Ecological Economics 217 (2024) 108070 Contents lists available at ScienceDirect Ecological Economics journal homepage: www.elsevier.com/locate/ecolecon Analysis How carbon pricing affects multiple human needs: An agent-based model analysis Joël Foramitti a,b , Ivan Savin c,a,d ,∗, Jeroen C.J.M. van den Bergh a,b,e,f a Institute of Environmental Science and Technology, Universitat Autònoma de Barcelona, Spain b Institute for Environmental Studies, Vrije Universiteit Amsterdam, The Netherlands c ESCP Business School, Madrid Campus, Spain d Graduate School of Economics and Management, Ural Federal University, Yekaterinburg, Russian Federation e ICREA, Barcelona, Spain f School of Business and Economics, Vrije Universiteit Amsterdam, The Netherlands ARTICLE INFO ABSTRACT Keywords: Climate policy is urgently needed to reduce emissions, but must also be evaluated in regard to its impact Climate policy on human quality of life. To identify policy approaches that are able to achieve effective mitigation together Human needs with high levels of well-being, multiple human needs must be taken into account. To this end, we present an Low-carbon infrastructure agent-based model that is able to describe the interaction between various economic sectors on the supply-side Transport and individuals with heterogeneous incomes and needs on the demand-side. Individuals make choices under Well-being both income and time constraints; the latter being needed for non-market activities and time-intensive forms of consumption such as low-carbon modes of transport. The following climate policy instruments are considered: a carbon tax, permit trading, direct regulation, and investments in low-carbon infrastructure. Impacts are analyzed in regard to three different types of mitigation: avoid, shift, and improve. Results show that to achieve emission reductions together with high levels of well-being, carbon pricing is best combined with effective improvements of low-carbon infrastructure; revenue recycling should be progressive; and unnecessary abatement costs should be avoided. A comparison is provided with traditional formulations of social welfare. 1. Introduction However, while some needs can be satisfied through specific types of consumption, others also require non-market activities. A central question of climate action is how emission reduction can According to Creutzig et al. (2021), there is significant potential be combined with a high quality of life for all people (O’Neill et al., to reduce emissions in a way that is consistent with high levels of 2018). On one hand, this depends on the economy’s emission intensity, well-being. To do so, they suggest that (1) polluting activities of little which describes the amount of consumption that can be realized per importance to well-being should be avoided; (2) activities of high unit of emission. However, not all consumption is of equal importance importance should, whenever possible, be shifted to low-carbon alter- to human well-being (Gough, 2017). Thus, it also matters which goods natives; and (3) improvements in carbon efficiency should be focused and services are produced and how effectively they are allocated for on activities where such alternatives are missing. An overview of this the aim of increasing human well-being (Roberts et al., 2020; Rao and avoid-shift-improve1 approach is given in Fig. 1. Wilson, 2021). How actual mitigation is achieved and ultimately impacts well- An essential aspect of human well-being is that it depends on being depends on which climate policies are used. While multiple multiple human needs, all of which must be met to enable a high instruments are available for designing climate policy, it is still some- quality of life (Jackson et al., 2004; Costanza et al., 2007; Gough, what contested which of them are most effective to reach the Paris 2017). Strong increases in well-being can be achieved through the climate goals (Cullenward and Victor, 2020; Penasco et al., 2021; satisfaction of needs that are deprived, while little can be gained from van den Bergh et al., 2021). Many researchers argue that carbon needs that are satiated (Sirgy, 2021). A high quality of life can thus be pricing should be the key instrument of climate policy (Newell and achieved by prioritizing those goods and services that are most crucial. ∗ Corresponding author at: Institute of Environmental Science and Technology, Universitat Autònoma de Barcelona, Spain. E-mail address: ivan.savin@uab.cat (I. Savin). 1 The decomposition of abatement into these three categories is a common idea that can be found under various names such as scale effects, substitution or composition, and efficiency improvements or technique effects (Larch and Wanner, 2017). https://doi.org/10.1016/j.ecolecon.2023.108070 Received 5 January 2023; Received in revised form 8 September 2023; Accepted 28 November 2023 Available online 8 December 2023 0921-8009/© 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). J. Foramitti et al. Ecological Economics 217 (2024) 108070 Stavins, 2003; Cramton et al., 2017; Boyce, 2018; Savin and van den Bergh, 2022). In this study we take it as exemplary instrument and consider its interaction with other policy measures. Carbon pricing can be implemented through a carbon tax, where emitters pay a fee per unit of caused emissions; or a permit market, where emitters can trade emission permits (Goulder and Schein, 2013). An alternative approach is to restrict emissions through direct regulation such as an emission quota, which also translates into an implicit carbon price. A key difference between these instruments is that the latter two can cause higher abatement costs due to inefficient selection of abatement options (Foramitti et al., 2021a). Another well-known difference re- gards windfall profits. This can happen under direct regulation, but also under permit trading if a grandfathering scheme is used to distribute permits for free (Foramitti et al., 2021b). There are further differences between these instruments that are not the focus of this study. For an overview, see Goulder and Schein (2013). Mitigation can further be aided by improvements of low-carbon infrastructure. The idea of this approach is to provide individuals with additional incentives for the use of low-carbon alternatives apart from carbon pricing. A particular way in which low-carbon alternatives can Fig. 1. The avoid-shift-improve framework represent different ways of achieving be made more attractive is by reducing the amount of time they require. emission reductions in a way that is consistent with high levels of well-being. The most relevant example for this is the case of mobility, where built infrastructure for walking, cycling, and public transport reduces the amount of time that is required to choose low-carbon modes of of an existing good or service. Outcomes are evaluated in regard to mobility. Similar potential for infrastructure improvements has also effectiveness, well-being, and inequality. been found regarding other parts of the economy (Creutzig et al., 2021). The remainder of this chapter is structured as follows. Section 2 The aim of this study is to examine the performance of a car- motivates the focus on carbon pricing. Section 3 describes the general bon pricing when multiple human needs are taken into account. It model. Section 4 introduces the numerical experiment that is con- uses the method of agent-based models (ABMs), which allows for the ducted with the model. Section 5 presents the results, while Section 6 representation of economic dynamics based on the interaction of au- concludes. tonomous agents with heterogeneous characteristics (Dawid and Delli Gatti, 2018; D’Orazio and Valente, 2019; Rengs and Scholz-Wäckerle, 2. Policy context 2020; Castro et al., 2020; Terranova and Turco, 2022; Savin