Abstract
Studying the so-called refugee crisis in Germany, this article asks about the effectiveness of crisis management by a large number of local administrations, each acting upon the same crisis impulse of a high number of asylum seekers who entered the country in 2015 and 2016. Instead of theorizing the exact administrative design features fit for an effective crisis response, the focus is on the ability of administrations to adjust. We conceptualize such shifts in administrative practices as informal and temporary (latent) deviations from routine action along two dimensions of organizational behavior typically dominant in private and nonprofit sector organizations, respectively: internal flexibility and citizen participation (hybridity). Novel survey data from 235 out of 401 German district authorities are reported. We test the effects of different forms of latent hybridization on administrative effectiveness using regression modeling. Findings indicate that changes in administrative practices towards more flexible and participatory action had a positive impact on self-reported crisis management effectiveness. The effect of flexible action was especially pronounced in districts that were allocated higher shares of asylum seekers. These findings advance theory on crisis management and bottom-up implementation, highlighting the ability of local agencies to shift practices as a key explanatory factor for effective administrative action in exceptional situations.
Introduction
Crises pose ambiguous challenges to societies. They can be triggered by natural events such as earthquakes, hurricanes or pandemics, by technical failure or by social processes, for example, massive migration movements. Although societies have always faced crises, researchers claim that because of the global interconnectedness and technological vulnerability of modern societies, crises “are the new normal” (Tierney 2014, 238). Despite multiple manifestations, the defining characteristic of a crisis is that it threatens fundamental societal structures, values, or norms (Boin, ’t Hart, and Kuipers 2018). Yet, to the extent that crises provoke tensions and distress, overcoming a crisis can also pave the way for something new, strengthening societal solidarity, identity, and resilience (Duit 2016; ‘t Hart 2014).
When a crisis hits, the question of successful crisis management is essentially a question of the effectiveness of public administration. Much is already known about the particular administrative challenges raised by crises (Ansell, Boin, and Keller 2010; Peters, Pierre, and Randma-Liiv 2011; Stark 2011). The literature on crisis management provides numerous recommendations for appropriate institutional structures and processes. These findings, however, tend to be drawn from single case studies or small-N comparisons. There is hardly any large-N research, including concepts that enable comparative study of administrative crisis management. To fill this gap, and to study how local administrations respond differently to a crisis that manifests with varying intensity in different regions of one country, we analyze novel survey data covering administrative action1 in the so-called “refugee crisis” of 2015/16 in Germany. The article asks how variation in the extent to which administrations shift from routine behavior to crisis action determines the effectiveness of the crisis response.
We assume that a crisis impulse leads to observable shifts in administrative practice or behavior. Our theoretical aim is to conceptualize such shifts from administrative routine to crisis action on the basis of foundational ideas in organization theory. At the most basic level, organizations can be classified according to their dominant behavioral principles as belonging to three sectors: the public, private, and nonprofit (sometimes third) sector. Public sector principles such as hierarchy and the rule of law appear problematic from the perspective of crisis management (Ansell et al. 2010; Boin and Lodge 2016; Christensen, Lægreid, and Rykkja 2016; Rosenthal, Boin, and Comfort 2001; Rosenthal, Charles, and ’t Hart 1989). By contrast, private sector organizations value flexibility in aligning internal resources to organizational goals. Nonprofit organizations seek to integrate voluntary labor and hence emphasize participatory principles in their outside interaction. We argue that informal and temporary (latent) shifts towards such behavioral principles from the private and nonprofit sector (hybridity)—more flexible internal operations and more participation in outside interaction—are well suited to match the specific politico-administrative challenges posed by crisis situations (Ansell et al. 2010). Together, these features of administrative behavior during crisis management can be termed latent hybridity.
We propose that the extent to which administrations become more hybrid in a crisis response situation has an impact on the output side of administrative action. We capture this effect by measuring both changes in administrative action during a crisis and administrations’ self-perceived effectiveness of the crisis response. We also understand shifts towards hybridity as gradual adaptations to the severity of crisis exposure, meaning that higher affectedness implicates a higher need for hybridization.
Empirical data are drawn from a comparative study of administrative action at the district-level (Landkreise and kreisfreie Städte) in the so-called “refugee crisis” in Germany in 2015 and 2016. Those particular years are referred to as a crisis because of the overburdening of the entire politico-administrative system (Bogumil et al. 2018). Our main data source is an original survey that was sent out to each district commissioner in 2018 and returned by 235 (58% response rate). This novel data allowed observing shifts in administrative action that occurred as a response to the crisis and to analyze whether these changes account for variation in administrative effectiveness during the crisis response. Empirical findings correspond with our main hypotheses: Shifts in administrative action towards flexibility and participation were associated with more effective crisis management. Furthermore, flexible action was most helpful in counties that had to cope with higher shares of asylum seekers, thus faced a more severe crisis impact.
In addition to its empirical contribution, the article makes theoretical and conceptual contributions to literature on public administration and crisis management. We introduce the new idea of latent hybridity, which delimits from more formal and standardized forms of hybridity because of its informal, temporary and ephemeral nature (e.g., Christensen and Lægreid 2011; Denis, Ferlie, and Van Gestel 2015; Seibel 2015; Smith and Besharov 2017) and propose a link with literature on crisis management (e.g., Ansell et al. 2010; Boin et al. 2018; Christensen et al. 2016). Our findings show that administrations’ ability to adjust their organizational behavior is an important precondition for effective crisis management. Because crises tend to unfold in unforeseen ways, this is arguably more useful than pre-conceived plans, which often turn out as “paper tigers” (Aken and Fenema 2014, 1). More broadly speaking, our work resonates with others who emphasize the importance of the ability of organizations to switch between practices or phases, to become more responsive or agile (Mergel, Ganapati, and Whitford 2020; Schakel, Fenema, and Faraj 2016).
Furthermore, literature on bottom-up implementation and street-level bureaucracy emphasizes the importance of individual discretion to deal with complex or “wicked” policy problems (Brodkin 2011; Keiser 1999; Lavee 2020; Pires 2011; Thomann, van Engen, and Tummers 2018). Our findings on latent hybridity at the organizational level call for more attention to the social processes that link individual action with organizational behavior and administrative output. Speaking lastly to literature on participatory governance, we offer new evidence on the instrumental value of participation to enhance administrative effectiveness (Baldwin 2019; Batory and Svensson 2020; Thomsen and Jensen 2019).
The article is structured as follows: The section below defines terms and specifies the main theoretical argument. In the next section, we discuss the expected impact of latent hybridity on crisis management, drawing on examples from the German case. This is followed by a specification of the hypotheses. Subsequent sections discuss case selection, operationalization and data gathering, present and discuss the results, and offer a conclusion.
