration will differentially determine adaptation. Because public administrators face varied combinations of macro- and meso-level institutional logics, we are able to investigate whether certain combinations are more or less adaptable to climate change.
Proposition 2: Some combinations of institutional logics will be more strongly associated with adaptation to climate change.
According to a configurational perspective, organizational actions take meaning from the amalgam of organizational attributes as a whole and cannot be understood in isolation (Meyer et al. 1993). Understanding how individuals navigate the multiplicity of institutional logics requires taking an integrative account of the configurational patterns of macro- and meso-institutions.
Institutions at the same level are likely to contradict each other, creating tensions and sometimes hindrance to adaptation. At the macro-level, scholars increasingly recognize and investigate the inconsistencies or conflicts among multiple institutional forces (Besharov and Smith 2014; Greenwood et al. 2011). By imposing incongruent objectives, norms, and performance metrics, the institutions place competing demands on an organization’s attention, strategic choices, and resource allocation (Thornton et al. 2012). In our current case, strains among government, market, and professional logics can pull organizations into different or opposite directions. How organizations simultaneously attend to the three logics can have direct bearing on their adaptation decisions and planning. Similarly, tension can emerge from incompatible meso-institutions. Organizations highly cognizant of the risks associated with climate change might falter at adaptation when low capacity-based logic is present. Conversely, strong capacity-based logic can be met with low risk-based logic, leading to complacency and missed opportunities for adaptation.
Tensions can also arise across institutional levels. Notwithstanding their coexistence, macro- and meso-institutions likely do not have equivalent influence on the organization and some may carry much greater relative weight than others. The relative priority of logics can create incongruence and contradictions that affect adaptation. For example, while meso-level institutional logics could justify authentic motivation to adapt, formidable resistance or animosity can come from macro logics that undercut adaptation planning and implementation. To illustrate, in their study on 135 critical infrastructure organizations, Wilson and colleagues (2010) find the ingrained market logic led preexisting dominant coalitions to control the organizational structure and block adaptation plans from inclusion in the strategic agenda. The same holds true for organizations that are prodded by macro-institutions to adapt but are nevertheless held back by low risk-based or capacity-based logics.
The tensions existing among the set of institutional logics within and across levels suggest that the same institutional logic can have varying effects on adaptation depending on the presence or absence of other logics in a given configuration.
Proposition 3: The effect of any particular logic on climate adaptation will depend on the combination of macro- and meso-institutional logics operating in the organization.
Data and Measurement
We focus the empirical analysis on extreme weather events as the manifestation of climate change and the primary stimuli of climate adaptation (Berrang-Ford et al. 2011). Public transit agencies under extreme weather provide a rich context to test the propositions. The agencies are already experiencing costly impacts from extreme weather, leading to damaged roads and vehicles, congested traffic, service disruptions, or system breakdown. Those challenges will escalate as extreme weather grows in frequency, magnitude, and scope. Agency performance under extreme weather also has broad social ramifications, particularly for socially vulnerable populations.
The study uses data from a 2019 national survey on the largest transit agencies in the United States, matched with agency profiles obtained from the National Transit Database (NTD), disaster declaration data from the Federal Emergency Management Agency (FEMA), and 2014–2018 demographic data taken from the US Census Bureau. The study’s target population includes all major US metropolitan fixed-route public transit agencies with an annual fare revenue of at least one million dollars in 2013 (N = 312). The survey includes lead managers from five departments: operations, maintenance, service planning, strategic planning, and engineering. To collect the names and contact information of the managers, the research team adopted a standard protocol that includes three primary methods: online searches on the agencies’ webpages, telephone calls to the agencies, as well as FOIA requests. Because some agencies refused to participate or were unreachable after repeated attempts and not every agency has all the five departments, the final sample frame includes 1,039 respondents from 291 agencies. The project was reviewed and approved by the Institutional Research Board at Arizona State University (STUDY #00003589).
The survey was administered online from April 9 to June 3, 2019. We conducted a pretest with seven managers across six agencies to validate the instrument before administering the survey to the entire sample. After removing ineligible, retired, and non-contactable individuals, the final adjusted sample includes 853 individuals from 278 agencies. A total of 313 individuals from 194 agencies completed the survey, yielding an individual-level response rate at 36.7% and agency-level response rate at 70.0%. A non-response bias test shows that the responding and non-responding agencies are comparable regarding geographical locations, organizational size (measured by total revenue, annual service provisions, and number of employees) as well as service area characteristics (measured by area size and population).
Applying the configurational approach, we treat each agency as the unit of analysis. We aggregated the individual-level measures to the organizational level. The aggregation is done after we have confirmed the measurement model in the structural equation model (SEM) and extracted the factor score (to be detailed shortly). The agency-level data include 193 observations: one agency was dropped because it was merged with another agency and the NTD stopped updating its operational data in 2017.
Table 1 presents measures for all key and control variables. The items for the adaptation variable encompass typical techno-engineering, administrative, and fiscal measures applied in climate adaptation (Jones et al. 2012). Prior research on institutional logics is mostly interpretive and based on specific cases, relying on field-studies, ethnography, or interviews to identify logics in organizational settings (e.g., Bévort and Suddaby 2016; Corbett et al. 2018; Jay 2013; Waring and Currie 2009). In this study, we see the utility of measuring institutional logics by using responses to closed-ended survey questions and institutional data to tap into top managers’ perceptions, interpretations, and experiences. With regard to government authority logic, Frumkin and Galaskiewicz (2004) offer a direct measure by asking respondents about the extent to which their agency is subject to coercive influence from governments (i.e., government review or license). We adapted their measure by asking managers about the influence of local, state, and federal governments on their organization. We followed Burruss and Giblin (2014) to measure professional logic as the managers’ responses to questions on the prevalence of professionalization (e.g., professionalization and listserv) and publications (e.g., industry standards and journal articles) in organizational decision making. Market logic draws on the NTD database and is measured as the proportion of contracted hours that transit vehicles travel for revenue generation. This measure reflects the degree of penetration of business models in an agency’s everyday operation routine, providing an appropriate proxy for the strength of market logic.