et al., 2022). ABMs have proven to be flexible enough to integrate elements Effective emission reduction is likely to require a price on carbon from different disciplines such as sociology, psychology and economics emissions (Baranzini et al., 2017). This approach increases the relative and therefore are most suitable for this purpose (Savin et al., 2022). The prices of goods and services in proportion to their carbon content, presented model builds upon the Needs & Limits framework introduced stimulating firms and individuals to take direct and indirect emissions by Foramitti (2023), where human individuals try to increase their into account in their decisions and switch to low-carbon alternatives. quality of life and are heterogeneous in regard to both their income This incentivizes emitters with the lowest marginal abatement costs to and needs. In the present study agents only interact in the labor market act first, which in turn reduces the overall costs of mitigation (Newell and final good markets affecting chances of their peers to find a job and Stavins, 2003). or satisfy their needs. Furthermore, as shown in Foramitti (2023), Within neoclassical economics, carbon pricing is considered the best more sophisticated interaction (e.g., imitation or status-comparison) approach to climate policy as it is able to achieve mitigation at the can be implemented in this approach. This demand-side is combined lowest possible costs to society, which in turn leads to optimal wel- with a supply-side consisting of sectors that produce different types fare (Pigou, 1920).2 Even without relying on the notion of optimality, of consumption goods and are able to reduce their emission intensity economists have argued that ’carbon pricing is more effective, at a through the adoption of low-carbon technology. The resulting model reasonable cost, in reducing emission than other approaches’ (Baranzini makes it possible to analyze how climate policy, through the dynamic et al., 2017). Recent models have further shown how revenue recycling behavior of various economic actors, translates into selective growth can bring additional benefits by reducing inequality and thus increasing and decline of different sectors and ultimately into well-being. well-being of low-income groups (Budolfson et al., 2021). Multiple numerical experiments are conducted to compare the per- These advantages of carbon pricing have been criticized from a formance of different policy approaches. First, we compare the effects perspective of human needs, as ‘it is only the level of mitigation costs of carbon pricing under varying price levels and different schemes for rather than the relevance to human well-being which determines which tax revenue recycling. Second, we analyze the impact of two poten- emissions will cease and which will continue’ (Huwe and Frick, 2022). tial side-effects: the presence of windfall profits and the existence of However, carbon pricing will not only affect industry but also individu- unnecessary abatement costs due to inefficient selection of abatement als — the latter will move away from high-carbon options unless these options. We further consider the impact of low-carbon infrastructure are essential for their well-being. That is, both individuals and produc- investments in combination with carbon pricing. Finally, the needs- ers will be forced to trade off all costs against all benefits (Baranzini based approach of this model is compared with traditional formulations et al., 2017; van den Bergh and Savin, 2021). of social welfare. Reductions in emissions are analyzed based on the Policy conclusions on this matter depend on the objective criteria relative share of the three strategies: avoid, shift, and improve. The used for the assessment of social welfare (Botzen and van den Bergh, first strategy, avoid – also called reduce – implies that the consumer chooses to use none of or less of a resource. The second strategy, shift – also called maintain – that the consumer switches from a less 2 While Pigou himself never considered climate policy or the externality sustainable method of consumption to a more sustainable one. The last of carbon emissions, this result is a special case of his work on environmental strategy — suggests that the consumer increases the resource efficiency externalities and policy. 2 J. Foramitti et al. Ecological Economics 217 (2024) 108070 2014). The conventional economic approach is to look at social welfare 3.3. Production in terms of the total utility that can be gained from the consumption of goods. Most models applied to the study of climate policy regard Sectors produce goods based on their available workforce 𝑁𝑠,𝑡 and only a single representative good. When multiple types of consumption labor intensity 𝜇𝑠 . are taken into account such as in Budolfson et al. (2021), utility is 𝑁𝑠,𝑡 𝑃 commonly defined through a constant elasticity of substitution (CES) 𝑞𝑠,𝑡 = (4) 𝜇𝑠 function. The emissions of each sector depend on the emission intensity 𝜀𝑠 : Here, the objective criterion is people’s quality of life (QOL), with the aim of policy maker being to reach high average levels with a 𝑃 𝑒𝑠,𝑡 = 𝑞𝑠,𝑡 𝜀𝑠 (5) low disparity. Particular emphasis is put on the fact that multiple human needs must be met to enable a high quality of life (Jackson The costs per unit of production in each sector depend on the average et al., 2004; Costanza et al., 2007; Gough, 2017). Based on the Needs wage 𝑤, labor intensity 𝜇𝑠 , emission intensity 𝜀𝑠 , and carbon price 𝜏. & Limits framework, this conceptualization of well-being is described 𝑐𝑠 = 𝑤 𝜇𝑠 + 𝜏 𝜀𝑠 (6) through an extended CES function that makes it possible to represent a wide range of psychological characteristics such as the satiability of The sales price is based on a fixed mark-up rate 𝑚 on these production needs (Foramitti, 2023). QOL is thus representing subjective well-being costs: in the broadest sense. Agents undertake actions in different life domains 𝑝𝑠 = 𝑐𝑠 (1 + 𝑚) (7) maximizing their overall satisfaction given their social, economic, and environmental experiences. The present study reports an application of 3.4. Income the QOL framework to climate policy contributing to earlier literature that integrates the perspective of needs satisfaction into economic Each round, individuals receive the wage 𝑤𝑖 as their first source research (Bogdanovych and Trescak, 2016; Jager, 2017; Kangur et al., of income. Wages are heterogeneous, and follow an exogenously given 2017). income distribution. In addition, individuals receive the recycled rev- enues from carbon pricing 𝑟𝑡 , which depend on each sector’s emissions 3. Model description and the part of the carbon price 𝜏 that has to be paid as a fee to the regulator. The windfall profit rate3 𝜒1 describes a share of the carbon The model consists of 𝑛𝐼 human individuals and 𝑛𝑆 economic sec- price that is not paid as a fee to the regulator, which will be explored tors that produce different types of economic goods. Sectors represent as a potential side-effect. different life domains in which agents fulfill their needs. Since agents ∑ 𝑟𝑡 = 𝑒𝑠,𝑡 𝜏 (1 − 𝜒1 ) (8) have limited resources (time, money, etc.), actions of agents face trade- 𝑠∈𝑆 offs in different domains. Individuals are denoted by the index 𝑖 ∈ 𝐼 = {1, … , 𝑛𝐼 }. Sectors and their respective goods are denoted by The model includes three different ways of recycling the revenue from carbon pricing: 𝑠 ∈ 𝑆 = {1, … , 𝑛𝑆 }. The simulation follows discrete time-steps (or rounds), denoted by 𝑡 ∈ 𝑇 = {1, … , 𝑛𝑇 }. The chain of events during • Neutral: Revenues are given out to individuals in proportion to each step 𝑡 is as follows: their income — which means that the Gini coefficient of the resulting income distribution remains unchanged. 