This is the accepted version of the article: Exadaktylos, Filippos; van den Bergh, Jeroen C.J.M. «Energy-related behaviour and rebound when rationality, self-interest and willpower are limited». Nature Energy, Vol. 6 (Octubre 2021), p. 1104-1113. DOI 10.1038/s41560-021-00889-4 This version is available at https://ddd.uab.cat/record/255225 under the terms of the license Energy-related behaviour and rebound when rationality, self-interest and willpower are limited Filippos Exadaktylosa * and Jeroen van den Bergha b c a Institute of Environmental Science and Technology, Universitat Autònoma de Barcelona, Spain b ICREA, Barcelona, Spain c School of Business and Economics & Institute for Environmental Studies, VU University Amsterdam, the Netherlands * Corresponding author Abstract The extent to which adopting energy-efficient technologies results in energy savings depends on how such technologies are used and how monetary savings from energy efficiency are spent. Energy rebound occurs when potential energy savings are diminished due to post-adoption behaviour. Here we review empirical studies on how six behavioural regularities affect three energy-relevant decisions and ultimately rebound: adoption of energy-saving products or practices, their intensity of use, and spending of associated monetary savings. The findings suggest that behaviours which reflect limited rationality and willpower may increase rebound, while effects of behaviours driven by bounded self-interest are less clear. We then describe how interventions associated with each of the behavioural regularities can influence rebound and thus serve to achieve higher energy savings. Future research ought to study energy-relevant decisions in a more integrated manner, with a particular focus on re-spending as this presents the greatest challenge for research and policy. (Main text) Energy rebound denotes that potential energy savings of adopting an energy-efficient technology or practice, possibly triggered by some policy, are offset by subsequent behavioural and systemic responses that increase energy use, resulting in diminished net energy savings1,2. There are three main types of rebound: direct rebound or intensity-of-use effect – a technology becomes more energy-efficient and thus less costly in its use, causing consumers or producers to use it more intensively; indirect rebound or re- spending effect – spending less due to using a more energy-efficient technology releases money that is subsequently spent on other products or services that use energy over their life cycle; and economy-wide rebound – more energy efficiency leads to many other changes in the economy, such as investments in expansion of production, impacts on capital and labour markets, and indirectly increases in consumption, all with consequences for energy use. Figure 1 illustrates these rebound types, and their joint impact on energy savings, for the case of switching to a more fuel-efficient car. While there is little doubt that rebound is an important issue that warrants serious attention in policy design, the exact magnitude of rebound effects is surrounded by uncertainty. This is partly the result of the difficulty of assessing rebound empirically, because of a lack of appropriate data and a clear counterfactual, and as rebound lacks unambiguous system and temporal boundaries3–5. Estimates for direct rebound range from 0 to 87%, differing between application areas, such as space heating (2-60%), cooling (0-50%) and transportation (5-87%)5. Evidence from 33 macrolevel studies shows that economy-wide rebound effects tend to exceed 50%, and may include backfire (>100%)6. Among economists, a conventional way to think about rebound is that it results from optimal adjustment to observable and perfectly known changes in marginal costs due to improvements in the energy-efficiency of a product or technology. In accordance with this, rebound can be estimated as the combination of price and income effects. However, rebound is the result of more than just changes in relative prices7–12. It also depends on behavioural reactions to policies, markets, technologies and social peers. A realistic representation of individual decision-making is therefore necessary to allow inferences to be drawn about when rebound is likely to occur, the magnitude of the effect, and how it can be curbed. Many studies of decision-making implicitly or explicitly assume rational and selfish agents, who consistently optimize a utility or profit function and are unaffected by behaviours or opinions of others. However, modern behavioural sciences demonstrate that people are bounded in their rationality, willpower and self-interest13. Bounded rationality means they do not always optimize decisions because they cannot perfectly calculate future costs and price effects; bounded willpower, also known as limited self-control, prevents agents from acting upon their preferences; and bounded self-interest indicates that factors such as concerns about others, self-image or social norms affect peoples’ behaviour. While bounded self-interest suggests that individuals have additional motivations beyond strictly selfish goals, bounded rationality and willpower have implications for the way individuals try to achieve these goals. Limits to rationality and 1 willpower often result in individuals not achieving their goals – irrespective of whether these are selfish or non-selfish. For example, potential financial benefits could motivate individuals to save energy by lowering the thermostat level of their heating system, but limited awareness about such benefits, or habits of wearing light clothing indoors, may prevent them to from doing so. In other words, bounded rationality and willpower moderate how motives affect behaviour. In Table 1, we present six behavioural regularities relevant for the case of rebound. These are well- documented behavioural patterns that emanate from the underlying behavioural assumptions of bounded rationality, willpower and selfishness, as explained in Box 1. In this Review we present evidence of how each of them influences the three energy-relevant decisions, with the aim of assessing whether they have a propensity to increase or decrease the rebound effect (see Figure 2). In particular, potential energy savings and thus rebound can be thought of as the result of individual energy-relevant decisions following energy-efficient technology adoption, namely intensity of use and re- spending. Intensity of use indicates how intensively adopters use the energy-efficient product as compared to pre-adoption use levels. This in turn determines direct rebound. Re-spending denotes that adopters spend monetary savings associated with initial energy savings on additional goods or services that in turn cause energy use. This gives rise to indirect rebound. Here we review studies from any field that relates rebound to behaviour, including the broader behavioural and social sciences. While a few studies have already employed insights from behavioural sciences to explain rebound7–12, here we provide a comprehensive review of the available evidence on energy-related decisions irrespective of whether the corresponding study explicitly described implications for rebound. We derive many of the insights regarding intensity of use and re-spending from studies on curtailment and low-carbon consumption, respectively. We draw on the general literature addressing energy-efficient behaviour to draw inferences about when direct and indirect rebound are expected to occur. We will discuss how this can inform improved policy design for, and future research on, effectively controlling or curbing rebound. Behavioural regularities, energy decisions and rebound Inattention and misconceptions People lack the cognitive capacity and time to properly analyse all the available information before reaching a decision. In some cases, they disregard relevant information and in others they misinterpret it, leading to biased beliefs or misconceptions. As a result, the effects of prices and income on the one hand, and moral and social motivations on the other, are moderated by inattention and misconceptions. Evidence for vehicle purchases suggests that agents pay limited attention to financial data14, or imperfectly assess these15. 