Variable name . | Measurement . | Source . |
---|---|---|
Adaptation | Survey items asking managers whether their agency did the following in the past two years to address extreme weather (Scale: 1 = Yes, 0 = No): | 2019 Transit survey |
1) Invested in new weather-smart equipment and technologies; | ||
2) Adopted stricter construction and engineering standards to address extreme weather; | ||
3) Abandoned or relocated transit infrastructure due to extreme weather impacts | ||
4) Set aside new funds dedicated to extreme weather; | ||
5) Submitted a grant application for projects to minimize weather impacts; | ||
6) Purchased additional insurance specifically for extreme weather events. | ||
Government authority logic | Survey items asking managers about the level of influence the following actors exert over their agency’s decision making (Scale 1–5: 1 = no influence, 5 = very strong influence): | 2019 Transit survey |
1) Elected or appointed local officials (Mayor, Mayor Council, etc.); 2) City planning commission; 3) City government agencies; | ||
4) State Department of Transportation; 5) Other State agencies; 6) Regional transit authority or council; 7) metropolitan planning organizations; 8) Federal agencies | ||
Professional logic | Survey items asking managers the extent to which their agency relies on sources below to increase its ability to manage extreme weather risks (Scale 1–5: 1 = not at all, 5 = very high extent): | 2019 Transit survey |
1) Professional mailing lists or newsletters; 2) Publicly available datasets; 3) Professional associations (e.g., APTA, TRB, AASHTO[1]); 4) Industry standards (e.g., engineering standards); 5) Publications in academic journals | ||
Market logic | The ratio of vehicle revenue hours an agency contracted to other organizations (over 90% private firms) in 2017. | 2017 NTD database |
Risk-based logic | Survey items asking about managers’ level of agreement (Scale 1–5: 1 = Strongly disagree, 5 = Strongly agree): | 2019 Transit survey |
1) My agency is increasingly concerned about the impact of extreme weather on our transit infrastructure; | ||
2) My agency is increasingly concerned about the impact of extreme weather on our transit operations; | ||
3) Most people in my agency recognize that extreme weather events are becoming more frequent. | ||
Capacity-based logic | Survey items asking about managers’ level of agreement (scale 1–5: 1 = Strongly disagree, 5 = Strongly agree): | 2019 Transit survey |
1) My agency has sufficient internal expertise to respond to future extreme weather challenges; | ||
2) My agency has the financial capacity to address future extreme weather challenges; | ||
3) My agency has the operational capacity to address future extreme weather challenges; | ||
4) My agency has the authority to address future extreme weather challenges | ||
Control variables | ||
Disaster experience | The total number of weather disasters that occurred to an agency’s county from 2014 to 2018 | FEMA |
Organization size | The natural log of total fund an agency received from all sources in 2017 | 2017 NTD database |
Special purpose authority | Dummy coded 1 if the agency is a special-purpose government and 0 if affiliated with a city or state government. | 2017 NTD database |
Director | Dummy coded 1 if one or more members of an agency’s leadership is politically appointed. | 2019 Transit survey |
Bus only | Dummy coded 1 if an agency provides only bus services exclusive of rail services. | 2017 NTD database |
Density | The natural log of the total population normalized by square miles of the service area | 2014-2018 US Census |
Median income | The natural log of median household income in the agency’s county | 2014-2018 US Census |
Commute time | The average number of minutes for commute | 2014-2018 US Census |
Democratic | Dummy coded 1 if the agency operates in a state that voted majority democratic and 0 otherwise in the 2016 general election. | New York Times |
Variable name . | Measurement . | Source . |
---|---|---|
Adaptation | Survey items asking managers whether their agency did the following in the past two years to address extreme weather (Scale: 1 = Yes, 0 = No): | 2019 Transit survey |
1) Invested in new weather-smart equipment and technologies; | ||
2) Adopted stricter construction and engineering standards to address extreme weather; | ||
3) Abandoned or relocated transit infrastructure due to extreme weather impacts | ||
4) Set aside new funds dedicated to extreme weather; | ||
5) Submitted a grant application for projects to minimize weather impacts; | ||
6) Purchased additional insurance specifically for extreme weather events. | ||
Government authority logic | Survey items asking managers about the level of influence the following actors exert over their agency’s decision making (Scale 1–5: 1 = no influence, 5 = very strong influence): | 2019 Transit survey |
1) Elected or appointed local officials (Mayor, Mayor Council, etc.); 2) City planning commission; 3) City government agencies; | ||
4) State Department of Transportation; 5) Other State agencies; 6) Regional transit authority or council; 7) metropolitan planning organizations; 8) Federal agencies | ||
Professional logic | Survey items asking managers the extent to which their agency relies on sources below to increase its ability to manage extreme weather risks (Scale 1–5: 1 = not at all, 5 = very high extent): | 2019 Transit survey |
1) Professional mailing lists or newsletters; 2) Publicly available datasets; 3) Professional associations (e.g., APTA, TRB, AASHTO[1]); 4) Industry standards (e.g., engineering standards); 5) Publications in academic journals | ||
Market logic | The ratio of vehicle revenue hours an agency contracted to other organizations (over 90% private firms) in 2017. | 2017 NTD database |
Risk-based logic | Survey items asking about managers’ level of agreement (Scale 1–5: 1 = Strongly disagree, 5 = Strongly agree): | 2019 Transit survey |
1) My agency is increasingly concerned about the impact of extreme weather on our transit infrastructure; | ||
2) My agency is increasingly concerned about the impact of extreme weather on our transit operations; | ||
3) Most people in my agency recognize that extreme weather events are becoming more frequent. | ||
Capacity-based logic | Survey items asking about managers’ level of agreement (scale 1–5: 1 = Strongly disagree, 5 = Strongly agree): | 2019 Transit survey |
1) My agency has sufficient internal expertise to respond to future extreme weather challenges; | ||
2) My agency has the financial capacity to address future extreme weather challenges; | ||
3) My agency has the operational capacity to address future extreme weather challenges; | ||
4) My agency has the authority to address future extreme weather challenges | ||
Control variables | ||
Disaster experience | The total number of weather disasters that occurred to an agency’s county from 2014 to 2018 | FEMA |
Organization size | The natural log of total fund an agency received from all sources in 2017 | 2017 NTD database |
Special purpose authority | Dummy coded 1 if the agency is a special-purpose government and 0 if affiliated with a city or state government. | 2017 NTD database |
Director | Dummy coded 1 if one or more members of an agency’s leadership is politically appointed. | 2019 Transit survey |
Bus only | Dummy coded 1 if an agency provides only bus services exclusive of rail services. | 2017 NTD database |
Density | The natural log of the total population normalized by square miles of the service area | 2014-2018 US Census |
Median income | The natural log of median household income in the agency’s county | 2014-2018 US Census |
Commute time | The average number of minutes for commute | 2014-2018 US Census |
Democratic | Dummy coded 1 if the agency operates in a state that voted majority democratic and 0 otherwise in the 2016 general election. | New York Times |
Variable name . | Measurement . | Source . |
---|---|---|
Adaptation | Survey items asking managers whether their agency did the following in the past two years to address extreme weather (Scale: 1 = Yes, 0 = No): | 2019 Transit survey |
1) Invested in new weather-smart equipment and technologies; | ||
2) Adopted stricter construction and engineering standards to address extreme weather; | ||
3) Abandoned or relocated transit infrastructure due to extreme weather impacts | ||
4) Set aside new funds dedicated to extreme weather; | ||
5) Submitted a grant application for projects to minimize weather impacts; | ||
6) Purchased additional insurance specifically for extreme weather events. | ||
Government authority logic | Survey items asking managers about the level of influence the following actors exert over their agency’s decision making (Scale 1–5: 1 = no influence, 5 = very strong influence): | 2019 Transit survey |
1) Elected or appointed local officials (Mayor, Mayor Council, etc.); 2) City planning commission; 3) City government agencies; | ||
4) State Department of Transportation; 5) Other State agencies; 6) Regional transit authority or council; 7) metropolitan planning organizations; 8) Federal agencies | ||
Professional logic | Survey items asking managers the extent to which their agency relies on sources below to increase its ability to manage extreme weather risks (Scale 1–5: 1 = not at all, 5 = very high extent): | 2019 Transit survey |
1) Professional mailing lists or newsletters; 2) Publicly available datasets; 3) Professional associations (e.g., APTA, TRB, AASHTO[1]); 4) Industry standards (e.g., engineering standards); 5) Publications in academic journals | ||
Market logic | The ratio of vehicle revenue hours an agency contracted to other organizations (over 90% private firms) in 2017. | 2017 NTD database |
Risk-based logic | Survey items asking about managers’ level of agreement (Scale 1–5: 1 = Strongly disagree, 5 = Strongly agree): | 2019 Transit survey |
1) My agency is increasingly concerned about the impact of extreme weather on our transit infrastructure; | ||
2) My agency is increasingly concerned about the impact of extreme weather on our transit operations; | ||
3) Most people in my agency recognize that extreme weather events are becoming more frequent. | ||
Capacity-based logic | Survey items asking about managers’ level of agreement (scale 1–5: 1 = Strongly disagree, 5 = Strongly agree): | 2019 Transit survey |
1) My agency has sufficient internal expertise to respond to future extreme weather challenges; | ||
2) My agency has the financial capacity to address future extreme weather challenges; | ||
3) My agency has the operational capacity to address future extreme weather challenges; | ||
4) My agency has the authority to address future extreme weather challenges | ||
Control variables | ||
Disaster experience | The total number of weather disasters that occurred to an agency’s county from 2014 to 2018 | FEMA |
Organization size | The natural log of total fund an agency received from all sources in 2017 | 2017 NTD database |
Special purpose authority | Dummy coded 1 if the agency is a special-purpose government and 0 if affiliated with a city or state government. | 2017 NTD database |
Director | Dummy coded 1 if one or more members of an agency’s leadership is politically appointed. | 2019 Transit survey |
Bus only | Dummy coded 1 if an agency provides only bus services exclusive of rail services. | 2017 NTD database |
Density | The natural log of the total population normalized by square miles of the service area | 2014-2018 US Census |
