Expert Perspectives

Anticipating Negative Feedback and Avoiding Premature Equilibria in a Low-Carbon Transformation

Limiting global temperature rise to 1.5 or 2 degrees Celsius (°C) and averting dangerous climate change will require sustained behavioral change at the individual and collective level among citizens, businesses, and governments. Practitioners and scholars broadly agree that fostering such behavioral change requires policies that gather momentum over time and induce a cascading effect at the systemic level (Cashore et al. 2016; Jordan et al. 2018; Schmidt et al. 2019). The focus of such studies, therefore, has been on identifying not only how emissions in a sector or region might ratchet lower but also the mechanisms through which the desired behavior can be ingrained (Meckling 2011).

What criteria should policies satisfy that aim to avert the climate crisis by fostering sustained behavioral change through a cascading effect? To address the “super wicked” problem of climate change, Levin et al. (2012) proposed that decision-makers ask (and answer) the following diagnostic questions (DQs):

  • DQ1: What can be done to create stickiness, making reversibility immediately difficult?
  • DQ2: What can be done to entrench support over time?
  • DQ3: What can be done to expand the population that supports the policy?

They argued that policies that respond to these questions can create a virtuous cycle through which, as time passes, the original objectives are reinforced through a series of “forward steps” (i.e., positive feedback), leading to significant and durable behavioral change. The initial task facing those who seek to create behavioral change was to find “easy to pull levers” that were “hard to reverse.”

It has become clear in the last decade that some policy initiatives designed with this approach in mind have fallen short of their designers’ hopes—such that behavioral change has been much more limited than expected. For example, although the German feed-in tariff changed the behavior of energy producers and consumers and facilitated investment in renewable energy technologies, some are concerned that it risks plateauing early and falling short of the behavioral changes needed to contribute to the 1.5 or 2°C goal. Similarly, although the British Columbia carbon tax, designed for entrenchment and heralded as a leading policy tool, emphasized coalition building through payments to municipalities and taxpayers, it appears to have plateaued far earlier than what would be needed to induce meaningful behavioral change. We argue that a critical shortcoming in each case was not simply that the policies were not designed well enough to cast forward multiple steps of policy “entrenchment” but also that they paid less attention to simultaneous “backward steps” (i.e., negative feedback) that undermined support for the intervention long before it could reach its objectives.

In this paper, we reflect on how to avoid policy reversals and identify new ways to “cast forward” innovative policy design strategies that trigger durable pathways fostering sustained behavioral change. This effort is timely for two reasons. First, even when no apparent backward steps are in play, existing studies and strategies have been unable to distinguish clearly whether the initiatives at hand, and steps involved, will lead to behaviors capable of meeting the 1.5 or 2°C target rather than, say, a 5°C increase. That is, although Levin et al. (2007, 2012) highlighted that only triggers capable of addressing the climate crisis ought to be activated and others avoided, the vast majority of research, analysis, and practitioner engagement in the last decade has failed to make this distinction theoretically or empirically.1 Second, the original focus on triggering “bottom-up” climate action from the private and nonprofit sectors, while important, may be inadequate to cause systemwide behavioral change that advances a low-carbon transformation (Rosenbloom et al. 2019).

As a corrective, we integrate the applied forward reasoning approach proposed by Levin et al. (2012) with more recent research in policy studies and policy design. We posit that strategists, practitioners, and scholars should view policy design as a multistep process. From this perspective, effecting sustained behavioral change also requires anticipating and addressing the influence of policy formulation and implementation on three related but distinct parts of the system: societal issues and problems, future policy alternatives, and ensuing politics (Kingdon 1984). We offer three additional diagnostic questions that can help identify creative alternatives for designing durable policies that initiate and sustain behaviors necessary for a low-carbon transformation.

What is policy durability?