1. Sectors decide about their objectives • Progressive: Revenues are split in equal shares between all indi- 2. Sectors and individuals interact on the labor market viduals — which comes down to a scheme of moderately progres- 3. Sectors produce goods and cause emissions sive revenue distribution (Konc et al., 2022). 4. Individuals receive income • Infrastructure improvements: A share 𝜍 of the revenue 𝑟𝑡 is 5. Individuals consume goods and perform activities used to finance infrastructure improvements from a specific sector 6. Sectors adopt technological improvements 𝑠′ at a price 𝑝𝑠′ .4 This reduces the amount of time 𝜓𝑎,𝑡 that is The following subsections describe each of these steps in more detail, needed to perform an activity 𝑎 (see Section 3.5). The parameter adhering to the same order. 𝜆𝑎 describes the effectiveness of these improvements. 𝜆𝑎 3.1. Objectives 𝜓𝑎,𝑡 = 𝜓𝑎0 − 𝜍 𝑟𝑡 (9) 𝑝𝑠′ 𝐺 to match their As a third and final source of income, remaining profit from both Each sector sets their desired production level 𝑞𝑠,𝑡 markup 𝑚 and windfall profits 𝜏 𝜒1 is distributed as dividends to expected demand, which is based on the sector’s experienced demand 𝐷 . Following Caiani et al. (2016), this each individual in proportion to their income, assuming in a simpli- from the previous round 𝑞𝑠,𝑡−1 fied manner that individuals hold company stocks in relation to their objective is further adjusted by a desired inventory share 𝜈 and the 𝐼 wealth (Caiani et al., 2016). existing inventory of unsold goods from the previous round 𝑞𝑠,𝑡−1 . Note that the carbon tax without any redistribution is neutral with 𝐺 𝐷 𝐼 respect to income inequality since we do not assume agents with lower 𝑞𝑠,𝑡 = 𝑞𝑠,𝑡−1 (1 + 𝜈) − 𝑞𝑠,𝑡−1 (1) income to have higher consumption of high-carbon good. However, as has been shown in Foramitti (2023), the tax causes an increase 3.2. Labor market in emissions inequality, which reduces emissions more among low- income agents than among high-income agents since the latter are The labor demand of each sector is based on the desired production 𝐺 and labor intensity 𝜇 . less price-sensitive in their demand. Moreover, without tax revenue level 𝑞𝑠,𝑡 𝑠 redistribution, average QOL is reduced by the tax. 𝐷 𝐺 𝑁𝑠,𝑡 = 𝑞𝑠,𝑡 𝜇𝑠 (2) The labor market is represented in a simplified manner, assuming 3 This is a well-known side-effect under permit trading if permits are that each sector manages to hire the share of available labor that is allocated through grandfathering. 4 proportional to their share of labor demand within the economy: These investments are reminiscent to investments into climate projects 𝐷 𝑛𝐼 (such as investments in public transport), which are considered in many 𝑁𝑠,𝑡 𝑁𝑠,𝑡 = ∑ (3) countries as the most supported option to spend carbon tax revenues by the 𝐷 general public (Maestre-Andrés et al., 2021). 𝑠′ ∈𝑆 𝑁𝑠′ ,𝑡 3 J. Foramitti et al. Ecological Economics 217 (2024) 108070 3.5. Consumption sector, including capital production, new material inputs, maintenance, or higher labor requirements in the production of end-use goods. The behavior and well-being of individuals is based on the Needs This cost of technological adoption is represented by the abatement & Limits framework from Foramitti (2023).5 In each round, individuals cost factor 𝛽𝑠 , which describes the marginal increase in labor intensity choose their behavior from a given set of possible activities 𝑎 ∈ 𝐴 = 𝛥𝜇 that results from a marginal reduction of the emission intensity {1, … , 𝑛𝐴 }. The extent to which each activity is undertaken by an indi- 𝛥𝜀.6 vidual during each time-step is described by the activity intensity 𝛼𝑖,𝑎,𝑡 . 𝛥𝜇 𝛽𝑠 = (12) The activity intensity describes the amount of each activity performed 𝛥𝜀 by an individual at a time. Sectors choose a level of abatement through technological adoption These activities lead to the satisfaction of human needs, which in 𝜙𝑠 ∈ [0, 1] in order to minimize their production costs. This means turn leads to fulfillment 𝑞𝑖,𝑑,𝑡 ∈ [0, 1] within different life domains 𝑑 ∈ that they increase 𝜙𝑠 up to the point where the marginal increase in 𝐷 = {1, … , 𝑛𝐷 }. The impact of each activity on each domain is given labor intensity would lead to equal costs as the marginal gains from by the impact factor 𝛿𝑎,𝑑 . Life domains are described as consisting of paying less carbon tax. Since production factors are kept constant in satiable needs, which means that an activity will add to little additional this model, this choice does not change over time. This optimal point fulfillment if the needs of that domain are already satiated to a high is derived in Appendix A. degree. The strength of this effect is represented by the satiation rate 𝑘𝑖,𝑑 . It is worth noting that unlike (Maslow, 1943), we assume no fixed 4. Numerical experiments order between different needs since their relative importance depends on their current level of deprivation (Jackson et al., 2004). Multiple numerical experiments are conducted in this study. The ∑ 𝑞𝑖,𝑑,𝑡 = 1 − 𝑒 − 𝑘𝑖,𝑑 𝑎∈𝐴 𝛼𝑖,𝑎,𝑡 𝛿𝑎,𝑑 (10) computational model is written in Python with the AgentPy pack- age (Foramitti, 2021). The code is available under an open-source The quality of life 𝑄𝑖,𝑡 ∈ [0, 1] is described as a bounded variable, license. representing a range from the worst possible life (𝑄𝑖,𝑡 = 0) to the The following describes the illustrative setting that is assumed for best possible life (𝑄𝑖,𝑡 = 1). The relationship between the separate life these experiments. An overview is given in Fig. 2. Parameter values are domains and the quality of life is described by a constant elasticity of provided in Appendix B. The model describes two sectors that produce substitution function, which is able to represent the interaction between goods for one specific life domain. The first produces a high-carbon different life domains. The parameters 𝜔𝑑 describe the weights of each good, and the second produces a low-carbon alternative. A third sector life domain and 𝜎 the substitution factor. represents the rest of the economy, which satisfies an aggregate life ∑( )1 domain for all other market-related needs. In addition, the rest of the 𝑄𝑖,𝑡 = 𝜔𝑑 𝑞𝑖,𝑑,𝑡 𝜎 𝜎 (11) economy can also produce and sell infrastructure improvements to the 𝑑∈𝐷 policymaker. Some activities regard the consumption of goods from a sector, in Individuals have two resources at hand: money and time. They can which case individuals are restricted by their income budget. Next to use these resources for four activities. The first three represent the budget constraints, consumers have a time constraint which notably consumption of goods from the three sectors, and thus require money. affects activities that use infrastructure. As a result, the availability of The consumption of the low-carbon good can be further characterized different types of infrastructure creates more flexibility to meet this as more time-costly. Without loss of generality we assume henceforth second constraint. Similarly, improvements in infrastructure can allow that the high-carbon good does not require consumption time. The reducing time use of specific activities, such as transport. The impact amount of time that is required can be reduced through low-carbon of each activity on time is given by 𝜓𝑎,𝑡 . infrastructure improvements that are paid for by the policymaker. During each time-step, individuals try to make the best choice of Finally, individuals can also use their time for non-market activities, 𝛼𝑖,𝑎,𝑡 in order to improve their quality of life. Following Foramitti which lead to the satisfaction of non-market needs. (2023), this is formalized through a numerical algorithm of constrained A real-world example of this setting would be to think of the specific optimization, using sequential least square programming. Agents start life domain as the need for mobility. Cars can then be seen as the from their choices of 𝛼𝑖,𝑎,𝑡−1 from the previous