Using data from the Summer 2020 NSEE wave, a national probability phone survey, and employing multinomial logistic regression, we explored the preferences of the American public for four policy options to combat climate change. We use this to understand a riddle found in the literature concerning under what social conditions and specific GE approaches moral hazard or reverse moral hazard beliefs might arise. Later we explore moral hazard as a slippery concept in need of further elaboration as a contingency model. We end by discussing the need for more social analysis, especially cross-national probability public opinion surveys with greater regional and global range.
We find that the American public has clear preferences for greenhouse gas reductions and reduced consumption over geoengineering or adaptation, which support previous research. We build on this understanding with our data, suggesting that among the American public, certain demographics, beliefs, and support for specific GE approaches generate conditions for moral hazard while other technologies may not. Among our findings, higher support for support for ambient air capture (AAC) and stratospheric aerosol injections (SAI) predicted greater prioritization of geoengineering over our multinomial logistic regression base of institutionally-focused reduction of greenhouse gases. However, for AAC we only found a significant difference between those who were somewhat opposed compared to those strongly opposed to AAC. A GE approach that might generate conditions for reverse moral hazard beliefs is ocean fertilization (OF), where we found evidence that those who were somewhat opposed compared to those strongly opposed to OF were more likely to prioritize GHG reductions over geoengineering. This suggests policymakers and scientists need to consider options carefully among the GE technologies. Public support for GE technologies is essential in addressing current climate change concerns.
6.1 Public Support for Major Policy Directions
When given the choice among four major policy directions, our US respondents selected as their top choice lifestyle changes, i.e. individual-level behavioral actions related to carbon reduction. This was followed by institutional-level (policy and market-based) mitigation efforts, and then by adapting to climate change. GE was the least preferred. Our data also clearly demonstrate that the American public remains unfamiliar with GE technologies generally, and most are opposed to their use, especially SRM approaches (Figure 1). If policymakers consider using any of these technologies in addressing climate change, especially SRM, they should be prepared for considerable skepticism if not a backlash. Our findings re-emphasize the American public’s lack of understanding of and enthusiasm for this policy action, which replicates recent similar research (Burns et al. 2016; Carlisle et al. 2020; Jobin and Siegrist 2020).
The American public’s selection of lifestyle changes as their top climate action choice was a surprise. We are uncertain how to interpret this result. It could be framed as a type of mitigation, but it is based on individual behavioral choices in reducing their carbon load. Previous work by Bellamy et al. (2016) found that UK citizens and climate specializes generally preferred voluntary low carbon living above all other climate options when judged against environmental, feasibility, economic, and other criteria (2016), so the American public may also support individual-level climate actions for the same reasons. This approach, however, is seen by experts as at best a partial solution to addressing climate change and other sustainability concerns unless the vast majority of individuals and households engage in the behavior which requires institutional support (e.g. Lorenzen 2018). Thus, although lifestyle changes were the most highly prioritized climate action, we focused our investigation on moral hazard beliefs on comparisons against those who prioritized reducing GHG emissions, as distraction from this action is of most concern.
6.2 Preferences for CDR over SRM; SAI is unpopular
Our research compared four specific geoengineering approaches in the US. Overall, our findings align with previous work that looked at beliefs in the US, UK, New Zealand, Australia, Germany, and other countries (Braun et al. 2018; Carlisle et al. 2020; Jobin and Siegrist 2020). Our results support those who found that SRM generally and SAI technology specifically remain very unpopular with the public including in the US, whereas CDR approaches experience larger levels of support. In our study, however, over a quarter to nearly a third of the American public were unable to state a preference. The continuation of low support and high uncertainty remains an important finding, especially as Congress and national scientific institutions give more attention and funding, to SAI specifically (Clifford 2022; Fialka 2020; National Academies of Sciences, Engineering, and Medicine 2021; National Research Council 2015). These findings confirm that if governance efforts hope to approach emerging climate engineering technologies legitimately and responsibly, they must “open up” deliberation and engagement to these publics with risk concerns, or risk failure in acceptance and efficacy (Bellamy 2016; Stilgoe 2015; Stirling 2008).
6.3 Building a Contingency Model for Moral Hazard Beliefs
We argue that moral hazard beliefs likely exist within certain social, political, and technical contingencies. Our findings begin to illustrate the contexts in which climate decision makers should watch out for moral hazard beliefs amongst the US public. The main ones we find are those related to political and environmental beliefs and priorities, and to individual GE approaches. Race also may play a significant role in the contingency model of moral hazard beliefs.
Our findings fall in line with previous literature that point to political beliefs and climate change beliefs as drivers for reverse moral hazard beliefs. We find that conservatives are more likely to prioritize geoengineering over institutional-level mitigation of GHG emissions. Our findings strengthen previous research that point towards conservatives as being more susceptible to moral hazard beliefs. Campbell-Arvai et al. (2017) found that political ideology moderates the effect of climate risk perception on mitigation policy support when learning about CDR. Raimi et al. (2019) find that conservatives and moderates who read multiple framings of geoengineering descriptions have weaker belief in anthropogenic climate change than those who do not read about climate change, and that conservatives and moderates had lower risk perceptions of climate change when viewing SRM as a major solution. Since we factored out climate change risk (through seriousness) in our model, our results strengthen the claim that moral hazard beliefs may come out more for conservatives than for liberals.
