Democratic legitimacy, accountability and performance in frontline implementation

Transcription

1 Democratic legitimacy, accountability and performance in frontline implementation Eva Lieberherr and Eva Thomann Paper prepared for the Panel Accountability and Reputation at the 2018 ECPR Joint Sessions. Nicosia April Abstract Draft Democratic legitimacy is often seen as exclusively implying political-administrative accountability, within hierarchical relationships. In street-level organizations, however, multiple accountabilities are practiced. Based on the literature on democratic legitimacy, we conceptualize the practice of streetlevel accountability as a constitutive dimension of the throughput legitimacy of state action. This conceptual clarification enables an empirical exploration of how the democratic (input) legitimacy of policy decisions affects the practices of street-level accountability (throughput legitimacy), as well as the consequences for performance (output legitimacy). The case at hand is the implementation of the Swiss Federal Forest Policy 2020 by subnational governments and private actors between We find that deficient input legitimacy of policies produces accountability dilemmas and thereby negatively affects output performance. 1

2 1. Introduction Rothstein (2014) argues that political scientists focus too much on the input side of the democratic process, on who wins elections and on party competition, and not enough on the state machinery and its performance or lack thereof. As Héritier (2016: 17) points out, this argument is particularly interesting in light of empirical findings which indicate that in public opinion good governance of the state is a more important source of democratic legitimation than the correct and fair democratic procedures as such. Conversely, in the literature on street-level bureaucracy, the accountability and performance of public actors and agencies are central concepts (e.g., Bovens et al. 2014; Brodkin 2008; Hupe and Hill, 2007; Hupe and Buffat, 2014; Hupe and van der Krogt, 2013), while issues of democratic legitimacy are often treated as upstream factors.. However, little effort has been made to connect the two literatures. As a result, for example, it is well known that the exercise of street-level discretion, if it results in deviations from centrally decided rules, can have consequences that threaten democratic principles, such as an unjust treatment of citizens (Cohen and Gershgoren 2016; Danziger and Seefeldt, 2003; Epp et al., 2016; Frederickson, 1990; Gooden and Portillo, 2011; Maynard-Moody and Musheno, 2012; Thomann and Rapp,2017) or the neglect of public service goals e.g., in a context of contracting-out (Dias and Maynard-Moody, 2006; Smith and Lipsky, 2009; Thomann and Sager 2017). Conversely, much less is known about the implications of democratic legitimacy of policy decisions at large for street-level discretion and the subsequent potential effects on performance. This is unfortunate because the notions of legitimacy and accountability are closely intertwined. This paper seeks to address this gap in street-level bureaucracy research. Democratic legitimacy is often seen as being given when democratically decided rules are followed by street-level bureaucrats. Following this general conceptualization of legitimacy, it is assumed that democratic accountability exclusively implies political-administrative accountability, within hierarchical relationships. In this view, democratically decided policy decisions should ultimately be implemented as intended at the street level. In street-level organizations, however, multiple accountabilities are practiced (Thomann et al. 2018). How has this insight been evidenced so far, and what does this mean for theory and research? To answer this question, this paper conceptually and empirically differentiates between multiple dimensions of legitimacy and links democratic participation in policy decisions (input legitimacy) with accountability (throughput legitimacy) and performance (implementation of policy goals as intended, i.e. output legitimacy). We first differentiate between input and output legitimacy and then conceptualize the practice of street-level accountability as a constitutive dimension of the throughput legitimacy of state action. Doing so is helpful to clarify the relation between input, throughput and output legitimacy and assess this interplay empirically. Our main expectation is that the democratic legitimacy of policy decisions 2

3 influences how street-level organizations experience and exercise different forms of accountability. Dilemmas between different accountabilities can result from lacking input legitimacy and negatively translate into street-level performance. The case at hand is the implementation of the Swiss Federal Forest Policy 2020 (Forest Policy 2020) by subnational governments and private actors between 2012 and The paper proceeds as follows. We first introduce the notions of input, throughput and output legitimacy and situate the extended accountability framework (Thomann et al. 2018) within the throughput dimension. We then outline the causal chain linking democratic legitimacy with accountability dilemmas and output performance. After introducing the methods and empirical background, we report our results. Our main conclusion is that deficient input legitimacy of policies produces accountability dilemmas and thereby negatively affects performance. 2. Input, throughput, output legitimacy and accountability regimes According to political philosophy, legitimacy refers to the recognition and acceptance of power, justified via democratic processes that are based on liberal democratic constitutions. To achieve this in modern, Western polities, such legitimating arguments tend to focus on institutional arrangements ensuring democratic participation, the accountability of governors, and safeguards against the abuses of governing powers (Scharpf, 2007, p. 7). In the governance literature, legitimacy as a general notion serves as an umbrella or meta-principle for setting standards in order for street-level bureaucrats to be regarded as appropriate and just by those over whom they exercise authority (Bekkers et al., 2007). To provide a more differentiated analysis of legitimacy, we distinguish between multiple aspects, namely input, throughput and output legitimacy (Scharpf, 1999). Input legitimacy is a quality of political decisions and policies. It focuses on what to do and how rather than on what to obtain. Institutions are legitimized via principles of democracy, representation and participation (Beisheim & Dingwerth, 2008; Considine & Lewis, 2003; Dingwerth, 2007; Verba, 1972; Wolf, 2002). This type of legitimacy assumes structures to be relatively resilient to the idiosyncratic preferences and expectation of individuals (March & Olsen, 1984, p. 741). For instance, public servants are assumed to operate within the framework of the state and therefore to adopt a democratically decided policy. A key aspect of input legitimacy is participation (Schmidt, 2013) and the ability of actors to influence decision-making. Accordingly, the quality of input legitimacy may depend on the degree to which the policy-making process is inclusive (Lieberherr et al., 2012). We expect that this quality can influence actors willingness to follow the rules within street-level organizations. Output legitimacy refers to the results of policies obtained through state action. It focuses on what to obtain or performance rather than focusing on the process. Hence institutions are legitimized 3

