Policy Research Working Paper 10125 Will the Developing World’s Growing Middle Class Support Low-Carbon Policies? Matthew E. Kahn Somik Lall Equitable Growth, Finance and Institutions Practice Group July 2022 Policy Research Working Paper 10125 Abstract As billions of people in the developing world seek to carbon pricing hinges on offsetting its perceived negative increase their living standards, their aspirations pose a income effects. Rising environmentalism in the developing challenge to global efforts to cut greenhouse gas emissions. world could also increase support for credible greenhouse The emerging middle class is buying and operating energy gas reduction policy. This paper quantifies these effects by intensive durables ranging from vehicles to air condition- estimating Engel curves of durables ownership, compar- ers to computers. Owners of these durables represent an ing the grid’s carbon intensity by nation, and studying the interest group with a stake in opposing carbon pricing. The demographic correlates of support for prioritizing environ- political economy of encouraging middle class support for mental protection. This paper is a product of the Equitable Growth, Finance and Institutions Practice Group. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://www.worldbank.org/prwp. The authors may be contacted at kahnme@usc.edu and slall1@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Will the Developing World’s Growing Middle Class Support Low-Carbon Policies? Matthew E. Kahn University of Southern California Somik Lall World Bank Keywords: climate change, international co-operation, energy consumption, environmentalism JEL Classification: F53, H87, Q54 Acknowledgments: The authors thank Ambar Narain and Minh Nguyen for sharing data on durable assets, and Milan Brahmbhatt, Tim Besley, Alan Fuchs, Robert Huang, and Ruth Hill for helpful discussions. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Summary for Policy Makers Rising middle-class incomes in low- and middle-income countries (LMICs) will lead to a surge in demand for consumer durables, leading to rapid increases in the demand for electricity and energy. Consumer durables, electricity, and energy are important for human welfare and a successful development strategy would include growing abundance of all such durables. Considering the high carbon intensity of existing energy systems in these countries, rising energy demand is likely to be accompanied by large increases in CO2 emissions, creating a dilemma for policy makers. Economists would recommend carbon pricing as an effective response, but this can lead to large increases in energy prices for an extended period. Careful design of carbon taxes and recycling of tax revenues back to the public could alleviate losses by energy consumers. Nevertheless, governments in LMICs remain wary of large-scale carbon pricing schemes, given the potential for political opposition from the growing and politically influential middle class. One possible way out of the dilemma is evidence that as incomes and education increase, people also become more supportive of environmental protection, and more willing to trade private consumption for a cleaner environment. However, even in high-income countries, there appears to be limited support for carbon pricing schemes. In environmentally conscious Europe, there are many subnational regions featuring high carbon average footprints, where the middle class elected officials are likely to oppose carbon pricing, pockets of “green resistance”. While carbon pricing has strong economic credentials, there appear to be significant political- economy obstacles to its adoption at the speed and scale needed to tackle the problem of climate risk. This suggests the need for policy makers to devote attention to complementary climate policies, including ones that work with rather than against the public's ingrained preference for abundant and cheap energy. International transfers and technology adoption that greens the local power grid will need to complement pricing instruments. The middle class in LMICs will support low carbon policies if they are compatible with their growing demand for consumer durables. 2 I. Introduction In November 2021, representatives from nations from all over the world attended the COP 26 meetings in Scotland where negotiators sought a pathway forward to achieve a global carbon budget to limit global warming to 1.5 degrees Celsius. Analysts who have studied the necessary carbon emissions reductions for achieving this goal posit that only 11% of the carbon budget is still available (Dasgupta, Lall, and Wheeler 2022). As observed at the COP 26 meetings, global cooperation to cap greenhouse gas (GHG) emissions faced pushback from many developing countries. A type of bargaining game is now playing out. Many academic economists have argued that there is a global Hicksian Pareto improvement possible if the developing world is compensated for not engaging in fossil fuel intensive economic development (see Kleinnijenhuis, Adrian and Bolton 2022). Yet, efforts to reduce GHG emissions require incurring upfront costs in return for uncertain future benefits of less climate change risk. The emerging middle class in the developing nations could bear much of these costs as they are now buying and operating energy intensive durables ranging from