Median income | The natural log of median household income in the agency’s county | 2014-2018 US Census |
Commute time | The average number of minutes for commute | 2014-2018 US Census |
Democratic | Dummy coded 1 if the agency operates in a state that voted majority democratic and 0 otherwise in the 2016 general election. | New York Times |
Variable name . | Measurement . | Source . |
---|---|---|
Adaptation | Survey items asking managers whether their agency did the following in the past two years to address extreme weather (Scale: 1 = Yes, 0 = No): | 2019 Transit survey |
1) Invested in new weather-smart equipment and technologies; | ||
2) Adopted stricter construction and engineering standards to address extreme weather; | ||
3) Abandoned or relocated transit infrastructure due to extreme weather impacts | ||
4) Set aside new funds dedicated to extreme weather; | ||
5) Submitted a grant application for projects to minimize weather impacts; | ||
6) Purchased additional insurance specifically for extreme weather events. | ||
Government authority logic | Survey items asking managers about the level of influence the following actors exert over their agency’s decision making (Scale 1–5: 1 = no influence, 5 = very strong influence): | 2019 Transit survey |
1) Elected or appointed local officials (Mayor, Mayor Council, etc.); 2) City planning commission; 3) City government agencies; | ||
4) State Department of Transportation; 5) Other State agencies; 6) Regional transit authority or council; 7) metropolitan planning organizations; 8) Federal agencies | ||
Professional logic | Survey items asking managers the extent to which their agency relies on sources below to increase its ability to manage extreme weather risks (Scale 1–5: 1 = not at all, 5 = very high extent): | 2019 Transit survey |
1) Professional mailing lists or newsletters; 2) Publicly available datasets; 3) Professional associations (e.g., APTA, TRB, AASHTO[1]); 4) Industry standards (e.g., engineering standards); 5) Publications in academic journals | ||
Market logic | The ratio of vehicle revenue hours an agency contracted to other organizations (over 90% private firms) in 2017. | 2017 NTD database |
Risk-based logic | Survey items asking about managers’ level of agreement (Scale 1–5: 1 = Strongly disagree, 5 = Strongly agree): | 2019 Transit survey |
1) My agency is increasingly concerned about the impact of extreme weather on our transit infrastructure; | ||
2) My agency is increasingly concerned about the impact of extreme weather on our transit operations; | ||
3) Most people in my agency recognize that extreme weather events are becoming more frequent. | ||
Capacity-based logic | Survey items asking about managers’ level of agreement (scale 1–5: 1 = Strongly disagree, 5 = Strongly agree): | 2019 Transit survey |
1) My agency has sufficient internal expertise to respond to future extreme weather challenges; | ||
2) My agency has the financial capacity to address future extreme weather challenges; | ||
3) My agency has the operational capacity to address future extreme weather challenges; | ||
4) My agency has the authority to address future extreme weather challenges | ||
Control variables | ||
Disaster experience | The total number of weather disasters that occurred to an agency’s county from 2014 to 2018 | FEMA |
Organization size | The natural log of total fund an agency received from all sources in 2017 | 2017 NTD database |
Special purpose authority | Dummy coded 1 if the agency is a special-purpose government and 0 if affiliated with a city or state government. | 2017 NTD database |
Director | Dummy coded 1 if one or more members of an agency’s leadership is politically appointed. | 2019 Transit survey |
Bus only | Dummy coded 1 if an agency provides only bus services exclusive of rail services. | 2017 NTD database |
Density | The natural log of the total population normalized by square miles of the service area | 2014-2018 US Census |
Median income | The natural log of median household income in the agency’s county | 2014-2018 US Census |
Commute time | The average number of minutes for commute | 2014-2018 US Census |
Democratic | Dummy coded 1 if the agency operates in a state that voted majority democratic and 0 otherwise in the 2016 general election. | New York Times |
We used top managers’ responses to survey questions to measure meso-level organizational logics, because managerial views are reflective of organization-level logics operating at any point (Daft and Weick 1984; Hambrick 2007). The measures for risk-based logic describe the managers’ understanding of the probability of extreme weather and its consequences if materialized. Measures for capacity-based logic capture top managers’ perceived organizational capacity for meeting the manifold technical, fiscal, and operational demands for conducting adaptation.
Methods and Analysis
Figure 2 presents the steps involved in the analysis, each producing an output that is entered as an input in the subsequent step. First, we conducted confirmatory factor analysis on the individual-level data—the measurement part of SEM—to verify the relationships between the latent variables and their indicators. Second, after achieving the measurement model fit, we extracted factor scores for the latent variables and aggregated all individual-level scores to obtain their organizational means. Third, we applied a latent profile analysis (LPA) to identify clusters of organizations based on the combination of the five logics. Finally, we conducted a multigroup SEM to test if the effects of each logic are significantly different across the clusters. Appendix A addresses potential common method bias involved in the analysis.
SEM Measurement Model
We included all the latent variables and estimated the measurement model in Mplus. The categorical indicators for adaptation led to the use of weighted least square means and variance adjusted (WLSMV) to extract the underlying continuous and normal factors. Specifically, the analysis identifies two factors among the eight items measuring government authority logic: local government influence consists of the first three items and state or federal government influence includes the latter five. Government authority logic is therefore constructed from two lower-order factors including city government influence and state or federal government influence.
Appendix B reports the standardized factor loading for each factor. The fit statistics demonstrate satisfactory model fit (RMESA = 0.024 with 90% CI [0.011,0.034], CFI = 0.944, TLI = 0.936, WRMR = 0.076). All the standardized factor loadings exceed the minimum threshold of 0.40 and are also bigger than twice their standard errors. A series of pairwise Wald tests assessing the correlations among the latent factors are all statistically significant, showing support for discriminatory reliability.
We then obtained factor scores of the latent variables standardized at a mean of zero. Appendix C shows the summary statistics for all variables at the organizational level. A higher factor score indicates a stronger unobserved characteristic or attribute. By way of illustration, a high score on adaptation suggests that the organization possesses strong latent attributes, in terms of capacity, autonomy, incentives or commitment, that permit it to enact adaption measures.
Latent Profile Analysis
The analysis takes a case-oriented view (Ragin 2009) by treating organizations as complex cases that embody combinations of the macro- and meso-level logics. The effect of each logic is to be understood from the whole and cannot be isolated to study its net effect (Meyer et al. 1993; Ragin 2009). The purpose of the analysis rules out regression techniques which are oriented toward individual variables (instead of cases) and often seek to understand the net effect of a given variable through partial correlation.
Instead, latent profile analysis (LPA) provides a fitting analytical technique to identify clustering of continuous profile indicators.1 The use of LPA to identify categorical groups is gaining popularity in multiple fields in recent years, including psychology, sociology, education, and management (e.g., Daimon and Atsumi 2018; Pastor et al. 2007; Pavlova and Silbereisen 2015; Spurk et al. 2020).
LPA is a type of latent variable mixture modeling to discover various subpopulations in an overall population, each with a distinct probability distribution (Morin et al. 2016). Compared with the commonly used cluster analysis, it is a model-based clustering method using a probabilistic model to describe the data distribution. The application of statistical theory and modeling allows LPA to provide goodness of fit statistics and guidelines for determining the number of clusters, thereby overcoming the subjectivity weaknesses associated with cluster analysis (Spurk et al. 2020).
Most of our profile indicators are latent variables. Full latent mixture models that incorporate indices of items to first estimate latent profile indicators and then extract profiles are few. A frequent alternative is to use the factor scores saved from preliminary measurement models as the profile indicators (Morin et al. 2016).
Configurations of Macro- and Meso-Level Institutional Logics
Figure 3 shows a graphical representation of the latent profile model wherein the squares denote observed variables and ovals denote latent variables. Following Spurk (2020), we used the maximum likelihood estimator with robust standard errors (MLR) to correct for non-normality in some profile indicators. Previous studies favor the Bayesian Information Criterion (BIC), sampled-adjusted BIC (SABIC), bootstrap likelihood ratio test (BLRT), and adjusted Lo-Mendell-Rubin (LMR) to choose the best fitting LPA model (Spurk et al. 2020). A smaller value on BIC and SABIC suggest a better fitting model, while a significant BLRT or adjusted LMR value means the model with one less cluster fits the data better. An entropy of 0.8 suggests acceptable classification quality (Celeux and Soromenho 1996). Appendix D presents the model statistics across LPA solutions with two to four profiles, all estimated with standardized indicators. The output for each model returns no error message and demonstrates successful replication of the loglikelihood value, suggesting no sign of local maxima.
All the solutions satisfy the 0.8 entropy threshold and all fit indices point to a four-profile solution. Nevertheless, since one of the four profiles has only 23 cases and this profile does not provide theoretically meaningful insight, we settled for three profiles due to parsimony and meaningfulness considerations (Lubke and Neale 2006). The systematic clustering of the five logics shows support for Proposition 1.
Validation of Configurations
One way to validate LPA profiles is to test mean differences across the clusters in relation to theoretically relevant outcomes (Spurk et al. 2020). We thus compare the mean differences in adaptation across the three clusters. Figure 4 graphs the spread of adaptation across the clusters, with the horizontal line in the boxplot indicating the cluster mean. The levels of adaptation vary significantly across the three clusters. Specifically, agencies in cluster 1 score noticeably higher on adaptation than in the other two clusters (p = .000), whereas those in cluster 2 and cluster 3 do not significantly differ in their levels of adaptation (p = .202). The significant differences of adaptation across the three clusters support Proposition 2.