Policy design consists of three elements nested within policy ends—high-level abstract goals, program-level objectives, and on-the-ground settings—and three nested within policy means—general intervention logic, specific policy tools, and the calibrations in which those tools are deployed (Cashore and Howlett 2007; Howlett and Cashore 2009). The means achieve the ends when they influence the behavior of individuals or collectives in a particular direction (Weaver 2009).


Table 1. Elements of a policy


Policy Level

High-level abstraction (policy orientation)

Operationalization (program)

On-the-ground specification (measures)

Policy content

Policy ends (aims)


What general types of ideas govern policy development?

e.g., environmental protection, economic development, social cohesion


What does policy formally aim to address?

e.g., saving wilderness or species habitat, reducing greenhouse gas emissions


What are the specific “on-the-ground” requirements of the policy?

e.g., size of protected areas, level of carbon tax

Policy means (instruments)

Intervention logic

What general norms guide policy instrument preferences?

e.g., coercive “command and control,” voluntary, markets, and neoliberal norms


What types of instruments are utilized?

e.g., tax incentives, loans, public enterprise, cap-and-trade carbon markets


What are the specific ways in which the instrument is applied?

e.g., qualification of tax incentives, rules governing cap-and-trade markets, such as specifics on leakage, allocation of resources, and approach to enforcement

Source: Adapted from Cashore and Howlett (2007, 536) and iterations since then.


Note:  an earlier version of the 2007 AJPS table (Cashore and Howlett 2007, 536) appeared in Cashore and Howlett (2004, 10) and Cashore and Howlett (2006, 150). This table expanded Hall’s (1993) first-, second-, and third-order typology of policy change into six “elements or components” (Cashore and Howlett 2007, 535). Cashore and Howlett’s (2007) application of these six “elements or components” uncovered historical patterns of endogenous and exogenous policy development that were inconsistent with Hall’s theories of policy learning and policy change. This table also fostered greater conceptual and empirical work about measuring policy change, and how to conduct forward-looking policy design, including identification of policy (Howlett et al. 2017; Peters 2018, 28).

Policy durability is different from policy stability in that it does not refer to the absence of policy change; instead, it refers to policy design (and redesign, if necessary) to ensure goal attainment. In the context of climate change, therefore, a durable policy or a pathway is one that effects and sustains the behavioral change necessary to meet the 1.5 or 2°C target despite, or due to, changes in on-the-ground settings and policy instruments.

While policy durability might be achievable in a single step, usually multiple steps are involved in which two logics exist: (1) distinct logics within each step and (2) the logics for moving from one step to the next to reinforce or expand the desired behavior. For example, some policy design choices can simultaneously draw support from seemingly opposing groups—such as “Baptists” who favor prohibition on alcohol and “bootleggers” who thrive on selling it illicitly—and open a window of opportunity to change behavioral norms in society. When such policy design is durable, it can lead to a series of incremental changes that create a “race to the top” and sustain desirable behavior (Vogel 1997). However, unless followed by subsequent change that reinforces the desired behavior, such arrangements can also unravel and undermine the original problem, result in unintended consequences, and/or cause a “race to the bottom” instead.


Table 2. Steps to policy change and durability


Number and size of steps

Few, large

Several, small


Durable change

(A new “equilibrium” is established)

Classic paradigmatic


Progressive incremental

Nondurable change

(Change is temporary, goes back to original position)

Faux paradigmatic

Classic incremental

Source: Adapted from Cashore and Howlett (2007, 536) and Durrant and Diehl (1989).


Designing, therefore, is typically a multistep endeavor for catalyzing “progressive incremental” behavioral change rather than a one-step solution to effect paradigmatic change (Schneider and Sidney 2009). Careful thought must be given to each element in the policy matrix, within a given step, as well as to the evolution of this matrix from each step to the next, as each element (and each step) has its own logic. Anticipating and addressing the consequences of policy design choices (that is, thinking about the relationship of a given step to the subsequent steps) can allow practitioners to reinforce policy durability and sustain low-carbon behavior through positive feedback (Levin et al. 2012) while identifying and addressing the “weakest link in the chain” to ensure resilience to negative feedback (Jordan and Matt 2014; Weaver 2010).