vehicles to air conditioners to computers. Owners of these durables represent an interest group with a stake in opposing carbon pricing. In the absence of targeted transfers, such a Pigouvian policy would introduce a negative income effect. The French Yellow Vest protests of 2018 may offer a preview of the future in the developing world if the emerging middle class views the low carbon policy agenda to be elitist and to lower their standard of living by raising the prices of energy and limiting their personal freedom. In this paper, we use several data sets to explore the consumption patterns, carbon emissions and environmental attitudes of people with different incomes and educational attainment levels. First, we use data from the World Bank’s Living Standards Measurement Surveys to estimate durables Engel curves. We provide new evidence on the extensive margin of demand for a variety of carbon intensive durables – cars, motorcycles, air conditioners, refrigerators, computers, televisions, washing machines and cell phones. If a nation’s electricity grid could be cheaply decarbonized then such consumers would not bear the incidence of a carbon tax. Owners of internal combustion engine motorcycles and vehicles would face the carbon tax burden as their asset’s resale value would decline and its operating cost would increase. 3 Research that studies the climate change externality challenge posed by income growth posits that no single durables owner or buyer has an incentive to internalize the social costs caused by fossil fuel energy consumption. As documented by Davis and Gertler (2015) and Gertler et al. (2018), there is a rising demand for durable goods across the developing world. For the case of air conditioning, Davis and Gertler (2015) estimate ownership increases by 2.7 percentage points per $1,000 of annual household income. They use their estimates combined with income growth projections, data on utilization, and an estimate of the grid’s future carbon intensity (i.e., tons of GHG emissions per MWh of power generation) to predict the impact of the world’s growing middle class on future GHG emissions. Biardeau et al. (2020) document how cooling degree days in developing nations with extremely hot summers translates into increased electricity consumption as air conditioning ownership rates increase. Despite dramatic improvements in energy efficiency over time, economic growth in LDCs is not less energy- intensive than past growth in developed countries (van Benthem 2015). We build on this past research by presenting new Engel curves estimates and we focus on what these Engel curves mean in terms of stakeholders in the status quo carbon intensive economy. We augment these data with national data on the emissions intensity of the grid to discuss how the social costs of operating these various durables differs across nations. In developed nations, we present new evidence on the political economy challenge posed by durable internal combustion engine vehicles and by high-carbon cities within nations. Using data on California vehicle registrations, we document the slow progress in increasing the share of electric vehicle ownership. For each European nation, we document the variation in the cross- city carbon emissions per-person. The results from these data sets document that even in rich nations there are many people whose carbon footprint is large and thus they bear more of the economic incidence of carbon pricing (Glaeser and Kahn 2010). The growth of the middle class in developing nations is fueled by urbanization and rising educational attainment. These secular changes raise the possibility of increased support for environmental protection over time. Richer people are willing to pay more than poorer people to avoid risk and climate change raises our risk exposure along many margins (Costa and Kahn 2004). U.S based research documents that the more educated are more patient and more likely to support environmental protection (Becker and Mulligan 1998, Kahn 2002). We explore the role of both income and individual education in determining one’s stated support for environmental 4 protection. Using data from the World Values Surveys, we find that more educated people state their support for prioritizing environmental protection over economic growth. In the case of local pollutants such as PM2.5 or water pollutants, the reduced form cross- national carbon dioxide Environmental Kuznets Curve shifts down and inward over time (Dasgputa et al. 2002). One explanation for this empirical finding is that as nations grow richer, they internalize the social costs caused by local pollutants. This fact raises the possibility that as developing nations grow richer their own people may support a “greening” of coal fired power plants in order to enjoy local air pollution gains. Our cross-national Engel curve estimates suggest that this optimism has not held in recent years for carbon dioxide emissions. The sheer scale of consumption increases in large lower-middle-income nations such as China and India means that these nations must make significant carbon intensity progress to offset the scale effects associated with increased consumption (Acemoglu et al. 2019). In richer nations, greenhouse gas emissions continue to rise but at a decreasing rate. The rest of this paper is organized as follows. Section II sets up the trade-offs