round, and try to adjust brown good and public transport or bicycles as the green good. In this the values of 𝛼𝑖,𝑎,𝑡 in order to reach higher levels of well-being while example, the three types of abatement would be as follows. Cars could staying within their given constraints of money and time. be improved, i.e. made more carbon efficient. Consumption could shift The consumption choices summed over all individuals becomes the to a different sector like walking, cycling, and public transportation experienced demand 𝑞𝑠,𝑡 𝐷 of each sector. When the total demand for a that is able to fulfill the same need for mobility. Improvements of low- sector exceeds the available supply for sale, the consumption of each carbon infrastructure can represent a wide range of measures in this agent is reduced proportionally. If time is freed up in this process, it case, including: vehicle stocks for shared mobility; charging stations is used for non-market activities. The resulting consumption becomes a for electric vehicles; or built infrastructure such as rails, bike lanes, and 𝑆. sector’s actual sales 𝑞𝑠,𝑡 pedestrian streets that would make these alternatives more convenient. Finally, the total amount of transportation could be reduced, since 3.6. Technological adoption not all transportation is essential to well-being. While this example of mobility is suited for illustration, the three strategies of avoid, shift, Sectors are able to adopt new technologies that will reduce their and improve are also relevant to other parts of the economy (Creutzig emission intensity. Since labor is the only production input in this et al., 2021). model, the cost of this improvement is represented as an increase of The described setting consists of 100 individuals, each representing labor intensity. This is because low-carbon technologies tend to be one percentile of a population. Their income rates are calibrated to more labor-intensive (Fankhauser et al., 2008; Bulavskaya and Reynes, 2018; Fragkos and Paroussos, 2018) and thus can cause additional labor requirements at different points within the supply chain of a 6 In the literature technological change is sometimes modeled in a more sophisticated way (Dawid and Delli Gatti, 2018; Lamperti et al., 2018; Rengs and Scholz-Wäckerle, 2020; Savin, 2021), but since our focus in this study is 5 Alternative formulations of social welfare for comparison are described in on the interaction of human needs and climate policy instruments, we leave Appendix D. This section provides only a short overview. extensions of this part to further research. 4 J. Foramitti et al. Ecological Economics 217 (2024) 108070 Fig. 2. Overview of agents’ resources, activities, and life domains. follow the global income distribution in 2016, which does not imply 2. The amount of mitigation that can be achieved by a given level that these agents represent the global population but merely display of carbon price. the same degree of inequality as in the real world. The satiation rates 3. Distributional impacts on income inequality. of each agent are drawn from a random distribution. The distribution Analysis is further aided by a decomposition of mitigation into the parameters as well as the domain weights are calibrated so that the relative contribution of the three categories: avoid, shift, and improve. well-being outcome follows a similar distribution as real-world values The derivation of these abatement shares is given in Appendix C. of global life-satisfaction. More detail on the parameter choices are Finally, the needs-based approach of this study is compared with given in Appendix B. traditional formulations of social welfare. As shown in Appendix D, The numerical experiments consist of repeated simulations of this social welfare is described in terms of aggregation of individual utilities setting for different policy parameters. The central parameter of in- that are derived from consumption. The following variations are con- terest is the carbon price 𝜏, with a value 𝜏 = 0 representing the no sidered. First, the measurement of welfare is based on either prices or policy or business as usual scenario. The impacts of carbon pricing are on fixed weights. And second, two different ways to aggregate welfare analyzed for each of the three revenue recycling methods described gains from different types of consumption are considered: in Section 3.4, namely neutral revenue recycling, progressive revenue recycling, and a mix of progressive revenue recycling and expenditure 1. Linear aggregation of welfare gains over all types of consump- for low-carbon infrastructure improvements (Maestre-Andrés et al., tion (perfect substitution). This can be seen as the implicit 2021). For the latter, two different cases are considered in which the assumption in models that assume only one aggregate consump- improvements have either a small or large effect on time-use. tion good, such as in D’Orazio and Valente (2019). The basic setting of this experiment represents the impacts of a 2. CES aggregation with limited substitution between different carbon tax. Deviations from this setting are considered to explore the types of consumption, such as in Budolfson et al. (2021). difference in performance if carbon pricing is realized through permit trading or direct regulation. Two effects are identified in Foramitti 5. Results et al. (2021a,b) as potential side-effects of quantity-based approaches. Instead of reproducing these in the model, we simplify the current 5.1. General results setting and reintroduce them exogenously. The two factors are as follows: We first look at the impact of carbon pricing for different price levels and recycling schemes. Fig. 3 presents the results. Let us first consider 1. The windfall profit rate 𝜒1 describes the case where a part of the difference between neutral and progressive recycling. The first plot the resulting carbon price does not become revenue but instead in Fig. 3 presents the outcome on well-being under different levels additional profit. of achieved emission reduction. Under a neutral recycling scheme, 2. The abatement costs error 𝜒2 (Appendix B) describes increased increased abatement leads to lower levels of well-being. This is because abatement costs due to inconsistent incentives that lead to an mitigation reduces the availability of high-carbon goods while the low- inefficient selection of the cheapest abatement options within carbon alternative competes with non-market activities for time. This each sector. This has been identified as a potential side-effect loss in well-being becomes steeper for higher levels of abatement as the of both permit trading and direct regulation.7 loss of consumption becomes more essential due to the deprivation of needs. The results of these experiments are analyzed along three different Results show that the decrease of well-being that is caused by evaluation criteria to study impacts on well-being, effectiveness, and mitigation can be counteracted through progressive revenue recycling equity: (note that with progressive revenue recycling QOL rises by about 25%). This recycling increases the purchasing power of low-income individ- 1. The quality of life that can be achieved at a given level of uals who are able to substantially improve their quality of life with emission reduction. small amounts of additional income. The negative effect of this change on high-income individuals is small in comparison as their needs are 7 already satiated to a high degree. Average well-being is thus higher This is because the prior literature has shown that when firms abate emissions through innovation, permit trading can lead to a fall in prices under progressive than under neutral recycling. that can drive green firms out of the market and make their low-emission The second plot in Fig. 3 demonstrates