4 based on their utility and effectiveness (Blatter, 2004; Peters & Pierre, 2010, p. 49; Pierre, 2009). Public sector reforms such as New Public Management rest on this type of legitimacy. Accordingly, a legitimate system of public policy implementation would not necessarily be based on consent through democratic institutions. Instead, the rationale of this approach is that legitimacy can be achieved through a focus on results and performance. While certain scholars define output legitimacy with responsiveness and accountability (Bekkers et al., 2007), we follow Schmidt (2013) and argue that the latter fall under throughput legitimacy. Hence, in terms of output legitimacy, our focus is on effectiveness, which relates to achieving goals to solve a policy problem (cf. Scharpf, 1999). The political science literature has typically focused on the perceived effectiveness rather than on actual problem-solving (Benz & Papadopoulus, 2006; Scharpf, 1999). Effectiveness can be assessed in terms of (1) output, i.e., the results of policy implementation; (2) outcome, i.e., the change in actors behavior in response to the implementation process; and (3) impact; i.e., the physical or material consequences of the outputs on, for example, service delivery quality, environmental protection (cf. Sager and Rüefli 2005; Scharpf, 1999). Throughput legitimacy is process-oriented and serves as an intermediary dimension between input and output legitimacy (Lieberherr, 2016). Following Schmidt (2013) throughput focuses on what goes on inside the black box of the policy process between input and output. Schmidt (2013) relates throughput to governance and includes accountability as one of its elements. Other governance scholars also address accountability, i.e. the holding of actors responsible for their decisions or practices, under the throughput dimension (Chatzopoulou, 2015; Lieberherr, 2015). In line with these authors, we specifically situate accountability under the throughput dimension, as it refers to a process between an actor and a forum (Bekkers et al., 2007, p. 45). Throughput has often been conflated with input or output legitimacy in the literature (Lieberherr et al., 2012; Scharpf, 1999; Schmidt, 2013). Indeed, scholars such as Bekkers et al (2007) have included accountability under output legitimacy, as they argue that the process relates to the actor s performance (cf. Bekkers et al., 2007). However, we argue that accountability relates to the governance process itself, i.e. how actors interact and are held to account, which may have implications for performance, but does not focus on outputs. Throughput can involve different qualities of legitimacy, depending on the accountability regime (i.e. forum) implementing agents are held to account and refer to. Following Schmidt (2013), disentangling throughput from input and output legitimacy [ ] has explanatory power about the legitimization (or lack thereof) of the other two dimensions. More throughput legitimacy [ ] may have little effect on public perceptions of legitimacy, while less of it [ ] may bring down the whole house of cards (Schmidt, 2013, p. 14). Figure 1 summarizes the three types of legitimacy in relation to state action, situates streetlevel accountability in the throughput dimension, and outlines our main causal expectation. We argue 4

5 that the quality of input legitimacy of policies affects how implementing actors exercise accountability at the street level. As we now outline, lacking input legitimacy may produce accountability dilemmas, between sources of state and that of the profession, market and society (explained below) and thereby, negatively affect output performance. Figure 1: Analytical model Source: own illustration. 2.1 Accountability regimes Honing in on throughput legitimacy, recent scholarship scrutinizes the interaction of institutional and informal accountability mechanisms. The institutionalized form of the relationship in which a public organization reports on its performance to a competent political actor who is able to judge and sanction the organization (Bovens 2007) serves to reduce information asymmetries, limit discretionary powers of non-majoritarian actors, and to give an incentive for better public service performance (McCubbins, Noll, and Weingast 1987). However, the one-sided focus on performance measurement, administrative oversight and control fails to do justice to essential aspects of informal practices in street-level discretion, such as responsiveness to clients and professional judgment (Brodkin 2008; e.g., Thomann 2015). This points to a missing link in our understanding of the behavioural effects of accountability mechanisms on individual actors: the felt or experienced accountability (Lerner and Tetlock 1999; Hall, Frink, and Buckley 2015; Schillemans 2016; Schillemans and Busuioc 2015). It is a central motivation in research on street-level bureaucracy (Lipsky 1980/2010) that agents take many decisions on the basis of what they find appropriate (Olsen 2013). We therefore rely on the sociological concept of accountability, including felt accountabilities, that underlies studies of street-level bureaucracies (Brodkin 2008). Accordingly, accountability refers to 5

