This is the accepted version of the article: Bergh, Jeroen C. J. M. van den; Castro Santa, Juana; Drews, Stefan; [et al.]. «Designing an effective climate-policy mix: accounting for instrument synergy». Climate Policy, (March 2021). DOI 10.1080/14693062.2021.1907276 This version is available at https://ddd.uab.cat/record/238738 under the terms of the license Designing an Effective Climate-Policy Mix: Accounting for Instrument Synergy J. van den Bergha,b,c,d, J. Castroa, S. Drewsa, F. Exadaktylosa, J. Foramittia,d, F. Kleina, T. Konca, I. Savina,e a Institute of Environmental Science and Technology, Universitat Autònoma de Barcelona, Spain b ICREA, Barcelona, Spain c School of Business and Economics, VU University Amsterdam, The Netherlands d Institute for Environmental Studies, VU University Amsterdam, The Netherlands e Graduate School of Economics and Management, Ural Federal University, Yekaterinburg, Russian Federation Contact: jeroen.bergh@uab.es Second revision after editorial suggestions, March 2021 Acknowledgements This study has received funding through an ERC Advanced Grant from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation program (grant agreement n° 741087). IS acknowledges financial support from the Russian Science Foundation (RSF grant number 19-18-00262). Abstract We assess evidence from theoretical-modelling, empirical and experimental studies on how interactions between instruments of climate policy affect overall emissions reduction. Such interactions take the form of negative, zero or positive synergistic effects. The considered instruments comprise performance and technical standards, carbon pricing, adoption subsidies, innovation support, and information provision. Based on the findings, we formulate climate-policy packages that avoid negative and employ positive synergies, and compare their strengths and weaknesses on other criteria. We note that the international context of climate policy has been neglected in assessments of policy mixes, and argue that transparency and harmonization of national policies may be key to a politically feasible path to meet global emission targets. This suggests limiting the complexity of climate-policy packages. Key policy insights: - Combining technical standards or targets, such as renewable-energy quota, or adoption subsidies with a carbon market can produce negative synergy, up to the point of adding no emissions reduction beyond the cap. For maximum emissions reduction, renewable energy policy should be combined with carbon taxation and target expensive reduction options not triggered by the tax. - Evidence regarding synergy of information provision with pricing is mixed, indicating a tendency for complementary roles (zero synergy). Positive synergy is documented only for cases where information provision improves effectiveness of price instruments, e.g. by stimulating social imitation of low-carbon choices. - We conclude that the most promising packages are combining innovation support and information provision with either a carbon tax and adoption subsidy, or with a carbon market. We further argue that the latter could have stronger potential to harmonize international policy, which would allow to strengthen mitigation policy over time. Keywords: instrument interaction, technical standards, carbon pricing, adoption and innovation subsidies, information provision, mitigation policy packages. 1 1. Introduction Many academic writings on climate policy are concerned with the performance of single instruments targeting greenhouse gas emission reduction. In reality, however, one typically finds an extensive set of policy instruments implemented simultaneously, often on different regulatory levels. There are many potential reasons for using multiple instruments. The instruments might be complementary or even create positive synergy in terms of the associated goal. More specifically, they might deal with distinct market failures (Jaffe et al., 2005; Freire-González, 2018). Another important reason for combining instruments is that they can accomplish multiple objectives, such as effectiveness, efficiency and equity. In more abstract terms, a policy mix can reflect a second-best (non-optimal) response to a first-best (theoretically optimal) single instrument not being feasible – because of monitoring of pollution being imperfect, the control span of policy being limited, or the existence of political constraints. The additional instruments then compensate for the non-optimal level of the main policy instrument (Bennear and Stavins, 2007). More practically, multiple instruments might arise from political compromises between stakeholders with distinct policy preferences, or from adding instruments to compensate for the insufficiency of already available instruments (Bouma et al., 2018). Finally, in line with the Tinbergen (1952) rule, a distinct type of policy mix results from the