全球变化知识资源中心

The international climate regime represented by the United Nations Framework Convention on Climate Change has been widely critiqued. However, 'new' dynamic forms of climate governing are appearing in alternative domains, producing a more polycentric pattern. Some analysts believe that the new forms will fill gaps in the existing regime, but this optimism is based on untested assumptions about their diffusion and performance. The advent of polycentric governance offers new opportunities for climate action, but it is too early to judge whether hopes about the effectiveness of emerging forms of climate governance are well founded.

It is a truism that humanity is struggling to govern climate change. In spite of all the resources invested in the regime centred on the United Nations Framework Convention on Climate Change (UNFCCC), emissions continue to rise, dramatically reducing the probability of remaining within 2 °C above pre-industrial temperatures¹. Achieving the emissions reductions that are factored into many low-concentration pathways arguably requires new and much more 'integrated and aggressive'² forms of governance (that is, modes and mechanisms to steer society)³. But where will these new forms originate, how will they diffuse, and what factors will shape their ability to perform as hoped?

Most analysts used to assume that the innovative thrust in governance would spring from a comprehensive global climate regime⁴. However, even before the failure of the 2009 Copenhagen conference, some international relations scholars had moved on from the idea of a single, monocentric regime to consider multiple, interlocking 'regime complexes', such as those focusing on trade, energy, and climate^{5, 6, 7, 8, 9}. What is striking about this strand of work is that while it hints at the potential of more pluralistic forms of governing⁹, its scale is still international and its underlying ontology remains essentially top down and state centric.

While this is clearly an important and flourishing perspective, there is a growing belief that it is only a partial one, and that the landscape of climate governance has extended beneath the international level¹⁰ through changes initiated by numerous actors from different backgrounds, such as business, local government, and civil society. Armed with less top-down, more governance-centred analytical frameworks, social scientists have started to chart the changing landscape of climate governance, now increasingly populated by novel forms, including emissions trading systems¹¹, offsetting standards, emissions registries, carbon-labelling schemes, and collaborations between cities^{4, 12}. These efforts have spilled back into the UNFCCC negotiations to some degree, with discussions on climate action pre-2020 engaging with non-state actors more deeply (for example, through so-called technical expert meetings and the UNFCCC Secretariat's web portal: Non-State Actor Zone for Climate Action; http://climateaction.unfccc.int).

The argument that to become more effective, climate governance in toto should become more diverse and multi-levelled is not new^{13, 14, 15}. Economists have long debated the theoretical merits of linking national and/or regional emissions trading systems¹⁶. Political theorists have also emphasized the advantages of governing from the 'bottom up' — more scope for experimenting, a better fit with local priorities, and so on — for considerably longer¹⁷. What has changed is that fresh empirical efforts are now revealing that the emergence of such forms of governing has a solid basis in empirical reality, and the overall landscape of climate governance has started to exhibit some of the characteristics of polycentricity foreseen by Elinor Ostrom, that is, more diverse, multi-levelled, and with a much greater emphasis on bottom-up initiatives¹⁸.

A vibrant and energetic debate is underway concerning the merits of a more polycentric approach to climate governance, only some of it reported in the IPCC's Fifth Assessment Report. In stark contrast to discussions centred on the UNFCCC, this debate is exciting precisely because it appears to offer empirical validation for a broader narrative of dynamism in a world disillusioned with the UNFCCC process^{4, 19}. However, we believe that the challenge confronting scholars of the new climate governance is to ensure that this positive narrative remains empirically informed and attentive to the tendency for over-enthusiasm to creep into studies of innovative activity²⁰.

In this Perspective, we directly address the challenge of ensuring that expectations remain informed by evidence by critically reflecting on the opportunities created by, and the limitations inherent in, the new forms of governing. We draw on insights from two even more recent strands of research, broadly covering the national (including sub-national) and the transnational domains, and relate them to the findings emerging from the first (and much older) strand of work on international/global governance noted above. We argue that much deeper connections between them are needed to understand better the opportunities and pitfalls of both the new and the older climate governance. To that end, we reveal that existing work on the new governance has uncovered many new and important insights, but is yet to comprehend the complex interconnections between the transnational, national, and international domains.