2 Similarly, consumers’ environmental concerns affect car purchase choices but this is moderated by inaccurate perceptions of emissions16. Inattention has been identified as a main factor hindering adoption of energy-efficient technologies. Consumers and firms tend to pay more attention to purchasing than to operating costs, causing them to underinvest in energy-efficient options17. Regarding intensity of use, evidence shows that this increases due to inattention. In general, consumers tend to underestimate energy costs, and as a result they increase energy use. For example, evidence from the UK refrigerator market shows that consumers underestimated future energy costs by 35%, which led to a 9,2% increase of average use compared to what perfectly rational consumers would consume18. This effect is more pronounced for energy-efficient products. One study found adopters of heat pumps to be unaware they were using them more intensively than the conventional heating, resulting in higher indoor temperature19. According to another study consumers incorrectly infer that high energy efficiency translates to low overall energy use, which may consequently lead them to use energy-efficient products more intensively20. Finally, in a quasi-experimental setting, consumers used significantly less electricity during a month following the receipt of a bill that crossed a salient threshold21. Indirect evidence suggests that inattention increases re-spending. According to the studies mentioned in the previous paragraph, consumers underestimate energy use after adoption and therefore overestimate energy and associated monetary savings, in turn translating into increased spending. In addition, people systematically misjudge information about energy use: they underestimate energy consumption and savings associated with high-energy activities and overestimate them for low-energy activities22,23. Taken together the evidence suggest that re-spending increases and is directed towards more energy-intensive or larger appliances. However, direct evidence is lacking. In summary, compared to what the rational-agent theory would predict, inattention leads to lower adoption rates and, in case of adoption, to higher intensity of use, which may translate to higher direct rebound. Effects of re-spending are uncertain. Mental accounting In the face of cognitive limitations and overwhelming information, individuals adopt heuristics to simplify their decisions. Mental accounting describes common ways people use to organize their budget which include coding, categorizing and evaluating economic outcomes24. It has various elements. Instead of considering their entire budget as fungible, individuals keep various smaller mental budgets which they consider in isolation. Examples of more specific accounts are clothing and entertainment expenses25. The marginal propensity to consume – i.e. how easily money is spent – varies across mental accounts. A common feature of all accounts is that they have a reference point, where the account is considered balanced in the minds of consumers. Exceeding this point is classified as a gain and being below it as a loss. The distinction is important as losses create stronger reactions than gains, known as loss aversion26. To illustrate, 3 when in 2001 the US government attempted to boost the economy by transferring $38 billion to taxpayers, instead of spending the received money, most people saved it. This arguably happened because the government called the transfer a “rebate”, a term that implied that money was initially owned by individuals and subsequently returned to them, restoring an imbalance27. In terms of mental accounts, people placed the transferred “losses” to a mental “savings” account, with a low propensity for consumption. Further research suggested that had the transfer been framed as a bonus, it might have stimulated spending39. Mental accounting has indeed been suggested to significantly shape energy-related behaviours28. Loss aversion hinders adoption of energy-efficient products. It leads people to focus disproportionally on immediate and certain investment costs versus potential long-term savings on energy expenditures. Survey data from across Europe places loss-aversion among the strongest predictors of not adopting energy-efficient technologies, such as hybrid vehicles, energy-efficient light bulbs, energy- efficient household appliances and house insulation29,30. In addition, more than half of households that decide to not switch to an alternative energy provider state that they do not do so out of fear that something might go wrong31. The effect of mental accounting on intensity of use depends on the specific mental accounts that consumers hold. Evidence suggests that people have ‘topical’ mental accounts, meaning that they classify money according to the topic of the decision. For instance, money saved in transportation will most likely to remain in this account32, leading to increased intensity of use. Direct evidence comes from the ‘UK winter payment’, a direct cash transfer to households, whose name led households to spent it disproportionately on heating33. At the same time people tend to keep separate mental accounts for one-shot investments and monthly energy bills. This means that after adoption, there will be more budget available on the second account, further increasing intensity of use. For example, when asked to calculate monthly car costs, only 29% of respondents included depreciation of the initial investment34. This is further confirmed by empirical evidence for travel35 and heating expenses11,19. In addition, the effect is likely to be exaggerated by the sunk cost effect, which is considered an expression of mental accounting36: having invested money on adopting an energy-efficient product, people feel compelled to use it more. Regarding re-spending, if energy savings are classified by consumers as gains, the propensity of spending the savings will be high25. There is evidence that in such cases consumers tend to buy goods that they do not normally consume37, which tend to be more energy-intensive38. However, this may depend on the particular area of energy conservation: while re-spending from vehicle and food measures has been directed to goods with higher emissions, savings from heating and lighting improvements have been directed to goods and services with lower emissions39. Overall, loss aversion seems to reduce adoption, while mental budgeting may lead to higher direct rebound by increasing intensity of use, and higher indirect rebound by directing re-spending to higher energy-intensive products. The latter effect may depend on the type of consumption. 4 Defaults and Habits Decisions by default and habitual behaviour are quasi-automatic processes that function at a low level of consciousness. A default denotes the option that individuals select if they do not make an active choice. Default bias refers to the observation that the probability of the default being chosen is disproportionally high40. This is due to inertia (avoid costs of searching for alternatives), loss aversion (avoid potentially inferior alternatives) or it being interpreted as the suggested option41. Habits are repetitions of past behaviours and determine individuals’ default decisions in the presence of certain stimuli42. They are considered a principal barrier for behavioural change43. Defaults and habits characterize many decisions that pertain to energy use – think of mobility, electricity use and eating. Defaults and habits hinder adoption of energy-efficient products since they both contribute to the status quo, weakening the connection between intentions and behaviour. For example, although the majority of people express that they favour a green-energy provider and are willing to pay a premium for it, only 2% end up selecting it, due to the default effect44. It is part of the reason why information-provision interventions signalling the availability of energy-efficient options having limited effect45. On the positive side, once people switch, the new product or practice becomes the novel status quo, causing default bias and habitual behaviour to stabilize it. For example, when the default in housing renovations changed from incandescent to CFL bulbs, the percentage of consumers choosing CFL went up from 56% to 80%46. Regarding use-intensity, adoption of an energy-efficient product is a source of habit disruption. For instance, switching to a hybrid car will often lead to distinct refuelling intervals. In theory, the final effect on use intensity will depend on the consumption type and on other details. The available evidence seems to suggest that the more habitual individuals are, the more intense their post-adoption use is. Individuals exhibiting habitual behaviour were found to overheat their apartments after insulating it11 or after purchasing a heat pump19, and to drive slightly more after buying an electric car.11 Regarding insulation, habitual behaviour is expected to have an influence only in the absence of thermostats. Evidence suggest that only 27% of household vary thermostat levels over time43. Overall, habitual behaviour tends to have a negative effect on intensity of use, but the evidence is not conclusive. Direct evidence is lacking regarding the specific effects of defaults and habits on re-spending behaviour. A model-based analysis showed that habitual behaviour deters consumers from moving to low- carbon alternatives47. One can argue that if people are habitual in their spending of unexpected monetary savings, such as on weekend trips or other short holidays involving long-distance travel, then this could create considerable rebound. However, if people are more habitual in the specific travel rather than the spending, rebound might be contained. To sum up, habits and status quo lead to decreased adoption rates; habitual behaviour may increase intensity of use, while effects on re-spending are situation-dependent and thus uncertain. 