presence of multiple objectives, such as climate change mitigation and limiting biodiversity loss (Braathen, 2007; Sterner et al., 2019). There are also several reasons to be careful about combining instruments into a policy mix. As will be illustrated for various cases in Section 3, policies may overlap or create negative synergies. This can even lead them to offset each other in terms of emissions reduction. In such cases, a policy mix would perform no better or even worse than a single instrument. Taking into account that each policy instrument generates an additional cost for the regulator or government in terms of expenditures and human resources – including transaction costs of political and policy processes until implementation, cost of monitor and control, and sometimes serious budgetary sacrifices (such as with subsidies) – policy-makers may want to limit the number of instruments in the policy mix. Moreover, given that policy instruments often cause unintended market distortions, employing multiple instruments runs the risk of introducing potentially multiple distortions into the economy. Finally, policy mixes complicate the comparison of policy stringency among regions and countries compared to single instruments. Schmidt and Sewerin (2019) demonstrate this for renewable-energy policy mixes in nine OECD countries. Difficulty to compare policies in turn confounds policy integration between distinct governance levels within a country or within a supra-national system like the European Union (Howlett et al., 2017). As we argue in Section 4, reducing the complexity of climate policy might increase the feasibility of international policy harmonization. In view of these contrasting arguments, this article examines the synergy of combining specific instruments in a policy mix aimed at effectively reducing greenhouse gas emissions. To this end we consider both theoretical arguments and empirical or experimental evidence from a variety of disciplines that have devoted attention to policy mixes, notably economics, psychology, and innovation and 2 transition studies (Jaffe et al., 2005; Bulkeley and Kern, 2006; Howlett and Rayner, 2007; Oikonomou and Jepma, 2008; Oikonomou et al., 2010; Rogge et al., 2017; Mundaca et al., 2019; Somathan et al., 2014, Section 15.7.3). This allows us to obtain a comprehensive picture of possible combinations of climate policy instruments that can achieve non-negative or even positive interactive effects on emissions reduction. Most of the aforementioned reviews do not focus on systematically assessing synergy of particular instrument combinations as we do here, nor do they include all the instrument combinations we address. Hence our study adds to the existing literature. While we focus on climate mitigation policy in the context of emission reduction, there are also reviews or synthetic studies of policy mixes for energy-efficiency (Boonekamp, 2006; Hood, 2013; Rosenow et al., 2017; Wiese et al., 2018), renewable energy (Pitelis, 2018), accelerating technological change (Rogge and Reichardt, 2016), or broader environmental issues (Lehmann, 2012). Following the logic of Bowles (2016, Appendix 1) for the relationship between incentives and social preferences, we distinguish between four cases of interaction between policy instruments, namely (i) no (zero) synergy, (ii) positive synergy, (iii) (moderately) negative synergy, and (iv) backfire. The first case indicates that the overall effect of a policy mix is the sum of the individual instrument effects, meaning there are no synergistic interaction effects, or the instruments are independent and complementary. This is also known as additive separability. The second case describes cases in which one instrument reinforces another, meaning a positive interaction effect is at stake. This case is sometimes referred to as super-additivity, super-modularity and crowding-in.1 An example is information provision creating awareness which in turn strengthens an incentive effect. The third case reflects that one instrument weakens another, such as when monetary incentives crowd-out intrinsic pro- environmental preferences. Here the interaction is negative, and the outcome is variably known as substitutability, sub-additivity or crowding-out. This happens, for instance, if instruments overlap in their impact on particular decisions by agents and associated emissions, so that the effect of the policy mix is lower than the sum of the isolated effects. The instruments are then (partial) substitutes of one another. An extreme version of this is the fourth case of ‘backfire’, denoting that one of the instruments offsets the effect of the other, causing the policy mix to perform worse than the best-performing instrument alone. This differs from moderate negative synergy, which is still reasonably effective in that it means that the combination of instruments reduces emissions more than a single instrument alone. The boundary between these two cases is what we will call “compensating negative synergy”: here negative synergy results in overall emissions reduction equalling B, i.e. the level achieved by the most effective instrument alone. Figure 1 illustrates these various potential outcomes of interactions between two policy instruments. 