The recent publication of the Sixth Assessment Report on Climate Change (AR6) by the Intergovernmental Panel on Climate Change (IPCC) (IPCC, 2021) reveals an increasingly difficult challenge: the nature and rate of change in the world's climate due to human activity is occurring at a pace faster than the best science and research is organized to report. Successive IPCC reports, for example, document changes that often are out of date by the time it takes the Panel to organize an assessment. There is growing concern in the research community that what we know today about the climate system is destined to understate our actual challenges and risks (Beck & Mahony, 2018; Hayhoe & Kopp, 2016; Xu et al., 2018). This raises serious problems with policy designs that rely on known current costs and benefits, which will almost certainly be wrong tomorrow. And, even more concerning, policy designs built on the knowns of social and economic valuation imply that climate change is an incremental process. From the available evidence, this assumption is clearly no longer valid. Policy designs need to immediately move away from approaches meant to balance incremental social and economic costs and benefits and face the reality of “compounding extremes” (a concept introduced in the latest IPCC report). Updated estimates in the 2021 report re-confirm facts described in previous assessments, but the damage and risks of escalating change are far worse than previously identified. For instance, AR6 once more underscores the “established fact” that climate change is due to human activity, but it then documents a “substantial increase” in global warming relative to previous reports and indicates that the required reduction in human-induced greenhouse gas releases is much deeper than we thought just a few years ago. AR6 now assesses our challenge as compounding, with “increased intensities, durations, and/or spatial extents unprecedented in the observational record” now accelerating (IPCC, 2021, p. SPM-35). These compounding processes appear to create “climate surprises” with potentially dramatic consequences and almost no time to mitigate, much less adapt to, them.1 The implication for policy is profound. Policy analysis and policy-making have all too often been far behind in conceiving, much less actually addressing the latest realities of climate change. In effect, policy analysis and policy-making have become a reactive process of following climate change rather than leading social transformation. The Paris Climate Agreement is an example of how this process plays out: the Parties to the Agreement widely understood that the ambitions and concessions made by nation-states at the international negotiation table are insufficient to address assessment targets based on physical changes recorded by science that is 5 years or older (e.g., Roelfsema et al., 2020). Promises of “ramping-up” ambition under the Paris Agreement architecture signal the incremental character of climate policy-making—a process found to be lacking even when considering multiple collective improvement scenarios (Geiges et al., 2020). This troubling condition of policy failure is directly traceable to how policy research thinks about the problem. Commonly practiced “business-as-usual” (BAU) benchmarking and search processes to find least-cost options can only deliver out-of-date and mostly wrong characterizations of costs and benefits when the underlying process of climate change has advanced several iterations beyond BAU.2 Policy research suffers from a deeply misguided understanding of its task as it scours the rear-view mirror of past change in hopes of finding the future. The flaw of incremental climate policy-making and the research that informs it is now obvious: using the cost of mitigation and adaptation as the ranking principle to guide decisions is a formula for doing nothing or very little. Positioning the cost of change as the primary guidepost for public action delivers policy hesitancy—a condition that is bound to be increasingly wrong in a rapidly transforming climate system of “compounding extremes” (IPCC, 2021). A climate system of compounding extremes precludes an assumption that overshoot can be repaired by belated but “serious” climate policy interventions. As a case in point, rampant wildfires in California are associated with compounding conditions of drought, extreme precipitation, and more. Their currently known cost is astonishing—California's annual wildfire-related costs are upwards of $50 billion in 2018 (Figure 1; see also AghaKouchak et al., 2020)—but it can hardly provide a usable benchmark when the damage and the mechanics creating it were heretofore inconceivable or dramatically understated. FIGURE 1Open in figure viewerPowerPoint Example of compounding climate extremes in California Recent findings in AR6 make clear that a rapid and far-reaching social transformation unprecedented in terms of scale is required (see also IPCC, 2018). Levels of technological and social change sufficient to overcome decades of policy hesitancy are urgently needed. A lesson from policy analysis failure is that BAU benchmarking and “lowest cost” design are useless today. The climate system, and therefore, societies and economies, are increasingly susceptible to dramatic decline and catastrophe than our benchmarks and lowest-cost metrics presume. Put bluntly, we can no longer be guided by a policy regime that depends on prior experience with previous circumstances that no longer apply.