6 norms and standards about how actors should behave and specifically in terms of who has an obligation to justify his actions vis-à-vis whom (Hill and Hupe, 2007). In their extended accountability framework, Thomann, Hupe and Sager (2018) wrap up the literature and differentiate four main sources of accountability with associated accountability regimes (Thomann et al. 2018; cf. Table 1) to which street level bureaucrats are subject to. Accountability regimes can be differentiated in relation to the state, the market, the profession and society (Klenk and Lieberherr, 2014; see also Behn 2001; Bovens et al. 2014; Busuioc and Lodge 2015; Deleon 1998; Koppell 2005; Mashaw 2005; Oberfield 2016; Pierre 2009; Poulsen 2009; Schillemans and Busuioc 2015). Action prescriptions are norms and demands about how street-level bureaucrats should preferably behave; they emerge at three levels: from institutionalized contexts; the organization and work circumstances; and individual characteristics. The resulting types of (subjective or objective) pressure add up to work pressure at the street level, which, depending on the accountability regime at play, translate to rule, incentive, vocational or societal pressure (cf. Table 1). Table 1: Four accountability regimes in street-level organizations Key source State Market Profession Society Accountability Politicaladministrative Customer- and shareholderoriented Vocational Participatory Action prescriptions Levels of aggregation: Individual, organization, system Formal rules Rule pressure Cost minimization Benefit maximization Customer satisfaction Shareholder value creation Competition Incentive pressure Professional values, norms, and attitudes Good practice Peer review Vocational pressure Societal expectations Perceived clients needs Societal pressure Source: Thomann et al. (2018: 303). XX request permission to reprint Street-level bureaucrats exercise accountability in their actions by referring to action prescriptions or pressures (cf. Table 1) stemming from these accountability regimes to differing degrees. Norm reference, meaning the intensity with which street-level bureaucrats allocate their attention to and identify with particular action prescriptions associated with the different accountability regimes, arguably influences performance (output legitimacy). This paper empirically scrutinizes this link. Moreover, our core argument is that the degree of democratic (input) legitimacy in a given policy 6

7 context may also influence the implementing actors reference to differing accountability regimes and ultimately affects performance. Key in this link is the prevalence of so-called accountability dilemmas. 2.2 Accountability dilemmas In principle, the use of street-level discretion should be unproblematic whenever accountability toward the state is given and overall legitimacy would be achieved. Conversely, if politicaladministrative accountability conflicts with other accountabilities, then results may not be achieved as intended. In this vein, the term dilemmas (Lipsky 1980/2010) refers to situations of conflicting and irreconcilable accountabilities, when some action prescriptions are at the expense of or incongruent with other demands (Thomann et al. 2016, 2018; e.g., Cohen et al. 2016; Dias and Maynard-Moody 2006; Sager et al. 2014). We now sketch three different types of dilemmas before linking the kind of accountability regime actors adhere to with input and output legitimacy. Following Thomann et al. (2018) we focus on horizontal accountability dilemmas between rule pressure on the one hand and incentive, vocational and societal pressure on the other hand, rather than on vertical dilemmas at different levels of aggregation (cf. Table 1). First, an example of a dilemma between rule and incentive pressure would be when a street-level bureaucrat cannot simultaneously achieve the goals of a policy and comply with output targets (Thomann et al. 2016; see also Dias and Maynard-Moody 2006). In more hybridized or privatized settings, such dilemmas also occur when e.g. inspection duties, lengthy decision-making paths, complicated standards decrease profits or frustrate customers (Thomann et al., 2016,2018). A second dilemma is rule and vocational pressure, which entails a clash when street-level bureaucrats perceive the demands of the policy to be incongruent with their professional values, norms, and attitudes (Tummers et al. 2012, 4, 5). Finally, the perceived lack of an added value of the policy to socially relevant goals or for the clients would indicate a third dilemma between rule pressure and societal pressure. The general expectation is that, firstly, when implementing actors are confronted by any of the three dilemmas sketched above, that this provides an explanation for deficient implementation: as street-level bureaucrats cannot comply with multiple demands at once, they need to take a side and, depending on their coping strategies (Tummers et al. 2015), this may be at the expense of publicly defined policy goals (Jilke et al. 2018). When making the link between the above dilemmas and input-output legitimacy, it is useful to think of the different accountabilities as being complementary as well as potentially conflicting. Drawing on the democratic dilemma debate (Dahl, 1994, p. 30), which posits that if there is more legitimacy as an input function, then output legitimacy, i.e., system effectiveness, is decreased, we make a counter argument: Drawing on such scholars as Grote & Gbikpi (2002), Heinelt (2002); Lindgren & Persson, 7