5 Present bias Present bias refers to the tendency of people to overvalue immediate costs and benefits compared to future ones that leads them to make time-inconsistent choices. In more technical terms, people discount costs and benefits at a non-constant rate, i.e. depending on when they discount them48. This leads individuals to procrastinate, or make decisions that they may regret in the future27. Present bias therefore prevents individuals from acting upon their preferences, values, social expectations and long-term interests, causing stronger reliance on habitual behaviour. A wide range of phenomena that involve trade-offs between short- run costs and long-run benefits are evidently sensitive to present bias, such as obesity, failure to quit smoking and insufficient saving for retirement49. The same holds for energy-relevant behaviour since adoption of energy-efficient products or practices involves trade-offs between present and future cost and benefits. Several studies find that present bias impedes adoption of fuel-efficient cars, home insulation, energy-efficient appliances, and curtailment behaviour18,50–52, while fewer cannot establish such a relation30,53. For example, car buyers consistently underestimate future fuel costs, leading to a lower likelihood of adopting high fuel efficiency cars54. Present bias is associated with higher intensity of use. Among participants in a goal-setting energy program in the US, those with hyperbolic time preferences had higher post-adoption intensity of use55. Importantly, awareness of present bias mitigated this effect. In addition, procrastination is associated with less interest in reducing indoor temperature56. While no studies have examined the potential role of present bias on re-spending decisions, it likely will contribute to inefficient choices characterized by relatively low up-front costs. Overall, present bias decreases adoption, tends to increase intensity of use and thus direct rebound, while the effect on re-spending decisions and indirect rebound is not clear. Pro-environmental values and moral licensing Pro-environmental values may emanate from concerns about the wellbeing of other people, other species or the environment57. Such preferences motivate individuals to behave so as to decrease their environmental impact58. However, although people intend to act upon such values, they do not always manage to do so because financial and contextual factors, such as available budget and infrastructure, may be constraining. This holds especially true for energy efficiency investments59–61. Additionally, people might choose alternative options that are more alluring. In an attempt to self- justify deviations from what their morals prescribe, and to avoid cognitive dissonance and associated feelings of guilt or loss of self-esteem, people may use pro-environmental behaviours at earlier times or in other domains as a moral excuse. This is known as moral licensing62 and has been employed to explain rebound9. For example, the purchase of an electric vehicle may be used as a moral excuse for subsequent intense use of it. It is as if individuals have a moral mental account which they try to keep balanced8. 6 Phenomena such as compensatory behaviours63, behavioural spillovers64 and moral hazard65 describe similar processes whereby engaging in one behaviour can trigger other behaviours that contribute to energy rebound. Additional reasons for the mismatch between preferences and behaviour include inattention, lack of knowledge regarding the effectiveness of potential solutions to environmental problems and self-control problems associated with individuals’ bounded willpower. Pro-environmental values increase the probability of adopting more energy-efficient products and practices. There is considerable of evidence in this respect regarding insulating houses66, buying a fuel efficient car, using energy-efficient light bulbs67 or participating in carbon offsetting68 and green-electricity programs69. Regarding intensity-of-use effects of moral licensing, the evidence is mixed. A study among Swedish drivers switching to more efficient cars estimates the direct rebound effect was on average 24%. However, among the subsample that switched to a green-labelled car, the direct rebound was found to be zero70. A potential interpretation is that whenever adoption was motivated by pro-environmental reasons, the intensity of use did not increase, even if driving had become cheaper. Similar evidence is found among German71 and Austrian11 e-car drivers. At the same time however, there is evidence for moral licensing among electric car adopters. In a study of rebound in Norway, the degree to which a driver accepts responsibility for, and is willing to mitigate, the negative outcomes of driving a car was found to significantly drop after adopting an electric efficient car72. In another study, individuals scoring higher in a moral licensing scale (that is, justifying compensatory behaviours) were found to have higher intensity of use after adopting an e-car and after insulating their house11. Regarding re-spending, there is a well-established link between environmentally significant consumption and pro-environmental values73. Controlling for income and other socio-demographics, people with strong pro-environmental values have a spending pattern characterised by relatively low carbon emissions74. In fact, consumers with pro-environmental values are willing to pay a premium for green products75. To the extent that re-spending and general consumption patterns are behaviourally similar, the positive effect extends to re-spending and thus to indirect rebound. Indeed, there is evidence that pro- environmental values are the strongest predictor of a reduced re-spending pattern after adoption11. Overall, pro-environmental values tend to increase adoption and divert re-spending to low-carbon options. Their effect on intensity of use and thus direct rebound is less clear, however, due to potential moral licensing effects. Peer influence Individuals’ behaviour is strongly affected by what other people think and do. Peer influence can take various forms: people comply with the norms to avoid social sanctions; they imitate others to fit in a group, or as a social heuristic when there is uncertainty about what is the right behaviour; and they signal socially 7 desirable personal traits through conspicuous consumption, in order to gain social status that comes with preferential treatment in social interactions76. In the domain of energy conservation, the effects of peer influence tend to be larger the more costly or effortful is a given behaviour, notably in the public (vs private) domain, where behaviours are more observable77,78. As such, the strength of peer effects is expected to be higher for energy-efficiency than curtailment behaviours. And within efficiency behaviours, installing solar panels is more relevant for social status than indoor house insulation, since the latter is not easily observable by third parties78. There is considerable evidence that peer influence encourages the adoption of more energy- efficient products and practices79. For example, purchase for the Toyota Prius is enhanced by social pressure and status80, while people are more likely to install solar panels and subscribe to energy-efficiency programs when their neighbours do so81. Regarding intensity of use, the evidence is limited. An indication of a positive effect comes from a series of experiments done with household electricity demand when households’ electricity consumption was compared to that of their neighbours. Results indicate that the effects of social norms on energy use are significant: on average they motivated consumers to reduce their energy use by 2%82. There is some evidence that such effects are long-lasting83. However, the latter interventions targeted total household energy use, not post-adoption intensity of use. In addition, status concerns might even lead to an increase in intensity of use. To illustrate, in an attempt to signal pro-environmental behaviour, owners of electric or hybrid vehicles might use them more frequently in order to increase visibility. Regarding re-spending behaviour, various controlled experiments have shown that individuals who perceive a strong descriptive or injunctive pro-environmental social norm, tend to switch to low-carbon consumption patterns84–86. However, in the context of industrialized societies, where high consumption is the norm, peer effects might trigger more consumption, inducing higher re-spending. Somewhat counterintuitively, the stronger the pro-environmental social norm, the higher the