Gaps between environmental goals and the state of the environment are not unstudied in environmental research. On the contrary, lack of goal achievement in public policy is researched since long in several disciplines. In natural and technical sciences, focus is placed on advancing the understanding of the problem at hand, thereby reducing uncertainty for decision-makers, and on engineering solutions of various kind. Environmental sociologists focus on e.g., societal change and how to characterise environmental risk, including for example, scholars who assert explanatory power to theories on Ecological or Reflexive Modernisation, sometimes constituting a basis for normative prescriptions (Karlsson 2005; Boström et al. 2017; Machin 2019). In environmental psychology, issues relating to, for example, behaviour and change are explored (Steg and Vlek 2009; Gifford and Nilsson 2014). Environmental economics and law rather study how to manage market failures (Baranzini et al. 2017; Shmelev 2017) and shortcomings in legal systems (Jóhannsdóttir 2009; Rakhyun and Bosselmann 2013).
Science and technology studies as well as environmental risk governance research give attention to interactions between science and natural systems on the one hand, and policy and social systems on the other (van Asselt and Renn 2011; Linke et al. 2013; Gilek et al. 2016). Studies on risk decision-making and environmental goal-setting are found in, for instance, regulatory toxicology and environmental philosophy (Sandin et al. 2002; Karlsson 2006; Edvardsson 2007). Well-developed theorising of relevance for this Perspectives article is carried out not least in the domain of political sciences. The field of policy studies includes the substance-focused branch of policy analysis (knowledge in policy) (Walker 2000; Runhaar et al. 2006), as well as policy process research (knowledge of policy) (Weible 2017). Within the latter, a number of theories are concerned with policy development, change and implementation, for example the Advocacy Coalition Framework (ACF), the Institutional Analysis and Development framework, the Multiple Streams Framework (MSF), and the Innovation and Diffusion Models (Ruseva et al. 2019). Sometimes, different approaches are applied in parallel to describe a complex environmental policy situation (Eriksson et al. 2010).
In the referred literature, an array of different interconnected, overlapping and sometimes unrelated theories, frameworks, models and concepts is used to describe and explain reasons for gaps, and barriers against bridging them. Common concepts include science denial (Edvardsson et al. 2017), scientific uncertainty and disagreement (Karlsson 2005; Saunders et al. 2017), non-rational goals (Edvardsson 2007), politicisation (Eriksson et al. 2010), institutional lock-in (Unruh 2000), governance barriers (Tynkkynen 2017), burden of proof requirements (Sandin et al. 2002; Karlsson 2005; Alfredsson and Karlsson 2016; van den Bergh 2017), mismatch (Gilek and Karlsson 2016), and implementation constraints (Hassler 2017). Broader explanations are given by not least the ACF, which centres on policy subsystems and actors (Jenkins-Smith et al. 2017). In the ACF, people operating within an advocacy coalition are considered to share common belief systems, including a stable deep normative core, a set of specific policy core beliefs, and related, and more changeable, secondary beliefs or aspects. In pursuing their respective interests within policy subsystems, coalitions negotiate and learn gradually from e.g., policy implementation results, but can also be influenced by internal or comparatively dramatic external events. These factors interact and can potentially lead to policy change, albeit after long time in the normal case. The MSF is also of relevance for delay, not least the insight that even when the three problem, policy and political streams couple, agenda change is often dependent on various windows of opportunity, which policy entrepreneurs may not always be able to influence (Herweg et al. 2017).