The Argument: Latent Hybridity in Administrative Crisis Management
Following a classic definition, crises are characterized by a real or perceived threat to central social values, high time pressure on decision-makers, and high levels of uncertainty concerning the appropriate response (Ansell et al. 2010; Boin et al. 2018). There is no general theory of the factors contributing to effective administrative crisis management. Public administration and crisis management literature, however, provide a number of concrete recommendations (Boin and Lodge 2016; Christensen et al. 2016; Comfort, Waugh, and Cigler 2012; Kapucu, Arslan, and Collins 2010; Rosenthal et al. 1989, 2001). These include, among others, an effective combination of ordered hierarchical emergency response in combination with functionally decentralized problem solving (Ansell et al. 2010; Boin et al. 2018; Moynihan 2009), crisis leadership (‘t Hart 2014; Van Wart 2013), the importance of surge capacities (Boin, Kelle, and Clay Whybark 2010), and effective horizontal or transboundary coordination networks (Nowell et al. 2018).
However, these and similar conclusions are essentially orientated toward the past, drawing conclusions from the analysis or evaluation of specific responses to yesterday’s crises. Yet, as Boin and Lodge (2016, 291, emphasis original) summarize the problem, “one can never be sure that available risk tools apply to all possible events.” Case studies have repeatedly shown a common gap between central plans and local challenges, when crisis manuals and contingency plans “prove to be “paper tigers” once the emergency is there: organizations are not well prepared, coordination of efforts is dismal, the emergency does not develop according to plan and the agreements and plans do not sufficiently allow for flexibility, responsiveness, resilience and improvisation” (Aken and Fenema 2014, 1). As such, it should be difficult, if not impossible, to specify ex ante what particular institutional design features or administrative actions will assure effective management for any given crisis.
Following Levitt and March’s (1988, 326) classical credo, that “during very bad times (…) no routine will lead to success,” we therefore suggest redirecting the focus of attention. Instead of analyzing optimal institutional designs, we study administrative adaptation; that is, to what extent administrations shift their modus-operandi from every-day routine work to satisfy the ambiguous demands posed by the crisis. The extent of this shift can be understood as gradual adjustment, dependent on the imperative of demands produced by the severity of the crisis.
Defining Latent Hybridity
In order to cope with dynamic and complex crises, administrations oftentimes mix or replace established bureaucratic routines with new procedures and practices. Drawing on foundational ideas about organization and public administration, we conceptualize such shifts in administrative behavior as latent hybridity, whereas the term latent refers to the informal and temporary nature of the change, and hybridity to the mix of new practices that are adopted.
Beginning with the latter, it is helpful to acknowledge that organizational practices develop in response to diverse expectations from the external environment (Meyer and Rowan 1977; Selznick 1949, 1957). Outside demands shape distinctive institutional logics that characterize and delimit certain types of organizations. These institutional logics have been described as relative stable sets of material practices and beliefs that shape organizational behavior (Besharov and Smith 2014; Thornton, Ocasio, and Lounsbury 2012). A seminal distinction is between organizations in the trichotomy of public, private, and nonprofit sectors, who each face distinctive outside demands and thus developed different dominant practices: Public sector agencies emphasize lawful decisions to adhere to legitimacy demands, private sector organizations nurture flexibility to leverage performance, and nonprofits specialize in volunteer participation to achieve shared idealistic goals.
In this context, the term hybridity accounts for the phenomenon that organizations combine behavioral patterns from more than one sector of the trichotomy (Smith and Besharov 2017). Inspired by the developments of new public management, some studies in public administration see hybrids as combining “a public orientation with a market orientation” (Joldersma and Winter 2002, 84). Christensen and Lægreid (2011, 410) include the nonprofit sector, describing hybrids as “quasi-governmental organizations that exist at the interface between the public and private sector, which may be either market or civil society organizations.” Billis (2010, 3) equally specifies that hybrid organizations “possess “significant” characteristics of more than one sector (public, private, and nonprofit)” (similar, Brandsen, van de Donk, and Putters 2005, 750; Denis et al. 2015; Seibel 2015). Examples of hybrid arrangements can be found in health care, social housing, education and others, typically in the form of public–private partnerships, state-owned enterprises, or government-funded think tanks (Brandsen et al. 2005). Hybrids are essentially fit-for-purpose entities, which “experiment with different combinations of features to come up with something that will achieve its objectives” (Rhodes and Donnelly-Cox 2014, 1637).
Whereas hybrid organizations are usually conceived of as permanent and stable settings (Besharov and Smith 2014; Mair, Mayer, and Lutz 2015), a few studies noticed that hybridity can also prevail as a latent, temporary or ad-hoc phenomenon. For instance, Mintzberg and McHugh (1985, 162) coined the term adhocracy as the “formation of emergent strategies” in contrast to the routine bureaucracy. Skelcher and Smith (2015, 436) describe hybridity as a “short-lived exceptions to the norm,” and Seibel (2015, 689) argues that “actual hybridity may originate from informal rather than formal arrangements.” Matinheikki, Aaltonen, and Walker (2019, 300, emphasis added) add the temporal dimension, defining “hybridization as a change process through which organizations aim to transfer from one organizational settlement (i.e., configuration of structural and cognitive elements of organizing) into a new one.”
Such an ephemeral understanding of hybridity is particularly suitable for the crisis context, when public sector organizations must adjust for a limited period of time, before they can return back to the status quo ante. We therefore use the term latent hybridity to account for the phenomenon that administrations adjust their behavioral practices in a crisis situation to cope with the new demands, to become fit-for-purpose. Similar to Starkey, Barnatt, and Tempest (2000, 300), who spoke of latent organizations as “emerging forms of organization,” we use the term latent in the sense of “existing but not developed or manifest” (Oxford English Dictionary) to delimit our concept from the typical understanding of hybridity as a formal and stable setup. We therefore define latent hybridity as informal and temporary shifts from routine administrative behavior to crisis action along behavioral principles that typically characterize the private and nonprofit sector, respectively.
Latent Hybridity as Changes Towards Flexibility and Participation
For public sector organizations, behavioral adjustments during a crisis are functionally necessary because the design and processes of public sector organizations typically aim to generate public legitimacy through predictable decision-making. The focus lies on hierarchy and the rule of law, which ensures impersonality and equal treatment (Weber 1922). Problematically, though, associated practices tend to undermine effective crisis management: “Serious threat, un-certainty and acute time pressures are conditions adverse to routine processing of information, compliant behavior, and functionally divided responsibilities” (Rosenthal et al. 1989, 2001, 212). By contrast, effective crisis management requires organizational adaptation: internally, to challenge and overcome standard operating procedures, and externally, to mobilize additional resources. As summarized in figure 1, flexibility and participation—the dominant behavioral principles of the private and nonprofit sector—promise to serve such ends.