To better understand the clustered patterns, figure 5 displays the estimated means of each cluster on the profile indicators. With the exception of capacity-based logic, all indicators manifest marked differences across clusters. Agencies in cluster 1 register consistently high on government authority logic, professional logic, and risk-based logic, and low on market logic. Almost in stark contrast, agencies in cluster 2 score low on government authority logic, professional logic, and risk-based logic. One salient feature for agencies in cluster 3 is the dominance of market logic, with all others hovering around the medium levels.
Our theory suggests that adaptation is positively associated with professional logic, risk-based logic, and capacity-based logic and negatively associated with market logic, while its relationship with government authority logic can work in both directions. When viewed together, the adaptation outcome (figure 4) and the combination patterns of the logics (figure 5) are in line with our theory and provide validation for the LPA solution: cluster 1 scores high on logics we expect to promote adaptation, whereas cluster 2 shows lower levels of adaptation due to the absence of those conditions and cluster 3 is dominated by market logic. We label the three clusters as forerunner, complacent, and market-oriented, respectively.
Differential Effects of Institutional Logics Across Configurations
We performed a multigroup SEM to investigate the variance of structural paths across the three clusters. Multigroup SEM estimates separate models in two or more concrete groups and conducts likelihood ratio comparisons to test equality of structural coefficients across groups (Hoyle 2012).
The results are reported in table 2. All the continuous variables are standardized for comparison. Since both the institutional logics and the dependent variables are continuous variables, interpretations of the coefficients follow those typical in multiple regressions. Take the example of professional logic. For each one standard deviation increase in professional logic, organizational adaptation rises by 0.397, 0.274, and 0.592 standard deviations in the forerunner, complacent, and market-oriented cluster, respectively.
. | Forerunner . | Complacent . | Market-oriented . |
---|---|---|---|
Government authority logic | −0.102 (0.109){0.350} | −0.011 (0.067){0.871} | −0.351 (0.109){0.001} |
Professionalism logic | 0.397 (0.070){0.000} | 0.274 (0.065){0.000} | 0.592 (0.086){0.000} |
Market logic | −0.013 (0.093){0.888} | −0.069 (0.073){0.344} | −0.086 (0.102){0.401} |
Risk-based logic | 0.223 (0.113){0.049} | 0.597 (0.082){0.000} | 0.506 (0.089){0.000} |
Capacity-based logic | 0.256 (0.091){0.005} | 0.156 (0.059){0.008} | −0.029 (0.087){0.734} |
Disaster experience | 0.160 (0.117){0.171} | −0.117 (0.088){0.181} | 0.072 (0.094){0.443} |
Midwest | −0.504 (0.375){0.179} | −0.114 (0.260){0.661} | 0.552 (0.331){0.096} |
South | −0.717 (0.437){0.101} | 0.019 (0.303){0.950} | 0.191 (0.246){0.439} |
West | −0.929 (0.439){0.034} | −0.415 (0.388){0.286} | 0.066 (0.290){0.820} |
Organization size | 0.586 (0.178){0.001} | 0.027 (0.125){0.826} | −0.091 (0.083){0.277} |
Authority | −0.214 (0.212){0.311} | −0.124 (0.195){0.524} | 0.381 (0.142){0.007} |
Director | 0.148 (0.236){0.529} | −0.176 (0.186){0.345} | −0.045 (0.219){0.836} |
Bus only | 0.752 (0.317){0.018} | −0.066 (0.247){0.789} | 0.063 (0.236){0.791} |
Density | −0.321 (0.151){0.033} | 0.015 (0.140){0.914} | −0.139 (0.093){0.134} |
Median household income | 0.012 (0.141){0.934} | 0.026 (0.089){0.774} | −0.106 (0.093){0.254} |
Commute time | 0.105 (0.116){0.365} | 0.027 (0.079){0.734} | −0.020 (0.134){0.883} |
Liberal | 0.449 (0.271){0.098} | 0.625 (0.217){0.004} | 1.039 (0.243){0.000} |
R-squared | 0.659 | 0.622 | 0.767 |
N | 61 | 80 | 52 |
. | Forerunner . | Complacent . | Market-oriented . |
---|---|---|---|
Government authority logic | −0.102 (0.109){0.350} | −0.011 (0.067){0.871} | −0.351 (0.109){0.001} |
Professionalism logic | 0.397 (0.070){0.000} | 0.274 (0.065){0.000} | 0.592 (0.086){0.000} |
Market logic | −0.013 (0.093){0.888} | −0.069 (0.073){0.344} | −0.086 (0.102){0.401} |
Risk-based logic | 0.223 (0.113){0.049} | 0.597 (0.082){0.000} | 0.506 (0.089){0.000} |
Capacity-based logic | 0.256 (0.091){0.005} | 0.156 (0.059){0.008} | −0.029 (0.087){0.734} |
Disaster experience | 0.160 (0.117){0.171} | −0.117 (0.088){0.181} | 0.072 (0.094){0.443} |
Midwest | −0.504 (0.375){0.179} | −0.114 (0.260){0.661} | 0.552 (0.331){0.096} |
South | −0.717 (0.437){0.101} | 0.019 (0.303){0.950} | 0.191 (0.246){0.439} |
West | −0.929 (0.439){0.034} | −0.415 (0.388){0.286} | 0.066 (0.290){0.820} |
Organization size | 0.586 (0.178){0.001} | 0.027 (0.125){0.826} | −0.091 (0.083){0.277} |
Authority | −0.214 (0.212){0.311} | −0.124 (0.195){0.524} | 0.381 (0.142){0.007} |
Director | 0.148 (0.236){0.529} | −0.176 (0.186){0.345} | −0.045 (0.219){0.836} |
Bus only | 0.752 (0.317){0.018} | −0.066 (0.247){0.789} | 0.063 (0.236){0.791} |
Density | −0.321 (0.151){0.033} | 0.015 (0.140){0.914} | −0.139 (0.093){0.134} |
Median household income | 0.012 (0.141){0.934} | 0.026 (0.089){0.774} | −0.106 (0.093){0.254} |
Commute time | 0.105 (0.116){0.365} | 0.027 (0.079){0.734} | −0.020 (0.134){0.883} |
Liberal | 0.449 (0.271){0.098} | 0.625 (0.217){0.004} | 1.039 (0.243){0.000} |
R-squared | 0.659 | 0.622 | 0.767 |
N | 61 | 80 | 52 |
Coefficients are followed by standard errors in parentheses and p-values in braces. Bold values indicates statistical significance at p < .05.