8 (2010), Wolf (2002, 2006) we argue that the two dimensions of legitimacy can re-enforce each other, depending on the role of throughput legitimacy. On the one hand, this follows the logic that participation is arguably a condition of system effectiveness ( Grote & Gbikpi, 2002), as input legitimacy is a pre-condition for effective problem-solving (Benz & Papadopoulus, 2006). Hence input legitimacy will increase output legitimacy (Bäckstrand, 2006a, p. 473) and vice versa (Bäckstrand, 2006b; Scharpf, 1999). The argument is that the more relevant actor participants are incorporated within/associated with a policy making process, the greater the chance that the measures taken will be voluntarily accepted and effectively implemented (Grote & Gbikpi, 2002, p. 18). More specifically, on the other hand we expect that input legitimacy affects the prevalence of accountability dilemmas. The democratic process should serve as an opportunity to integrate differing demands from market actors, professional associations and target groups into policy decisions. If this opportunity is missed, then the policy should arguably be less congruent with the demands from the market, the profession, and society. This in turn may create situations for street-level bureaucrats when rule pressure stands at odds with market, professional and societal accountabilities. Our second expectation is thus that low input legitimacy should trigger the presence of (some) accountability dilemmas and hence negatively affect results. As a result, low input legitimacy may actually lead to low output legitimacy. 3. Empirical background and methods 3.1 Forest Policy 2020 We explore how democratic (input) legitimacy affects the prevalence of accountability dilemmas and how this empirically affects street-level performance (output legitimacy). To this end, we have chosen Swiss forest policy for our empirical analysis and assess whether and how the responsible actors of the Forest Policy 2020 have implemented their assigned measures as intended by the public principal (the Confederacy). The Forest Policy 2020 includes two phases: from 2012 to 2015 and from 2016 to We analyze the first phase ( ). The Forest Policy 2020 replaces the Swiss National Forest Program of The main objective of the this policy is guaranteeing of sustainable forest management and the creation of favourable conditions for an efficient and innovative forestry and timber sector (BAFU, 2013, p. 13). This overarching objective is concretized in additional eleven goals. For each objective, the Forest Policy 2020 defines strategic directions and concrete measures on the part of the Confederation (cf. Figure 2 for an over of the Forest Policy 2020). The Forest Policy 2020 sets measures to be implemented at the federal level (the Federal Office for the Environment). As the forest policy is a joint task of the 8

9 Confederation and subnational government (cantons), the implementation of the federal measures is carried out with the cantons and coordinated with the central stakeholders (private actors, e.g. forestry and timber), who play a key role in the implementation of many measures. Hence, the policy additionally specifies roles for the subnational governments as well as private actors, i.e. from the forestry and wood sectors, who have for-profit logics, as well as actors from the environmental and educational sectors (see Appendix 1 for a list of the implementing actors). This makes the case a likely candidate for observing all types of accountability regimes mentioned above, where the actors face rule, incentive, vocational and societal pressure. While all measures are specified for the federal level, not all measures involve a role of the subnational public or the private implementing actors. (cf. Appendix 1). These roles contain the Confederacy s expectation about what the subnational governments and additional actors are supposed to contribute to achieve the policy s goals. For an example of the roles, as specified in the policy, see Figure 2. For each goal, indicators are specified with a required bio-physical target value e.g. of forest growth that is supposed to be achieved. The institutional accountability mechanisms (oversight, monitoring, enforcement) in this policy are generally the same for both public and private actors. The goals and measures of the national policy are binding guidelines for only the federal administration (namely the Office for the Environment), not the subnational and private actors. However, several measures specify that the Confederacy has to engage with the cantons in program agreements for certain topics, e.g.in the domain of biodiversity in forests. This then entails that the Confederacy provides funding so that the cantons can achieve the agreed upon tasks. The program agreements also involve more detailed monitoring than measures in the policy not linked to these. Figure 2: Structure of Forest Policy 2020 Main goal (time horizon 2020) Goal 1: potential offered by sustainably harvestable wood shall be exhausted Strategic direction1.1: Forest managers shall be provided with expert decision support for optimal forest management Goal 2: Climate change: mitigation and Strategic direction 2.1: The impacts adaptation of climate change (storms, drought shall be etc.) on the forest shall be studied guaranteed and the suitability of the silviculture methods used shall be examined Source: adapted from Wilkes-Allemann et al., Goal 11: Measure for Confederacy: The Confederation shall demonstrate the possible harvesting potential in the forest Measure for Confederacy: The Confederation shall complete the second phase of the Forest and Climate Change research programme and shall ensure the publication and communication of the results, and the provision 9 of corresponding continuing education and training. Role of sub-national govt: Provision of information bases, interpretation of studies and scenarios in the regional context, dissemination of information, offering of advice Role of sub-national govt: Dissemination of information, interpretation of the results in the regional context, planning and implementation of measures Role of other actors: provision of Provision information bases, interpretation of studies and scenarios in specific contexts, offering of advice Role of other actors: Education, Training of forestry experts (initial and continuing education and training), implementation of measures