re-spending11. In sum, while peer influence tends to lead to higher adoption rates, its effects on intensity of use and re-spending are less clear and can go either way87. Figure 3 summarizes the effects of the six regularities applied to energy-relevant decisions. Behavioural interventions Policy-makers can directly address the behavioural regularities in order to minimize their distortions on the effectiveness of policy instruments by employing instruments that recognize peoples’ limits on rationality, willpower and self-interest. The use of such behavioural interventions is becoming widespread in the field of energy conservation due to their cost effectiveness and political feasibility. There is now considerable evidence on what works and under which conditions45,88–93. Similar interventions can be used in the case of rebound. However, given that the end goal of the policy is not to curb rebound per se, but rather to reduce 8 total energy consumption, the real challenge is to promote adoption and decrease rebound or at least not increase it. Thus the effective policy will have a dual intention: encourage adoption and provide incentives to discourage rebound. Below we briefly describe behavioural interventions that can address the various behavioural regularities and describe how they affect the three energy-relevant decisions and hence rebound: adoption which will help closing the energy gap, intensity of use to address direct rebound, and re-spending to curb indirect rebound. Table 2 provides a summary. Inattention & misconceptions Regarding adoption, energy labels have been used to succinctly communicate energy efficiency and lifecycle energy consumption of durables in a visually friendly way employing colours and scales. However, their effectiveness has been found to be rather limited88. Consumers’ responsiveness is found to be higher when efficiency is expressed in monetary terms94. The effects of inattention on intensity of use can be addressed by providing feedback regarding consumption levels. In the case of household consumption, feedback is given traditionally in the form of disaggregated energy consumption information at the end of the month (i.e. billing). However, feedback given this way does not permit consumers to calculate intensity of use of a particular source and even less to compare pre- and post-adoption levels. Smart meters can better address the temporal and spatial disassociation by providing frequent, near real-time and detailed resolution feedback, thus allowing consumers to observe and calculate energy improvements95,96. For example, the following personalized message “If you reduced the thermostat temperature in your house one degree you would save 11 kWh; this is equivalent to £1.43” reduced indoor temperature from 22.4 to 21.7 on average97. Modern disaggregation technologies can indeed provide energy feedback at the appliance level. Creative examples include light bulbs that change colour after prolonged use27, or a polar bear standing on a melting ice floe during a shower98. In summary, to promote adoption and minimize rebound, energy labels and lifecycle energy information can be combined with real-time post-adoption feedback. Mental accounting A recent meta-analysis of studies using gain vs loss frames suggests that the latter are more effective in promoting pro-environmental behavior99. In the case of promoting adoption of more efficient options this means that the best way to promote adoption is to focus on the negative aspects – forgone long term monetary incentives or negative environmental consequences – of the less efficient options. To do so, the policy maker needs to assure that individuals combine adoption and intensity of use in the same mental account. In addition, an integrated mental budget can provide a key mechanism for curbing intensity of use. Consumers wanting to keep their mental budgets balanced will assume a frugal post-adoption behaviour in 9 order to progressively balance it. This will essentially distort the core economic mechanism of direct rebound. In terms of policy intervention, the aim is to make consumers think of the initial investment as part of the topical mental budget, such as “heating”, instead of a separate one, such as “one-time investments”. The ideal behavioural intervention will do so without discouraging adoption in the first place. In the case of heating, one might use, for example, a smart meter (illustrated in Figure 4) to assist consumers become aware of the time left for the initial investment until full payback. This nudges households to consider adoption and intensity of use in tandem. In addition, consumers can be given the option to adjust their daily energy consumption level according to a desired payback time, which can serve as a clear reference point. This will address inattention, and activate goal-setting mechanisms. What is more, this intervention of mental-account management will have direct implications for re-spending, since adopters will perceive no monetary savings until the end of the payoff period. Negative savings will then keep re-spending to a minimum. Defaults & habits Policy-makers may employ status quo bias to promote adoption. There is plenty of evidence that setting an energy-saving option as an opt-out, rather than as opt-in, significantly increases adoption of energy-efficient products, practices and energy-saving programs41. On the other hand, breaking less efficient habits is much more difficult to do in practice100, especially given that adoption oftentimes includes large one-off purchases. That said, there are a few successful examples of disruption, such as temporality closing a highway101, giving one-month free bus tickets102 to promote use of public transportation, or exchanging car keys for a one-month free electrical bike103. In addition, there is evidence that interventions to promote sustainable behaviours are more effective shortly after important changes in peoples’ lives, such as moving house104. Leveraging on status quo bias has been less exploited in the case of intensity of use but presents considerable potential. Lower temperature defaults of thermostats105 and washing machines106 have been shown to be effective in reducing energy use. Therefore, if after adoption intensity of use is set automatically to the pre-adoption levels (think of a thermostat), it is expected to have a significant effect. Simultaneous application of pre- and post-adoption defaults can therefore both promote adoption and curb rebound. Present bias To overcome present bias and promote adoption, policy-makers can draw the consumer’s attention to the long-term implications of the energy-efficient choice107. In addition, long term benefits can be broken down and communicated in shorter periods intervals. However, as noted, such interventions have not always led to significantly higher adoption rates53,108 10 Commitment and goal-setting programs provide additional opportunities for addressing present bias, procrastination and lack of commitment. The basic idea is that consumers after adoption commit to an energy-reduction target regarding a particular electrical device or overall electricity use109. These programmes are designed to direct consumers’ attention to the activities at hand, to engage them and, crucially for addressing present bias, to motivate them to prolong their commitment110. In addition, such programmes offer the opportunity to define the way potential benefits can be spent by tailoring them with particular activities, or eco-friendly shops and products, potentially linking to white certificates and complementary currencies111. This resembles commercial loyalty projects that involve earning points, which can be redeemed in eco-friendly products and services (e.g., free public transportation), such as the “NU-Spaarpas” and “Zet Milieu op de Kaart” projects in the Netherlands and Belgium, respectively. Pro-environmental values and moral licensing Policies and behavioural interventions using moral suasion and social marketing techniques have been used extensively to promote adoption of energy-efficient products and practices, frugal energy use, and low- carbon consumption85. They leverage on peoples’ pro-environmental values and they have been proven to be even more effective than financial incentives in certain occasions. For example, a randomized control trial for electricity reduction revealed that communicating public health risks outperformed communicating financial gains112. The challenge of such interventions, however, is to minimize potential negative spillovers due of moral licensing. For example, eco-labels aimed at inducing ‘greener’ consumption can stimulate overconsumption113. The issue is even more pronounced in the case of adoption since it is easier for individuals to consider adoption in tandem with the subsequent post-adoption use, which provides the ideal environment for moral licensing to take place9. That said, moral licensing is less