In seeking to encourage a more holistic and reflexive understanding of climate governance, we examine how far the three strands address three important, cross-cutting topics, relating to:

• Distribution: What forms of governing are emerging, when, and in which sectors and/or countries? How new are they?
• Initiation, origins and invention: Why are the new forms of governing emerging and through which mechanisms are they diffusing and/or scaling up?
• Performance: What do they actually add up to, for example, in terms of emissions reduced?²¹ More broadly, are they filling 'gaps' in the regime²² or reproducing what is already there?

We reveal that work on the new climate governance is addressing the first of these topics, but should do more to tackle the other two. We find that the three strands (international, transnational, national) are broadly complementary in their approach, but much deeper collaboration, organized around shared terms and concepts, is required to produce a more holistic picture. In the final section, we explore the most important policy-relevant research gaps that emerge from our analysis.

One of the most dynamic strands of research activity concerns transnational forms of governance²³. For Abbott²⁴, these span national borders, dissolve the traditional analytical divide between public and private spheres, and are decentralized. Analysis of a number of databases^{4, 23, 24, 25} reveals that these transnational initiatives are numerous and highly diverse in form (ranging from setting rules to sharing information) and scale (from global down to city level). Most are relatively new (post-2005) and are mainly mitigation focused²³. Most have been initiated by actors in industrialized countries, albeit with the active participation of actors from the global South^{23, 26}. For example, six countries (Bangladesh, Canada, Ghana, Mexico, Sweden, and the United States) and the UN Environment Programme launched the Climate and Clean Air Coalition in 2012 to promote action on short-lived climate pollutants by state and non-state actors. Projects are being implemented to reduce levels of black carbon, methane, and hydrofluorocarbons across a membership (in 2013) that has expanded to 43 states and 53 non-state partners^{27, 28}.

Another equally dynamic strand of research focuses on the public policy-making activities of nation states, including local governments. Until recently, the literature on national policy offered only broad overviews of whole countries and/or non-cumulative case studies of specific polices and instruments. But attempts are now being made to assemble a fuller and more detailed picture, also based on large databases. These reveal much greater dynamism than many originally assumed. For example, The Globe's database reveals that there were 487 climate change-related laws and policies in 66 countries in 2013, up from only 40 in 1997²⁹. Around 30 new policies are being adopted each year, with non-Annex 1 countries being especially active. Similar findings have been uncovered in a comparable database produced by Dubash et al.³⁰, which confirms that adoptions are occurring faster in non-Annex 1 countries and indicates that the share of world population covered by national policy doubled between 2007 and 2012. Moreover, states are not only actively innovating in relation to mitigation — the number of new national adaptation strategies has also grown spectacularly in the past decade³¹.

The rediscovery of the state as a dynamic site and catalyst of governing is a little unexpected, especially for those who identify states as a primary cause of governance failure³². These findings tie in with Green's suggestion³³ that global climate governance is a positive-sum game where governance efforts by state and non-state actors grow simultaneously and in a mutually reinforcing manner. Together, these findings emphasize the need to work across all three strands of the new climate governance literature.

These two strands of research suggest that there is a new climate governance emerging (that is, dating mostly from the mid-2000s), although we certainly do not wish to over-stress its novelty (some national policies predate the UNFCCC)^{34, 35, 36, 37, 38}. In fact, we firmly believe that now is the right time to debate the most appropriate analytical categories to measure the distribution of the new climate governance.

At national level, attempts to open up the analytical category of 'policy' — which tends to be 'black boxed' by those who focus on international processes — reveal that it can be characterized in multiple ways: some policies are legally binding whereas others are not; some are adopted by national policymakers, others by sub-national actors such as local governments and city mayors; some are explicitly labelled as climate policy whereas others are primarily seen as relating to older policy areas (finance, transport, housing, or forestry)^{39, 40}.

Although these distinctions may appear subtle, the multidimensionality of policy is important, perhaps revealing underlying political motives. Dubash et al.³⁰, for example, have found that non-binding strategies are being adopted at a much faster rate than legally binding policies. Variation is also evident in the sense that countries with very similar emission reduction targets routinely employ different mixes of implementing instruments that are harnessed by these policies⁴¹. Variations in policy type may also reveal politicians' expectations about implementation and, eventually, performance⁴². Although the existing distribution and internal characteristics of policy have probably not had a very significant effect on net emissions (relative to non-policy effects — see below), they may nevertheless open up new political opportunities to collaborate, such as by linking similar instruments (such as emissions trading) in different countries.