individual voters face in determining their support for low carbon policies. Section III describes the empirical work that estimates Engel curves, the adoption of electric vehicles in the United States, within-country variation in carbon emissions in Europe, and the correlates of support for environmental policies. Section IV concludes. II. Understanding Voter Preferences over Introducing Carbon Taxes A voter’s support for carbon pricing will be based on comparing the costs of a short run increase in energy prices versus the ongoing benefits of facing less climate change risk and gains from improved local air pollution (Weitzman 2014, 2017). In the discussion below, we focus on cost heterogeneity. Those who are terrified by future fat tail risks associated with climate change will place a great benefit on enacting a carbon tax because of its insurance benefits (Wagner and Weitzman 2016). 5 If people have a smaller carbon footprint in their consumption and if their job is not threatened by carbon pricing then they are more likely to support low carbon policies (Cragg et al. 2013). Even in progressive Washington State, the introduction of a carbon tax has failed to be enacted (Anderson, Marinescu and Shor 2019). Research investigating the political economy issues that arise in implementing carbon mitigation incentives has branched out into several subfields. Public finance economists have shown great creativity in exploring strategies for addressing distributional concerns. Metcalf (2007) proposes lowering the labor tax in return for raising the carbon tax. Public finance research has used computable general equilibrium models to simulate how different consumers across the income distribution would be affected by a carbon tax. Recent research has documented that European nation economic growth has not been slowed by carbon pricing (Metcalf and Stock 2020). There is an emerging literature that documents that voters respond negatively to salient price increases (Douenne and Fabre 2022). Sallee (2021) emphasizes the lack of precision in targeting and compensating losers as a rational choice explanation for why Pigouvian policy reforms are difficult to implement. The details of recycling the revenue will matter here. If the revenue is given back per households such as in Canada, then the Coasian transfer will be smaller than if the government keeps the revenue to use for other purposes. A key issue here pertains to trust in the national government. In the developing world, will the national government recycle the tax revenue back to the people? If the middle class doubts this, then such time consistency concerns will reduce voter support for Hicksian Pareto improving policies (Acemoglu 2003). Black and Heine (2019) highlight the challenges with program design. In October 2019, Ecuador tried to abruptly remove fuel subsidies to consolidate the government budget. Sufficient compensation was not offered, leading to protests by transportation unions, students and indigenous people, leading to reversal of the policy after two weeks. Recent research has documented that misperceptions also play a role in explaining the carbon policy divide. Douenne and Fabre (2020) use a survey approach to study the support for a carbon tax and dividend policy based on a sample of French people. These authors find that the survey respondents reject a tax and dividend policy where the revenue is equally redistributed to 6 adults. They claim that people incorrectly believe that the policy is regressive and do not believe that the policy will achieve its stated environmental goals. A recurring theme from recent research on media economics is that social media and the news have a causal role in influencing people’s perceptions of the fairness of a new policy (DellaVigna and La Ferrara 2015, Gentzkow and Shapiro 2008). Such misperceptions are more likely to play out in a setting where people do not have previous experience with the introduction of carbon taxes. Stavins (2022) echoes this point as he emphasizes the importance of building political acceptance for pricing instruments through influencing public perceptions. The recent economics literature demonstrates why voters in developing nations are likely to be skeptical about the short run benefits to them of supporting carbon pricing. Higher fossil fuel prices reduce their short run material well-being and represent a loss of their implicit property right to access cheap fossil fuel fired energy sources. The reaction to the summer 2022 spikes in global gas prices highlight the adjustment challenge. Of course, this gas price spike was a surprise. Voters in the developing world are even less likely to support the carbon policies than voters in richer nations because the former will bear a larger marginal cost, as a fraction of their total income, and they may value the benefits of reduced climate risk less as their value of a statistical life is likely to be lower (Costa and Kahn 2004). Transfers from the North to South could compensate here. The transfer that people will require is a decreasing function of their concern about climate change and increasing function of whether the carbon tax would raise their unemployment risk or lower their consumption opportunities by raising the cost of purchasing and operating durables. Lower income people are the most likely to be at the margin. If