These findings add another dimension to what we know about how political beliefs affect perceptions on geoengineering. As previously noted, Carlisle et al. (2020) found in their US sample that self-identified Republican Party members exhibited more negative preferences towards individual GE technologies. So, on one hand, conservatives are less likely to support individual GE approaches than liberals. On the other hand, conservatives are more likely to prefer GE over GHG mitigation strategies to address climate change. Lower climate risk perceptions could explain some but not all of this pattern, as we factored out belief in the seriousness of global warming.
The forced-choice aspect of our moral hazard question might illuminate this puzzle and add a dimension to the contingency of moral hazard. When given the option, conservatives may indicate low support for any approach or technology that is associated with addressing climate change. But when presented with a forced choice, they will choose the option that is more in line with conservative ideals: technological solutions that do not impact existing economic industries like fossil fuels. This falls in line with existing concerns that GE, especially CDR, might help corporations dependent on fossil fuels evade responsibility for climate change and broader environmental pollution (Morrow et al. 2020).
This dynamic of moral hazard beliefs will be very important for how policymakers frame comprehensive climate change policies. If climate change becomes a top policy priority for the American public, including conservatives, we will similarly face a “forced-choice” situation in deciding how much resources to allocate to each type of climate action. Conservatives may feel that geoengineering, often framed both as cheap and relatively easy to implement, can take priority. If climate change remains a low priority for the American public, though, then moral hazard amongst conservatives may be less of a concern since they will not support any policies regarding climate change anyway.
We also found evidence of reverse moral hazard beliefs for those with higher beliefs in the seriousness of climate change and in the risk of environmental harm from geoengineering; these respondents were more likely to prioritize GHG mitigation over geoengineering as the most important climate action. Our results add nuance to earlier research that found that the climate change concern was positively related to greater support for GE technologies (Braun et al. 2018; Carlisle et al. 2020; Mercer et al. 2011; Merk et al. 2015; Pidgeon et al. 2012). This extends the pattern just described for political beliefs: factors that predict greater support for GE technologies also predict reverse moral hazard beliefs. These findings strengthen the theory that moral hazard concerns may be most relevant for those who have lower climate risk perceptions when they are forced to decide on climate action. Thus, policymakers may need to judge climate and environmental harm risk perceptions amongst themselves and the public before considering geoengineering within a portfolio of climate policies. Hopefully, as climate action becomes a top priority for policymakers and the public, then it will be imperative to keep an eye out for efforts to prioritize geoengineering over mitigation.
One of our most important findings is that confidence in scientists to solve climate change support using GE technologies predicts moral hazard beliefs among US respondents. The findings suggest moral hazard beliefs are likely to be found among techno-optimists (see Keary 2016). They believe technology provide solutions to most of society’s wicked problems. Previous research on GE and public confidence in science similarly showed a positive relationship with CCS, a CDR technology, afforestation, and SRM generally (Braun et al. 2018; Merk et al. 2015; Terwel et al. 2009). These previous studies were in European countries; our findings validate this pattern in the US context. Our findings support the idea that those who believe in scientists and science, techno-optimists, are more likely to support specific GE technologies over mitigation efforts, which might lead to moral hazard. Given the recent increase in federal funding for geoengineering research, it will be important to also track public perceptions of climate scientists and engineers and understand how their trust and confidence in them is related to possible moral hazard beliefs.
Our findings on the effects of race are also striking and have not previously been found to be relevant to GE beliefs. Our sample size required us to only differentiate between white and non-white respondents, but our finding that non-whites are much more likely to prioritize geoengineering over mitigation indicates the need for further research. Considering that non-whites are more likely to negatively experience the effects of climate change (Harlan et al. 2015), this finding and further qualitative research may illuminate how climate inequalities could influence moral hazard beliefs. This strengthens the contingency model, especially on an international scale as communities and racial groups differently affected by climate change engage in climate policy prioritization.
6.4 Contingency Upon GE Approaches
The relationship we found between moral hazard beliefs and support for SAI is our most obvious evidence for an approach-specific condition for moral hazard. Our findings for the CDR approaches, OF and AAC are more difficult to understand. Only those who somewhat opposed AAC, compared to those who strongly opposed AAC, were more likely to prioritize geoengineering over mitigation. And those who were somewhat opposed to OF, compared to those strongly opposed to OF, were more likely to prioritize mitigation over geoengineering. We investigated whether any of the other beliefs that predicted moral hazard beliefs were explaining these findings or suppressing relationships to other levels of support by removing them from the models, but they did not change the results. We also looked for possible interaction effects with other variables but found none. These relationships are not straightforward, and our models cannot explain why they exist. However, they provide a clue as to the contingent nature of moral hazard beliefs. Not all geoengineering approaches, both CDR and SRM, will necessarily invoke moral hazard responses from the public, and they may strengthen mitigation prioritization for some. Further research is needed to determine how the public perceives the differences between GE approaches and how that influences their prioritization of climate actions.