Two important challenges remain for those working on national policies. First, scholars must find better ways to define and track policy adoptions and innovations — an under-appreciated challenge when policy-making activity is constant and no single actor is responsible for collecting comparable data. As the international regime shifts to a more bottom-up architecture, with each party pledging and reviewing its own policy commitments (in UNFCCC speak: 'intended nationally determined contributions'), we expect this hitherto largely academic task to become considerably more policy relevant. Second, having collected comparable data, explanations for what is causing observed differences can be sought — a precondition for altering, through purposive steering, the future orientation of governance. Again, experience suggests that explanations will take time to derive. Decades of comparative policy analysis suggest that cross-national variations are likely to arise from a complicated nexus of factors that are internal and/or external to particular countries⁴¹.

Those working on transnational governance have also begun to reflect on the distribution of non-state initiatives. As territorial categories are assumed to be less relevant, other categorizations have been invented. Abbott²⁴, for example, reveals that transnational initiatives perform myriad functions, including some (such as rulemaking and implementation) that are usually considered to be the preserve of states. Two, however, stand out: capacity building and information sharing¹² — governing functions where the state's comparative advantage is relatively low, especially at a restricted spatial scale. This finding suggests that actors may be self-organizing around a mutually beneficial division of labour, as envisaged in Ostrom's conceptualization of polycentricity^{18, 43, 44, 45}.

Finally, much like the databases of policy activity, most (but not all⁴⁶) studies of transnational governance offer rather static snapshots that struggle to account for the dynamic processes of evolution, diffusion, and performance. For example, how do they spread across borders and in which countries are they most active? What factors help to bring about the mutual learning from these processes of experimentation, without which the oft-claimed advantages of polycentrism⁴⁷ might not emerge? And who remains actively involved once a scheme is up and running, for how long, and why?

The two new strands of literature would seem to bear out the general claim that climate governance has become more polycentric^{48, 49}. Why it has become so is much less clear, not least because both strands offer partial perspectives. This constitutes a significant research gap. In environmental social science, two broad categories of motivation to engage in multi-actor governance are normally cited: financial and non-financial⁵⁰. How well do these carry across to the new climate governance? As regards to transnational governance, most scholars are still identifying potential sub-categories of motivation, including moral concerns, fear of new regulation (or the opportunity to secure first-mover advantages by shaping it), the pursuit of direct financial rewards, indirect or 'non-climate' benefits (for example, reputational enhancement), and the satisfaction of consumer expectations^{4, 23, 24, 33}.

Studies of the emergence of new national policies have also hypothesized — but not yet fully tested for — similar meta-motivations. Dubash et al.³⁰, for example, mention the need to comply with UNFCCC requirements, the desire to reap competitive advantages, and/or indirectly empower pro-environmental political actors (see also ref. 51). Studies of specific policy innovations have tried to disentangle these motivations using large n statistical techniques^{52, 53}. We have already noted how such studies tend to gloss over the subtle but important differences between and within individual policies; analysts are also becoming more aware of their insensitivity to slow processes of refinement as policies diffuse and take root in particular jurisdictions⁵⁴. Case study analyses have shown that groups advocating particular policy instruments (such as emissions trading¹¹) drive these processes, often in collaboration with policy entrepreneurs^{55, 56}.

Centre stage in these policy adoption processes are (sub)national politicians — a distinct actor category all too often ignored by transnational and international scholars. One of the enduring puzzles in public policy analysis is what motivates politicians to address climate change: to claim credit by adopting successful and innovative policies, to avoid blame for things that go wrong, or to generate a long-term policy legacy⁵⁷? Howlett⁵⁸ argues that when it comes to long-term problems such as climate change, where the causal chains connecting specific policy interventions and impacts are convoluted, politicians will normally opt to do nothing (or at the most, very little) rather than something bold (such as adopt a binding medium-term emissions target) for which they might eventually be blamed by powerful interest groups and/or voters.

But if this blame-avoidance motivation is really as common as Howlett suggests, what is driving the new policy activity noted above? Are politicians engaging in complex forms of political risk management, for example, by emulating and learning from what other countries are doing⁵⁹ — a form of diffusion that might in turn enable greater polycentrism? Or are they, as Dubash et al. seem to indicate³⁰, engaging in a more negative form of policy innovation, which is symbolic and/or simply intended to capture funding from abroad⁵⁸? Researchers have no