Conceptualization of Shifts Towards Latent Hybridity in Administrative Action During a Crisis Response.
Compared to the public sector, private sector organization are highly flexible in aligning their structures and processes to the primary goal of survival and competition with other firms. Their behavioral principle is flexibility. Depending on the particular functional need, private sector organizations adapt decision authorities and discretion, staff recruitment, and operating procedures. In public administration, the term flexibility has often been used as the antithesis to rule-bound operations, red-tape, or regulatory constraints (Brodkin 1997; Feeney and Rainey 2009). In particular, literature on street-level bureaucracy and bottom-up policy implementation equalize flexibility with discretion and responsiveness to local needs (Brodkin 2011; Keiser 1999; Pires 2011). Such discretion again has been linked to greater motivation by public officials to achieve policy goals (Thomann et al. 2018). Conceptually, two levels of flexibility can be distinguished. At the level of the employee, flexibility refers to decision latitude; at the level of the organization, flexibility refers to latitude in standardized procedures and the use of resources to adapt to changing environmental demands (Dettmers, Kaiser, and Fietze 2013; Reilly 1998). As we discuss below in more detail, administrative crisis management should benefit from shifts towards more flexibility at both these levels. As Christensen et al. (2016, 887) argue, “[f]lexibility and adaptation are key assets” for a well-functioning governmental crisis management system.
Organizations belonging to the nonprofit sector are not dependent on profit maximization and adaptation but are driven by socially meaningful mandates. These organizations typically aim to enable voluntary participation to achieve their idealistic ends via communication, openness, and integration. Public administration literature on co-production and participatory governance sees such integrative elements as a means to ensure that policy reflects the preferences and needs of citizens (Nabatchi 2012). Recent studies also emphasize the instrumental value: “public managers can design participatory processes to better achieve desired outcomes” (Baldwin 2019, 1; Thomsen and Jensen 2019). But while the term collaborative governance denotes a “collective decision-making process that is formal” (Ansell and Gash 2008, 544, emphasis added), latent hybridity emphasizes ad-hoc and informal ways of participation that emerge in a crisis response. Indeed, past crises have often triggered significant momentum for voluntary societal help, which is a resource public agencies must be able to absorb (Schmidt 2019). Nonprofit organizations typically rely on grass roots engagement and self-organization to enable large numbers of volunteers to contribute to the crisis mitigation efforts (Scanlon, Helsloot, and Groenendaal 2014). Such behavioral patterns should thus enable the administration to better communicate and interact with the public in general and with volunteers in particular.
Overall, we expect that temporary shifts towards more flexible and more participatory action during a crisis response should enable agencies to better deal with the specific politico-administrative challenges that characterize a crisis situation. Note that the starting point is not necessarily an extreme case Weberian administration, as agencies can already have incorporated elements of the other sectors prior to a crisis impulse.2 We next explain how exactly shifts towards more flexible and participatory action should impact the ability to manage a crisis.
Latent Hybridity to Cope With the Internal and External Challenges of Crisis Management: Insights From the German Refugee Crisis 2015/16
In crisis management, administrations face challenges regarding their internal operations and their external interaction with the outside world, as Ansell et al. (2010) wrote in one of the most cited papers in crisis management literature (see also, Boin et al. 2018; Christensen et al. 2016; Webb and Chevreau 2006, 66). Compared to administrative action in routine times, changes towards more flexible internal operations (private sector) and towards more external participation (nonprofit sector) should enable administrations to cope more effectively with crises. Drawing on crisis management literature as well as interview and survey data on administrative action during the German refugee crisis, collected by a polling institute between November 2015 and January 2016 (IfD 2016),3 this section discusses the nature of the internal and external administrative challenges of the 2015/16 refugee crisis in Germany and how flexibilization and participation served as coping mechanisms.
In Germany’s decentralized political system, administrative emergency and disaster management responsibilities are primarily allocated at the level of counties and district-free cities (Landkreise and kreisfreie Städte, henceforth district level) (Kuipers, Boin, Bossong, and Hegemann 2015). Between 2014 and 2016, around 1.5 million individuals entered the country (BAMF 2020, 5). Each month throughout the peak year 2016, 62,500 refugees needed to be registered and accommodated. In each of Germany’s 401 districts, around 155 newcomers arrived per month (with significant variation between districts depending on their population size).4 As one local official explained at the time: “These are growth figures which we had maybe during the war but never since” (IfD 2016, 35).5 Overall, each district received 3,832 asylum seekers on average in 2016 (Min: 155; Q1: 1,422; Q3: 3,985; Max: 82,490) (Destatis 2016). The main services provided by local administrations include the registration of incoming refugees,6 health checks, and the provision of food, accommodation and general living supplies, as well as more individual solutions for unaccompanied children and those traumatized and sick (Bogumil, Hafner, and Kastilan 2017; Speth and Becker 2016).
Internal Challenges and More Flexible Operations
Just as described by Ansell et al. (2010), this situation led to internal administrative challenges: the necessity to “cope with uncertainty” (Ansell et al. 2010, 197f.) about the potential evolution of the crisis; and the requirement to “organize a response” (Ansell et al. 2010, 198f.) in such an uncertain context. District administrations clearly experienced these challenges, including both uncertainty about the number of incoming refugees and an abundance of associated organizational challenges (Gesemann and Roth 2016; IfD 2016). Specifically, the provision of temporary shelter proved challenging because there was simply not enough living space for every newcomer, especially in more urban districts in Western Germany. One senior district official testified at the time that his or her administration was “stretched to its limits” (IfD 2016, 12). Another said: “We permanently face a critical situation. We do what we can to provide accommodation to refugees assigned to us. But it is difficult to get ahead of the situation, to produce some stockpile.” (IfD 2016, 12).
Flexibility at the individual and organizational level appears promising to cope with such internal challenges. Three specific areas of flexibilization stand out. The first is that administrations must flexibilize decision procedures at the individual level by decentralizing discretionary powers and allowing for pragmatism with regards to existing law in the crisis context (Christensen et al. 2016; Hahlen and Kühn 2016; Meyer 2016). In Germany, officials mentioned the discrepancy between the “normal” structure and duration of administrative planning procedures and authorization processes on the one hand, and the need to act fast on the other, as one of the most pressing challenges (IfD 2016, 18). In an early survey among administrative officials in 2016, 80% said that the legal framework obstructed their response (IfD 2016, 50). But facing the demands of the crisis, officials coped by enhancing discretion for employees and by applying a more “flexible” understanding of the law. One district commissioner testified:
“You are indeed walking on thin ice if you are building an accommodation right now. I told my staff, all legal regulations are gone, except fire safety…it is still better to circuit proper German building laws for a while rather than to have people camp in tents in the winter” (IfD 2016, 50).