1 Note that cases (i) and (ii) are sometimes also referred to as instruments being complementary. To avoid confusion, we use in this paper the term ‘complementary’ strictly for case (i). 3 Figure 1. Potential outcomes of instrument interactions Note: Adapted from Drews et al. (2020). The horizontal dashed line on top separates between positive and negative synergy, and the one underneath between backfire and other negative synergy. The remainder of this article is organized as follows. Section 2 discusses the approach, consisting of identifying main categories of instruments and assessing possible interactions between these. Section 3 examines the performance of specific instrument combinations in terms of emissions reduction. Section 4 summarizes the findings and generalizes these for more than two instruments, discussing relative advantages and disadvantages of a set of potential climate policy mixes. Further, it pays attention to the international context of climate policy and its implications for policy mixes. Section 5 concludes. 2. Approach There are many classifications of environmental and climate policy instruments aimed at altering behaviour of firms and households to achieve reductions in carbon emissions (Sterner, 2002; Bulkeley and Kern, 2006; Goulder and Parry, 2008; Somathan et al., 2014). Here we focus on a set of instruments that has received considerable attention in the literature on policy mixes: 1. Performance and technical standards. Performance standards can take various forms, such as a quota on a sector’s emissions, a renewable energy portfolio target (e.g., a certain share of renewables in electricity generation), phasing out fossil fuels (e.g., coal), or banning use of high- emission cars in cities. Technical standards specify minimum criteria for consumer or production technologies, such as fuel-efficiency standards for cars or best available technologies for pollution abatement. 4 2. Carbon pricing means incorporating the direct and indirect external costs of CO2(-eq) emissions in the prices of fuels, resources, intermediate products, and final goods and services. This can be done through a carbon tax or emissions trading (market creation or tradable permits), which do not necessarily generate identical outcomes (Goulder and Schein, 2013; Foramitti et al., 2021).2 3. Adoption subsidies encourage the adoption and diffusion of energy-efficient or low-carbon products by financially rewarding the purchase of energy-efficient or low-carbon products (e.g., hybrid or electric cars). 4. Innovation support aims to increase private investment in, and success of, research and development (R&D) on energy-efficient and low-carbon technologies and products. A common way to do this is through R&D subsidies. Many other instruments fall into this broad category (Dolfsma and Dongback, 2013): e.g., green procurement, funding basic research at universities and in research centres, or legislation on intellectual property rights. 5. Information provision and nudges address lack of information – such as limited awareness of environmental challenges, solutions or associated behavioural options – and behavioural barriers – such as limited attention, self-control issues and status concerns. Instruments of the first type include providing basic information, often in a simplified form (e.g., eco-labels and energy certificates). Instruments of the second type, targeting specific behaviours, entail reminders of excessive energy use, commitment devices, and feedback about choices of others (e.g., smart electricity meters). One can distinguish symmetric from asymmetric instruments mixes: while in the former case all sectors are subject to the instruments, in the latter case some sectors receive different treatment than others. For example, renewable electricity is subject to future targets and feed-in-tariffs, EU-ETS is limited to large industrial and energy power firms, a carbon tax is often focused on non-export sectors, and concrete technical standards apply to particular sectors – such as fuel-efficiency standards in car manufacturing. The approach we followed to select studies for review consists of two parts. The first is a snowball method: we looked for relevant studies in existing reviews which offered a good entry point into the literature. From there we skimmed the reference lists of all identified papers until we did not find additional suitable papers. The second part involved a search in Scopus, using the following search terms: 2 A cap-and-trade system can itself already be regarded as a mix of policy instruments, namely a quantity-based instrument (i.e. a cap to emissions, defining the sum of all emission permits) and a price-based instrument (i.e. variable price due to trade of permits). Such a hybrid policy has various advantages (Hepburn, 2006; Grüll and Taschini, 2011). Specific additional elements add further benefits, such as limiting volatility of the carbon price through a minimum price or price floor, and avoiding unsurmountable costs for emitters through a maximum price or “safety valve” (Jacoby and Ellerman, 2004; Philibert, 2009). 