10 3.2 Hybrid system in the environmental field The implementing agents are public and private street-level organizations at national and subnational levels. As we are analyzing a federalist system, we have two layers of outputs, since we have implementation of measures first at the national and then the subnational level (cf. Sager and Rüefli, 2005). We are also operating in a hybrid system, as we have both public and private implementing agents (Thomann and Sager 2017). Following Seva (2015) street level organizations in the field of environmental management may differ from Lipsky s traditional definition, as they do not necessarily have direct contact with target groups on a daily basis. In addition, the people working at these street level organizations are typically not professionals uniformly trained like e.g. police officers or teachers, as they tend to varied educational backgrounds. However, following Seva and Jagers (2013, p. 1062) It is nevertheless reasonable to assume that also environmental bureaucrats at the frontline have more or less discretionary power and autonomy similar to street-level bureaucrats. 3.3 Data collection and analysis The output data at the federal level are based on the appraisal of the Swiss Federal Office for the Environment: The degree of achieving the measures was assessed and documented by the responsible actors at the Swiss Federal Office for the Environment every half year between 2012 and We used this data provided to us by the Swiss Federal Office for the Environment and compared the specified degree by the federal office with the level defined in the forest policy. We verified this data through interviews with the responsible actors. The output data at the subnational public as well as the private implementing actor level include online surveys conducted at the level of the subnational governments (from 19. July 2016 until 17. August 2016) and additional private actors (from 21. July 2016 until 17. August 2016). The surveys contained questions about the degree to which the actors have implemented the measures (based on the roles allocated to them in the Forest Policy 2020 and using the language of the forest policy). 25 (of 26) subnational governments responded to the survey and 16 (of 20) additional actors responded. Regarding the latter, the following actors groups were represented: forestry sector (3 people), wood sector (5 people), environmental sector (1 person), additional associations and organizations (1 person), research and education (2 people), recreation and leisure (1 person), insurance (1 person), agricultural sector (1 Person), transport sector (1 person). In addition to multiple-choice questions about the degree to which actors have implemented the measures, space was provided for comments. These qualitative inputs provide the main basis for the analysis of the accountability dilemmas in the results section. The comments were sorted by goal, interpreted based on the accountability dilemmas and translated by the authors. 10

11 Finally, to analyze impact, i.e. the achievement of the indicators (which relate to the physical or material consequences of the output e.g. environmental protection) as specified in the Forest Policy 2020, we assessed the national forest inventory and statistical data provided to us by the Swiss Federal Office for the Environment and compared the degree to which the bio-physical indicators have been achieved in relation to the level specified in the forest policy. For example, for the indicator mixed forest areas the defined level is a 10% increase in mixed forest areas until We then compared this to the percentage to which this forest area has increased until A key aspect here is that the indicator levels have been defined for the time period of 2020 (not 2015). Hence, we have conducted a progress-report of the extent to which the intended outcome is on-track. A key challenge in the outcome analysis was that for several indicators no data were available or that the specified level could not be assessed. For 40% (19) of the indicators no data were available. Of the 28 indicators for which data were (partially) available 21% were not verifiable (reasons include: the target value) is not in accord with the data collected or that the defined level was not precise enough and thus could not be measured). 4. Analysis To assess how input legitimacy interplays with accountability (throughput legitimacy) and how this then affects the implementation of policy goals (output legitimacy) we first present results on the policymaking of the Forest Policy 2020 (input). We focus on the subnational public actors and the private implementing actors, rather than the federal actors, as we address the street-level organizations to understand the accountability dilemmas. Finally, we take a look at output legitimacy in terms of the outputs at the federal as well as the subnational governmental level and that of the private implementing actors as well as impact in terms of our preliminary data on the achievement of the indicators specified in the forest policy. Table 2 presents a summary of the results, aggregated at the level of the goals; data were collected at the level of the measures (compare Figure 2 and methods section). 11