the effectiveness of car- technology into stranded assets (Foramitti et al., 2021a). Therefore, some firms bon pricing: it shows that amount of emission reduction that can be may abate too much and then go bankrupt while other firms abate too little achieved (and thus damage to the environment that can be avoided) because of this uncertainty. under a given carbon price level. The lower this curve, the lower the 5 J. Foramitti et al. Ecological Economics 217 (2024) 108070 Fig. 3. Comparison of different carbon tax levels and recycling schemes. The first row shows well-being outcomes, effectiveness, and distributional impacts. The second row shows the shares of abatament through the factors avoid, shift, and improve. carbon price that is required to achieve a same amount of abatement. In Finally, we examine how these results change under the side-effects other words, carbon pricing is seen as more effective if it can achieve outlined in Section 4 that are meant to represent potential side-effects the same level of abatement with a lower carbon price level. Results of realizing a carbon price, at least partly, through quantity-based show that the marginal effect of carbon pricing is reduced for higher approaches like a permit market or direct regulation. The progressive price levels, as cheaper opportunities for abatement become exhausted. recycling scheme applies to all of the following results. The recycling scheme is shown to have almost no impact with respect The first side-effect regards the case of windfall profits. Results are to this measure of effectiveness. presented in Fig. 4. The main effect of windfall profits is that they The third plot in Fig. 3 presents the Gini coefficient of the income increase income inequality, and thus reduce well-being. As already distribution. The higher this coefficient, the higher income inequality. discussed, higher inequality causes a shift towards low-carbon alterna- By design, the neutral recycling scheme does not affect this measure. tives to already happen under lower levels of abatement. Notably, this The progressive recycling scheme, in contrast, leads to a reduction in change in income distribution has almost no impact on the carbon price income inequality. This is counteracted by the fact that the revenue that is needed to achieve a given level of abatement. Outcomes under from carbon pricing is based on the amount of emissions and thus windfall profits are thus similar to outcomes under a less progressive declines when emissions are reduced. The reduction of inequality is recycling scheme. thus slower than the increase in carbon price that can be seen in the The second side-effect regards increased abatement costs. Results second plot. are presented in Fig. 5. Higher abatement costs reduce the effectiveness The second row of Fig. 3 regards how emission reduction is of carbon pricing, which means that a higher carbon price is necessary achieved, decomposed into the three categories of avoid, shift, and to reach a given level of abatement. The increased carbon price means improve. Shift is shown to happen earlier under neutral recycling. This that there is also more revenue from carbon pricing, which reduces is because there are more individuals with a low income in this case, income inequality. However, this does not result in higher levels of who are more willing to use their time for the low-carbon alternative well-being as the quality of life is at the same time negatively affected since they benefit more from saving money than from saving time. A by the requirement of achieving emission reductions through a higher higher share of shift leads to lower shares of avoid and improve since share of ‘shift’ and ‘avoid’. the three shares of abatement have to sum up to the total amount of The sensitivity of the presented results towards changes in pa- abatement (Appendix C). rameter values for sectors’ labor intensity and emission intensity is Next, we look at the effect of using part of the revenue from presented in Appendix E. The sensitivity analysis shows that the relative carbon pricing for low-carbon infrastructure improvements while the importance of the three factors (avoid, shift, and improve) can change rest is still distributed through progressive recycling. Fig. 3 shows under different parameter values. This means that – in addition to the that such improvements increase the shift to low-carbon alternatives, mechanisms described above – the impact of carbon pricing strongly which in turn increases the effectiveness of a carbon price in reducing depends on the applied setting and particular industrial characteristics. emissions.8 Inequality is higher in this case as less revenue is available for progressive recycling. The impact of infrastructure improvements 5.2. Comparison of social welfare functions on well-being depends on the ratio between their effectiveness and their costs. If the reduction of time-use per unit of revenue used is low, Finally, we compare the well-being outcomes under progressive they lead to lower well-being, since labor is taken away from forms of recycling (Fig. 6a) with traditional formulations of social welfare production that serve human needs. In the opposite case, well-being can (Fig. 6b). The top row of Fig. 6b shows the case of perfect substitution, be increased as there is more time available for non-market activities. calculated as the sum of welfare gains over all types of consumption. The bottom row shows social welfare that is based on a CES func- tion, where there is limited substitution between different types of 8 These results are reminiscent of the study by Acemoglu et al. (2016). consumption (Appendix D). 6 J. Foramitti et al. Ecological Economics 217 (2024) 108070 Fig. 4. Policy performance under different levels of windfall profits. Fig. 5. Policy performance under different levels of abatement costs. The left column of Fig. 6b presents results when welfare is measured 6. Conclusions based on current prices. In this case, abatement leads to higher welfare under both perfect and limited substitution. This is because the price of This paper has presented an agent-based model to explore the polluting goods is increased through carbon pricing in two ways. First, demand-side performance of different climate policies, taking into ac- because the carbon price is added to the price of goods. And second, count that there are multiple dimensions to human well-being. The because the adoption of low-carbon technology can increase production model describes the interaction between various economic sectors and costs. heterogeneous individuals who are trying to increase their quality of In the right column of Fig. 6b, the welfare gains from different life. Policy performance is measured through the criteria of well-being types of consumption are calculated based on fixed weights. Under impacts, effectiveness, and inequality — as well as the way in which perfect substitution, abatement leads to a mostly linear decrease of abatement is achieved. social welfare. This is because all types of consumption matter equally. Our results show that the welfare impact of carbon pricing strongly Under limited substitution, the results appear similar to the needs-based depends on the way in which revenues are recycled. This is in line description of quality of life shown in Fig. 3. Amongst the tested social with recent literature, which argues that progressive recycling can welfare functions, the CES aggregation with fixed weights can thus be benefit well-being (Klenert and Mattauch, 2016; Budolfson et al., 2021; seen as the best proxy for the needs-based approach presented in this Zhao et al., 2022). Here, we add to the literature by demonstrating study. how progressive recycling can avoid negative impacts on well-being by 7 J. Foramitti et al. Ecological