probable when adoption is costly in monetary or in other terms114; when it is motivated by underlying pro-environmental values and identity115; and when individuals conceive adoption as part of a process and not the end of it, therefore creating a feeling of progress and self-efficacy64,101,116–118. The policy-making implications of such findings are that in order to avoid moral licensing when promoting adoption, monetary incentives, if in place, should not the main focus of the campaign or intervention. Additionally, it should be clearly communicated that adoption is only the first of a two-step procedure and will be insufficient unless combined with subsequent prudent intensity of use. Adding adoption and post-adoption use in the same mental account might be sufficient. Peer effects 11 Policies involving peer comparisons, status concerns, and communication of social norms can promote adoption by making the object of adoption visible119. Examples include the United Kingdom requiring registration plates of low fuel-consumption vehicles to be green, the Project Sunroof that allows homeowners to find out who of their neighbours have installed solar panels through aerial photos illustrating clustering120, and other visible symbols that serve as a welcome gift after adoption, such as sticker, doorplate, email signature or magnet121,122. These interventions may, however, backfire if not combined with post-adoption policies or nudges. To illustrate, in an attempt to signal pro-environmental behaviour, owners of electric or hybrid vehicles might use them more frequently in order to increase visibility. Regarding intensity of use, the effects of social norms is expected to be more limited since most of the times, the behaviour is ultimately private. The energy company O-power in the US, nevertheless, employed social norms in order to motivate households to moderate their energy use. In a series of experiments, households received messages regarding the mean energy use of their neighbours. On average, consumption is reduced, but below-average users increase their consumption after they learned about the norm. This negative effect can be overcome if the descriptive norm is accompanied by normative messages conveying approval, such as messages mentioning “Great” or smiley emojis123. An alternative way to employ peer influence is via community programs, such as the as the EcoTeam program or the Global Action Plan, where community members discuss ways of promoting energy conservation in general. Their effectiveness is not guaranteed, however124. Conclusions The extent to which adopting energy-efficient technologies and practices results in energy savings depends on how such technologies are used and how any associated savings are spent. The review reveals that bounded rationality and willpower tend to increase overall energy use as they magnify the rebound effect through higher intensity of use and re-spending on energy-intensive products. This involves moderation of the negative effects of economic factors and the positive effects of non-economic (bounded self-interest) factors. While the latter tend to increase adoption rates, their effect on direct rebound is uncertain due to moral licensing and social norms that encourage consumption. While bounded rationality is often employed to explain the puzzling energy gap – namely as a failure to engage in rational behaviour – rebound does not represent a puzzle, since it is consistent with both rational and limited-rational behaviour. The latter merely reduces or increases the magnitude of ‘rational rebound’. In this respect, the Review finds that effects on re-spending are the most uncertain, as they are difficult to identify and quantify given that they require observing behaviour across the whole set of consumption decisions by individuals. This poses a huge challenge for research on rebound. In terms of policy, we find that many behavioural interventions may be effective in promoting adoption and controlling intensity of use – thus curbing direct rebound – but not so much regarding re- 12 spending behaviour. We provide examples of interventions that can be combined to address both adoption and post-adoption behaviours. We have identified smart meters as a key tool to limit negative effects of inattention, habits and inertia on intensity of use. Using insights from the literature on mental accounting, we suggested that an additional feature of a smart meter can help to curb direct and indirect rebound. However, such interventions should not serve as a substitute for, but rather as a complement to, instruments that are known to effectively address rebound, such as carbon and energy pricing125,126. The extent to which a combination of interventions has positive or negative synergetic effects deserves further investigation. In particular, future research is needed on investigate how behavioural regularities affect the performance of pricing instruments like carbon taxes or energy subsidies, and the extent to which they interact with behavioural interventions in terms of rebound effects127. The review showed that evidence for certain behavioural regularities is quite mixed and sometimes scarce or missing. Most studies zoom in on the effect of a particular behavioural regularity on only one of the three rebound-relevant decisions. Studying them simultaneously will lead to a better picture of their joint effect on rebound while also providing insight into suitable combination of interventions to address adoption and post-adoption behaviours in tandem. Further research is needed to improve our understanding of rebound due to re-spending. A key question here is to what extent consumers are aware of energy-related savings, and if they are, whether this affects the way they spend these. Competing interests The authors declare no competing interests. 13 DISPLAY ITEMS Table 1. Behavioural regularities emanating from bounded rationality, willpower and self-interest Standard economic Insights from behavioural Behavioural regularities assumptions sciences relevant to rebound Unlimited cognitive abilities Limited abilities & time Inattention & misconceptions (Bounded) Optimizing behaviour Satisfying & heuristics Mental accounting Rationality Well-defined preferences Important “irrelevant” factors Defaults & habits (Bounded) Perfect self-control Limited self-control Present bias Willpower Other-regarding preferences Pro-environmental values Self-regarding preferences (Bounded) & moral values & moral licensing Self-interest Socially isolated agents Social interactions Peer influence 14 Table 2. Behavioural interventions and their impacts on energy-relevant decisions. The suggested interventions in each row should be used in tandem to promote adoption and control intensity of use and re-spending. Behavioural Adoption Intensity of use Re-spending regularity (energy gap) (direct rebound) (indirect rebound) Inattention & Energy-efficient labels and Real-time feedback misconceptions lifecycle information employing smart meters Loss frames for less Encourage a mental budget that combines adoption & Mental accounting energy-inefficient options intensity of use Energy defaults to energy- Setting defaults to pre- Defaults & habits efficient options & adoption levels energy-habit disruption Move incentives of (Gamified) goal setting Present bias energy-efficient options Tailor savings to eco-products programmes closer to the present Pro-environmental Moral & social framing values Avoid moral licensing by making identity motivations salient and communicating that & moral licensing adoption is not enough Public, status-enhancing Peer comparison of energy Peer influence ‘green signals’ consumption 15 Figure 1. Graphical illustration of the three types of rebound. Potential energy reduction after switching to a more fuel-efficient car is not fully realized due to three post-consumption behavioural reactions that lead to increased energy demand. These correspond to the three kinds of rebound, depicted in the figure by the red arrows: direct rebound results from using the car more often due to the reduced costs of driving (intensity of use); indirect rebound results from spending the monetary savings on other products and services whose production and consumption require energy (re-spending); and economy-wide rebound results from increased economic activity, due to various macroeconomic processes, such as higher economic growth or new economic activities. The figure is adapted from reference128 which illustrates rebounds for the case of switching to more energy-efficient light bulbs. 