12 Forest Policy 2020goals Input Throughput Output legitimacy Policy decisions Accountability dilemma Output at federal level Output at subnational govt & private actor level) Goal 1: The potential offered by 70% of measures not implemented Mixed implementation* (all actors) sustainably harvestable wood shall as intended be exhausted Goal 2: Climate change: mitigation and adaptation shall be guaranteed Goal 3: The protective forest service shall be guaranteed Goal 4: Biodiversity shall be conserved and improved in a targeted way Goal 5: Forest area shall be conserved Goal 6: The economic viability of the forestry sector shall be improved Goal 7: Forest soil, drinking water and the vitality of the trees shall not be endangered Goal 8: The forest shall be protected against harmful organisms Goal 9: A balance shall exist between the forest and wildlife Goal 10: The use of the forest for leisure and recreation shall be respectful Goal 11: Education, research and knowledge transfer shall be guaranteed Confederacy developed policy with little influence by subnational public actors and private implementing agents; democratically legitimated by Federal Council (top down) rule vs. incentive (public actors only) & rule vs. vocational (public & private actors) rule vs.incentive (public actors only) & rule vs. vocational (public & private actors) None rule vs. vocational (public & private actors) & rule vs. societal(public actors only) rule vs. vocational (public actors only) & rule vs. societal (public actors only) rule vs. vocational (public & private actors) rule vs. vocational (public actors only) rule vs. vocational (public actors only) None rule vs. vocational (public actors only) & rule vs. societal (public actors only) rule vs. vocational 60% of measures not implemented as intended 40% of measures not implemented as intended Ca. 50% of measures not implemented as intended 40% of the measures not implemented as intended 60% of measures not implemented as intended Ca. 50% of measures not implemented as intended Ca. 50% of measures not implemented as intended 0% of measures not implemented as intended Ca. 50% of measures not implemented as intended Ca. 30% of the measures not implemented as intended Low implementation public actors** Implementation on-track at level of the private actors Implementation as intended (all actors) Implementation on-track (public actors) Mixed implementation at level of private actors* Mixed implementation public actors *No role defined for private actors Mixed implementation * (all actors) Low implementation byy public actors** Implementation on-track at level by private actors Mixed implementation by public actors* Implementation on-track by private actors Mixed implementation* (all actors) Mixed implementation by public actors* Implementation on-track by privatel actors On-track (all actors) Impact (on physical environment) The 1 (of 2) indicators that could be assessed has not achieved intended level Due to a lack of data none of the indicators could be assessed. Of the 3 (of 6) indicators that could be assessed, none have achieved intended level One indicator has fully achieved level as intended and 3 have partially achieved level.. One indicator achieved level as intended, one partially achieved and one not yet achieved. Of the 3 (of 5) indicators that could be assessed, one has achieved level as intended the other two not Of the 2 (of 4) indicators that could be assessed, one has achieved level and one has not (partially on-track) Due to a lack of data none of the indicators could be assessed. Of the 2 (of 4) indicators that could be assessed, one has achieved level as intended and the other not (partially on-track) 1 (of 5) indicators could be assessed and has not achieved level as intended Of the 3 (of 5) indicators that could be assessed 2 have achieved level and one has not 12

13 Table 1: Summary of all goals Bold = implementation deficit = more than 60% of measures or indicators not implemented as intended. Italics = lack of data; * mixed = several measures not implemented while others are; ** low = all measures (under a goal) not or only minimally implemented. (Source: adapted from Wilkes-Allemann et al. 2017) 13

14 4.1 Input legitimacy We focus on the implementation at the subnational public level and on the private actors (rather than on the federal level). Hence the input legitimacy aspects of inclusion and influence play a more critical role than democratic legitimation by elected actors. The Forest Policy 2020 is in principle a non-legally binding strategy developed by the Confederacy in 2011, formally endorsed by the Federal Council The Confederacy developed the policy, with a minimum amount of consultation with the implementing actors (both the subnational governments and private implementing actors). The Confederacy assigned measures to the subnational governments and private implementing actors, while neither giving them a legal mandate to do so, nor including them in the decision-making process. Hence, the policy-making process was exclusive (Zabel and Lieberherr, 2016). This contrasts with the previous national forest strategy (in place between 2004 and 2012), which rested on a highly participatory approach (Schulz, Lieberherr, Zabel, in press). The policy making of the previous forest strategy lasted four years and involved over 130 public and private actors, including all those responsible for the implementation (Zingerli & Zimmermann 2006) Throughput legitimacy We find that the implementing actors at the subnational governmental level and the private implementing actors are subject to multiple accountability regimes (i.e. fora) and refer to differing action pressures. Specifically, both the public and private implementing actors are confronted with conflicts between the political-administrative accountability (rule pressure) with the other accountabilities. We illustrate this by presenting how the implementing actors experience the three dilemmas: rule vs. incentive pressure, rule vs. vocational pressure and rule vs. societal pressure. The results are summarized in Table Rule vs. incentive pressure Only the subnational public actors, and not the private implementing actors, note a conflict between rule vs. incentive pressures. The subnational public actors have a relatively strong norm reference to 1 However, in comparison to the previous forest strategy, which was only adopted at the level of the public administration, and not by politically elected actors, the current policy is hierarchically at a higher level than the previous policy and enjoys a higher democratic legitimation, which is meant to give forest an increased standing vis-à-vis other sectoral policies (Steinmann et al., 2017). If our analysis had focused on the implementation at the federal level and the ability to prioritize forestry measures at the expense of other national sectoral policy measures, and this due to the top-down legitimation of the Forest Policy 2020 as it has been approved by the Federal Council, then we might be able to paint a very different picture. 14