Moving to the organizational level of flexibility, a second major challenge in crisis situations is the necessity to make the use of personnel resources more flexible, either through recruitment or by redirecting and coordinating existing personnel (Ansell et al. 2010; Meyer 2016). Indeed, in the German crisis, local-level officials mentioned the lack of personnel resources as another highly pressing challenge (Gesemann and Roth 2016; IfD 2016, 18). To cope with the shortfall, districts lowered legal requirements for employment, speeded up the allocation of resources for new positions, asked staff to work extra hours, and shifted staff internally, for example, by temporarily borrowing from divisions that were under less pressure. Already in early 2016, counties had hired 35 new employees on average (82 in larger cities), pushing the number of staff concerned with the refugee crisis to 92 (212 in larger cities) on average (IfD 2016, 31).
Finally, studies typically highlight the need for creativity and innovation with respect to standardized operational processes as administrations must develop coping strategies for the extraordinary situation (Christensen et al. 2016; Schuppan and Köhl 2016; Webb and Chevreau 2006). This was also the case in German districts, which appeared flexible in adjusting or establishing new structures and processes. For instance, interviewees mentioned new coordinative structures, such as “asylum coordinators” who were created in a large number of districts to coordinate between the district administration and the large number of other offices at district level, such as the welfare agency, health department, immigration office, and the housing department, among others. This multi-actor structure of the crisis required a particularly well organized and coordinated response (Bogumil et al. 2017, 19ff.). Districts also came up with new processes designed to handle the processing of refugees. One senior official explained that her or his district invented a new “relocation management,” designed to speed up the transfer of people from collective accommodations to regular housing (IfD 2016, 20; see also, Ruge et al. 2016).
External Challenges and More Participation
As discussed by Ansell et al. (2010), administrations must also manage distinctive external challenges in a crisis. This includes to “gather surge capacities” (Ansell et al. 2010, 198) in terms of finances, equipment, supplies, and human resources, and to “communicate with the public” (Ansell et al. 2010, 200). Both these challenges were also critical in the German case, with districts responding with a range of participatory measures.
As a first response, studies emphasize that communication with the public is particularly important in crisis situations to establish an official storyline as opposed to rumors or half-truths that spread through social networks (Boin et al. 2018). In Germany, 90% of all local officials who answered an early survey perceived that “anxieties and sorrows” prevailed among the population in their districts, albeit mostly in relation to the “general development” rather than the “local situation in their district” (IfD 2016, 40). In response, they emphasized the importance of engaging closely with the general public, by interacting with civil society organizations, but also by doing more public relations work. One district commissioner said that anxieties usually disappear once “one engages in conversation” (IfD 2016, 45). Another explained about his or her response:
“We need to do very, very much public relations work. Very, very much enlightenment. I personally attend town hall meetings quite frequently, also in cities and communities, and present the situation myself, as a “matter for the boss” [Chefsache], and my staff even more so. I have one employee who is hoofing from one city council meeting to the other. Very, very much enlightenment.” (IfD 2016, 46).
Partly linked to these communicative responses, administrations in a crisis response can seek to integrate volunteers as a surge capacity (Boin et al. 2010). Surge capacities can, to some extent, be gathered internally, such as through staff mobilization. But crisis situations also tend to create significant societal willingness to help. Such voluntary resources must be absorbed by the administration, which mostly takes place through ad-hoc problem solving and spontaneous network coordination (Beck 2016; Ferguson, Schmidt, and Boersma 2018; Whittaker, McLennan, and Handmer 2015). In the German refugee crisis, public willingness to help was indeed overwhelming. In a survey among officials in 2016, 86% said the work done by volunteers was “very crucial,” 14% said it was “crucial” (IfD 2016, 35). Most importantly, voluntary work and donations served to fill financial and personnel gaps on the side of the administration. Drawing on narrative accounts, services provided by volunteers included consulting and support with asylum and integration issues, translation, childcare, and numerous auxiliary tasks; donations included food, clothes, toys, furniture, and other equipment. Officials reported that these services were “invaluable” (IfD 2016, 36), that “without volunteers we would not have been able to cope with the situation” (IfD 2016, 35) and that the absence of donations and assistance would have meant a “significant additional financial burden” (IfD 2016, 36). In order to successfully incorporate these volunteers, an enormous coordination effort was required by the district administrations (Gesemann and Roth 2016). Many districts created regional action alliances [Aktionsbündnisse] to establish linkages with relevant private initiatives and welfare organizations (IfD 2016, 20; Ruge et al. 2016). Other districts employed volunteer coordinators [Ehrenamtskoordinatoren], to represent the citizen perspective and to enable smooth information flows in both directions (Ruge et al. 2016).
As these accounts show, there are indeed a number of informal and temporary adjustments in administrative action in response to the crisis challenges. In the German refugee crisis, flexible problem-solving seems to have become an important internal principle, superimposing the logic of hierarchy that is traditionally dominant in public administration. Externally, administrations seem to have enhanced means for citizen participation. This style of hybridization can be described as latent because it does not rely on stable and formal long-term structures and arrangements but rather on informal features, such as discretionary leeway and short-term collaborative arrangements.
The Impact of Latent Hybridity on the Output Side of Administrative Action: Hypotheses
This section develops hypotheses for a more comprehensive assessment of the theoretical argument. In line with the latent hybridity as a means to become fit-for-purpose, we expect a discernable impact on the output-side of administrative action. For the purpose of this study, we capture such effects by studying administrative effectiveness. Effectiveness at the level of administrative action or organizational macro-behavior refers to the “organizational ability to attain the goals set by itself, the organization’s ability to function well as a system and the organization’s ability to satisfy its stakeholders” (Donaldson 2001, 6). These considerations inform our two main hypotheses:
H1: The more an administration becomes flexible in its internal operations pursuant to a crisis impulse, the better its ability to effectively manage a crisis response.
H2: The more an administration becomes participatory in its outside interaction pursuant to a crisis impulse, the better its ability to effectively manage a crisis response.
Furthermore, shifts in administrative action should constitute gradual adaptations, that are interconnected with the severity of crisis exposure. Consequently, it is only natural that not all administrations feel the same pressure to shift their crisis management behavior to a high extent, especially if they are not equally affected by the severity of the crisis. Hence, the relationship established in H1 and H2 should be moderated by the crisis impact. In other words, the positive effect hybridity should have on crisis management effectiveness should be more pronounced if an administration is hit harder by the crisis than in areas less severely affected. H3 and H4 capture this expectation separately for the two variables flexibility and participation.