How can one systematically anticipate and identify issues that threaten to undermine a low-carbon transformation? One promising approach is to integrate applied forward reasoning, offered by Levin et al. (2012), with a multiple-streams view of the policy process (Kingdon 1984) and propose that practitioners and scholars reflect on how policy design choices influence societal issues or their perception (i.e., the problem), policy alternatives, and politics in the future.2 Such a view allows us to distinguish distinct but related activities and behaviors in the policy process.

“Streams” of influence

As is well known, Kingdon (1984) viewed the policy process as an interplay among three streams: problem, policy, and the politics. The problem stream is the cumulation of perceptions of societal conditions while the policy stream witnesses the development of policy proposals. Distinct from these, the politics stream consists of factors such as public mood, party ideology, and interest groups. Each stream, thus, represents different sets of activities and associated behaviors. The problem stream, for example, is influenced by the behavior of “policy targets” while the politics stream is influenced by the behavior of politicians, interest groups, and citizens. Initiating and sustaining behavioral change among policy targets requires adequate “coupling” among the streams until policy ends are achieved (i.e., policy durability); while positive feedback reinforces alignment, negative feedback threatens to decouple streams through its effect on the problem stream, the policy stream, and/or the politics stream. A better understanding of the type of feedback and the stream(s) it influences can, therefore, facilitate more durable policy design.

Problem stream

Endogenous of exogenous changes that threaten to undermine problem durability can decouple the problem stream. These include not only factors that ameliorate societal conditions but also those that simply alter the salience or the emotive strength of the problem. Unintended consequences or whack-a-mole effects created by policy implementation can also decrease problem durability (Sewerin et al. forthcoming). Consequently, the policy can face pushback or noncompliance instead of triggering the desired behavioral change. An example of such an influence was observed in the case of the European Union’s biofuels policy, in which the influence shifted from self-reinforcing to self-undermining due to increasing awareness about the environmental impacts of biofuels (Skogstad 2017). Another example of how a decrease in problem salience can undermine low-carbon behavior is observed in the case of energy-efficiency improvement, where the deployment of a technological solution can result in rebound effects (Giest and Mukherjee 2018).

Policy stream

The policy stream consists of a “mix” of various policy tools that are layered over time to address issues within or across policy areas. The durability of individual policy instruments depends not only on criteria such as technical feasibility, value acceptability, public acquiescence, and receptivity among elected representatives (Kingdon 1984) but also on how the ideas and meanings attached to them change or evolve over time (Wilder and Howlett 2014; Dudley 2003). To initiate and sustain behavioral change, policy tools need to fit with the policy context (Jiao and Boons 2017), make accurate assumptions regarding the existing behavior of “policy targets” (Howlett 2018), and explicitly indicate the direction of desired behavioral change (Thomann 2018). Further, the durability of policy tools is also influenced by feedback from other tools deployed as part of the policy “mix” (Edmondson et al. 2018; Hoppmann et al. 2014; Schmidt et al. 2017). Policy design can be more durable when self-reinforcing or coevolutionary policy tools are mixed into the policy. For example, China ensured diffusion of its sustainable industrial parks by combining two programs, one that promoted ecoindustrial parks and another that promoted circular economy industrial parks, in a self-reinforcing manner (Jiao et al. 2018). An “adaptive” policy that offers design flexibility is also likely to be more durable in the face of uncertainties associated with climate change (Walker et al. 2001).