5 (“policy interaction” OR “policy synergy” OR “instrument combination” OR “instrument mix” OR “instrument interaction” OR “instrument synergy” OR “instrument combination” OR “instrument mix”) AND (“climate” OR “energy” OR “emission*”). This generated a list of 273 studies of which a little over ten percent were considered relevant to our specific purpose of assessing instrument synergy in regard to emission reduction. Most of these studies are from environmental economics, while additional ones come from energy studies, environmental science, psychology and other policy areas. We excluded studies that did not focus on instrument synergy, did not provide a clear basis for their claims, or did not relate directly to climate policy and emission reduction. 3. Specific instrument combinations for emissions reduction 3.1 Carbon pricing combined with performance or technical standards First, we consider combinations of pricing instruments, such as taxes or cap-and-trade, and direct regulation, such as quotas/targets or standards. As this topic has already been covered well in a previous IPCC report (Somathan et al., 2014), we will keep it relatively short. With cap-and-trade the addition of specific sector targets, technical standards (Thurber et al., 2015; Beckenbach et al., 2018) or a forced phase-out of an energy source, such as coal (Osorio et al., 2020), will certainly result in negative synergy. Theoretical studies indicate this (Roberts and Spence, 1976; Christiansen and Smith, 2012, 2014). The reason is that the target and the standard reduce emissions in sectors that then need fewer permits, leading to lower permit prices and higher emissions in other sectors subject to the cap but not the target or standard. This is known as the “waterbed effect”: emissions are not reduced but just appear in another place (Perino, 2018). It means that the carbon market price is determined not only by marginal abatement costs of emitting firms but possibly also by the other policy instrument, namely if it stimulates or obliges emissions reduction with a relatively high marginal cost (i.e. beyond the permit price in an isolated carbon market). The overall abatement cost will come out higher (Sijm, 2005; OECD, 2011). This holds equally for countries and multilevel systems. Well- intended energy policies by states or countries, respectively, can undercut the consistent approach aimed for by a national or supra-national policy such as the EU-ETS, frustrating the effectiveness and efficiency of emissions reduction. An assessment for the EU by Böhringer and Rosendahl (2011) finds more than 60% cost increase of achieving 25% CO2 reductions when a renewable energy quota is set 10% points higher than the endogenous renewable energy deployment level under the EU-ETS, and that the permit price falls from €41 to €16 per ton of CO2. Table 1 clarifies how synergy resulting from adding renewable energy policy (REP) in the form of standards or targets to carbon pricing differ between a carbon tax and carbon market creation (such as emission trading). With a carbon market, the outcome is compensating negative synergy (on the boundary of backfire; see discussion of Figure 1) making the second instrument unnecessary, while with 6 a carbon tax synergy can range from negative to zero (complementarity). For maximum effectiveness REP is thus best combined with a carbon tax rather than carbon market creation. Moreover, as explained by the difference between the two cases in the columns, such policy should avoid targeting relatively cheap options that are already triggered by the carbon tax.3 Table 1. Emissions reduction (and synergy) of combining carbon pricing and renewable energy policy Combined effect (emissions reduction and synergy) with renewable energy policy MAC > p MAC < p Carbon market emissions reduction X emissions reduction X implying synergy -Y implying synergy -Y Carbon tax emissions reduction X+Y emissions reduction X implying synergy 0 implying synergy -Y Notes: X denotes emissions reduction of carbon pricing on its own (isolated carbon market and tax are equivalent in terms of emissions reduction, i.e. tax=market price=p); Y is emissions reduction of renewable energy policy (REP) on its own. MAC denotes the marginal abatement costs of renewable energy investment. Of course, a part (say a, with 0