16 Figure 2. A schematic representation of the review. Selfish, moral and social motivations, moderated by bounded rationality and willpower, affect energy-relevant decisions, which in turn influence direct and indirect rebound. The elements of bounded rationality, willpower and self-interest explain the behavioural regularities listed in Table 1. For each regularity, we review empirical evidence for its impact on the three energy-relevant decisions. Figure 3. Dominant effects of behavioural regularities on energy-relevant decisions and rebound. Behavioural regularities that reflect limited rationality and willpower tend to decrease adoption and increase rebound, while those driven by bounded self-interest tend to increase adoption while their effects on rebound are less clear. 17 Figure 4. Smart meter adapted to indicate the days left for the initial investment on house insulation to be paid back. The additional feature nudges consumers to include the initial investment and post-adoption use in the same mental account, in turn encouraging them to reduce intensity of use. 18 Box 1. Behavioural dimensions of bounded rationality, willpower and self-interest that give rise to behavioural regularities Bounded rationality Rationality can be decomposed into three behavioural assumptions129: people possess good cognitive skills, have well-defined preferences, and exhibit optimizing behaviour. Bounded rationality describes that oftentimes individuals lack the cognitive capacity or simply the time to collect and analyse all relevant information before reaching a decision130. In the face of these limitations, they are not looking for strictly optimal solutions, but are satisficing instead, meaning that they settle for ‘good enough’ solutions. To this end, they employ heuristics or simple rules of thumb that allow them to make decisions in a fairly efficient way131. This, however, leads to cognitive biases, i.e. systematic errors in judgment and decision making. In addition, under heuristics the choice environment strongly affects peoples’ choices: apart from prices and preferences, other factors – supposedly irrelevant according to rational choice theory – determine people’s decisions. As an implication, policy-makers can nudge people to behave in a particular way by applying subtle modifications to the choice environment27. Bounded willpower Rational agent theory assumes that an action taken by an individual is the most preferred option in her choice-set. This ignores that people often lack the willpower to follow their preferences, instead being driven by temptations. Willpower is positioned in the zone between preferences and behaviours132. Bounded willpower involves a trade-off between long- and short-term goals, or between moral and selfish considerations. Due to self-control issues and attractiveness of short-term and selfish gains, people often opt for the latter. Bounded self-interest In a narrow sense, self-interest means focusing on personal outcomes, and is thus independent of the social context. Bounded self-interest recognizes that people care about the well-being of others and about moral values that act as internal constraints on behaviour133. Economists refer to these as “other-regarding preferences” or “identity concerns” while sociologists and psychologists call them “moral values”. These values in turn are shaped through social interactions, by what other people do or think, as captured by notions such as social norms, imitation or status- seeking134. These factors do not rule out that the underlying motives can be ultimately selfish. For instance, people often act pro-socially because they want to maintain a positive self-image or they follow a social norm just to avoid social punishment. In addition, bounded self-interest does not exclude the influence of narrowly selfish motives. Indeed, most of the time behaviour is the outcome of an interaction between selfish, moral and social motives. 19 References 1. Berkhout, P. H. G., Muskens, J. C. & W. Velthuijsen, J. Defining the rebound effect. Energy Policy 28, 425–432 (2000). 2. Brookes, L. The greenhouse effect: the fallacies in the energy efficiency solution. Energy Policy 18, 199–201 (1990). 3. Gillingham, K., Kotchen, M. J., Rapson, D. S. & Wagner, G. Energy policy: The rebound effect is overplayed. Nature vol. 493 475–476 (2013). 4. Azevedo, I. M. L. Consumer end-use energy efficiency and rebound effects. Annual Review of Environment and Resources vol. 39 393–418 (2014). A review article on the various definitions of rebound, the research gaps in the literature, and the importance of scope in estimating rebound. 5. Sorrell, S., Dimitropoulos, J. & Sommerville, M. Empirical estimates of the direct rebound effect: A review. Energy Policy 37, 1356–1371 (2009). 6. Brockway, P. E., Sorrell, S., Semieniuk, G., Heun, M. K. & Court, V. Energy efficiency and economy-wide rebound effects: A review of the evidence and its implications. Renewable and Sustainable Energy Reviews vol. 141 110781 (2021). 7. Peters, A. & Dütschke, E. Exploring rebound effects from a psychological perspective. in Rethinking climate and energy policies 89–105 (Springer International Publishing, 2016). doi:10.1007/978-3-319-38807-6_6. 8. Girod, B. & De Haan, P. Mental Rebound. https://www.research- collection.ethz.ch/bitstream/handle/20.500.11850/152329/eth-2118-01.pdf (2009). 9. Dütschke, E., Frondel, M., Schleich, J. & Vance, C. Moral licensing—another source of rebound? Front. Energy Res. 6, 38 (2018). A review suggesting that consumers may feel morally licensed to consume more energy after adopting a more energy-efficient technology or making an energy-conservation decision. 10. Santarius, T. & Soland, M. How technological efficiency improvements change consumer preferences: towards a psychological theory of rebound effects. Ecol. Econ. 146, 414–424 (2018). A study integrating rational-choice and psychological behavioural theories for the study of rebound and identifying multiple channels through which rebound can arise. 11. Seebauer, S. The psychology of rebound effects: Explaining energy efficiency rebound behaviours with electric vehicles and building insulation in Austria. Energy Res. Soc. Sci. 46, 311–320 (2018). 12. Font Vivanco, D., McDowall, W., Freire-González, J., Kemp, R. & van der Voet, E. The foundations of the environmental rebound effect and its contribution towards a general framework. Ecol. Econ. 125, 60–69 (2016). 13. Jolls, C., Sunstein, C. R. & Thaler, R. A Behavioral Approach to Law and Economics. Stanford Law Rev. 50, 1471 (1998). 14. Turrentine, T. S. & Kurani, K. S. Car buyers and fuel economy? Energy Policy 35, 1213–1223 (2007). 15. Allcott, H. Consumers’ perceptions and misperceptions of energy costs. Am. Econ. Rev. 101, 98– 104 (2011). 20 16. Wang, S., Fan, J., Zhao, D., Yang, S. & Fu, Y. Predicting consumers’ intention to adopt hybrid electric vehicles: using an extended version of the theory of planned behavior model. Transportation (Amst). 43, 123–143 (2016). 17. Gerarden, T. D., Newell, R. G. & Stavins, R. N. Assessing the energy-efficiency gap. J. Econ. Lit. 55, 1486–1525 (2017). 18. Cohen, F., Glachant, M. & Söderberg, M. Consumer myopia, imperfect competition and the energy efficiency gap: Evidence from the UK refrigerator market. Eur. Econ. Rev. 93, 1–23 (2017). 19. Halvorsen, B., Larsen, B. M., Wilhite, H. & Winther, T. Revisiting household energy rebound: Perspectives from a multidisciplinary study. Indoor Built Environ. 25, 1114–1123 (2016). 20. Waechter, S., Sütterlin, B. & Siegrist, M. The misleading effect of energy efficiency information on perceived energy friendliness of electric goods. J. Clean. Prod. 93, 193–202 (2015). 21. Keefer, Q. & Rustamov, G. Limited attention in residential energy markets: a regression discontinuity approach. Empir. Econ. 55, 993–1017 (2018). 22. Attari, S. Z., DeKay, M. L., Davidson, C. I. & De Bruin, W. B. Public perceptions of energy consumption and savings. Proc. Natl. Acad. Sci. U. S. A. 107, 16054–16059 (2010). 23. Camilleri, A. R., Larrick, R. P., Hossain, S. & Patino-Echeverri, D. Consumers underestimate the emissions associated with food but are aided by labels. Nat. Clim. Chang. 9, 53–58 (2019). 24. Thaler, R. H. Mental accounting matters. J. Behav. Decis. Mak. 12, 183–206 (1999). 25. Antonides, G., Manon de Groot, I. & Fred van Raaij, W. Mental budgeting and the management of household finance. J. Econ. Psychol. 32, 546–555 (2011). 26. Kahneman, D. & Tversky, A. Prospect theory: An analysis of decision under risk. Econometrica 47, 263–292 (1979). 27. Thaler, R. & Sunstein, C. Nudge: Improving decisions about health, wealth, and happiness. (Penguin, 2009). 28. Hahnel, U. J. J., Chatelain, G., Conte, B., Piana, V. & Brosch, T. Mental accounting mechanisms in energy decision-making and behaviour. Nat. Energy 5, 952–958 (2020). A perspective article on the ways through which mental accounting can affect energy-related behaviour, from which implications for rebound are drawn. 29. Schleich, J., Gassmann, X., Meissner, T. & Faure, C. A large-scale test of the effects of time discounting, risk aversion, loss aversion, and present bias on household adoption of energy- efficient technologies. Energy Econ. 80, 377–393 (2019). 30. Heutel, G. Prospect theory and energy efficiency. J. Environ. Econ. Manage. 96, 236–254 (2019). 