15 incentive pressure, as they try to cope with their limited budget by minimizing costs and taking an entrepreneurial stance. This shows how relevant the market type of accountability is even for analyzing public organizations in times of new and hybrid governance modes (Thomann et al. 2018). Several subnational public actors find that the current Forest Policy 2020 lacks entrepreneurial strategies and that the governance structure is inefficient. For instance, one such actor questions the formal rules promoting conventional management of young forests (goal 2), as this hinders cost-effective strategies: conventional young forest management is promoted. This is generally questionable. A serious evaluation would probably show its low direct use and the high negative side-effects (upkeep of inefficient structures). Public actors see a central need to adapt the policy to mitigate the conflict between inefficient practices promoted by the policy and a competitive forest sector, where the management of young forests is not an end in itself, but rather part of the business concept. In line with this, another subnational public actor holds the view that federal, bureaucratic measures specified in the policy are useless (goal 1). Instead, a functioning entrepreneurial forest economy with economic criteria to assess measures promoting wood from deciduous trees is necessary: We should rather focus on making the current wood processing more economically competitive. According to this actor, a state-directed economy has hindered forest cultivation and is counter-productive Rule vs. vocational pressure Both the public and private implementing actors have a strong norm reference to vocational pressure and note a conflict with rule pressure. The subnational public actors see themselves as already fulfilling the goals or having been implementing related measures for years, regardless of the national forest policy. These actors view the measures as standard implementation duties, and hence they do not per se implement the measures precisely as specified in the federal forest policy, as either this would lead to additional transaction costs or they are not completely aligned with their perceived professional norms. For instance a subnational public actor says that the wood use potential (goal 6) has been exploited in his jurisdiction, regardless of the federal policy. Other public actors say that they have been providing basic information, or implementing biological rationalization (goal 4), irrespective of the federal policy. Or that the biological measures (goal 4) have already been implemented since 1994 and that the subnational government did not have to implement any additional measures set by the forest policy. In line with this, another subnational public actor notes that the impetus for action has not been the biodiversity goals (goal 4) of the Confederacy, but rather the interests of the subnational government since Such actors hence regard the Forest Policy 2020 as largely superfluous, as e.g. there is no 15

16 need for a new concept or regulation at the subnational level as this is already covered in other legislation. In addition, certain subnational public actors find a misfit between the formal measures of the policy and professional norms and good practices fit for their jurisdiction. E.g. some public actors note that their perimeter is too small for implementing biodiversity measures (goal 4) or improving performance the economic viability of the forestry sector (goal 6). Also in relation to the goal that forest soil, drinking water and the vitality of the trees shall not be endangered (goal 7), a subnational administrative actor says that the measure is not applicable given the bio-physical characteristics in his jurisdiction. Also with respect to implementing the measure forest area conservation (goal 5), a subnational public actor explains that they have had no interest to implement this as creating static forest areas does not make sense in certain regions and would lead to competency issues, as ingrown forests would no longer be legally defined as forest. This dilemma was also reflected in the finding that certain actors were not familiar with certain terminology (e.g. Cleantech ) used in the Confederacy s policy and others were also not informed about certain measures that should have been established at the federal level and then transferred to the subnational level. As one public actor asks: "When and how were the cantons informed about the task-force on national forest damage? (goal 8). Certain subnational public actors also find that the measures specified by the Confederacy are either outdated or only work in theory (e.g. combatting Rotbandkrankheit in moorlands, goal 8). Finally, the public actors seem to see their role or function differently than the Confederacy: specifically in relation to leisure and recreation (goal 10) subnational public actors tend to delegate aspects like the valorization of forest services directly to the forest owners or the promotion of wood use to private companies or associations while the Confederacy expects them to implement these measures more directly. The private implementing actors also provide some explanations about the dilemma rule vs. vocational pressure. For instance, an actor from the wood sector notes that in relation to the exploitation of sustainably usable wood (goal 1), there is a misfit between the market estimates of the wood processors and the Federal Office for the Environment (goal 1). We have politically pushed for doing more for beech trees. However the Federal Agency continues with the current practices The market estimates of the wood industry and that of the Federal Agency diverge. According to this actor, the policy does not promote best practices known in the field. He also notes that wood sector is skeptical of the measures specified in relation to climate change (goal 2): the wood sector fears ideological influences (attitude against pinewood and Douglas fir), self-interest (significance for the forest service and forest research) and making things more complicated and expensive for timber harvesting. Also in relation to climate change (goal 2), actors from the forestry and wood sectors state that they are hoping for pragmatic suggestions, or are starting to make such themselves, to promote 16

17 climate fitness via more and regular interventions in the forest and shorter turnover time, as this is what they regard is critical for a well-functioning forest and wood sector. Finally, a leisure and recreation actor made a general statement that a federal policy can neither be implemented nor assessed at the national level, as each subnational jurisdiction has its own geographic specificities and own forest policies; the problems are local and should be addressed individually. Further arguments that arose from the private implementing actors, which link to the rule vs. vocational pressure dilemma, relate to goal ambiguity and perceived non-applicability of measures. For instance, an actor from research and education notes that the first objective of the Forest Policy 2020 is unclear and does not cohere with best practice: What is exactly meant here? Promotion and the highest exploitation of the resource wood or about optimal management, which calibrates the individual goals? Another private actor states that they only implement measures that cohere with our goals. Several private actors noted a misfit between their function in the sector and their role specified in the forest policy. Indeed, especially in relation to biodiversity (goal 4), a key explanatory factor was that the private actors do not feel that they are addressed by the measure, hence the frequent response option that was chosen in the online survey was not applicable. Particularly associations note that their role is to advise and network, but not to implement measures set by the Confederacy. Another factor here is that the exact roles of the subnational governments and additional actors are not clearly defined in the federal policy Rule vs. societal pressure Only subnational public actors experience a dilemma between rule versus societal pressure. In relation to biodiversity (goal 4) we find that there seems to be a misfit between the rules and the private owners needs and hence the subnational public actors have difficulty to implement the measures as specified in the federal forest policy, when there are many private forest owners. Also in terms of leisure and recreation (goal 10) there seems to be a misfit with respect to the measure related to protection zones for wild animals and societal needs: for instance, a subnational public actor implemented this measure by drafting a regional law, but the voters in the jurisdiction rejected it. And in another jurisdiction the regional government rejected a similar law. A subnational public actor also notes that the application of clearance compensation (goal 5) are simply not feasible in agglomerations, on land with intensive agriculture or in mountainous areas. Finally, a subnational public actor says that the implementation of such the measure related to protection zones for wild animals (goal 10) in the Swiss plateau is an illusion, as it is simply not societally accepted. 17