Politics stream

Among the three streams, political durability has probably received the most attention from scholars. Lockwood (2013), for example, argued that although the United Kingdom Climate Change Act of 2008 created new institutions to lock in commitment, it did not sufficiently “displace the power of existing ones.” As suggested by Levin et al. (2012), one way to ensure political durability is to create a coalition in favor of low-carbon behavior and expand its base over time. However, it is equally important to pay attention to negative feedback depending on the power dynamics within and outside such a coalition. Haelg et al. (2018), for instance, found that coalition membership changed during the policy process depending on the distribution of benefits and the costs of policy design choices. Similarly, in the case of the German feed-in-tariff, the deployment of renewable energy has altered the coalitions of interest, and the government has introduced an auctioning regime not only to control cost but also to protect conventional energy industries (Leiren and Reimer 2018). It is therefore important to anticipate whether a coalition for low-carbon behavior is vulnerable to realignment of internal interests or countermobilization of external interests.

The approach offered here does not imply that problem durability, policy (design) durability, and political durability are independent of each other. In fact, feedback that undermines the durability of one is quite likely to undermine the durability of the remaining two over time. Instead, we should focus on understanding and anticipating how negative feedback “backward steps” are likely to influence long-term behavioral change, and then take action such that the unraveling of policies never occurs.

Extending applied forward reasoning for durable policies and sustained behavioral change

We propose the following additional diagnostic questions to shield desirable policies against negative feedback and sustain low-carbon behavior:

  • DQ4: What can be done to ensure problem salience over time?
  • DQ5: What can be done to maintain policy design flexibility in the future?
  • DQ6: What can be done to prevent countermobilization or the coalitions’ realignment based on other interests?

To apply these questions in practice, strategists should carefully articulate the mechanisms of their policy tools and why they are likely to cause behavioral change. They should approach policy design as a multistep process and envision how the first step will influence the problem, policy, and politics streams. This exercise should involve not only thinking about the “forward steps” or positive feedback but also anticipating and avoiding “backward steps” or negative feedback. These exercises then pave the way for assessing how each stream might undermine the alignment, or “coupling,” of the three streams. Depending on whether the weakest link is identified in the problem stream (DQ4), policy stream (DQ5), or politics stream (DQ6), creative solutions to address the respective diagnostic questions can be identified; the process should be repeated with this solution as part of the policy design. Uncovering the multistep logics of each stream, and figuring out whether, when, and how to link them, can help policy entrepreneurs identify windows of opportunity to reinforce low-carbon policies and guard against influences that undermine durability and hinder long-term behavioral change.


Cashore, Benjamin, Graeme Auld, Steven Bernstein, and Kelly Levin. 2016. “Paris Could Be Different: But It Requires Policy Makers Apply Path Dependency Analysis to the ‘Super Wicked Problem’ of Climate Change.” Macmillan Center, Yale University.

Cashore, Benjamin, and Michael Howlett. 2004. “Punctuating What Equilibrium? Institutional Rigidities and Thermostatic Properties in Pacific Northwest Forest Policy Dynamics.” New Haven, CT: Yale School of Forestry and Environmental Studies.

Cashore, Benjamin, and Michael Howlett. 2006. “Behavioural Thresholds and Institutional Rigidities as Explanations of Punctuated Equilibrium Processes in Pacific Northwest Forest Policy Dynamics.” Chapter 6 of Punctuated Equilibrium and the Dynamics of U.S. Environmental Policy, edited by Robert Repetto. New Haven, CT: Yale University Press.

Cashore, Benjamin, and Michael Howlett. 2007. “Punctuating Which Equilibrium? Understanding Thermostatic Policy Dynamics in Pacific Northwest Forestry.” American Journal of Political Science 51 (3): 532–51.

Dudley, G. 2003. “Ideas, Bargaining and Flexible Policy Communities: Policy Change and the Case of the Oxford Transport Strategy.” Public Administration 81 (3): 433–58. doi:10.1111/1467-9299.00355.

Durrant, Robert F., and Paul F. Diehl. 1989. “Agendas, Alternatives, and Public Policy: Lessons from the U.S. Foreign Policy Arena.” Journal of Public Policy 9: 179–205.