31. Ofgem. Energy Supply Probe–Initial Findings Report. (2008). 32. Antonides, G. & Ranyard, R. Mental accounting and economic behaviour. in Economic Psychology (ed. Ranyard, B.) 123–138 (Wiley, 2017). 33. Beatty, T. K. M., Blow, L., Crossley, T. F. & O’Dea, C. Cash by any other name? Evidence on labeling from the UK Winter Fuel Payment. J. Public Econ. 118, 86–96 (2014). 34. Andor, M. A., Gerster, A., Gillingham, K. T. & Horvath, M. Running a car costs much more than people think — stalling the uptake of green travel. Nature vol. 580 453–455 (2020). 21 An empirical study showing that a majority of car drivers do not consider initial purchase costs of a new car as part of - i.e. in the same mental account as - total car costs. 35. de Haan, P., Mueller, M. G. & Peters, A. Does the hybrid Toyota Prius lead to rebound effects? Analysis of size and number of cars previously owned by Swiss Prius buyers. Ecol. Econ. 58, 592–605 (2006). 36. Cunha, Jr, M. & Caldieraro, F. Sunk-cost effects on purely Behavioral investments. Cogn. Sci. 33, 105–113 (2009). 37. Henderson, P. W. & Peterson, R. A. Mental accounting and categorization. Organ. Behav. Hum. Decis. Process. 51, 92–117 (1992). 38. Milkman, K. L. & Beshears, J. Mental accounting and small windfalls: Evidence from an online grocer. J. Econ. Behav. Organ. 71, 384–394 (2009). 39. Chitnis, M., Sorrell, S., Druckman, A., Firth, S. K. & Jackson, T. Who rebounds most? Estimating direct and indirect rebound effects for different UK socioeconomic groups. Ecol. Econ. 106, 12–32 (2014). 40. Kahneman, D., Knetsch, J. L. & Thaler, R. H. The endowment effect, loss aversion, and status quo bias. in Experiments in Environmental Economics vol. 1 243–256 (Taylor and Francis Inc., 2018). 41. Sunstein, C. R. & Reisch, L. A. Greener by Default. Trinity Coll. Law Rev. 21, (2018). 42. Verplanken, B. & Aarts, H. Habit, attitude, and planned behaviour: Is habit an empty construct or an interesting case of goal-directed automaticity? Eur. Rev. Soc. Psychol. 10, 101–134 (1999). 43. Huebner, G. M., Cooper, J. & Jones, K. Domestic energy consumption - What role do comfort, habit, and knowledge about the heating system play? Energy Build. 66, 626–636 (2013). 44. Pichert, D. & Katsikopoulos, K. V. Green defaults: Information presentation and pro- environmental behaviour. J. Environ. Psychol. 28, 63–73 (2008). 45. Abrahamse, W., Steg, L., Vlek, C. & Rothengatter, T. A review of intervention studies aimed at household energy conservation. J. Environ. Psychol. 25, 273–291 (2005). 46. Dinner, I., Johnson, E. J., Goldstein, D. G. & Liu, K. Partitioning default effects: Why people choose not to choose. J. Exp. Psychol. Appl. 17, 332–341 (2011). 47. Janssen, M. A. & Jager, W. Stimulating diffusion of green products - Co-evolution between firms and consumers. J. Evol. Econ. 12, 283–306 (2002). 48. Laibson, D. Golden eggs and hyperbolic discounting. Q. J. Econ. 112, 443–478 (1997). 49. Baumeister, R. & Vohs, K. Willpower, choice, and self-control. in Time and decision: Economic and psychological perspectives on intertemporal choice (eds. Loewenstein, G., Read, D. & Baumeister, R.) 201–216 (Russell Sage Foundation, 2003). 50. Bradford, D., Courtemanche, C., Heutel, G., McAlvanah, P. & Ruhm, C. Time preferences and consumer behavior. J. Risk Uncertain. 55, 119–145 (2017). 51. Fuerst, F. & Singh, R. How present bias forestalls energy efficiency upgrades: A study of household appliance purchases in India. J. Clean. Prod. 186, 558–569 (2018). 52. Tsvetanov, T. & Segerson, K. Re-evaluating the role of energy efficiency standards: A behavioral economics approach. J. Environ. Econ. Manage. 66, 347–363 (2013). 53. Allcott, H. & Taubinsky, D. Evaluating behaviorally motivated policy: Experimental evidence from the lightbulb market. Am. Econ. Rev. 105, 2501–2538 (2015). 22 54. Allcott, H. & Wozny, N. Gasoline Prices, Fuel Economy, and the Energy Paradox. Rev. Econ. Stat. 96, 779–795 (2014). 55. Harding, M. & Hsiaw, A. Goal setting and energy conservation. J. Econ. Behav. Organ. 107, 209– 227 (2014). 56. Lillemo, S. C. Measuring the effect of procrastination and environmental awareness on households’ energy-saving behaviours: An empirical approach. Energy Policy 66, 249–256 (2014). 57. Wesley Schultz, P. The structure of environmental concern: Concern for self, other people, and the biosphere. J. Environ. Psychol. 21, 327–339 (2001). 58. Mohana, R., Turaga, R., Howarth, R. B., Borsuk, M. E. & Rosenwald, J. Pro-environmental behavior: Rational choice meets moral motivation. Ann. N.Y. Acad. Sci2 1185, 211–224 (2010). 59. Black, J. S., Stern, P. C. & Elworth, J. T. Personal and Contextual Influences on Household Energy Adaptations. J. Appl. Psychol. 70, 3–21 (1985). 60. Wolske, K. S. & Stern, P. C. Contributions of psychology to limiting climate change. in Psychology and Climate Change: Human Perceptions, Impacts, and Responses 127–160 (Elsevier, 2018). doi:10.1016/B978-0-12-813130-5.00007-2. 61. Kollmuss, A. & Agyeman, J. Mind the gap: why do people act environmentally and what are the barriers to pro-environmental behavior? Environ. Educ. Res. 8, 239–260 (2002). 62. Monin, B. & Jordan, A. The dynamic moral self: A social psychological perspective. in Personality, Identity, and Character: Explorations in Moral Psychology (eds. Narvaez, D. & Lapsley, D.) (Cambridge university, 2009). 63. Hope, A. L. B., Jones, C. R., Webb, T. L., Watson, M. T. & Kaklamanou, D. The role of compensatory beliefs in rationalizing environmentally detrimental behaviors. Environ. Behav. 50, 401–425 (2018). 64. Truelove, H. B., Carrico, A. R., Weber, E. U., Raimi, K. T. & Vandenbergh, M. P. Positive and negative spillover of pro-environmental behavior: An integrative review and theoretical framework. Glob. Environ. Chang. 29, 127–138 (2014). 65. Wagner, G. & Zizzamia, D. Green Moral Hazards. SSRN Electron. J. (2019) doi:10.2139/ssrn.3486990. 66. Fischbacher, U., Schudy, S. & Teyssier, S. Heterogeneous preferences and investments in energy saving measures. https://www.econstor.eu/handle/10419/121763 (2015) doi:10.5282/UBM/EPUB.24894. 67. Di Maria, C., Ferreira, S. & Lazarova, E. Shedding light on the light bulb puzzle: the role of attitudes and perceptions in the adoption of energy efficient light bulbs. Scott. J. Polit. Econ. 57, 48–67 (2010). 68. Harding, M. & Rapson, D. Does absolution promote sin? A conservationist’s dilemma. Environ. Resour. Econ. 73, 923–955 (2019). 69. Clark, C. F., Kotchen, M. J. & Moore, M. R. Internal and external influences on pro- environmental behavior: Participation in a green electricity program. J. Environ. Psychol. 23, 237– 246 (2003). 70. Andersson, D., Linscott, R. & Nässén, J. Estimating car use rebound effects from Swedish microdata. Energy Effic. 12, 2215–2225 (2019). 23 An empirical study showing that when drivers switch to more efficient cars that are green-labelled direct rebound is null. 71. Matiaske, W., Menges, R. & Spiess, M. Modifying the rebound: It depends! Explaining mobility behavior on the basis of the German socio-economic panel. Energy Policy 41, 29–35 (2012). 72. Klöckner, C. A., Nayum, A. & Mehmetoglu, M. Positive and negative spillover effects from electric car purchase to car use. Transp. Res. Part D Transp. Environ. 21, 32–38 (2013). 73. Gatersleben, B., Steg, L. & Vlek, C. Measurement and Determinants of Environmentally Significant Consumer Behavior. Environ. Behav. 34, 335–362 (2002). 74. Vita, G. et al. Happier with less? Members of European environmental grassroots initiatives reconcile lower carbon footprints with higher life satisfaction and income increases. Energy Res. Soc. Sci. 60, 101329 (2020). 75. Laroche, M., Bergeron, J. & Barbaro-Forleo, G. Targeting consumers who are willing to pay more for environmentally friendly products. J. Consum. Mark. 18, 503–520 (2001). 76. Wolske, K. S., Gillingham, K. T. & Schultz, P. W. Peer influence on household energy behaviours. Nature Energy 1–11 (2020) doi:10.1038/s41560-019-0541-9. 77. Brick, C., Sherman, D. K. & Kim, H. S. “Green to be seen” and “brown to keep down”: Visibility moderates the effect of identity on pro-environmental behavior. J. Environ. Psychol. 51, 226–238 (2017). 78. Uren, H. V., Roberts, L. D., Dzidic, P. L. & Leviston, Z. High-status pro-environmental behaviors: Costly, effortful, and visible. Environ. Behav. 001391651988277 (2019) doi:10.1177/0013916519882773. 79. Griskevicius, V., Tybur, J. M. & Van den Bergh, B. Going green to be seen: Status, reputation, and conspicuous conservation. J. Pers. Soc. Psychol. 98, 392–404 (2010). 80. Sexton, S. & Sexton, A. The Prius halo and willingness to pay for environmental bona fides. J. Environ. Econ. Manage. 67, 303–317 (2014). 81. Bollinger, B. & Gillingham, K. Peer effects in the diffusion of solar photovoltaic panels. Mark. Sci. 31, 900–912 (2012). 82. Farrow, K., Grolleau, G. & Ibanez, L. Social norms and pro-environmental behavior: A review of the evidence. Ecological Economics vol. 140 1–13 (2017). 