Obviously, there are advantages and drawbacks associated with all of the referred concepts, frameworks and theories, which all help in analysing the degree of goal achievement in environmental policy. For example, it can be argued that the ‘barrier’ concept, while pointing out key challenges to work with to promote environmental goals, can also give the impression of a total common standstill from one policy stage to another, e.g., between policy formulation and implementation. While such gridlocks are not unheard of in environmental governance, the more usual situation, for example, in marine environmental governance in the Baltic Sea (Karlsson and Gilek 2018), is rather that of delayed progress. Other concepts, like ‘scientific uncertainty’ or ‘non-operational goals’, can more be seen as basic challenges or problems in science and policy, respectively. Our aim is rather to present a framework that can be used as an analytical lens in research and not least policy practice, to target mechanisms in specific cases of delay, and to potentially compare these. While the ACF, MSF and other policy process theories and frameworks have much to offer, they direct the focus mainly to the policy sphere and processes on a rather generic and long-term level. In many situations, we would argue that the actual mechanisms of delay are linked to more specific aspects in both science and policy. However, to what extent the approaches described have explanatory value is often contextual. Considering this, our intention is neither to pen down when a specific approach is applicable or not, nor to review and synthesise this broad literature and upon that present new theory. Instead, we want to highlight the existence of delay mechanisms as a complement to other approaches, and in doing so also help to analyse them in practice to ultimately improve the understanding of how delay can be counteracted.
A number of empirical studies show that delay may occur over the entire science-policy field, which should be kept in mind in any analysis. An early example is the study by Thelander and Lundgren (1989), who identify several delay mechanisms linked to the scientific discovery, communication and politics of environmental problems, for example concerning various hazardous chemicals. Nolin (1995) similarly describes the more than a decade long process from scientific discovery to political measures concerning ozone depleting substances. Two comprehensive reports from the European Environment Agency (EEA 2001, 2013), moreover, investigate late policy lessons from early scientific warnings, in a number of case studies—on, for example, various hazardous chemicals, climate change and ozone depletion—where delay, for various reasons, has been significant. While these previous studies indeed are informative, we believe that even more knowledge could have been generated if the findings would have been placed into a general framework that allows systematic and comparative analysis.
To concretise this, we will next propose a new framework over central delay mechanisms that ought to be further analysed and compared in environmental governance research and practice (Fig. 1). Thereafter, we exemplify the framework with two, out of a wider set of, delay mechanisms from the science and policy domain, respectively—denial of science and decision thresholds. While we, based on previous research, consider these mechanisms to be of central interest, we do not imply that other potential delay sources highlighted in Fig. 1 necessarily are less significant.
Developing a framework for analysing delay in environmental governance
The presented framework is based on the recognition that (i) delay can be caused by several different mechanisms in both the science and the policy domain; (ii) multidirectional interactions between and within the science and policy spheres have fundamental importance for both the emergence and mitigation of delay, and (iii) a multitude of delay mechanisms can be at play simultaneously on several levels, and that the playing field may differ substantially from one case to another. This means that, while we agree with e.g. Varjopuro et al. (2014), who explicitly explores delay linked to eutrophication in the Baltic Sea, that delay may be manifested at different places, we argue that contrary to their rather linear separation of decision-making followed by implementation delay, it is important to acknowledge the multiple sources of delay, as well as the interactions among these. We, moreover, recognise the fundamental differences between the system-to-be-governed and the governing system (Jentoft and Chuenpagdee 2015), where delay in the former in terms of natural processes (e.g., internal nutrient dynamics, environmental persistence of hazardous chemicals, etc.) is seldom governable per se, in contrast to the latter, where governance decisions are made (e.g. on measures to mitigate total nutrient loads as well as chemical pollution). Identifying inherent ecosystem mechanisms that can delay recovery from e.g. eutrophication and chemical pollution (e.g. nutrient release from sediments, and persistence of substances, respectively) is obviously important when elaborating adequate governance strategies. Such strategies, however, still need to be developed outside the ecosystems, in the governing system where delay has another character.
The framework we propose is built around potential delay mechanisms in the two spheres of science and policy (Fig. 1). It is important to note that our aim is to draw attention in environmental governance research and practice to these mechanisms and we, therefore, rather pragmatically focus on a set of issues of importance for goal achievement, based on previous empirical studies in the field. Obviously, also other mechanisms may contribute to delay, and we hope to stimulate further studies in that respect, as well as of potential underlying factors. However, the subsequent section presents some illustrative examples relating to climate, chemicals, biodiversity and marine environmental policy, being four areas where severe problems persist despite democratically agreed and ambitious policies.