Edmondson, D. L., F. Kern, and K. S. Rogge. 2018. “The Co-evolution of Policy Mixes and Socio-technical Systems: Towards a Conceptual Framework of Policy Mix Feedback in Sustainability Transitions.” Research Policy, March 28. doi:10.1016/j.respol.2018.03.010.

Giest, Sarah, and Ishani Mukherjee. 2018. “Behavioral Instruments in Renewable Energy and the Role of Big Data: A Policy Perspective.” Energy Policy 123: 360–66. doi:

Haelg, Leonore, Tobias S. Schmidt, and Sebastian Sewerin. 2018. “Technology-related actors and policy design choices: the case of the Swiss renewable energy feed-in tariff.” International Workshop on Public Policy, Pittsburgh, June 26–28.

Hall, Peter. 1993. “Policy Paradigms, Social Learning, and the State: The Case of Economic Policymaking in Britain.” Comparative Politics 25 (3) (April): 275–96.

Hoppmann, Joern, Joern Huenteler, and Bastien Girod. 2014. “Compulsive Policy-Making: The Evolution of the German Feed-In Tariff System for Solar Photovoltaic Power.” Research Policy 43 (8): 1422–41.

Howlett, Michael. 2018. “Matching Policy Tools and Their Targets: Beyond Nudges and Utility Maximisation in Policy Design.” Policy & Politics 46 (1): 101–24.

Howlett, Michael, and Benjamin Cashore. 2009. “The Dependent Variable Problem in the Study of Policy Change: Understanding Policy Change as a Methodological Problem.” Journal of Comparative Policy Analysis 11 (1): 33–46.

Howlett, Michael, Ishani Mukherjee, and Jeremy Rayner. 2017. “The Elements of Effective Program Design: A Two-Level Analysis.” In Handbook on Policy Formation, edited by Michael Howlett and Ishani Mukherjee, 129–46. Cheltenham, UK: Edward Elgar.

Jiao, W., and F. Boons. 2017. “Policy Durability of Circular Economy in China: A Process Analysis of Policy Translation.” Resources, Conservation and Recycling 117: 12–24. doi:10.1016/j.resconrec.2015.10.010.

Jiao, W., F. Boons, G. Teisman, and C. Li. 2018. “Durable Policy Facilitation of Sustainable Industrial Parks in China: A Perspective of Co-evolution of Policy Processes.” Journal of Cleaner Production 192: 179–90. doi:10.1016/j.jclepro.2018.04.226.

Jordan, A., and E. Matt. 2014. “Designing Policies That Intentionally Stick: Policy Feedback in a Changing Climate.” Policy Sciences 47 (3): 227–47. doi:10.1007/s11077-014-9201-x.

Jordan, Andrew, Dave Huitema, Harro van Asselt, and Johanna Forster, eds. 2018. Governing Climate Change: Polycentricity in Action? Cambridge: Cambridge University Press.

Kingdon, John W. 1984. Agendas, Alternatives, and Public Policies. Boston: Little, Brown.

Leiren, Merethe Dotterud, and Inken Reimer. 2018. “Historical Institutionalist Perspective on the Shift from Feed-In Tariffs towards Auctioning in German Renewable Energy Policy.” Energy Research amd Social Science 43: 33–40. doi:

Levin, Kelly, Benjamin Cashore, Steven Bernstein, and Graeme Auld. 2007. “Playing It Forward: Path Dependency, Progressive Incrementalism, and the ‘Super Wicked’ Problem of Global Climate Change.” International Studies Association Convention, Chicago, February 28–March 3.

Levin, Kelly, Benjamin Cashore, Steven Bernstein, and Graeme Auld. 2012. “Overcoming the Tragedy of Super Wicked Problems: Constraining Our Future Selves to Ameliorate Global Climate Change.” Policy Sciences 45 (2): 123–52.