such individuals view the carbon tax to be elitist and threatening their material ambitions then they will be even more likely to oppose such a tax. They are likely to oppose these policies unless credible “cap and dividend” provisions are built into the policies. III. Empirics Our empirical work first reports estimates of durable goods Engel curves as we study the relationship between durables ownership shares and per-capita income. We then use country specific data to estimate the carbon dioxide implications of durables ownership and use. We 7 report cross-national carbon dioxide Engel curves. We next present data from California and Europe to document that these wealthy areas feature significant numbers of consumers creating a large average carbon footprint. In our last piece of empirical work, we turn to micro survey data to estimate the association between education and support for environmental protection. Together, these various pieces of evidence allow us to provide a consistent explanation for the slow progress in negotiating sharp greenhouse gas emissions reductions. Estimating Engel Curves Table 1 lists the nations for which the World Bank has assembled data on household durables ownership. The table also reports the first year the survey data are available. Data on asset ownership come from the World Bank’s Global Monitoring Database (GMD). The original source of the data are country-level household survey data that measure household income or consumption in each country. World Bank teams work closely with National Statistical Offices (NSOs) to ensure that household survey data are of good quality and that technical calculations are robust and aligned with international best practices. However, data on durables ownership are not collected in all national household surveys. Further, the World Bank does not have access to micro data from China’s National Bureau of Statistics; thus, China does not appear in the data set. For each of these nation/year pairs we have five data points for each durable good. These five data points correspond to the quintiles of the income distribution in that nation/year. The World Bank data also provides a measure for each nation/year/income quintile of either the average daily personal consumption or average daily income measured in PPP $ 2011. For 66% of the sample, we observe the income and for 33% we observe the consumption measure. This variable is the key explanatory variable in our regressions below. In results available on request, we have run our Engel curve estimates separately (so using log(consumption) for 33% of the sample and log(income) for the other 66% of the sample). The estimated income effects are quite similar so we pool our results below but the separate regressions are available on request. Our econometric specification is presented in equation (1). The dependent variable is the share of households in income quintile i in nation j at time t that owns a specific durable. 8 ℎ = + 1 ∗ log ( ) + (1) We include nation/year fixed effects in each regression and the regressions are population weighted. The data source for population is the World Development Indicators. In estimating equation (1), it is important to note that the within nation/year variation in durables ownership and the income at the specific quintiles allows us to estimate 1 . The nation/year fixed effects control for nation specific durables prices. In Tables 2 and 3 we report our estimates and the lowest income category is the omitted category. Consider column (1) where we report the automobile ownership regression results. A doubling of personal real income increases the probability of owning a vehicle by .139*log(2) or 9.6 percentage points. This functional form embodies diminishing returns to scale and highlights that the growth in durables ownership is most affected by low-income nations growing richer. We find similar magnitude income effects for air conditioners and refrigerators, washing machines, cell phones, computers, and televisions. The Environmental Implications of the Growth of Middle Class Consumption Rising electricity consumption does not result in more greenhouse gas emissions if a nation’s electricity grid is quite green. To measure each nation’s grid we use data from 2014 from the World Development Indicators database. We take U.S. emissions factors for coal fired power plants, natural gas and oil fired power plants and weight these by nation specific shares to yield the national emissions factor. In Table 4, the electricity grid emissions factor is measured in 1,000 pounds of carbon dioxide per MWh. The formula for this variable is based on 2.21*share of power from coal + 0.91*share of power from natural gas + 2.13 * share of power from oil. While these emissions factors are based on U.S. data, there is little reason to believe that they sharply vary across the world. We use the carbon emissions factors for coal, natural gas and oil to create a single index of a given nation’s grid carbon intensity. Table 4 shows that the largest nations in the world often feature a dirty grid and this means that the rising durables 9 consumption will translate into higher carbon emissions than would take place if the grid was cleaner. In Figure 1, we take the World Development Indicators database for the year 2018 and graph the log of carbon dioxide emissions per-capita