H3: Administrative flexibility improves the ability to effectively manage a crisis response most strongly in counties where the severity of the crisis is highest.
H4: Administrative participation improves the ability to effectively manage a crisis response most strongly in counties where the severity of the crisis is highest.
Case Selection and Operationalization
We study the increased influx of asylum seekers to Germany in 2015/16 at the level of counties and district-free cities. This situation has been characterized as featuring the typical crisis characteristics, mostly because routine behavior was not adequate for a well-functioning response (Hahlen and Kühn 2016; Meyer 2016). Compared to more frequent emergencies, for example, local natural hazards or accidents, for which contingency command-and-control-structures exist, appropriate response mechanisms to deal with the massive influx of refugees did not exist in Germany at the time (Bogumil et al. 2017; Schuppan and Köhl 2016). The unit of analysis is the county and district-free city-level, with a total of 401 administrative entities in Germany. This is the administrative level above local communities but below the 16 German Bundesländer. Overall, the literature accentuates the good performance of local administrations in the crisis, essentially claiming that they prevented “state failure” (Bogumil et al. 2017). However, administrative actions and response effectiveness still varied across the country.
Survey on Administrative Action During the Crisis Response
Data on both dependent and independent variables were gathered using a questionnaire-based online survey distributed in 2018 among county and district-free city administrations. The survey data was collected between June and August 2018. Overall, the sample comprises the results of 235 counties and district-free cities (a response rate of 58.6%), more or less equally distributed among the Old and New Länder (respectively 53% and 59%).
With the survey approach, we follow one of the most frequently used measures of administrative action and performance, which is self-evaluation by public service managers (Emery and Giauque 2003; Lægreid, Roness, and Rubecksen 2006; Worrall, Cooper, and Campbell-Jamison 2000). The logic for using such a measure is that public servants, especially heads of agencies, “can be considered as experts and have first-hand and in-depth knowledge” of the performance of their agency (Andrews and Van de Walle 2013, 766). Facing a trade-off between collecting many responses from few agencies and covering all agencies, we opted for the latter and sought one survey response from a senior official of each administration (district commissioners, directors, the asylum coordinator, or similar).7 To avoid biases in responses due to social desirability, the survey invitation emphasized that responses would be treated anonymously and analyzed only in aggregated form. We collaborated with the German district administration association (Deutscher Landkreistag)—the association recommended their members participation in the survey—to increase both the response rate and respondents’ trust in data handling. To find the best way to measure variables of interest, we relied not only on academic literature, but also on a series of expert interviews and expert workshops. Participants were several senior district officials, a representative from the Landkreistag, representatives from nonprofit organizations, and academics.8
Operationalization of Variables
We measured the independent variables and the dependent variable with the same survey, following recommendations mentioned in the literature to avoid common-method biases (e.g., Spector 2006). To avoid common-method variance between dependent and independent variables (Spector 2006), we developed an index that combined observable items for each of the conceptualized deviations from routine practice towards crisis response. The latent construct changes in flexibility consists of seven items (v1; Cronbach’s α = 0.73), and the latent construct changes in participation consists of five items (v2; Cronbach’s α = 0.75). All are scored on a scale from 0 (least flexibility / participation) to 10 (most flexibility / participation). The specification and selection of items was based on input received during expert interviews and the workshop mentioned above, and literature on administrative action in crisis management. Figure 2 provides an overview of both independent variables and the 12 survey items.
Operationalization of the Independent Variables. Note: Own illustration based on the survey items used to operationalize and measure latent hybridity in crisis management. Full questions can be found in the Supplementary Appendix.
Changes in flexibility (v1) refers to shifts towards principles more dominant in private sector, where we identified internal decision-making procedures, the mobilization of administrative staff, and innovative processes as most important aspects to tackle the internal challenges of crisis management in the conceptualization above. The index is composed of seven items, two of them referring to individual level flexibility, five of them to organizational level flexibility: First, to enable employees to cope more flexibly with crisis challenges, administrations can adjust their internal decision-making procedures by invoking changes in decision authority (1.1) and by increasing discretionary power (1.2) for executive officials. Second, at the level of the organization, we distinguish flexibilization in the use of personnel resources (enhanced internal staff mobility (1.3) and the facilitation of new recruitment (1.4)), and the flexibilization of standardized processes through development of adequate coping strategies (1.5), swift organizational adaptation (1.6), and innovative processes and procedures (1.7).
With regard to changes in participation (v2), we used items capturing the extent to which the administration communicated with the public and the extent to which volunteers were integrated in the crisis response. The first three items refer to communication. We measured changes in administrative communication efforts via three items: intensification (2.3), prioritization (2.4), and immediacy of communication (2.5). The other two items refer more directly to the participation of volunteers in crisis management. Our survey items to capture these dynamics are increased cooperation with volunteers (2.1) and increased integration of volunteers (2.2). The two indicators were formulated in a very general way to comprise the wide spectrum of formal and informal activities and relationships with different civil society actor groups, including spontaneous volunteers as well as those with closer ties to welfare organizations (for more details on the items, see Supplementary Appendix C.1–3).
Similar to previous studies on administrative performance or effectiveness (Emery and Giauque 2003; Lægreid et al. 2006; Worrall et al. 2000), we measured the dependent variable as district commissioners’ perceived effectiveness in the crisis response. Although the concept of effectiveness is “grounded in the values and preferences of evaluators” (Cameron 2015, 1), there seems to be a coherent understanding among practitioners who linked it to the ability to prevent state failure and avoid extreme forms of distress in internal operations (Bogumil et al. 2018). Such an understanding is well in line with our definition of goal attainment and the ability of the organization to function well (Donaldson 2001), as discussed above.9 Given that the term effectiveness in itself is most intuitive to capture the concept, we used a single question asking about the overall effectiveness of administrative action in the refugee crisis. The variable was coded from 0 (not effective at all) to 10 (extremely effective) with a mean of 8.07 and a left-skewed distribution. The overall rather positive self-evaluation of district managers is in line with the positive performance assessment reported by other studies on local administrative action in the German refugee crisis (Bogumil et al. 2017; Hafner 2019; Meyer 2016). For the analysis, we centered the distribution by the range and then used the squared values to account for the more pronounced effects at the extremes (cf. Supplementary Appendix D.2 and D.3).