First, in relation to the science domain in Fig. 1, delay can occur in the production, interpretation and appraisal of knowledge, as well as in the generation of science-based advice in relation to environmental goals, and in scientific monitoring and the evaluation of policies. Here, practical challenges are linked to, for example, generation, understanding and evaluation of data, knowledge, and scientific uncertainty, various theory-based and methodological controversies between different scientific approaches, lack of sufficient consensus-promoting arrangements, as well as lack of sufficient resources and incentives for policy-relevant research. The interpretation and use of scientific knowledge and science-based advice may be quite challenging in the policy context, if scientists strongly disagree or are unable to reach consensus (Haas 1992, 2004; Saunders et al. 2017), and is particularly problematic if various stakeholders deny science and employ science-denying strategies (Edvardsson et al. 2017). This means that delay can occur due to science denial, scientific controversies, scientific uncertainty, and normal scientific challenges, as illustrated in Fig. 1. Compared to Varjopuro et al. (2014) we consequently broaden what they label as the monitoring and evaluation step in delay, to the entire scientific field. Our assertion that delay can occur also in the two-way interactions between science and policy similarly broadens the picture and it underlines that processes may be far from rational and sequentially ordered. In the science to policy direction, it is clear that processes of building consensus within epistemic communities may increase the influence of science-based advice and thereby reduce delay (Haas 1992, 2004). To exemplify interactions in the other policy to science direction, it is often observed that politicisation of the science domain, e.g., in terms of how scientific panels or guidelines for risk assessments are set up, can lead to e.g., scientific controversies and loss of trust in science (Eriksson et al. 2010).
Second, within the policy domain, we contend that delay can be caused by a number of different mechanisms, at different places. As referred to above, policy processes can be described by a number of theories, frameworks and models. Evidently, these have developed over the years, from e.g., the policy cycle model to the ACF, and from the garbage can model to the MSF, and there is still a need for advancing the research in the field (Weible 2017). While the policy cycle model, with its criticised assumed ordered and rational links between different policy stages, can still allow for some analysis, the flow diagram of the ACF says more about the nature of policy change and explains why it often takes quite a long time to achieve in contested cases (Jenkins-Smith et al. 2017). The MSF similarly describes how agenda change and decision-making commonly presumes that a number of independent circumstances and factors coincide in practice. In addition to these and other common policy process research approaches (Ruseva et al. 2019), our new framework focuses more directly on a set of potentially important delay mechanisms and on their interactions. In Fig. 1, these include unclear norms and goals, decision thresholds, policy formulation challenges, and implementation deficits, the latter being a comparatively common research area. There are also important linkages to explore; if the operationalisation of goals, for example, does not meet adequate rationality criteria (Edvardsson 2007), the formulation of policy instruments may suffer, as will probably the subsequent implementation of these. Another delay mechanism highlighted in the framework in Fig. 1, being of particular importance and in need of additional attention in both governance research and practice, is decision thresholds, for instance various burden of proof requirements that are theoretically impossible or practically difficult to meet (Sandin et al. 2002; Karlsson 2005; Alfredsson and Karlsson 2016), and other types of obstacles (e.g. Keskitalo and Pettersson 2012).
In the following, we exemplify the new framework with one central delay mechanism per sphere in Fig. 1—denial of science and decision thresholds, respectively—to give some detail on issues in need of more attention, while still keeping the text concise. Even though some previous studies have focused on these two mechanisms they have not, to our knowledge, been placed within a broader framework for analysis of delay in environmental governance. In addition, previous research on environmental science denial has primarily centred on climate policy (Edvardsson et al. 2017), whereas much research on decision-making thresholds in environmental governance has focused on, for example, chemicals policy and regulation (Karlsson 2005). Against the background of a recent research project, we highlight a set of different cases selected to illustrate different situations in which delay occurs due to several factors, namely the four wicked (Alford and Head 2017) problems of climate change, chemical pollution, eutrophication and biodiversity loss, where goals have been shown to be difficult to reach (EEA 2018; SEPA 2019).