Economics 217 (2024) 108070 Fig. 6. Policy performance under different formulations of social welfare. Left panel (a) shows the needs-based approach of this study (see Section 3.5). Right panel (b) shows traditional social welfare functions (see Appendix D). preventing the deprivation of needs. Another contribution of this paper welfare is aggregated over all individuals, which makes it difficult to is assessing the role of carbon pricing not only for needs fulfilled by represent the deprivation of needs of single individuals. And third, market goods and services but also other needs (fulfilled by informal non-market factors of welfare like time-use are not taken into account. activities). No mechanism is found to support the argument of Huwe The presented research has a number of limitations. First of all, and Frick (2022) that carbon pricing’s focus on abatement costs leads model complexity results in long computation times and difficulty to to adverse effects in relation to human needs. calibrate and validate its results. The latter can be mitigated through The impact of two potential factors has been tested to represent the use of detailed empirical data sets like in Kangur et al. (2017) potential side-effects of establishing a carbon price through permit and van Praag and Ferrer-i-Carbonell (2004). It should be noted that trading or direct regulation instead of a carbon tax. The first regards our aim in this study was not to reproduce exactly any particular windfall profits. They are shown to increase inequality as they reduce country, industry or society. Instead, we focus on a stylized model the amount of revenue that can be recycled — in turn leading to lower with a small number of agents for illustrative purposes. Further re- levels of well-being. The second regards increased abatement costs. This search is needed to test whether these results will hold in various negatively impacts effectiveness, as it reduces the share of emission real-world settings. Extensions of the model could test how impacts reduction that can be achieved through technological adoption. vary for diverse parts of the economy that have different possibilities These results suggest that the best way to achieve emission reduc- to avoid, shift, and improve. Realism can further be increased through tion as well as a higher level of well-being is through either a carbon calibration of production factors and elasticity of substitution between level of infrastructure and time to empirical data. Finally, a comparison tax or a well-designed form of permit trading that is able to avoid with other welfare approaches can help to understand how different windfall profits and provides clear incentives for abatement to prevent conceptual approaches to well-being can affect policy conclusions. unnecessary abatement costs. Finally, we find that the performance Another limitation of the present model is that any quantitative of carbon pricing can be enhanced through improvements of low- approach to human well-being is a simplification of the human experi- carbon infrastructure. Moreover, we demonstrate that spending carbon ence. As Joshanloo et al. (2019) noted earlier, every existing measure tax revenues on infrastructure improvements can create valuable syn- of well-being provides an incomplete picture. Moreover, since there are ergy effects to reach a higher QOL, something that has been only multiple philosophical perspectives on well-being, these views might argued earlier (Creutzig et al., 2021) without providing a formal proof. sometimes contradict each other. While we tried to be as careful as The infrastructure investments accelerate the shift towards low-carbon possible in this regard, these problems cannot be dismissed completely. alternatives by enhancing policy effectiveness as well as generating higher levels of well-being, provided the time-saving effect of these Declaration of competing interest improvements is sufficiently high. Finally, results are compared with traditional formulations of social The authors declare that they have no known competing finan- welfare. We find that the needs-based approach presented in this study cial interests or personal relationships that could have appeared to resembles the traditional description of social welfare with a CES aggre- influence the work reported in this paper. gation and fixed weights since both demonstrate that abatement leads to lower welfare. However, there are three key differences. First, the Data availability welfare function has no upper bound, which makes it difficult to allow for a qualitative interpretation of the numerical values. Second, social No data was used for the research described in the article. 8 J. Foramitti et al. Ecological Economics 217 (2024) 108070 Acknowledgment Table 1 Parameter values. Parameter Symbol Values This study has received funding through an ERC Advanced Grant from the European Research Council (ERC) under the European Union’s Number of individuals 𝑛𝐼 100 Number of sectors 𝑛𝑆 3 Horizon 2020 research and innovation programme (grant agreement Number of rounds 𝑛𝑇 10 N◦ 741087). Number of activities 𝑛𝐴 4 Number of life domains 𝑛𝐷 3 Appendix A. Technological adoption choice Initial labor intensity 𝜇𝑠0 1, 1, 1 Initial emission intensity 𝜀0𝑠 1, 0, 1 Abatement costs 𝛽𝑠 2, 0, 2 A sector’s emission intensity 𝜀𝑠 can be decomposed into a default Utility weights 𝜂𝑠 1, 1, 1 value 𝜀0𝑠 and a difference 𝛥𝜀𝑠 that is achieved through the adoption of Domain weightsa 𝜔𝑑 0.12, 0.16, 0.73 low-carbon technology. Satiation rates (mean)a 𝑘𝑑 1.13, 1.30, 1.04 Satiation rates (std)a 1.96, 1.84, 0.57 𝜀𝑠 = 𝜀0𝑠 + 𝛥𝜀𝑠 (13) Activity impacts on CoA needs 𝛿𝑎,1 1, 1, 0, 0 Activity impacts on RoE needs 𝛿𝑎,2 0, 0, 1, 0 As explained in Section 3.6, improvements of emission intensity come Activity impacts on non-market needs 𝛿𝑎,3 0, 0, 0, 1 at the cost of labor intensity. Applying the abatement cost factor 𝛽𝑠 Activity impacts on time 𝜓𝑎0 0, −1, 0, −1 Substitution rate 𝜎 0.1 from Eq. (12), a sector’s labor intensity can be defined as follows: Desired inventory share 𝜈 0.1 𝜇𝑠 = 𝜇𝑠0 + 𝛥𝜇 = 𝜇𝑠0 + 𝛥𝜀𝑠 𝛽𝑠 (14) Markup rate 𝑚 0.1 Note: Multiple values are given for parameters with indices, indicating that this The definition of a sector’s production costs 𝑐𝑠 from Eq. (6) can then be parameter has different values for each sector (s), life domain (d), or activity (a). rewritten as follows: a Calibrated parameters. 𝑐𝑠 (𝛥𝜀𝑠 ) = 𝑤 (𝜇𝑠0 + 𝛽𝑠 𝛥𝜀𝑠 ) + 𝜏 (𝜀0𝑠 + 𝛥𝜀𝑠 ) (15) Each sector wants to choose a level of technological adoption so that progressive recycling. Two cases are considered where the effectiveness this cost is minimized, which is described by the following condition. of these improvements on low-carbon consumption (𝜆𝑎 ) is set to either 𝑑 0.005 or 0.01. The parameters 𝜒1 and 𝜒2 are further varied from 0 to 𝑐 (𝛥𝜀) = 𝜏 − (𝛽𝑠 𝑤)((𝜇𝑠0 + 𝛥𝜀𝑠 𝛽𝑠 )2 ) = 0 (16) 𝑑𝛥𝜀𝑠 𝑠 0.6. The latter changes the parameters 𝛽𝑠 so that a value of e.g. 𝜒2 = 0.1 This describes the point where the marginal benefit from a reduction increases the value of 𝛽𝑠 by 10 %. Finally, the parameters 𝜇𝑠 and 𝜀𝑠 are of emission intensity equals the marginal costs from an increase in varied by ±30 % within the sensitivity analysis in Appendix E. labor intensity. The condition can be rewritten to find the optimal abatement level 𝜙𝑠 , which is defined as the relative reduction of the Appendix C. Decomposition of abatement default emission intensity: 𝛥𝜀𝑠 𝜙𝑠 = − (17) The economy’s total abatement 𝐴 is defined as the amount of saved 𝜀0𝑠 emissions in comparison to the setting where 𝜏 = 0, which is indicated The optimal abatement level then becomes: as BAU (business as usual). The values of each variable represent the ( √ ) state of that variable in the last time-step of the simulation. 