. | Forerunner . | Complacent . | Market-oriented . |
---|---|---|---|
Government authority logic | −0.102 (0.109){0.350} | −0.011 (0.067){0.871} | −0.351 (0.109){0.001} |
Professionalism logic | 0.397 (0.070){0.000} | 0.274 (0.065){0.000} | 0.592 (0.086){0.000} |
Market logic | −0.013 (0.093){0.888} | −0.069 (0.073){0.344} | −0.086 (0.102){0.401} |
Risk-based logic | 0.223 (0.113){0.049} | 0.597 (0.082){0.000} | 0.506 (0.089){0.000} |
Capacity-based logic | 0.256 (0.091){0.005} | 0.156 (0.059){0.008} | −0.029 (0.087){0.734} |
Disaster experience | 0.160 (0.117){0.171} | −0.117 (0.088){0.181} | 0.072 (0.094){0.443} |
Midwest | −0.504 (0.375){0.179} | −0.114 (0.260){0.661} | 0.552 (0.331){0.096} |
South | −0.717 (0.437){0.101} | 0.019 (0.303){0.950} | 0.191 (0.246){0.439} |
West | −0.929 (0.439){0.034} | −0.415 (0.388){0.286} | 0.066 (0.290){0.820} |
Organization size | 0.586 (0.178){0.001} | 0.027 (0.125){0.826} | −0.091 (0.083){0.277} |
Authority | −0.214 (0.212){0.311} | −0.124 (0.195){0.524} | 0.381 (0.142){0.007} |
Director | 0.148 (0.236){0.529} | −0.176 (0.186){0.345} | −0.045 (0.219){0.836} |
Bus only | 0.752 (0.317){0.018} | −0.066 (0.247){0.789} | 0.063 (0.236){0.791} |
Density | −0.321 (0.151){0.033} | 0.015 (0.140){0.914} | −0.139 (0.093){0.134} |
Median household income | 0.012 (0.141){0.934} | 0.026 (0.089){0.774} | −0.106 (0.093){0.254} |
Commute time | 0.105 (0.116){0.365} | 0.027 (0.079){0.734} | −0.020 (0.134){0.883} |
Liberal | 0.449 (0.271){0.098} | 0.625 (0.217){0.004} | 1.039 (0.243){0.000} |
R-squared | 0.659 | 0.622 | 0.767 |
N | 61 | 80 | 52 |
. | Forerunner . | Complacent . | Market-oriented . |
---|---|---|---|
Government authority logic | −0.102 (0.109){0.350} | −0.011 (0.067){0.871} | −0.351 (0.109){0.001} |
Professionalism logic | 0.397 (0.070){0.000} | 0.274 (0.065){0.000} | 0.592 (0.086){0.000} |
Market logic | −0.013 (0.093){0.888} | −0.069 (0.073){0.344} | −0.086 (0.102){0.401} |
Risk-based logic | 0.223 (0.113){0.049} | 0.597 (0.082){0.000} | 0.506 (0.089){0.000} |
Capacity-based logic | 0.256 (0.091){0.005} | 0.156 (0.059){0.008} | −0.029 (0.087){0.734} |
Disaster experience | 0.160 (0.117){0.171} | −0.117 (0.088){0.181} | 0.072 (0.094){0.443} |
Midwest | −0.504 (0.375){0.179} | −0.114 (0.260){0.661} | 0.552 (0.331){0.096} |
South | −0.717 (0.437){0.101} | 0.019 (0.303){0.950} | 0.191 (0.246){0.439} |
West | −0.929 (0.439){0.034} | −0.415 (0.388){0.286} | 0.066 (0.290){0.820} |
Organization size | 0.586 (0.178){0.001} | 0.027 (0.125){0.826} | −0.091 (0.083){0.277} |
Authority | −0.214 (0.212){0.311} | −0.124 (0.195){0.524} | 0.381 (0.142){0.007} |
Director | 0.148 (0.236){0.529} | −0.176 (0.186){0.345} | −0.045 (0.219){0.836} |
Bus only | 0.752 (0.317){0.018} | −0.066 (0.247){0.789} | 0.063 (0.236){0.791} |
Density | −0.321 (0.151){0.033} | 0.015 (0.140){0.914} | −0.139 (0.093){0.134} |
Median household income | 0.012 (0.141){0.934} | 0.026 (0.089){0.774} | −0.106 (0.093){0.254} |
Commute time | 0.105 (0.116){0.365} | 0.027 (0.079){0.734} | −0.020 (0.134){0.883} |
Liberal | 0.449 (0.271){0.098} | 0.625 (0.217){0.004} | 1.039 (0.243){0.000} |
R-squared | 0.659 | 0.622 | 0.767 |
N | 61 | 80 | 52 |
Coefficients are followed by standard errors in parentheses and p-values in braces. Bold values indicates statistical significance at p < .05.
The chi-square tests on structural path coefficients demonstrate variant effects of government authority logic (p = .029), professional logic (p = .013), risk-based logic (p = .027), and capacity-based logic (p = .065) across the clusters. The effects of market logic do not significantly differ (p = .849). Specifically, government authority logic has a significantly more negative effect on adaptation in the market-oriented cluster than in the complacent cluster (p = .008). Professional logic exerts a significantly more positive effect among the market-oriented than the complacent organizations (p = .003). Compared with the forerunner cluster, risk-based logic figures more prominently as a positive predictor in the complacent cluster (p = .008) and the market-oriented cluster (p = .049). Finally, capacity-based logic promotes adaptation to a significantly larger extent among the forerunners than among their market-oriented counterparts (p = .024). The findings on variant path coefficients for both macro- and meso-institutional logics support Proposition 3.
For ease of comparison, figure 6 visualizes the effects of the five logics on adaptation across the clusters of organizations. Among the macro-institutions, government authority logic has a significant and negative effect in the market-oriented cluster. Professional logic is a consistently positive predictor for adaptation across all three clusters. Although market logic does not show significant effects across the clusters, the market-oriented organizations consistently experience low levels of adaptation (figure 4). This might imply a saturated degree of market logic has been achieved to suppress adaptation.
With regard to meso-level institutions, risk-based logic is shown to foster adaptation in all three clusters and its motivating effect stands out in the complacent and market-oriented cluster. Compared with risk-based logic, capacity-based logic can promote adaptation in the forerunner and complacent organizations, but not in market-oriented ones.
Discussion
Despite the marked distinctions between the complacent and market-oriented configurations, they both register similarly low on adaptation. Those findings speak of equifinality in a configurational inquiry wherein dissimilar combinations of the parts can produce similar functional outcomes (Ragin 2009). When both macro- and meso-level logics synergistically contribute to complacency, there is little room for institutional awareness to contemplate or act on alternatives. In organizations where the market logic prevails, it can effectively counteract the moderate pull toward adaptation created by other logics, such as professional and risk-based logics.
The differential effects of each logic across clusters suggest institutional work at play. The results demonstrate that professional logic and risk-based logic consistently lead to higher levels of adaptation across all the three combinations of logics, with pronounced impacts on organizations that face strong inconsistencies or opposition from other logics. Specifically, the effects of professional logic are most salient in market-oriented organizations, implying managers’ doubled efforts to prioritize professionalism to counter opposing market forces and spur substantive adaptation. Similarly, in the complacent cluster, a strong risk-based logic provides managers with the opportunity to disrupt the macro institutions and mobilize new logics to meet the adaptation demand. The risk-based logic also manifests a bigger positive effect size in the market-oriented than in the forerunner cluster, further reinforcing the importance of institutional work to reconcile inconsistent logics in such a way as to prioritize climate risks and justify adaptive behavior. In contrast, the effects of the capacity-based logic appear to be aligned with the macro-institutional logics, thus providing no windows of opportunity for institutional work.
Note that institutional contradictions do not always lead to adaptive outcomes. Government authority logic is found to be maladaptive in market-oriented organizations. We suspect that when market logic dominates, managers facing strong authority logic tend to reconcile the inconsistencies through hasty fixes to satisfy short-term government demands, thereby preempting adaptive decision and actions. Although it is a common practice for government to increase scrutiny and pressure for public organizations to demonstrate accountability, particularly in the aftermath of extreme weather events (Birkland 2009), our findings point to the need of heeding the organizations’ institutional environment to avoid counterproductive outcomes.
A few limitations should be considered in future research. To begin with, this study applies the meta-theory of institutional logics to characterize the macro- and meso-level logics and connect them to climate adaptation. This institutionally focused framework means we were not able to explicitly incorporate other factors that also matter to adaptation, such as event characteristics (Giordono et al. 2020; Zhang 2022), network effects (Hovik et al. 2015), and policy learning (Birkland 1997). Our inclusion of institutional logics is not exhaustive either. For instance, the study does not include policy and political logics (Hustedt and Danken 2017), or public sector logic (Saz-Carranza and Longo 2012) due to their marginal relevance. Second, we are cognizant of possible limitations present in our measurement of the institutional logics. Those institutions are likely multidimensional while our data allow us to capture only one dimension. Future qualitative and quantitative research can add complexity and nuances to those measures. Third, our empirical focus on public transit agencies might indicate limited generalizability to other sectors. Given the novelty of this study with no a priori evidence about combinations of the macro- and meso-level institutional logic, there is uncertainty about whether our profile solution can replicate in other contexts—a direction for future studies. However, the focus on transit agencies is important in its own right, because this sector, despite its provision of essential public service in everyday and extreme situations, has received scant attention in the public administration literature as well as climate adaptation research (Dovers and Hezri 2010). The shared characteristics of transit with other critical infrastructure sectors, such as water, telecommunication and energy, also help alleviate concerns about applicability.
Conclusion
Our study contributes to the public administration literature by examining organizational response to adverse climatic changes. As much of the literature on climate change in public organizations has focused on mitigating organizational impacts on the climate through decarbonization (e.g., Clarke and Ordonez-Ponce 2017; Krause 2011; Pitt 2010), we reorient attention to reducing climate-induced physical impacts on organizations. We also break away from long-held assumptions about the stability, linearity, and predictability of climate conditions by considering the profound threats climate change brings about. A systematic view of climate change takes a critical first step toward building a conceptual foundation for public organization adaptation to climate change. Meanwhile, by empirically focusing on extreme weather, this study is favorably positioned to capture the impacts of extreme climate variability and unravel its implications for organizations, while circumventing the political controversies around climate change and climate actions. Through linking knowledge and theories from public administration to climate adaptation research, it makes a much-needed contribution to identify, theorize, and elucidate the origin and causal mechanisms of hinderance to climate adaptation practices (Dovers and Hezri 2010; Dupuis and Knoepfel 2013).