with respect to the log of GDP per-capita. For the 181 nations, the income elasticity estimate is 1.26. The figure’s dots are weighted by national population and India and China stand out for their contribution to emissions. In this cross-sectional figure, higher income nations have a less positively sloped elasticity than middle-income nations. In Figure 2 for 174 nations, we graph the percentage change in per- capita carbon dioxide emissions from the year 2000 to 2018 with respect to the log of national per-capita GDP in 2000. The slope is -.34 in this case. The poorer nations that are urbanizing and industrializing have a larger growth in carbon emissions than richer nations. Together, these two figures highlight the challenge that lower-middle-income nations pose for creating a low carbon emissions coalition. As these nations grow richer, their carbon emissions increase. One possible offsetting force is the recognition of the co-benefits of reducing reliance on coal fired power. Barrows, Garg and Jha (2019) quantify the air pollution externality associated with coal burning in India. Cesur, Tekin and Ulker (2017, 2018) use data from Turkey to document the local air pollution gains as the nation increases its reliance on natural gas for generating power and closed coal fired power plants. In developing nations, the value of a statistical life increases as economic development takes place (Costa and Kahn 2004). This means that the Pigouvian damage to India from air pollution created by its own coal fired power plants increases as the nation grows richer and more people live in a vicinity of the power plants. At least up until this point, this local benefit of substituting away from coal has not been a sufficient incentive. Electric Vehicles Adoption in California In the developing world, people are purchasing durable products that increases their carbon footprint. In the United States, a promising trend is that more and more people are purchasing electric vehicles with wealthier people purchasing vehicles such as the Tesla. If the 10 power grid is green, then this purchase decision can decouple consumption gains from greenhouse gas production. The state of California provides unique zip code level data on vehicle registrations. 1 The data are available from October 2018 and January 2020. In each of these two cross-sections, the data report each zip code’s count of vehicles by fuel type and by model year. In total, the data set includes over 30 million vehicles. We have merged year 2018 Internal Revenue Service zip code data on the household income of tax filers. 2 For roughly 1,900 zip codes in California, we merge in data to identify the 25% highest income zip codes in the state. The zip codes are sorted by the percentage of tax filers with incomes over $200,000. In Table 5, we report the empirical distribution of vehicle fuel types in 2018 and 2020 and we report these tabulations for rich and non-rich zip codes. The first key finding is that even in rich, educated and progressive California, the percentage of electric vehicles in the fleet is very low. The second fact is that the share of electric vehicles is growing over time. At each point in time, richer zip codes have a much larger share of electric vehicles. The California vehicle registration data also reports the model year distribution of the vehicles by zip code. In the bottom of Table 5, we report the empirical distribution of vehicles by three categories; vehicles built before 2010; vehicles built between 2010 and 2015 and “new” vehicles. The key point that emerges here is the durability of the capital stock. The vehicle stock is old and predominantly gasoline based. Given that cars can live on for decades, even California stakeholders have an incentive to oppose higher gas taxes. In this section, we have presented new evidence documenting the growth of EVs but the very small market share of this product even for rich people. We have also emphasized the durability of these products. The long lived nature of the capital stock slows down the adjustment process and creates vested interest groups with a stake in maintaining the status quo. Within-Nation Carbon Geography: Evidence from Europe 1 https://data.ca.gov/dataset/vehicle-fuel-type-count-by-zip-code 2 https://www.irs.gov/statistics/soi-tax-stats-individual-income-tax-statistics-zip-code-data-soi 11 We have documented that in the developing world millions of people are purchasing energy intensive durables. In the previous section, we documented that even California features a large durable set of fossil fueled vehicles. In this section, we present additional evidence from Europe documenting that within each of the continent’s nations there are geographic clusters of higher carbon footprint places. High emissions regions have incentives to lobby their local officials to protect them. Previous research on the United States has documented how local carbon emissions influence carbon politics. Eyer and Kahn (2020) document efforts by coal states such as West Virginia to use local policies to protect coal miner jobs. Cragg et al. (2013) document that Congressional voting on enacting Cap and Trade legislation is associated with the area’s local per-capita carbon footprint. Representatives are more likely to vote in favor of carbon pricing if their district’s emissions are lower. Within nations, there