Lockwood, M. 2013. “The Political Sustainability of Climate Policy: The Case of the UK Climate Change Act.” Global Environmental Change 23 (5): 1339–48. doi:10.1016/j.gloenvcha.2013.07.001.Meckling, J. 2011. Carbon coalitions: Business, climate politics, and the rise of emissions trading. MIT Press.

Peters, B. Guy, Giliberto Capano, Michael Howlett, Ishani Mukherjee, Meng-Hsuan Chou, and Pauline Ravinet. 2018. Designing for Policy Effectiveness: Defining and Understanding a Concept: Cambridge: Cambridge University Press.

Rosenbloom, Daniel, James Meadocroft, and Benjamin Cashore. 2019. “Stability and Climate Policy? Harnessing Insights on Path Dependence, Policy Feedback, and Transition Pathways.” Energy Research and Social Science 50: 168–78.

Schmidt, Tobias s., Nicolas Schmid, and Sebastian Sewerin. 2019. “Policy Goals, Partisanship and Paradigmatic Change in Energy Policy – Analyzing Parliamentary Discourse in Germany over 30 Years.” Climate Policy (in press)

Schmidt, Tobias S., Sebastian Sewerin, and Blair Bateson. 2017. “Does policy design predict a policy mix’s future outlook? A new approach to analysing path-dependency.” The Third International Conference on Public Policy (ICPP3), Singapore.

Schneider, A., and M. Sidney. 2009. “What Is Next for Policy Design and Social Construction Theory?” Policy Studies Journal 37 (1): 103–19. doi:10.1111/j.1541-0072.2008.00298.x.

Sewerin, Sebastian, Daniel Béland, and Benjamin Cashore, eds. Forthcoming. “Designing Policy for the Long Term: Integrating Agency and Policy Feedback for Uncovering Durable Pathways of Change.” Accepted for publication by Policy Sciences.

Skogstad, G. 2017. “Policy Feedback and Self-Reinforcing and Self-Undermining Processes in EU Biofuels Policy.” Journal of European Public Policy 24 (1): 21–41. doi:10.1080/13501763.2015.1132752.

Thomann, Eva. 2018. “‘Donate Your Organs, Donate Life!’ Explicitness in Policy Instruments.” Policy Sciences 51 (4): 433–56. doi:10.1007/s11077-018-9324-6.

Vogel, David. 1997. “Trading Up and Governing Across: Transnational Governance and Environmental Protection.” Journal of European Public Policy 4 (4): 556–71.

Walker, Warren E., S. Adnan Rahman, and Jonathan Cave. 2001. “Adaptive Policies, Policy Analysis, and Policy-Making.” European Journal of Operational Research 128 (2): 282–89. doi:

Weaver, K. 2010. “Paths and Forks or Chutes and Ladders? Negative Feedbacks and Policy Regime Change.” Journal of Public Policy 30 (2): 137–62. doi:10.1017/S0143814X10000061.

Weaver, Kent. 2009. “If You Build It, Will They Come? Overcoming Unforeseen Obstacles to Program Effectiveness.”  The Tansley Lecture - University of Saskatchewan.

Wilder, Matt, and Michael Howlett. 2014. “The Politics of Policy Anomalies: Bricolage and the Hermeneutics of Paradigms.” Critical Policy Studies 8 (2): 183–202. doi:10.1080/19460171.2014.901175.

1 Levin et al. (2012, 26) make clear that their attention to “progressive incremental’” pathways incorporates, rather than bypasses, the first feature of super wicked problems: that time is running out. We see interventions as needing to both foster one or more path-dependent processes and contain a clear rationale for sufficient and timely behavioral change consistent with the scale of the problem.

2 While originally criticized for being descriptive rather than explanatory, Kingdon’s multiple streams analysis is germane for practitioners as it identifies the importance of a policy entrepreneur in not only nurturing the three streams but also in connecting them during appropriate windows of opportunity.

All the interpretations and findings set forth in this expert perspective are those of the authors alone.