83. Allcott, H. & Rogers, T. The short-run and long-run effects of behavioral interventions: Experimental evidence from energy conservation. American Economic Review vol. 104 3003– 3037 (2014). 84. Demarque, C., Charalambides, L., Hilton, D. J. & Waroquier, L. Nudging sustainable consumption: The use of descriptive norms to promote a minority behavior in a realistic online shopping environment. J. Environ. Psychol. 43, 166–174 (2015). 85. Peattie, K. Green Consumption: Behavior and Norms. Annu. Rev. Environ. Resour. 35, 195–228 (2010). 86. Jackson, T. Motivating sustainable consumption: A review of evidence on consumer behaviour and behavioural change. (2005). 87. Biswas, A., Mukherjee, A. & Roy, M. Leveraging Factors for Consumers’ Car Purchase Decisions-A Study in an Emerging Economy. J. Manag. Policies Pract. 2, 99–111 (2014). 24 88. Nisa, C. F., Bélanger, J. J., Schumpe, B. M. & Faller, D. G. Meta-analysis of randomised controlled trials testing behavioural interventions to promote household action on climate change. Nat. Commun. 10, (2019). A recent meta-analysis finding that the size of emissions reduction of various behavioural interventions is not as big as previously thought. 89. Abrahamse, W. & Steg, L. Social influence approaches to encourage resource conservation: A meta-analysis. Glob. Environ. Chang. 23, 1773–1785 (2013). 90. Karlin, B., Zinger, J. F. & Ford, R. The effects of feedback on energy conservation: A meta- analysis. Psychol. Bull. 141, 1205–1227 (2015). 91. Delmas, M. A., Fischlein, M. & Asensio, O. I. Information strategies and energy conservation behavior: A meta-analysis of experimental studies from 1975 to 2012. Energy Policy 61, 729–739 (2013). 92. Andor, M. A. & Fels, K. M. Behavioral economics and energy conservation – A systematic review of non-price interventions and their causal effects. Ecol. Econ. 148, 178–210 (2018). A systematic review of social comparison, commitment devices, goal setting, and labelling as behavioural interventions aimed at achieving reductions in household energy use. 93. Iweka, O., Liu, S., Shukla, A. & Yan, D. Energy and behaviour at home: A review of intervention methods and practices. Energy Research and Social Science vol. 57 101238 (2019). 94. Camilleri, A. R. & Larrick, R. P. Metric and Scale Design as Choice Architecture Tools. J. Public Policy Mark. 33, 108–125 (2014). 95. Sintov, N. D. & Schultz, P. W. Unlocking the potential of smart grid technologies with behavioral science. Front. Psychol. 6, 410 (2015). 96. Darby, S. The effectiveness of feedback on energy consumption. (2006). 97. Mogles, N. et al. How smart do smart meters need to be? Build. Environ. 125, 439–450 (2017). A study showing that to increase their effectiveness, smart energy meters should provide context to the feedbacks they provide and help to improve energy literacy of households. 98. Tiefenbeck, V. et al. Overcoming salience bias: How real-time feedback fosters resource conservation. Manage. Sci. 64, 1458–1476 (2018). 99. Ropret Homar, A. & Knežević Cvelbar, L. The effects of framing on environmental decisions: A systematic literature review. Ecological Economics vol. 183 106950 (2021). A systematic review regarding the effects of framing on environmental decisions, which describe the conditions under which loss versus gain frames are more effective in promoting behavioural change. 100. Hermsen, S., Frost, J., Renes, R. J. & Kerkhof, P. Using feedback through digital technology to disrupt and change habitual behavior: A critical review of current literature. Computers in Human Behavior vol. 57 61–74 (2016). 101. Fujii, S., Gärling, T. & Kitamura, R. Changes in Drivers’ Perceptions and Use of Public Transport during a Freeway Closure. Environ. Behav. 33, 796–808 (2001). 102. Fujii, S. & Gärling, T. Development of script-based travel mode choice after forced change. Transp. Res. Part F Traffic Psychol. Behav. 6, 117–124 (2003). 103. Bamberg, S. Is a Residential Relocation a Good Opportunity to Change People’s Travel Behavior? 25 Results From a Theory-Driven Intervention Study. Environ. Behav. 38, 820–840 (2006). 104. Thomas, G. O., Poortinga, W. & Sautkina, E. Habit discontinuity, self-activation, and the diminishing influence of context change: evidence from the UK understanding society survey. PLoS One 11, (2016). 105. Brown, Z., Johnstone, N., Haščič, I., Vong, L. & Barascud, F. Testing the effect of defaults on the thermostat settings of OECD employees. Energy Econ. 39, 128–134 (2013). 106. McCalley, L. T. From motivation and cognition theories to everyday applications and back again: The case of product-integrated information and feedback. Energy Policy 34, 129–137 (2006). 107. Johnson, E. J. et al. Beyond nudges: Tools of a choice architecture. Mark. Lett. 23, 487–504 (2012). 108. Allcott, H. & Knittel, C. Are consumers poorly informed about fuel economy? Evidence from two experiments. Am. Econ. J. Econ. Policy 11, 1–37 (2019). 109. Osbaldiston, R. & Paul Schott, J. Environment and Behavior Meta-Analysis of Proenvironmental Behavior Experiments Environmental Sustainability and Behavioral Science. Environ. Behav. 44, 257–299 (2012). 110. Locke, E. A. & Latham, G. P. Building a practically useful theory of goal setting and task motivation: A 35-year odyssey. Am. Psychol. 57, 705–717 (2002). 111. Deconinck, G. et al. An approach towards socially acceptable energy saving policies via monetary instruments on the smart meter infrastructure. in 3rd International Conference on Next Generation Infrastructure Systems for Eco-Cities, INFRA 2010 - Conference Proceedings (2010). doi:10.1109/INFRA.2010.5679226. 112. Asensio, O. I. & Delmas, M. A. Nonprice incentives and energy conservation. Proc. Natl. Acad. Sci. U. S. A. 112, E510–E515 (2015). 113. Bougherara, D., Grolleau, G. & Thiébaut, L. Can labelling policies do more harm than good? An analysis applied to environmental labelling schemes. Eur. J. Law Econ. 19, 5–16 (2005). 114. Gneezy, U., Imas, A. & Madarász, K. Conscience Accounting: Emotion Dynamics and Social Behavior. http://dx.doi.org/10.1287/mnsc.2014.1942 60, 2645–2658 (2014). 115. Thøgersen, J. & Crompton, T. Simple and Painless? The Limitations of Spillover in Environmental Campaigning. J. Consum. Policy 2009 322 32, 141–163 (2009). 116. Nilsson, A., Bergquist, M. & Schultz, W. P. Spillover effects in environmental behaviors, across time and context: a review and research agenda. Environ. Educ. Res. 23, 573–589 (2017). 117. Nash, N. et al. Climate-relevant behavioral spillover and the potential contribution of social practice theory. Wiley Interdiscip. Rev. Clim. Chang. 8, e481 (2017). 118. Maki, A. et al. Meta-analysis of pro-environmental behaviour spillover. Nat. Sustain. 2, 307–315 (2019). 119. Bollinger, B., Gillingham, K., Kirkpatrick, A. J. & Sexton, S. Visibility and Peer Influence in Durable Good Adoption. SSRN Electron. J. (2019) doi:10.2139/ssrn.3409420. 120. Sunter, D. A., Castellanos, S. & Kammen, D. M. Disparities in rooftop photovoltaics deployment in the United States by race and ethnicity. Nat. Sustain. 2, 71–76 (2019). 121. Font Vivanco, D., Kemp, R. & van der Voet, E. How to deal with the rebound effect? A policy- oriented approach. Energy Policy 94, 114–125 (2016). 26 A review article that analyses strategies for curbing rebound, suggesting that economic instruments might be the most effective. 122. Hanimann, R. Consumer Behaviour in Renewable Electricity: Can Identity Signaling Increase Demand for Renewable Electricity? (Uppsala University, 2013). 123. Allcott, H. Social norms and energy conservation. J. Public Econ. 95, 1082–1095 (2011). 124. Bardsley, N. et al. Domestic thermal upgrades, community action and energy saving: A three-year experimental study of prosperous households. Energy Policy 127, 475–485 (2019). 125. Freire-González, J. Energy taxation policies can counteract the rebound effect: analysis within a general equilibrium framework. Energy Effic. 13, 69–78 (2020). 126. van den Bergh, J. C. J. M. Energy conservation more effective with rebound policy. Environ. Resour. Econ. 48, 43–58 (2011). 127. Drews, S., Exadaktylos, F. & van den Bergh, J. C. J. M. Assessing synergy of incentives and nudges in the energy policy mix. Energy Policy 144, 111605 (2020). 128. Sigmavert, C. Fiscal Interventions to Change Energy Behaviour. A review of the literature. (2019). 129. Gary, B. The economic approach to human behavior. (University of Chicago Press, 1976). 130. Simon, H. A. A behavioral model of rational choice. Q. J. Econ. 69, 99 (1955). 131. Kahneman, D., Tversky, A. Choices, values, and frames. (Cambridge University Press, 2000). 132. White, M. D. Does homo economicus have a will? in Economics and the Mind (ed. Barbara Montero, M. D. W.) 143–158 (2007). 133. Sen, A. Rational Fools: A Critique of the Behavioral Foundations of Economic Theory. Philos. Public Aff. 6, 317–344 (1977). 134. Henrich, J. et al. ‘Economic man’ in cross-cultural perspective: Behavioral experiments in 15 small-scale societies. Behav. Brain Sci. 28, 795–855 (2005). 27