Campell-Arvai et al. (2017) found that learning about CDR in general, and BECCS and direct air capture specifically, generally dampens perceived risks of climate change and therefore support for mitigation policies in the US. Our findings may differ simply based on how we gave information about and framed each of the technologies. The NSEE also asked every respondent about their support for all four GE approaches, so there may have been a framing effect from learning about both CDR and both SRM approaches, influencing their moral hazard beliefs. Regardless, our results point towards the importance of the type of GE in setting expectations around moral hazards. Investigating more closely the relationship between risk and trust factors, demographic factors, and moral hazard beliefs in the context of specific types of GE technologies will continue to be key.
6.5 Reducing Moral Hazard Beliefs
The moral hazard of geoengineering cannot be ignored as a climate justice issue; even if the US did have robust mitigation and decarbonization policies in place now, we have found evidence that continued research and possible development of specific GE approaches may lead to re-prioritization of climate policies without careful framing and consideration. Given this possibility, should GE technologies be used to augment current climate goals? One framing measure that can address this issue is continued emphasis of geoengineering as “peak shaving,” “offsetting overshoot,” or a “stop-gap measure” (Buck et al. 2020; MacMartin et al. 2018; Niemeier and Tilmes 2017; Tilmes, Sanderson, and O’Neill 2016). For example, in a recently proposed roadmap of SRM research, the authors advocate for key research into SRM approaches that would act as “rapid, temporary, additive measure[s] to reduce catastrophic impacts from anthropogenic climate change, not as a substitute for aggressive GHG mitigation” (Wanser et al. 2022). Instead of a last resort or “Plan B,” GE technologies become additional means to keep global temperatures within the 1.5 degrees C limits set by the 2015 Paris Agreement (UNFCCC 2015).
Pursuing this line of reasoning suggests GE as a supplement to on-going climate mitigation measures, as opposed to a substitute. The moral hazard questions become more specific: will the public and policymakers see GE as a means for fulfilling the Paris Agreement? If so, another critical question arises about what technologies are best to address the climate problem from a techno-policy-econ-public acceptance perspective. Under this context, GE becomes not a replacement for failed mitigation efforts but as a supplement to on-going efforts. Will the public and policymakers understand the distinction? Our research indicates that this distinction will be important to reduce prioritization of geoengineering, especially if or when climate change becomes a top policy priority for the entire population and even those unconcerned about climate risks are forced to choose.
Our different findings both within and between CDR and SRM approaches regarding associations with moral hazard beliefs also highlight that policymakers must carefully consider how these approaches will fit in to a broader portfolio of climate change policies in a way that prioritizes mitigation and decarbonization. Our initial survey research cannot explain why certain approaches are more associated with moral hazard beliefs. It could be because those approaches seem more realistic or possible, or it could be because those approaches trigger lower risk perceptions. While further explanation is needed, it is clear GE researchers and climate policymakers will need to keep an eye not just on the differences in feasibility and efficacy of GE approaches, but also on the differences in their capacity to induce moral hazard beliefs. This is vital to the overall risk calculation of climate policies.
6.6 Future Research
More systematic studies of the public are required given the wide-ranging regional to global consequences of GE technologies and how those risks and beliefs vary. This is especially true if GE is to become an essential contributor to a broader policy portfolio of current climate strategies. One of our most glaring confirmations has been the lack of globally comprehensive social research on this critical topic. Others have noted this gap (e.g. Aldy et al. 2021; Burns et al. 2016). Most current social research on GE has focused on the US, UK, Germany, Australia, Canada, New Zealand, and a handful of other OECD countries. A number of them have been insightful experimental studies useful in understanding social processes, e.g. decision-making processes. However, these are not nationally representative and cannot make generalizable statements at the national level. There have been some attempts to explore GE preferences in several important Asian countries, OECD and non-OECD countries (Cummings and Rosenthal 2018). To our knowledge, researchers have not studied Sub-Saharan African countries, those from the Middle East-North Africa, and few if any from Latin America and Asia, outside of Singapore. A soon to be released public opinion study of the European Union is welcome news but other parts of the world require attention. As our contingency model indicates, social and political contexts will be highly important in understanding moral hazard and more general GE beliefs, thus emphasizing the necessity of geographic range in research.
While one may argue that given the required GE technology for addressing climate change, the effort rests largely with the wealthy OECD countries and others with large economies like India and China that have the technical capabilities and resources. While this may be true, the consequences from employing GE technologies are likely to affect non-wealthy countries the most and have fewer resources to deal with the unexpected consequences (Rahman et al. 2018). The residents and policymakers of these countries must be included and have a voice in these critical discussions. An ambitious social research program awaits and is urgently needed if the international community is to pursue successfully a GE supplemental, peak-shaving approach to addressing climate change beyond climate emergencies.