We extend previous research on organizational adaptation to climate change (e.g., Zhang et al. 2018) that highlights the role of risk perception through directly addressing the institutional contexts surrounding adaptation. This study contributes to the emerging scholarship on organizational adaptation to climate change by characterizing the multiplicity of institutional logics and exploring heterogeneity of organizational response. The meta-theory of institutional logics permits us to incorporate the role of institutionally aware top managers in navigating institutional complexity in response to climate change. Our analysis shows institutional work at play that creates, maintains, or disrupts institutions through reflection, interpretation, translation, editing, and transformation. This is evident from the pronounced effects of professional and risk-based logic on adaptation when facing incongruence or opposition from other logics. In the meantime, the average low levels of adaptation in the complacent and market-oriented organizations suggest that the institutional work is advanced within the limit of the overall institutional environment and is successful only to a certain extent.
Moreover, we apply a configurational approach to combine macro- and meso-level logics between and across levels, and advance insights on managers’ institutional work with a multigroup SEM probing how the effects of each logic differ across clusters. The holistic perspective on the co-occurrence and interplay of logics uncovers new insights. Notably, our meta-theoretic configuration approach shows that the most adaptive organizations combine macro-level professional logic and meso-level risk-based logic, an insight unachievable through variable-oriented techniques applied in previous studies (Miao et al. 2018b; Zhang et al. 2018). As such, the study confirms theoretical and analytical gains in explaining climate adaptation with a configurational approach. We argue that the configurational approach assumes applicability to other issue contexts using the lens of institutional logics. Particularly, given the current limited application of institutional logics in public administration, we see this study’s potential to advance this line of research tackling critical questions on hybridization in new organizational forms, interorganizational partnerships, public values, openness and transparency, equity or rule orientation, to name a few. Moreover, we also envision this approach’s utility for the broader study of public organizations. A prime area of application is the behavioral research in public administration. This strand of research is criticized for leaving out the institutions in which organizations are embedded and is suspicious of neglecting “big questions” (Moynihan 2018). The configuration approach offers a vehicle to combine micro-level cognitions, backgrounds, and behaviors of individuals actors with the macro institutions, thereby bridging the schism between micro and macro public administration (Moynihan 2018).
Our theory suggests that in the field of public organizations, macro-level logics can potentially influence meso-level logics. For example, the macro-level government authority logic might dampen the extent to which public organizations recognize or accentuate meso-level risk-based logic. Similarly, as public organizations institutionalize market mechanisms over time, this macro-level logic could undermine meso-level risk-based and capacity-based logic, thereby eroding the legitimacy of adaptive decisions and actions. While we do not investigate these linkages empirically, we believe they are important priorities for future research.
Practically, the study enriches understanding about organizational climate adaptation by uncovering distinct configurations of the macro- and meso-institutions: forerunner, complacent, and market-oriented. The complacent cluster and market-oriented cluster are both maladaptive to climate change, collectively accounting for 69% of the entire sample. Those patterns shed important light on the predominantly reactive and ad hoc response to climate change in the public sector (Bierbaum et al. 2013; Zhang and Maroulis 2021). While complacency is long known as a fundamental barrier to adaptation (O’Brien et al. 2006), the findings on the market-oriented cluster serve as an awakening call for the public sector that has relied increasingly or excessively on market-based instruments, practices, and structures. A strong market logic can lead to goal displacement, resource squeezing, long-term underinvestment, as well as chronic structural and cultural weaknesses (Inderberg 2011). As climate change continues to worsen, adherence to this logic would come at the expense of maladaptation and increased vulnerability (Zhang 2021), the cumulative costs of which would outweigh the short-term efficiency gains from market-based mechanisms. Consequences of the hollowing out are abundantly evident from public organizations’ incompetence and halting response at the onset of the COVID-19 outbreak (Balz 2020). Since climate adaptation and pandemic preparedness share similarities in terms of the needs and approaches to resilience, an integrative approach is needed to plan for the “new normal” in the post-COVID19 era, starting with problematization of the market logic amid a changing climate.
Moreover, the consistently stimulating effects of professional and risk-based logics across all organizations suggest points of interventions to advance climate adaptation. Public organizations can strengthen professional logic through subscription to communication from professional associations, continuous engagement in professional networks, recruitment efforts, or interorganizational exchange activities. Governments can also play an instrumental role by garnering ideas, diffusing knowledge and practices, and providing technical assistance to facilitate climate adaptation. With respect to risk-based logic, appropriate measures need to be taken to train institutionally aware and climate observant managers while overcoming the constraints of public bureaucracies. The changes can start with developing monitoring systems that can detect and amplify climate anomalies, archiving, communicating, and stressing lessons learned from complacency toward climate change, as well as involving all organizational units (instead of just emergency or risk management personnel) in scanning, monitoring, interpreting, and enacting on climate risks.
Acknowledgments
This research was made possible through generous support by the Federal Transit Administration, US Department of Transportation.
Data Availability Statement
The data and code underlying this article are available at https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/3UHIOD.
Conflict of Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Footnotes
Its equivalent is latent class analysis (LCA) when the profile indicators are categorical variables.
Appendix A: Common Method Bias
We carried out a series of procedural and statistical remedies to address potential common method bias (CMB). First, we pretested the survey instrument with seven individuals who provided detailed comments and feedback on specific questions to avoid ambiguity or confusion in wording and eliminate unfamiliar or inappropriate terms (Podsakoff et al. 2003). Each survey question provided verbal labels for the response categories rather than only the end points. Measurements of the predictors and the criterion variable were placed physically apart to minimize stylistic response or contextual effects (Podsakoff et al. 2012). We also conducted the Harman’s single-factor test to probe the presence of CMB. The single factor model shows a poor model fit (df = 299, χ2 = 884,335, p-value = .0000, RESEA = .079, CFI = 0.378, TLI = .342, and SRMR = .139) and the common factor accounts for 17.5% of the total variance. Given that the common method variance needs to be relatively high (above 70%) to sufficiently bias relationships and alter substantive conclusions (Fuller et al. 2016), the results from the Harman’s test can mitigate those concerns.
While we recognize the remedies are inadequate to rule out CMB (Podsakoff et al. 2003), we note that it would be difficult to collect data about the institutional logics of public organizations at the national level to capture a broad range of heterogeneity. Previous research also supports the validity of self-reported data to measure organizational characteristics and experiences (Moynihan and Pandey 2005; Pandey and Wright 2006).