In addition to the effect of changes in flexibility and participation on effectiveness (H1 and H2), we also hypothesized more pronounced effects of hybridization in counties that were confronted with more severe forms of crisis affectedness (H3 and H4). We used the relative share of protection seeking persons residing in the district as a proxy for the interaction variable, ranging between 0.2% and 4.3% with a mean of 1.7% (record day 31 December 2016). The data was provided by the German Federal Statistical Office (Destatis). In cooperation with the Federal Office for Migration and Refugees (BAMF) and the Central Register of Foreigners (AZR) Destatis gathers information of all persons seeking protection recognized by the German local immigration authorities, enabling evaluations with breakdowns to regional and local levels (Destatis 2016). Furthermore, we included as control variables the GDP per capita, tax, and debt level of the counties, which were provided by the German Federal Statistical Office (Regionalstatistik 2018). The Central Register of Foreigners (AZR) publishes data on the relative share of asylum seekers, unemployment rates, and foreigner ratios (Destatis 2016), and the Federal Office for Building and Regional Planning provided data on the differentiation between district types according to settlement structures (BBSR 2018).10
Results
Firstly, we find observable variation in administrative action among counties. Depicted along the dimensions flexibility and participation in administrative action, the results show that some administrations became significantly more flexible and participatory than others did. In addition, variation in administrative action was observed with respect to the number of asylum seekers entering counties in East and West Germany (figure 3).
Between District Variation With Regards to Changes in Flexibility and Participation. Note: The scatterplot shows how the different counties are positioned along the two dependent variables changes in flexible and participatory behavior. Participation and flexibility are standardized, indicating the mean at 0, while positive and negative signs indicate whether there is more or less latent hybridity compared to the mean of all districts. Eastern and Western states are depicted with different forms, while the color indicates how severely the counties were affected by the crisis (measured via the standardized relative shares of asylum seekers, indication standard deviations from least affected (negative) to most severely affected counties (positive). Data = HybOrg Survey, N = 231.
Next, we performed regression analyses to analyze the effect of variation in hybridization during the crisis on the effectiveness of the administrative response. In line with H1 and H2, more flexible and participatory action were the main explanatory variables in the models.11 We calculated several models, including only the control variables (Model 0, only reported in Supplementary Appendix), the explaining variables plus several control variables (Model 1), and the two interaction models for the interaction effect of more flexible action and relative share of asylum seekers (Interaction Model 1), as well as more participatory action and relative share of asylum seekers (Interaction Model 2). Among the controls were counties’ tax revenue per capita, public debt level per capita, GDP per capita, the share of staff in the public sector, foreigner ratios, unemployment rates, and the differentiation of the district types (district-free cities, urban districts, rural districts, sparsely populated districts). Following Gelman and Hill (2006), we standardized all numerical explaining and control variables.12 As for the goodness of fit, all models showed statistical significance of the overall model (significant F-statistics with rejectable F-tests), and quite good Adjusted R2 (Model 1: 0.227 / Interaction Model 1: 0.232 / Interaction Model 2: 0.223), accounting for the variance that was explained by the model, while penalizing the number of terms. Hence, the models explained about 22% of the variance in effective crisis management (full regression output in the Supplementary Appendix E.1).
Figure 4 presents the coefficient plots for all models estimating the hypothesized effects of changes in administrative action on the effectiveness of the local crisis response. Additionally, to the same structural control variables, Interaction Models 1 and 2 include the interaction terms with the relative share of asylum seekers as moderating variable.
Coefficient Plot for Model 1 and Interaction Models 1 and 2. Note: The full regression output can be found in Supplementary Appendix E.1; in all models: N = 231.
Model 1 tested the effects without interaction terms, allowing a straightforward interpretation of the regression coefficients in order to find an answer to the first two hypothesized effects (H1 and H2). The coefficients for changes in flexibility and participation were both significant according to the t-tests of the regression output and the coefficient plot (figure 4). The predictors indicated that more flexible and more participatory administrative action both showed positive effects on the effectiveness of crisis management. As all predictors were standardized, this means that each unit increase in the standard deviation of changes in flexibility, resulted in a positive difference of 1.04 ) on effectiveness, ceteris paribus. For changes in participation, the coefficient indicated a positive difference of 1.56 () on the effectiveness for each increase in units of standard deviations of participation, ceteris paribus.13 Consequently, our analysis supported the first two hypotheses, as we observe more flexible and participatory administrative action leading to a more effective crisis management.14 Additionally, the control variables unemployment and district type showed some significant results, indicating that higher unemployment rates in a district have a negative impact on effectiveness, while less populated districts generally perceive their effectiveness to be higher than in urban districts, which is the reference category in the analysis. This finding is plausible because urban districts have higher population shares and thus received more asylum seekers as discussed above. Because administrative resources in a district are not linearly linked to population size, this seems to have amplified the impact of the crisis in urban districts.
Next we present the interaction models in more detail. In the theoretical section, we hypothesized that latent hybridity in crisis management is especially relevant when the administration is most severely affected by the crisis (H3 and H4). We expected that counties with the highest shares of asylum seekers relative to the population should have the largest changes towards more flexible and participatory actions for effectively managing the crisis response. In order to measure this moderating effect of the relative share of asylum seekers on the effects of latent hybridity on effectiveness, the two interaction models (Interaction Models 1 and 2) were included. The interaction models used the same variables as Model 1 but added the respective interaction effects between the changes in administrative action and the relative share of asylum seekers as additional covariates. We also plotted the marginal effects for those interaction effects, as statistical significance itself does not sufficiently inform the actual associations of the terms (Berry, Golder, and Milton 2012; Brambor, Clark, and Golder 2006).
In the graphic depiction of the marginal effects for the Interaction Models 1 and 2 (figure 5), we observe a significant moderating effect of the share of asylum seekers on the effect of changes in flexibility on effectiveness. In other words, the higher the share of asylum seekers, the more positive is the coefficient of flexibility with respect to the effectiveness of the crisis response. Figure 5 shows these predicted values of effectiveness for changes in flexibility when the share of asylum seekers is low (blue), average (red), and high (green), for better visibility and more straightforward interpretation.
Plots for Interaction Models 1 and 2. Note: The full regression output can be found in Supplementary Appendix E.1; both models N = 231.
What becomes more easily observable in figure 6 is that changes in flexibility have especially positive effects on the effectiveness in those counties with high shares of asylum seekers, as the green line is the most pronounced. Simultaneously, the effects were the least noticeable in counties with low shares of asylum seekers. Our analyses therefore support hypothesis 3, that flexible administrative action improves the ability to effectively manage a crisis response most strongly in counties where the assigned shares of asylum seekers were the highest.
Regression Lines for Different Values of Shares of Asylum Seekers. Note: Regression line for changes in flexibility and effectiveness in Model 3. Share of asylum seekers seperated for: median, +1 standard deviation, −1 standard deviation. N = 231.
However, the same effect was not observed for Interaction Model 2 regarding more participatory administrative action. The marginal effects (figure 4) show that the coefficient of changes in participation on effectiveness remains constant across the different relative numbers of asylum seekers in counties. Hence, no conditional effects were observed in the second interaction model, and hypothesis 4, that participatory administrative action improves the ability to effectively manage a crisis response most strongly in counties where the severity of the crisis is highest, has to be rejected.