0 1 𝜇𝑠 𝑤 𝜙𝑠 = − (18) 𝜀0𝑠 𝛽𝑠 𝛽𝑠 𝜏 ∑ 𝑁 ∑ 𝑁 ( 𝐵𝐴𝑈 ) 𝐴=− 𝛥𝑒𝑠 = 𝑒𝑠 − 𝑒𝑠 (19) 𝑠=1 𝑠=1 Appendix B. Parameter values The achieved amount of emission reduction are decomposed into three Table 1 presents the parameter values of the model. Values are different abatement shares, which can be related to the avoid-shift- chosen in order for the illustrative setting to display visible effects in improve framework as follows: regard to all three types of mitigation (shift, avoid, and improve). The 1. Avoid: A change in overall production of the sector due to a wages 𝑤𝑖 follow the world’s income distribution in the year 2016, as reduction of output. reported in World Inequality Lab (2021). These wages are normalized 2. Shift: A compositional change within the sector due to a shift so that the total income of all agents sums up to one. At the beginning 𝐼 = 0 and expect equal from high- to low-carbon firms. of a simulation, sectors start with no inventory 𝑞𝑠,0 𝐷 = 1∕𝑛𝑆 . 3. Improve: A change in emission intensity due to adoption of shares in demand 𝑞𝑠,0 low-carbon technology. The satiation rates 𝑘𝑖,𝑑 for each of the three life domains are heterogeneous amongst agents, and drawn randomly from a normal The decomposition is given by the following equation, as derived distribution that is truncated to include only positive values. The mean in Foramitti (2023): and standard deviation of these distributions, as well as the domain 𝑁 ( ∑ ) 𝑁 ( ∑ ) ∑𝑁 weights 𝜔𝑑 , are calibrated to be in a similar range as global self- ( ) 𝐴 = −𝐺 𝛥𝜌𝑠 𝜉 𝑠 − 𝛥𝐺 𝜌𝑠 𝜉 𝑠 − 𝛥𝜉𝑠 𝑔 𝑠 (20) reported life-satisfaction between 2014–2016, as reported in Helliwell 𝑠=1 𝑠=1 𝑠=1 et al. (2017). The calibration procedure is the same as in Foramitti The factor 𝑔𝑠 describes the amount of production in each sector. To (2023). Within the numerical experiments, several parameters are varied. make the goods of different sectors comparable, their amount is mea- Values are chosen in order to contribute to the visibility of the discussed sured in relation to their prices in the BAU scenario. mechanisms. The tax rate 𝜏 is varied between 0 and 0.3. The recycling 𝑔𝑠 = 𝑞𝑠𝑃 ∗ 𝑝𝐵𝐴𝑈 (21) 𝑠 scheme is varied between ‘neutral’ and ‘progressive’ – as described in Section 3.4. The infrastructure parameters 𝜍 and 𝜆𝑎 are generally set Similarly, an adjusted emission intensity has to be used: to zero, except for two scenarios where a share 𝜍 = 0.15 is used for 𝑒𝑠 low-carbon infrastructure improvements while the rest is still used for 𝜉𝑠 = (22) 𝑔𝑠 9 J. Foramitti et al. Ecological Economics 217 (2024) 108070 Fig. 7. Policy performance under different values of labor intensity. The factor 𝐺 describes the total amount of production. The factor 𝛥 refers to the difference of a variable in comparison to the BAU scenario: ∑ 𝑁 𝐺= 𝑔𝑠 (23) 𝛥𝑥 = 𝑥 − 𝑥𝐵𝐴𝑈 (25) 𝑠=1 The factor 𝜌 describes each sector’s relative share of production: And a bar refers to the average between these two values: 𝑔𝑠 𝑥 + 𝑥𝐵𝐴𝑈 𝜌𝑠 = (24) 𝑥= (26) 𝐺 2 10 J. Foramitti et al. Ecological Economics 217 (2024) 108070 Fig. 8. Policy performance under different values of emission intensity. Appendix D. Social welfare functions other, while there is limited substitution between these two sectors and the rest of the economy (𝑠 = 3). The following social welfare formulations are considered for com- ( )1 𝑈𝑡2 (𝑢𝑠,𝑡 ) = (𝑢1,𝑡 + 𝑢2,𝑡 )𝜎 + (𝑢3,𝑡 )𝜎 𝜎 (30) parison with the needs-based approach described in Section 3.5. In line with traditional economic literature, welfare is described in terms of Based on these definitions, four different formulations of social welfare utility that is derived from consumption. A common way of measuring can be described: the welfare gains from each type of consumption is in terms of its current price: 1. 𝑈𝑡1 (𝑢1𝑠,𝑡 ): Based on prices, with perfect substitution 2. 𝑈𝑡1 (𝑢2𝑠,𝑡 ): Based on weights, with perfect substitution 𝑢1𝑠,𝑡 = 𝑞𝑠,𝑡 𝑆 𝑝𝑠 (27) 3. 𝑈𝑡2 (𝑢1𝑠,𝑡 ): Based on prices, with limited substitution Alternatively, the welfare gains from each type of consumption can be 4. 𝑈𝑡2 (𝑢2𝑠,𝑡 ): Based on weights, with limited substitution described based on fixed weights 𝜂𝑠 : Appendix E. Sensitivity analysis 𝑢2𝑠,𝑡 = 𝑞𝑠,𝑡 𝑆 𝜂𝑠 (28) We consider two ways in which the welfare gains from different types This appendix examines the sensitivity of the model towards dif- of consumption are aggregated. The first is a simple sum of all utilities, ferent parameter values of labor intensity 𝜇𝑠 and emission intensity which implies perfect substitution between all types of consumption. 𝜀𝑠 . Note that these parameters are different for each sector, with the ∑ three sectors of the model being a high-carbon sector, a low-carbon 𝑈𝑡1 (𝑢𝑠,𝑡 ) = 𝑢𝑠,𝑡 = 𝑢1,𝑡 + 𝑢2,𝑡 + 𝑢3,𝑡 (29) alternative to the high-carbon sector, and the rest of the economy 𝑠∈𝑆 (Section 4). The progressive recycling scheme applies to all of the The second way is to use a constant elasticity of substitution (CES) following results. function. In line with the presented setting (see Fig. 2), we assume that Fig. 7 presents the impact of variations in labor intensity. For the the high-carbon and low-carbon sectors (𝑠 ∈ {1, 2}) can substitute each high carbon sector (Fig. 7a), a reduction in labor intensity results in 11 J. Foramitti et al. Ecological Economics 217 (2024) 108070 higher levels of technological adoption. This reduces the overall emis- D’Orazio, P., Valente, M., 2019. The role of finance in environmental innovation sion intensity and thus also the need for ‘shift’ and ‘avoid’, increasing diffusion: An evolutionary modeling approach. J. Econ. Behav. Organ. 162, both well-being and effectiveness. However, it also leads to higher 417–439. Fankhauser, S., Sehlleier, F., Stern, N., 2008. Climate change, innovation and jobs. inequality as there is less revenue from carbon pricing. Climate Policy 8 (4), 421–429. A lower labor intensity in the low-carbon sector (Fig. 7b) has a Foramitti, J., 2021. Agentpy: A package for agent-based modeling in python. J. Open smaller effect, since the choice to consume this good does not just Source Softw.. depend on prices but also on time use. A lower labor intensity is shown Foramitti, J., 2023. A framework for agent-based models of human needs and ecological to increase the share of ‘shift’ and ‘improve’, while there is less need limits. Ecol. Econom. 204 (Part A), 107651. for ‘avoid’. Once there has been a shift towards this sector, a higher Foramitti, J., Savin, I., van den Bergh, J.C., 2021a. Emission tax vs. permit trading carbon price is needed to reach the same level of abatement. There is under bounded rationality and dynamic markets. Energy Policy 148, 112009. Foramitti, J., Savin, I., van den Bergh, J.C., 2021b. Regulation at the source? Comparing thus more revenue from carbon pricing and less inequality. upstream and downstream climate policies. Technol. Forecast. Soc. Change 172, Changes in labor intensity within the rest of the economy (Fig. 7c) 121060. have similar effect as for the high-carbon sector. Again, a lower level Fragkos, P., Paroussos, L., 2018. Employment creation in EU related to renewables of labor intensity leads to increases in effectiveness, well-being, and expansion. Appl. Energy 230, 935–945. inequality. A key difference between the two cases can be found in Gough, I., 2017. Recomposing consumption: defining necessities for sustainable and the abatement shares, where changes of labor intensity in the rest equitable well-being. Phil. Trans. R. Soc. A 375 (20160379). of the economy do not affect the share of ‘shift’. Improvements from Goulder, L.H., Schein, A.R., 2013. Carbon Taxes Versus Cap and Trade: A Critical Review. Climate Change Econ. 04 (03), 1350010. technological adoption instead lead to a larger reduction in the share Helliwell, J., Layard, R., Sachs, J., 2017. World Happiness Report 2017. Sustainable of ‘avoid’. Development Solutions Network, New York. Fig. 8 presents results regarding variations of emission intensity. Huwe, V., Frick, M., 2022. Far from optimal? Exploring the normative premises and The low-carbon sector is excluded here as it causes no emissions. In politics of carbon pricing. Energy Res. Soc. Sci. 86, 102458. both of the other sectors, a higher emission intensity results in lower Jackson, T., Jager, W., Stagl, S., 2004. Beyond insatiability – needs theory, consumption effectiveness as well as a lower quality of life. While technological and sustainability. In: The Ecological Economics of Consumption. Edward Elgar adoption is initially higher in this case, this is reversed for higher levels Publishing, Cheltenham, UK. of abatement. Again, the main difference between the high-carbon Jager, W., 2017. Enhancing the realism of simulation (EROS): On implementing and developing psychological theory in social simulation. J. Artif. Soc. Soc. Simul. 