Appendix B: Measurement Model
. | Estimate . | S.E. . | p-Value . | R-squared . |
---|---|---|---|---|
City government influence | ||||
City government influence 1 | 0.531 | 0.057 | 0.000 | 0.282 |
City government influence 2 | 0.834 | 0.058 | 0.000 | 0.696 |
City government influence 3 | 0.706 | 0.045 | 0.000 | 0.498 |
State/federal government influence | ||||
State/federal government influence 1 | 0.661 | 0.049 | 0.000 | 0.437 |
State/federal government influence 2 | 0.624 | 0.050 | 0.000 | 0.389 |
State/federal government influence 3 | 0.466 | 0.061 | 0.000 | 0.217 |
State/federal government influence 4 | 0.570 | 0.049 | 0.000 | 0.325 |
State/federal government influence 5 | 0.467 | 0.056 | 0.000 | 0.218 |
Government authority logic | ||||
City government influence | 0.641 | 0.079 | 0.000 | 0.411 |
State/federal government influence | 0.888 | 0.098 | 0.000 | 0.789 |
Professional logic | ||||
Professional logic 1 | 0.848 | 0.029 | 0.000 | 0.719 |
Professional logic 2 | 0.718 | 0.039 | 0.000 | 0.516 |
Professional logic 3 | 0.685 | 0.041 | 0.000 | 0.469 |
Professional logic 4 | 0.683 | 0.042 | 0.000 | 0.466 |
Professional logic 5 | 0.710 | 0.038 | 0.000 | 0.504 |
Risk-based logic | ||||
Risk-based logic 1 | 0.791 | 0.048 | 0.000 | 0.626 |
Risk-based logic 2 | 0.759 | 0.046 | 0.000 | 0.576 |
Risk-based logic 3 | 0.762 | 0.039 | 0.000 | 0.581 |
Capacity-based logic | ||||
Capacity-based logic 1 | 0.579 | 0.047 | 0.000 | 0.335 |
Capacity-based logic 2 | 0.559 | 0.045 | 0.000 | 0.312 |
Capacity-based logic 3 | 0.838 | 0.044 | 0.000 | 0.702 |
Capacity-based logic 4 | 0.618 | 0.045 | 0.000 | 0.382 |
Adaptation | ||||
Adaptation 1 | 0.698 | 0.109 | 0.000 | 0.487 |
Adaptation 2 | 0.714 | 0.110 | 0.000 | 0.510 |
Adaptation 3 | 0.619 | 0.108 | 0.000 | 0.383 |
Adaptation 4 | 0.568 | 0.139 | 0.000 | 0.323 |
Adaptation 5 | 0.862 | 0.077 | 0.000 | 0.743 |
Adaptation 6 | 0.763 | 0.093 | 0.000 | 0.582 |
. | Estimate . | S.E. . | p-Value . | R-squared . |
---|---|---|---|---|
City government influence | ||||
City government influence 1 | 0.531 | 0.057 | 0.000 | 0.282 |
City government influence 2 | 0.834 | 0.058 | 0.000 | 0.696 |
City government influence 3 | 0.706 | 0.045 | 0.000 | 0.498 |
State/federal government influence | ||||
State/federal government influence 1 | 0.661 | 0.049 | 0.000 | 0.437 |
State/federal government influence 2 | 0.624 | 0.050 | 0.000 | 0.389 |
State/federal government influence 3 | 0.466 | 0.061 | 0.000 | 0.217 |
State/federal government influence 4 | 0.570 | 0.049 | 0.000 | 0.325 |
State/federal government influence 5 | 0.467 | 0.056 | 0.000 | 0.218 |
Government authority logic | ||||
City government influence | 0.641 | 0.079 | 0.000 | 0.411 |
State/federal government influence | 0.888 | 0.098 | 0.000 | 0.789 |
Professional logic | ||||
Professional logic 1 | 0.848 | 0.029 | 0.000 | 0.719 |
Professional logic 2 | 0.718 | 0.039 | 0.000 | 0.516 |
Professional logic 3 | 0.685 | 0.041 | 0.000 | 0.469 |
Professional logic 4 | 0.683 | 0.042 | 0.000 | 0.466 |
Professional logic 5 | 0.710 | 0.038 | 0.000 | 0.504 |
Risk-based logic | ||||
Risk-based logic 1 | 0.791 | 0.048 | 0.000 | 0.626 |
Risk-based logic 2 | 0.759 | 0.046 | 0.000 | 0.576 |
Risk-based logic 3 | 0.762 | 0.039 | 0.000 | 0.581 |
Capacity-based logic | ||||
Capacity-based logic 1 | 0.579 | 0.047 | 0.000 | 0.335 |
Capacity-based logic 2 | 0.559 | 0.045 | 0.000 | 0.312 |
Capacity-based logic 3 | 0.838 | 0.044 | 0.000 | 0.702 |
Capacity-based logic 4 | 0.618 | 0.045 | 0.000 | 0.382 |
Adaptation | ||||
Adaptation 1 | 0.698 | 0.109 | 0.000 | 0.487 |
Adaptation 2 | 0.714 | 0.110 | 0.000 | 0.510 |
Adaptation 3 | 0.619 | 0.108 | 0.000 | 0.383 |
Adaptation 4 | 0.568 | 0.139 | 0.000 | 0.323 |
Adaptation 5 | 0.862 | 0.077 | 0.000 | 0.743 |
Adaptation 6 | 0.763 | 0.093 | 0.000 | 0.582 |
. | Estimate . | S.E. . | p-Value . | R-squared . |
---|---|---|---|---|
City government influence | ||||
City government influence 1 | 0.531 | 0.057 | 0.000 | 0.282 |
City government influence 2 | 0.834 | 0.058 | 0.000 | 0.696 |
City government influence 3 | 0.706 | 0.045 | 0.000 | 0.498 |
State/federal government influence | ||||
State/federal government influence 1 | 0.661 | 0.049 | 0.000 | 0.437 |
State/federal government influence 2 | 0.624 | 0.050 | 0.000 | 0.389 |
State/federal government influence 3 | 0.466 | 0.061 | 0.000 | 0.217 |
State/federal government influence 4 | 0.570 | 0.049 | 0.000 | 0.325 |
State/federal government influence 5 | 0.467 | 0.056 | 0.000 | 0.218 |
Government authority logic | ||||
City government influence | 0.641 | 0.079 | 0.000 | 0.411 |
State/federal government influence | 0.888 | 0.098 | 0.000 | 0.789 |
Professional logic | ||||
Professional logic 1 | 0.848 | 0.029 | 0.000 | 0.719 |
Professional logic 2 | 0.718 | 0.039 | 0.000 | 0.516 |
Professional logic 3 | 0.685 | 0.041 | 0.000 | 0.469 |
Professional logic 4 | 0.683 | 0.042 | 0.000 | 0.466 |
Professional logic 5 | 0.710 | 0.038 | 0.000 | 0.504 |
Risk-based logic | ||||
Risk-based logic 1 | 0.791 | 0.048 | 0.000 | 0.626 |
Risk-based logic 2 | 0.759 | 0.046 | 0.000 | 0.576 |
Risk-based logic 3 | 0.762 | 0.039 | 0.000 | 0.581 |
Capacity-based logic | ||||
Capacity-based logic 1 | 0.579 | 0.047 | 0.000 | 0.335 |
Capacity-based logic 2 | 0.559 | 0.045 | 0.000 | 0.312 |
Capacity-based logic 3 | 0.838 | 0.044 | 0.000 | 0.702 |
Capacity-based logic 4 | 0.618 | 0.045 | 0.000 | 0.382 |
Adaptation | ||||
Adaptation 1 | 0.698 | 0.109 | 0.000 | 0.487 |
Adaptation 2 | 0.714 | 0.110 | 0.000 | 0.510 |
Adaptation 3 | 0.619 | 0.108 | 0.000 | 0.383 |
Adaptation 4 | 0.568 | 0.139 | 0.000 | 0.323 |
Adaptation 5 | 0.862 | 0.077 | 0.000 | 0.743 |
Adaptation 6 | 0.763 | 0.093 | 0.000 | 0.582 |
. | Estimate . | S.E. . | p-Value . | R-squared . |
---|---|---|---|---|
City government influence | ||||
City government influence 1 | 0.531 | 0.057 | 0.000 | 0.282 |
City government influence 2 | 0.834 | 0.058 | 0.000 | 0.696 |
City government influence 3 | 0.706 | 0.045 | 0.000 | 0.498 |
State/federal government influence | ||||
State/federal government influence 1 | 0.661 | 0.049 | 0.000 | 0.437 |
State/federal government influence 2 | 0.624 | 0.050 | 0.000 | 0.389 |
State/federal government influence 3 | 0.466 | 0.061 | 0.000 | 0.217 |
State/federal government influence 4 | 0.570 | 0.049 | 0.000 | 0.325 |