Discussion
Our analysis showed that German counties responded with certain forms of latent hybridization to cope with the crisis situation of 2015/16. We could also show that these changes applied to varying extents in different counties. We expected that changes towards more flexible and participatory administrative action—what we conceptualize as latent hybridity—should be associated with higher values regarding the effectiveness of crisis management. We measured concepts on the side of independent and dependent variables with a survey instrument. There are some limitations in the extent to which such perception data correspond with the theoretical quantities of interest, which we address in more detail below. Bearing such challenges in mind, empirical findings support our main expectations. In line with hypotheses 1 and 2, we were able to show the positive effects of latent hybridity on effective crisis management: When district administrations enacted changes towards more latent hybridity, administrative respondents reported higher levels of effectiveness of their crisis management activities.
Going beyond survey data alone, the analysis also accounts for interaction effects with the severity of the crisis, that is, how many refugees needed to be processed by each district administration. This followed the expectation that administrations adjust gradually, depending on crisis impact. Our measure was the relative share of protection seeking persons per district. The analysis supported hypothesis 3, that flexible administrative action improves the ability to effectively manage a crisis response most strongly in counties where the severity of the crisis is highest. Another relevant result was the insignificant interaction effect of participatory action and crisis severity and the rejection of hypothesis 4.
Our measurements bear certain limitations that should be recognized. The first is the use of perception data in general, which is however without alternative given our interest in internal administrative procedures. A further limitation is the low number of perceptions collected from each entity. Facing a practical trade-off between surveying few administrations in great detail by collecting many perspectives and surveying many administrations but collecting only one response, we opted for the latter by addressing only the district leadership, such as head of counties or heads of units. We controlled for biases in their responses but could not find significant effects of the respondent’s position within the administration (Supplementary Appendix F.1).
Another challenge is that the dependent variable measure asks directly about respondents’ perceived effectiveness in crisis management. The disadvantage is the possibility of overinflated answers due to strategic answering. The advantage however is a simple concept that respondents could intuitively understand. In expert interviews and workshops, it became clear that the few available alternative proxies also have limitations. Local media reporting or population perceptions, for instance, are prone to political biases and outside observers are also unable to grasp the internal dimensions of effectiveness (Donaldson 2001). Citizen evaluation data in particular run to all other sorts of problems when evaluating administrative effectiveness, particularly their dependence on demographic factors and citizen expectations (Morgeson and Petrescu 2011). Consequently, asking officials directly about the perceived effectiveness remains a valid strategy to capture the output side of administrative action in a crisis response. After all, public servants are experts and work with an appropriate basis of background knowledge to evaluate the performance of their agency (Andrews and Van de Walle 2013, 766). The testimonies cited in Section “Latent Hybridity to Cope With the Internal and External Challenges of Crisis Management: Insights From the German Refugee Crisis 2015/16” further indicate that district officials in Germany are quite frank in their assessments, even when responses are recorded and transcribed by a polling institute.
A final limitation is about case selection. While we deem the theoretical concept as relevant for other cases/ crisis situations, the exact items used to measure the theoretical constructs (seven items for flexibilization and five items for changes in participation; see figure 2) were developed for the case of the German refugee crisis of 2015/16 and based on input by practitioners. There might be restrictions regarding the generalization of these items for other types of crises or for other countries.
Despite such limitation, findings provide no reason to reject the main theoretical propositions. This conclusion derives from the empirical regression analysis in association with a thorough review of the literature, including narrative expert accounts collected during the crisis (IfD 2016). Many qualitative studies in crisis research describe administrative action in crisis management with features similar to those included in our concept. Our conceptual contribution is to link such single observations within one concept that draws on foundational ideas of organizational behavior. The operationalization of main concepts may bear limitations, but it also lacks real alternatives. For now, the finding is that the empirical observations correspond with our main theoretical expectations. We conclude that the beneficial effect of internal flexibilization (changes in decision authority, staff recruitment, and operational processes) and enhanced outside participation (changes in communication and interaction with volunteers) in administrative action should also bear positive effects in related crisis situations beyond the German refugee crisis of 2015/16.
Another implication from the data analysis is that we must better understand the mechanisms linked to citizen participation in crisis management. The null-effect observed for hypothesis 4 means that the “surge” potential of volunteers did not play out in counties most affected by the crisis. One explanation for this finding is a potential trade-off linked to the inclusion of volunteers. Participation by volunteers is also associated with high organizational costs and losses in efficiency (Moynihan 2009): Promotion, networking, and coordination of volunteers require administrative resources, and in Germany, many counties needed to develop such skills from scratch (Beck 2016; Raffer and Kairies-Lamp 2016). This means that in counties that were previously overwhelmed by the crisis situation, the administration may have been unable to absorb voluntary engagement in such a way that it served as surge capacity and improved the effectiveness of the crisis response. An alternative explanation could be that civil engagement developed independent of the counties’ asylum seekers numbers. This would dissociate the additional administrative effort associated with managing high volunteer numbers from the actual number of asylum seekers those counties dealt with, and thus the severity of the crisis. Citizen demand for participatory administrative action, including collaboration, coordination, and communication may have become relevant in all counties, and not just in those affected by high shares of asylum seekers. More research is necessary to understand these issues, including the question of whether administrations can better prepare for citizen engagement during crises.
Conclusion
Measuring and evaluating the impact of variation in administrative behavior on crisis management performance is difficult. So far, quantitative data on administrative crisis management is scarce. The first contribution of this article is the presentation of comparative survey data on administrative action in 235 out of 401 German counties and district-free city authorities (59%), each acting upon the same crisis impulse of a high number of asylum seekers who entered the country in 2015 and 2016. The data showed that there was variation between district administrations, both in terms of the relative share of asylum seekers that they processed, their changes in flexibility and participation, and how managers described the effectiveness of the crisis response.
Theoretically, we propose that the effectiveness of administrative crisis management depends on administrations’ ability to switch practices. Our assumption was that in order to cope with dynamic and complex crises, administrations adjust by mixing or replacing established bureaucratic routines with new procedures and practices drawn from the private and the nonprofit sector. Such shifts are functionally driven because routine operations by the public sector tend to undermine effective crisis management (Rosenthal et al. 1989, 2001). We conceptualized these behavioral shifts as latent because they are informal rather than formal, and temporary and ephemeral rather than standardized and sustainable. The term hybridity highlights that the new type of latent behavior combines significant characteristics of more than one sector. Bearing in mind limitations of the measurement, the empirical finding is that district administrations indeed became more flexible to better cope with the uncertainty of the crisis, to speed up decisions processes, and to organize a more creative response. They also fostered participatory methods to improve communication with the public and to integrate volunteers as a surge capacity. These changes were positively associated with the perceived effectiveness of the crisis response.