20 sector (Fig. 8a) and the rest of the economy (Fig. 8b) is that the latter (3). does not affect the share of ‘shift’ and thus has a bigger impact on the Joshanloo, M., Jovanović, V., Taylor, T., 2019. A multidimensional understanding of share of ‘avoid’. prosperity and well-being at country level: Data-driven explorations. PLoS One 14 (10), e0223221. References Kangur, A., Jager, W., Verbrugge, R., Bockarjova, M., 2017. An agent-based model for diffusion of electric vehicles. J. Environ. Psychol. 52, 166–182. Klenert, D., Mattauch, L., 2016. How to make a carbon tax reform progressive: The Acemoglu, D., Akcigit, U., Hanley, D., Kerr, W., 2016. Transition to clean technology. role of subsistence consumption. Econom. Lett. 138, 100–103. J. Political Econ. 124 (1), 52–104. Konc, T., Drews, S., Savin, I., van den Bergh, J., 2022. Co-dynamics of climate policy Baranzini, A., van den Bergh, J.C.J.M., Carattini, S., Howarth, R.B., Padilla, E., Roca, J., stringency and public support. Global Environ. Change 74, 102528. 2017. Carbon pricing in climate policy: seven reasons, complementary instruments, and political economy considerations. Wiley Interdiscip. Rev. Clim. Change 8 (4). Lamperti, F., Dosi, G., Napoletano, M., Roventini, A., Sapio, A., 2018. Faraway, so close: coupled climate and economic dynamics in an agent-based integrated assessment Bogdanovych, A., Trescak, T., 2016. Generating needs, goals and plans for virtual model. Ecol. Econom. 150, 315–339. agents in social simulations. In: Traum, D., Swartout, W., Khooshabeh, P., Kopp, S., Scherer, S., Leuski, A. (Eds.), Intelligent Virtual Agents. In: Lecture Notes in Larch, M., Wanner, J., 2017. Carbon tariffs: An analysis of the trade, welfare, and Computer Science, Springer, Cham. emission effects. J. Int. Econ. 109, 195–213. Botzen, W., van den Bergh, J.C., 2014. Specifications of social welfare in economic Maestre-Andrés, S., Drews, S., Savin, I., van den Bergh, J., 2021. Carbon tax accept- studies of climate policy: overview of criteria and related policy insights. Environ. ability with information provision and mixed revenue uses. Nature Commun. 12 Resour. Econ. 58 (1), 1–33. (1), 1–10. Boyce, J.K., 2018. Carbon pricing: Effectiveness and equity. Ecol. Econom. 150, 52–61. Maslow, A.H., 1943. A theory of human motivation.. Psychol. Rev. 50 (4), 370. Budolfson, M., Dennig, F., Errickson, F., Feindt, S., Ferranna, M., Fleurbaey, M., Newell, R.G., Stavins, R.N., 2003. Cost heterogeneity and the potential savings from Klenert, D., Kornek, U., Kuruc, K., Méjean, A., et al., 2021. Climate action with market-based policies. J. Regulat. Econ. 23 (1), 43–59. revenue recycling has benefits for poverty, inequality and well-being. Nature Clim. O’Neill, D.W., Fanning, A.L., Lamb, W.F., Steinberger, J.K., 2018. A good life for all Change 11, 1111–1116. within planetary boundaries. Nat. Sustain. 1 (2), 88–95. Bulavskaya, T., Reynes, F., 2018. Job creation and economic impact of renewable Penasco, C., Anadon, L.D., Verdolini, E., 2021. Systematic review of the outcomes and energy in the netherlands. Renew. Energy 119, 528–538. trade-offs of ten types of decarbonization policy instruments. Nature Clim. Change Caiani, A., Godin, A., Caverzasi, E., Gallegati, M., Kinsella, S., Stiglitz, J.E., 2016. 11 (3), 257–265. Agent based-stock flow consistent macroeconomics: Towards a benchmark model. Pigou, A.C., 1920. The Economics of Welfare. Macmillan and Co, London. J. Econom. Dynam. Control 69, 375–408. Rao, N.D., Wilson, C., 2021. Advancing energy and well-being research. Nat. Sustain.. Castro, J., Drews, S., Exadaktylos, F., Foramitti, J., Klein, F., Konc, T., Savin, I., van Rengs, B., Scholz-Wäckerle, J., 2020. Evolutionary macroeconomic assessment of den Bergh, J., 2020. A review of agent-based modeling of climate-energy policy. employment and innovation impacts of climate policy packages. J. Econ. Behav. Wiley Interdiscip. Rev. Clim. Change e647 (May 2019), 1–26. Organ. 169, 332–368. Costanza, R., Fisher, B., Ali, S., Beer, C., Bond, L., Boumans, R., Danigelis, N.L., Dickin- Roberts, J.T., Steinberger, J.K., Dietz, T., Lamb, W.F., York, R., Jorgenson, A.K., son, J., Elliott, C., Farley, J., Gayer, D.E., Glenn, L.M.D., Hudspeth, T., Mahoney, D., Givens, J.E., Baer, P., Schor, J.B., 2020. Four agendas for research and policy McCahill, L., McIntosh, B., Reed, B., Rizvi, S.A.T., Rizzo, D.M., Simpatico, T., on emissions mitigation and well-being. Global Sustain. 3 (e3), 1–7. Snapp, R., 2007. Quality of life: An approach integrating opportunities, human Savin, I., 2021. On optimal regimes of knowledge exchange: a model of recombinant needs, and subjective well-being. Ecol. Econom. 61 (2–3), 267–276. growth and firm networks. J. Econ. Interact. Coord. 16, 497–527. Cramton, P., MacKay, D., Ockenfels, A., Stoft, S., 2017. Global Carbon Pricing: The Savin, I., Creutzig, F., Filatova, T., Foramitti, J., Konc, T., Niamir, L., Safarzynska, K., Path To Climate Cooperation. The MIT Press. van den Bergh, J., 2022. Agent-based modeling to integrate elements from different Creutzig, F., Niamir, L., Bai, X., Cullen, J., Díaz-José, J., Figueroa, M., Grübler, A., disciplines for ambitious climate policy. WIREs Climate Change 14 (2), e811. Lamb, W., Leip, A., Masanet, E., et al., 2021. Demand-side solutions to climate Savin, I., van den Bergh, J., 2022. Tired of climate targets: Shift focus of ’IPCC scenarios’ change mitigation consistent with high levels of wellbeing. Nature Clim. Change. from targets to policies. Ann. New York Acad. Sci. 1517 (1), 5–10. Cullenward, D., Victor, D., 2020. Making Climate Policy Work. Polity Press, Cambridge. Sirgy, J., 2021. The Psychology of Quality of Life - Wellbeing and Positive Mental Dawid, H., Delli Gatti, D., 2018. Agent-based macroeconomics. In: Hommes, C., Health, third ed. Springer Nature Switzerland. LeBaron, B. (Eds.), Handbook of Computational Economics, Vol. 4. Elsevier, pp. Terranova, R., Turco, E.M., 2022. Concentration, stagnation and inequality: An 63–156. agent-based approach. J. Econ. Behav. Organ. 193, 569–595. 12 J. Foramitti et al. Ecological Economics 217 (2024) 108070 van den Bergh, J., Castro, J., Drews, S., Exadaktylos, F., Foramitti, J., Klein, F., van Praag, B.M.S., Ferrer-i-Carbonell, A., 2004. Happiness Quantified: A Satisfaction Konc, T., Savin, I., 2021. Designing an effective climate-policy mix: Accounting Calculus Approach. Oxford University Press. for instrument synergy. Climate Policy 21 (6), 745–764. World Inequality Lab, 2021. World inequality database. https://wid.world/. (Accessed: van den Bergh, J., Savin, I., 2021. Impact of carbon pricing on low-carbon innovation 3 June 2021). and deep decarbonisation: controversies and path forward. Environ. Resour. Econ. Zhao, Y., Wang, C., Cai, W., 2022. Carbon pricing policy, revenue recycling schemes, 80 (4), 705–715. and income inequality: A multi-regional dynamic CGE assessment for China. Resour. Conserv. Recy. 181, 106246. 13