can also be geographic areas that oppose carbon pricing because the local economy is carbon intensive. A region can have a large carbon footprint because of the industries concentrated there, the area’s low population density, and the types of power generation used to generate power. To explore this point, we use data from jurisdictions within European nations. The raw data are from https://openhgmap.net/data. 3 In Table 6, we report the average carbon footprint and the coefficient of variation across cities within the same nation in Europe. World carbon dioxide emissions per-capita in 2018 was 4.5 tons. 4 These results indicate that there are high emitting regions even in relatively low carbon nations such as France and Germany. Poland features a high coefficient of variation indicating that there are high carbon areas. The people and the political leaders of these regions will be less likely to support the low carbon agenda and they are likely to argue that Pigouvian policies are elitist unless there are spatial transfers to such regions. The United States has wrestled with how to compensate coal miners for their anticipated dislocation (Eyer and Kahn 2020). 3 This project maps CO2 emissions across Europe. The aim is to estimate an emissions inventory for each of the ~116 000 administrative jurisdictions across Europe and the UK. The model spatially disaggregates each country's official (Eurostat) CO2 emissions inventory to places using OpenStreetMap. Vehicle emissions are attributed across fuel stations, train emissions at stations, aviation bunker fuel emissions at airports, and so on. Industrial source emissions are located at the registered address where these emissions physically occur or are legally controlled. Data are for the year 2018. 4 https://data.worldbank.org/indicator/EN.ATM.CO2E.PC 12 The Demographic Correlates of Support for Prioritizing Environmental Protection While developing nations feature a growing middle class who value their increased consumption of durables that often operate using a carbon intensive grid, these same individuals are growing richer because they are urbanizing and obtaining more education. More educated people are more likely to prioritize the environment, understand their own health production function and be more likely to make long run trade-offs even if they incur costs today (Becker and Mulligan 1998, Costa and Kahn 2004, Kahn 2002). We turn to the World Values Survey Wave 7 and use these microdata to study self- reported environmentalism. We study how such prioritization of protecting the environment is associated with a person’s age, education, and income. We document that more educated people (who are richer) are more likely to be pro-environment. Table 7 lists the nations included in the WVS wave from 2020. We use these data to estimate a linear probability model for person i in nation j. = + ∗ + (2) Table 8 reports three estimates of equation (2). In columns (1) and (2), we report linear probability models. In column (1), the survey question focuses on whether protecting the environment is the respondent’s priority rather than economic growth. In the regression, we include nation fixed effects. We find that a respondent’s education is monotonically associated with greater support for protecting the environment. Relative to a person with very little education, a college graduate is 13 percentage points more likely to prioritize environmental protection. The mean of the dependent variable is 0.55. Younger people and rural people are more likely to prioritize environmental protection. More liberal people are also more likely to prioritize environmental protection. We fail to reject the hypothesis that self- reported personal income has no effect on support for environmental protection. From the coefficients on these survey responses, we view the education results and the age effects to be 13 the most important. A college educated 30-year-old is more than 16 percentage points more likely to prioritize the environment than a 60-year-old with a high school degree. In column (2) of Table 8, we explore how one’s propensity to vote is associated with individual level attributes. Older, more educated people are much more likely to vote than younger, less educated people. Finally, in column (3) we explore how personal ideology on a liberal to conservative scale is associated with individual attributes. Richer people are less likely to be right wing. Urbanites and more educated people are less likely to be right wing. Older people are much more likely to be right wing. IV. Conclusion Without placing an explicit price on carbon emissions, billions of people are not internalizing the social costs of their consumption choices. In richer nations, millions of people are deeply concerned about the climate change challenge and are willing to pay to slow down this global public bad. In the developing world, billions of people are seeking to achieve middle class status and prioritize their own material gains. A core political economy challenge arises as richer nations seek to convince poorer nations to adopt the textbook Pigouvian policy solution. In this paper, we have presented new empirical work that indicates that implementation is becoming more challenging over time as there are more stakeholders due to the growth of the middle class in developing nations. This