State/federal government influence 5 | 0.467 | 0.056 | 0.000 | 0.218 |
Government authority logic | ||||
City government influence | 0.641 | 0.079 | 0.000 | 0.411 |
State/federal government influence | 0.888 | 0.098 | 0.000 | 0.789 |
Professional logic | ||||
Professional logic 1 | 0.848 | 0.029 | 0.000 | 0.719 |
Professional logic 2 | 0.718 | 0.039 | 0.000 | 0.516 |
Professional logic 3 | 0.685 | 0.041 | 0.000 | 0.469 |
Professional logic 4 | 0.683 | 0.042 | 0.000 | 0.466 |
Professional logic 5 | 0.710 | 0.038 | 0.000 | 0.504 |
Risk-based logic | ||||
Risk-based logic 1 | 0.791 | 0.048 | 0.000 | 0.626 |
Risk-based logic 2 | 0.759 | 0.046 | 0.000 | 0.576 |
Risk-based logic 3 | 0.762 | 0.039 | 0.000 | 0.581 |
Capacity-based logic | ||||
Capacity-based logic 1 | 0.579 | 0.047 | 0.000 | 0.335 |
Capacity-based logic 2 | 0.559 | 0.045 | 0.000 | 0.312 |
Capacity-based logic 3 | 0.838 | 0.044 | 0.000 | 0.702 |
Capacity-based logic 4 | 0.618 | 0.045 | 0.000 | 0.382 |
Adaptation | ||||
Adaptation 1 | 0.698 | 0.109 | 0.000 | 0.487 |
Adaptation 2 | 0.714 | 0.110 | 0.000 | 0.510 |
Adaptation 3 | 0.619 | 0.108 | 0.000 | 0.383 |
Adaptation 4 | 0.568 | 0.139 | 0.000 | 0.323 |
Adaptation 5 | 0.862 | 0.077 | 0.000 | 0.743 |
Adaptation 6 | 0.763 | 0.093 | 0.000 | 0.582 |
Appendix C: Agency-Level Summary Statistics
. | Mean . | Std . | Min . | Max . |
---|---|---|---|---|
Adaptation | 0.06 | 0.44 | −0.81 | 1.82 |
Government authority logic | −0.01 | 0.26 | −0.72 | 0.63 |
Professional logic | −0.02 | 0.59 | −1.06 | 1.71 |
Market logic | 0.35 | 0.39 | 0 | 1 |
Risk−based logic | 0.02 | 0.59 | −1.74 | 1.36 |
Capacity-based logic | −0.01 | 0.40 | −1.05 | 1.10 |
Disaster experience | 0.95 | 0.84 | 0 | 3 |
Northeast | 0.12 | 0.47 | 0 | 1 |
Midwest | 0.22 | 0.42 | 0 | 1 |
South | 0.32 | 0.46 | 0 | 1 |
West | 0.33 | 0.47 | 0 | 1 |
Organization size | 17.69 | 1.44 | 14.75 | 21.70 |
Authority | 0.63 | 0.48 | 0 | 1 |
Director | 0.75 | 0.44 | 0 | 1 |
Bus only | 0.79 | 0.40 | 0 | 1 |
Density | 7.86 | 0.43 | 6.79 | 8.85 |
Median household income | 10.99 | 0.23 | 10.49 | 11.58 |
Commute time | 25.31 | 4.91 | 15.70 | 39.10 |
Liberal | 0.51 | 0.50 | 0 | 1 |
. | Mean . | Std . | Min . | Max . |
---|---|---|---|---|
Adaptation | 0.06 | 0.44 | −0.81 | 1.82 |
Government authority logic | −0.01 | 0.26 | −0.72 | 0.63 |
Professional logic | −0.02 | 0.59 | −1.06 | 1.71 |
Market logic | 0.35 | 0.39 | 0 | 1 |
Risk−based logic | 0.02 | 0.59 | −1.74 | 1.36 |
Capacity-based logic | −0.01 | 0.40 | −1.05 | 1.10 |
Disaster experience | 0.95 | 0.84 | 0 | 3 |
Northeast | 0.12 | 0.47 | 0 | 1 |
Midwest | 0.22 | 0.42 | 0 | 1 |
South | 0.32 | 0.46 | 0 | 1 |
West | 0.33 | 0.47 | 0 | 1 |
Organization size | 17.69 | 1.44 | 14.75 | 21.70 |
Authority | 0.63 | 0.48 | 0 | 1 |
Director | 0.75 | 0.44 | 0 | 1 |
Bus only | 0.79 | 0.40 | 0 | 1 |
Density | 7.86 | 0.43 | 6.79 | 8.85 |
Median household income | 10.99 | 0.23 | 10.49 | 11.58 |
Commute time | 25.31 | 4.91 | 15.70 | 39.10 |
Liberal | 0.51 | 0.50 | 0 | 1 |
Agency-level N = 193.
. | Mean . | Std . | Min . | Max . |
---|---|---|---|---|
Adaptation | 0.06 | 0.44 | −0.81 | 1.82 |
Government authority logic | −0.01 | 0.26 | −0.72 | 0.63 |
Professional logic | −0.02 | 0.59 | −1.06 | 1.71 |
Market logic | 0.35 | 0.39 | 0 | 1 |
Risk−based logic | 0.02 | 0.59 | −1.74 | 1.36 |
Capacity-based logic | −0.01 | 0.40 | −1.05 | 1.10 |
Disaster experience | 0.95 | 0.84 | 0 | 3 |
Northeast | 0.12 | 0.47 | 0 | 1 |
Midwest | 0.22 | 0.42 | 0 | 1 |
South | 0.32 | 0.46 | 0 | 1 |
West | 0.33 | 0.47 | 0 | 1 |
Organization size | 17.69 | 1.44 | 14.75 | 21.70 |
Authority | 0.63 | 0.48 | 0 | 1 |
Director | 0.75 | 0.44 | 0 | 1 |
Bus only | 0.79 | 0.40 | 0 | 1 |
Density | 7.86 | 0.43 | 6.79 | 8.85 |
Median household income | 10.99 | 0.23 | 10.49 | 11.58 |
Commute time | 25.31 | 4.91 | 15.70 | 39.10 |
Liberal | 0.51 | 0.50 | 0 | 1 |
. | Mean . | Std . | Min . | Max . |
---|---|---|---|---|
Adaptation | 0.06 | 0.44 | −0.81 | 1.82 |
Government authority logic | −0.01 | 0.26 | −0.72 | 0.63 |
Professional logic | −0.02 | 0.59 | −1.06 | 1.71 |
Market logic | 0.35 | 0.39 | 0 | 1 |
Risk−based logic | 0.02 | 0.59 | −1.74 | 1.36 |
Capacity-based logic | −0.01 | 0.40 | −1.05 | 1.10 |
Disaster experience | 0.95 | 0.84 | 0 | 3 |
Northeast | 0.12 | 0.47 | 0 | 1 |
Midwest | 0.22 | 0.42 | 0 | 1 |
South | 0.32 | 0.46 | 0 | 1 |
West | 0.33 | 0.47 | 0 | 1 |
Organization size | 17.69 | 1.44 | 14.75 | 21.70 |
Authority | 0.63 | 0.48 | 0 | 1 |
Director | 0.75 | 0.44 | 0 | 1 |
Bus only | 0.79 | 0.40 | 0 | 1 |
Density | 7.86 | 0.43 | 6.79 | 8.85 |
Median household income | 10.99 | 0.23 | 10.49 | 11.58 |
Commute time | 25.31 | 4.91 | 15.70 | 39.10 |
Liberal | 0.51 | 0.50 | 0 | 1 |
Agency-level N = 193.
Appendix D: Statistical Indices for LPA Models
Model . | Classes . | # Parameters . | BIC . | BLRT . | Adjusted LMR . | SABIC . | Entropy . | Cluster Sizes . |
---|---|---|---|---|---|---|---|---|
1 | 2 | 16 | 2650.262 | 0.000 | 0.000 | 2599.578 | 0.969 | 141,52 |
2 | 3 | 22 | 2616.145 | 0.000 | 0.010 | 2546.455 | 0.808 | 80,61,52 |
3 | 4 | 28 | 2537.187 | 0.000 | 0.061 | 2448.490 | 0.972 | 23,40,82,48 |
Model . | Classes . | # Parameters . | BIC . | BLRT . | Adjusted LMR . | SABIC . | Entropy . | Cluster Sizes . |
---|---|---|---|---|---|---|---|---|
1 | 2 | 16 | 2650.262 | 0.000 | 0.000 | 2599.578 | 0.969 | 141,52 |
2 | 3 | 22 | 2616.145 | 0.000 | 0.010 | 2546.455 | 0.808 | 80,61,52 |
3 | 4 | 28 | 2537.187 | 0.000 | 0.061 | 2448.490 | 0.972 | 23,40,82,48 |
The selected solution is highlighted in bold.
Model . | Classes . | # Parameters . | BIC . | BLRT . | Adjusted LMR . | SABIC . | Entropy . | Cluster Sizes . |
---|---|---|---|---|---|---|---|---|
1 | 2 | 16 | 2650.262 | 0.000 | 0.000 | 2599.578 | 0.969 | 141,52 |
2 | 3 | 22 | 2616.145 | 0.000 | 0.010 | 2546.455 | 0.808 | 80,61,52 |
3 | 4 | 28 | 2537.187 | 0.000 | 0.061 | 2448.490 | 0.972 | 23,40,82,48 |
Model . | Classes . | # Parameters . | BIC . | BLRT . | Adjusted LMR . | SABIC . | Entropy . | Cluster Sizes . |
---|---|---|---|---|---|---|---|---|
1 | 2 | 16 | 2650.262 | 0.000 | 0.000 | 2599.578 | 0.969 | 141,52 |
2 | 3 | 22 | 2616.145 | 0.000 | 0.010 | 2546.455 | 0.808 | 80,61,52 |
3 | 4 | 28 | 2537.187 | 0.000 | 0.061 | 2448.490 | 0.972 | 23,40,82,48 |
The selected solution is highlighted in bold.
References
———.
———.
———.
———.
———.