Critical readers may question the significance of this finding, given that flexibility and participation are oftentimes seen as generally desirable features of administrative action. However, there is a clear trade-off linked to hybridization: To the extent that public organizations depart from Weberian principles, they may become more prone to arbitrary action or fraud because flexibilization goes along with reduced oversight and accountability. For instance, a number of anecdotal and media reports about problematic cases of administrative procurement in the refugee crisis have become public up until now. They refer to the acceptance of advantages by civil servants in exchange for preferential treatments of bids by certain service providers, such as private security or manufacturing companies.15 While these may be exceptions only, the possibility of such trade-offs emphasizes the necessity to shift back to a more Weberian operational principles after the crisis has ended.
Our findings offer new insights about the nature of administrative responses to crisis situations, emphasizing the ability of the organization to adjust and to shift practices as a critical trait for effective crisis management (e.g., Ansell et al. 2010; Boin et al. 2018; Christensen et al. 2016). More broadly speaking, this finding resonates with others who emphasize the importance of organizations’ ability to switch between practices or phases, to become more responsive or agile (Mergel et al. 2020; Schakel et al. 2016). Speaking to bottom-up implementation and street-level bureaucracy, our findings confirm the importance of discretion, especially when dealing with complex policy problems or exceptional situations (Brodkin 2011; Keiser 1999; Pires 2011; Thomann et al. 2018). Future research may benefit from further exploring the processes that link individual discretion with flexibility at the aggregate level of organizational action. Furthermore, more ambiguous findings on participation highlight the instrumental value of participatory governance (Baldwin 2019; Batory and Svensson 2020; Thomsen and Jensen 2019), but also that there is need for more nuanced research to understand the associated mechanisms and potential trade-offs.
This research also has practical implications for the way administrations should think about their crisis preparedness. Instead of planning for certain types of crises or crisis scenarios, as currently the focus of both administrative practice and crisis management, more attention should be payed to features enabling shifts in administrative practice. Our research provides insights about the dimensions and sub-dimensions that may be relevant for such shifts and that may be beneficial for organizational learning and crisis preparedness. Furthermore, while participation as such had a positive effect on effectiveness, we observed no differences between highly and lowly crisis-affected counties in Germany. One practical implication may be that administrations should try to incorporate volunteers early (before the crisis) to avoid the problem of overwhelming district administrations in times of stress.
Data Availability Statement
The data underlying this article cannot be shared publicly because the data collection process warranted respondents’ anonymity while the covariates in the dataset reveal the identity of counties. The data will be shared on reasonable request to the corresponding author. The Supplementary Appendix provides detailed information on the data collection process and data analysis.
References
———.
———.
The authors are grateful for instructive guidance by editor Mary K. Feeney during the review process and would like to express their thanks to the reviewers. We also thank colleagues at the University of Munich and Konstanz, as well as participants of respective 2019 ECPR and 2019 EGPA panels for their feedback. Mark Bovens, Paul t’Hart, and Arjen Boin were so kind as to host us for a workshop at the Utrecht University School of Governance (USBO) and commenting on our work, which we gratefully acknowledge. Finally, we thank our partners from the German Red Cross, the German County Association, and several county and city administrations for their support during the research process. This article is a product of the HybOrg research project (https://www.hyborg-projekt.de/en/), funded by the German Federal Ministry of Education and Research. The lead authors of this article are Steffen Eckhard and Alexa Lenz. Wolfgang Seibel, Florian Roth, and Matthias Fatke are co-leaders of the research project and provided input on project outline, theory, survey design as well as the article. We also acknowledge important support by Lorenz Wiese, the project coordinator, as well as invaluable research assistance by Franziska Graf, Laurin Friedrich, Emily Chmiel, and Darja Ljubownikow.
Footnotes
With administrative action, we refer to the macro-behavior of the organization., such as its processes and interaction with the external environment (e.g., March 1996).
We expect that this starting point is rather consistent within but varying between countries, for instance, between countries that apply continental European legal systems based on Roman Law versus Anglo-Saxon systems based on Common Law; or dependent on the extent to which new public management principles have been implemented (Kuhlmann and Wollmann 2019).
The polling institute Allensbach published a report with excerpts from 25 qualitative phone interviews and a survey with 349 respondents, collected among public officials at the district and community level. A “light analysis” was applied, by identifying main challenges and trends.
Decisions on the distribution of incoming refugees throughout Germany were made based on a quota system called the Königsteiner Schlüssel.
All quotes are originally in German, translation by the authors.
Whereas the Federal office for migration and refugees (BAMF) was in charge of executing registration, local administrations provided the infrastructure.
There is no directory of employees in all German counties, which prevented addressing many employees per agency. Instead, we collected district commissioners’ names and e-mail addresses from the internet and addressed them personally with an invitation to participate in the survey.
See Supplementary Appendix for further information.
This implies that we discard the third dimension contained in Donaldson’s definition, namely stakeholder satisfaction. The reason is that such an outside perspective implies a fundamentally different measurement strategy that went beyond the scope of this study.
Information on all variables can be found in the Supplementary Appendix D.1. All numerical explanatory and control variables in the regression analysis are standardized (using the column means and standard deviations).
Scatterplots of the relation between dependent and independent variables can be found in the Supplementary Appendix D.4.
Scale function in R (centering via subtraction of column mean; scaling via standard deviations) (Becker, Chambers, and Wilks 1988).
We tested the significance levels of the predictors with an ANOVA (Supplementary Appendix F.2). We also checked for non-linearities and multicollinearity (Supplementary Appendix F.3–F.6 and F.7), where no conspicuous results were found.
We ran several robustness checks for Model 1, including fixed effects regression for the multilevel federal-state structure and ordered logistic regression and found no substantial alterations in the results (Supplementary Appendix F.1, F.8, and F.9).
Discussions of such cases can be found in the record of the parliamentary assembly of Bavaria (Dachau) (Document number 17/9061, 18 December 2015), and the local media, for example, Rheinische Posthttps://rp-online.de/nrw/staedte/haan/staatsanwaltschaft-es-geht-auch-um-bestechung_aid-33775707 or the Stadtspiegelhttps://www.lokalkompass.de/bochum/c-politik/kommunaler-aufklaerungsausschuss-soll-fehlverhalten-der-stadtverwaltung-bei-der-beauftragung-von-leistungen-zur-betreuung-und-unterbringung-von-fluechtlingen-aufdecken_a885420 (all last accessed in June 2020).