18 4.3 Output legitimacy Overall we find that output legitimacy is deficient, as the degree of output performance of the Forest Policy 2020 during the first phase is low. As shown in Table 2, all goals (except goal 11 education, research and knowledge-transfer ) suffer from an implementation deficit with respect to at least one of the dimensions of output and impact. Notable is the particular implementation deficit in terms of all three dimensions of goal 1 the potential of sustainably usable wood is utilized and goal 6 the economic performance of the forest sector is improved. Goal 2 climate change: mitigation and adaption are secured and goal 10 leisure activities and recreation occur without damage have an implementation deficit in terms of two dimensions: Goal 2 in terms of output and impact and goal 10 in terms of impact and outcome. Important to note is the dependency of subnational governments on the implementation of certain measures by the Confederacy. In cases where the implementation at the level of federal outputs is deficient, the subnational governments were also not able to implement the measures (e.g. goals 2 and 6). We find that the implementation of goals 1 and 2 suffer particularly due to the dilemma between rule and incentive pressure, as the public-administrative actors at the subnational level have not implemented measures as intended. The implementation deficit of goals 1, 2, 4, 6, 7, 8, 10 and 11 can be attributed to the rule vs. vocational pressure dilemma. Finally, challenges in the implementation of measures related to goals 4, 5 and 10 can be explained, in part, by the dilemma between rule vs. societal pressure. Interestingly, goal 9 experiences deficient implementation at the level of output (subnational public and private actors) despite no identification of a dilemma. It must be noted, though, that for the first phase only one measure was set for this goal (policy failure rather than implementation failure). 5. Discussion We find that the public and private implementing actors at the subnational level are confronted by the three dilemmas drawn from the literature: rule vs. incentive pressure, rule vs. vocational pressure and rule vs. societal pressure, which provide explanations for the deficient implementation at the level of output and impact, as discussed below. Our findings thus support our first expectation that when implementing actors are confronted by any of the three dilemmas sketched above, that this, in part, provides an explanation for deficient implementation. The most explanation was found for the rule vs. vocational pressure dilemma. We have found that many of the implementing actors do not apply the Forest Policy 2020 directly due to the following reason and dilemmas: - they find the measures inefficient or useless (rules vs. incentive pressure dilemma) 18

19 - they find a misfit between the measures specified in the policy and best practices (rule vs. vocational dilemma) - they have already been engaging in similar activities (as standard practices in the field) as those specified in the federal policy; they see no added-value of the federal policy and hence rather stick to their previous routines as they consider these to be closer to the state-of-the-art in their field than those of the policy (rule vs. vocational dilemma). - the roles of the subnational governments and private actors have not been clearly defined in the policy, or they have been interpreted differently by the implementing actors than what the Confederacy intended (rule vs. vocational dilemma) - they have tried to implement the measures by e.g. proposing regional laws, but these have been rejected by decision-makers and the public (rule vs. societal dilemma) In line with previous evidence (Thomann et al. 2018), we find that professional norms seem to play a key role for understanding output performance regardless of whether the actors at the street level stem from the public or private sector. We find differences between the dilemmas experienced by public and private actors. While the subnational public actors have faced all the dilemmas, the private actors noted fewer conflicts in general and only in relation to rule vs. vocational pressure. This finding contradicts the results from earlier studies arguing that hybrid implementation arrangements can increase the accountability dilemmas particularly of private street-level bureaucrats (Thomann et al. 2016, 2018). However, we must take into consideration that the subnational government actors are ascribed a larger role in the Federal Policy 2020 than the private implementing actors. It is, however, particularly interesting that it is the subnational public actors who note conflicts between rule vs. incentive pressure and not the private actors from e.g. the wood sector etc., as the latter are typically expected to have a stronger norm reference to incentive pressure than public actors. We also find evidence supporting our second expectation that low input legitimacy triggers the presence of (some) accountability dilemmas and hence negatively affects results. We find that the Forest Policy 2020 has deficient input legitimacy, as it has been developed at the federal level with minimal inclusion of and influence by a wide range of actors, e.g. the implementing actors at the subnational public level as well as the private implementing actors. Indeed, the fact that the Forest Policy 2020 is a policy of the Confederacy and not of the Confederacy and the subnational governments and private implementing actors seems to affect the implementation of the policy in several ways and due to the three dilemmas. We have found that this low input legitimacy has led to accountability dilemmas. For instance, we find that the implementing actors have pushed for a different approach regarding young forests, but that the federal administration has taken a different path. This has led to a rule vs. incentive conflict 19