emerging middle class seeks to purchase energy intensive durable assets such as vehicles, air conditioners, and computers. These assets increase standards of living, allowing more people to access jobs, amenities, and information, as well as alleviate the hardships from rising global temperatures. However, rising demand for durables will increase greenhouse gas emissions if developing nations have a carbon intensive electricity grid and transportation sector. This consumption growth causes a political economy challenge due to the carbon intensive lock-in effect that stiffens resistance to pricing emissions; the middle class in LMICs will support low carbon policies if they are compatible with their growing demand for consumer durables. Our empirical work highlights the offsetting factors that emerge in an urbanizing nation where incomes are rising. Rising incomes cause greater consumption but are associated with 14 increased educational attainment. We have used several data sets to explore how these two factors influence support for addressing a global externality. By estimating international Engel curves for energy intensive durables, we document that the growth of the urban middle class will increase electricity demand and that the grid around the world continues to be carbon intensive. Using data from California, we have documented that even in this rich, progressive state the vast majority of vehicles are gasoline fueled. Even in richer zip codes, the electric vehicle percentages are quite low. Unlike for local pollutants such as PM2.5, there is little evidence of an inverted “U” shape between carbon dioxide emissions and income. Richer people produce more greenhouse gas emissions in the absence of a global carbon treaty. Within Europe, we document that there are many cities featuring a high carbon average footprint. Such area’s middle class is likely to oppose carbon pricing. Those high carbon areas will feature elected officials who will be unlikely to embrace the green economy agenda. In this sense, our paper pinpoints pockets of “green resistance”. In the final section of the paper, we use global micro data from the WVS to explore the correlates of environmental concern. Younger, more educated people are more likely to prioritize environmental protection. Such politically involved demographic groups help to overcome the free-rider problem that no one voter feels that her own activism matters. Future research should explore how a low-carbon coalition engages in persuading the emerging middle class to join their coalition. Stavins (2022) posits that more social education on the risks of climate change can increase demand for carbon mitigation policies. The role of social media and salient events such as natural disaster shocks in stimulating short run demand for public safety investment merits more research (DellaVigna and La Ferrara 2015, Deryugina 2013). 15 References Acemoglu, Daron. "Why not a political Coase theorem? Social conflict, commitment, and politics." Journal of comparative economics 31, no. 4 (2003): 620-652. Acemoglu, Daron, Ufuk Akcigit, Douglas Hanley, and William Kerr. "Transition to clean technology." Journal of political economy 124, no. 1 (2016): 52-104. Akerlof, George A., and Rachel E. Kranton. "Economics and identity." The quarterly journal of economics 115, no. 3 (2000): 715-753. Anderson, Soren T., Ioana Marinescu, and Boris Shor. Can Pigou at the polls stop us melting the poles?. No. w26146. National Bureau of Economic Research, 2019. Barrows, Geoffrey, Teevrat Garg, and Akshaya Jha. "The health costs of coal-fired power plants in India." Available at SSRN 3281904 (2019). Becker, Gary S., and Casey B. Mulligan. "The endogenous determination of time preference." The Quarterly Journal of Economics 112, no. 3 (1997): 729-758. Besley, Timothy, and Torsten Persson. "JEEA-FBBVA lecture 2017: The dynamics of environmental politics and values." Journal of the European Economic Association 17, no. 4 (2019): 993-1024. Bhandary, Rishikesh Ram, and Kelly Sims Gallagher. "What drives Pakistan’s coal-fired power plant construction boom? Understanding the China-Pakistan Economic Corridor’s energy portfolio." World Development Perspectives 25 (2022): 100396. Biardeau, Léopold T., Lucas W. Davis, Paul Gertler, and Catherine Wolfram. "Heat exposure and global air conditioning." Nature Sustainability 3, no. 1 (2020): 25-28. Bistline, John. Household Environmental Kuznets Vurves: Evidence from Passenger Transport Emissions. EPRI Working Paper 2022. Bruckner, Benedikt, Klaus Hubacek, Yuli Shan, Honglin Zhong, and Kuishuang Feng. "Impacts of poverty alleviation on national and global carbon emissions." Nature Sustainability (2022): 1- 10. 16 Carattini, S., M. Carvalho, and S. Fankhauser. 2017. How to make carbon taxes more acceptable. Policy re-port, Grantham Research Institute on Climate Change and the Environment and Centre for ClimateChange Economics and Policy, London School of Economics and Political Science, London. Cesur, Resul, Erdal Tekin, and Aydogan Ulker. "Air pollution and infant mortality: evidence from the expansion of natural gas infrastructure." The economic journal 127, no. 600 (2017): 330-362. Cesur, Resul, Erdal Tekin, and Aydogan Ulker. "Can natural gas save lives? Evidence from the deployment of a fuel delivery system in a developing country." Journal of health economics 59 (2018): 91-108. Chakravarty, Shoibal, and E. Somanathan. "There is no economic case for new coal plants in India." World Development Perspectives 24 (2021): 100373. Chen, Yuyu, Avraham Ebenstein, Michael Greenstone, and Hongbin Li. "Evidence on the impact of sustained exposure to air pollution on life expectancy from China’s Huai River policy." Proceedings of the National Academy of Sciences 110, no. 32 (2013): 12936-12941. Costa, Dora L., and Matthew E. Kahn. "Changes in the Value of Life, 1940–1980." Journal of risk and Uncertainty 29, no. 2 (2004): 159-180. Cragg, Michael I., Yuyu Zhou, Kevin Gurney, and Matthew E. Kahn. "Carbon geography: the political economy of congressional support for legislation intended to mitigate greenhouse gas production." Economic Inquiry 51, no. 2 (2013): 1640-1650. Davis, Lucas W., Alan Fuchs, and Paul Gertler. "Cash for coolers: evaluating a large-scale appliance replacement program in Mexico." American Economic Journal: Economic Policy 6, no. 4 (2014): 207-38. Davis, Lucas W., and Paul J. Gertler. "Contribution of air conditioning adoption to future energy use under global warming." Proceedings of the National Academy of Sciences 112, no. 19 (2015): 5962-5967. Dasgupta, Susmita, Benoit Laplante, Hua Wang, and David Wheeler. "Confronting the environmental Kuznets curve." Journal of economic perspectives 16, no. 1 (2002): 147-168. DellaVigna, Stefano, and Eliana La Ferrara. "Economic and social impacts of the media." In Handbook of media economics, vol. 1, pp. 723-768. North-Holland, 2015. Deryugina, Tatyana. "How do people update? The effects of local weather fluctuations on beliefs about global warming." Climatic change 118, no. 2 (2013): 397-416. 17 Douenne, Thomas. "The vertical and horizontal distributive effects of energy taxes: A case study of a french policy." The Energy Journal 41, no. 3 (2020). Douenne, Thomas, and Adrien Fabre. "Yellow vests, pessimistic beliefs, and carbon tax aversion." American Economic Journal: Economic Policy (2020). Eyer, Jonathan, and Matthew E. Kahn. "Prolonging coal’s sunset: Local demand for local supply." Regional Science and Urban Economics 81 (2020): 103487. Gentzkow, Matthew, and Jesse M. Shapiro. "Competition and Truth in the Market for News." Journal of Economic perspectives 22, no. 2 (2008): 133-154. Gertler, Paul J., Orie Shelef, Catherine D. Wolfram, and Alan Fuchs. "The demand for energy- using assets among the world's rising middle classes." American Economic Review 106, no. 6 (2016): 1366-1401. Glaeser, Edward L., and Matthew E. Kahn. "The greenness of cities: Carbon dioxide emissions and urban development." Journal of urban economics 67, no. 3 (2010): 404-418. Goulder, Lawrence H., Marc AC Hafstead, GyuRim Kim, and Xianling Long. "Impacts of a carbon tax across US household income groups: What are the equity-efficiency trade- offs?." Journal of Public Economics 175 (2019): 44-64. Graff Zivin, Joshua, and Matthew Neidell. "Environment, health, and human capital." Journal of Economic Literature 51, no. 3 (2013): 689-730. Grossman, Michael. The relationship between health and schooling: What’s new?. No. w21609. National Bureau of Economic Research, 2015. Grossman, Gene M., and Alan B. Krueger. "Economic growth and the environment." The quarterly journal of economics 110, no. 2 (1995): 353-377. Holian, Matthew J., and Matthew E. Kahn. "Household demand for low carbon policies: Evidence from California." Journal of the Association of Environmental and Resource Economists 2, no. 2 (2015): 205-234. Jacobsen, Grant D. "The Al Gore effect: an inconvenient truth and voluntary carbon offsets." Journal of Environmental Economics and Management 61, no. 1 (2011): 67-78. 18 Kahn, Matthew E. "Demographic change and the demand for environmental regulation." Journal of Policy Analysis and Management: The Journal of the Association for Public Policy Analysis and Management 21, no. 1 (2002): 45-62. Kahn, Matthew E. "Environmental disasters as risk regulation catalysts? The role of Bhopal, Chernobyl, Exxon Valdez, Love Canal, and Three Mile Island in shaping US environmental law." Journal of Risk and Uncertainty 35, no. 1 (2007): 17-43. Kahn, Matthew E. "The death toll from natural disasters: the role of income, geography, and institutions." Review of economics and statistics 87, no. 2 (2005): 271-284. Kahn, Matthew E., and Matthew J. Kotchen. "Business cycle effects on concern about climate change: the chilling effect of recession." Climate Change Economics 2, no. 03 (2011): 257-273. Kleinnijenhuis, Alissa, Tobias Adrian, and Patrick Bolton. The Great Carbon Arbitrage. No. 2022-07. Institute for New Economic Thinking at the Oxford Martin School, University of Oxford, 2022. Knittel, Christopher R., and Shinsuke Tanaka. Driving behavior and the price of gasoline: evidence from fueling-level micro data. No. w26488. National Bureau of Economic Research, 2019. Metcalf, Glibert. An equitable tax reform to address global climate change. Discussion Paper 2007–12 (The Hamilton Project, The Brookings Institution 2007), 2007. Metcalf, Gilbert E. "On the economics of a carbon tax for the United States." Brookings Papers on Economic Activity 2019, no. 1 (2019): 405-484. Metcalf, Gilbert E., and James H. Stock. "Measuring the macroeconomic impact of carbon taxes." In AEA Papers and Proceedings, vol. 110, pp. 101-06. 2020. Moran, Daniel, Peter-Paul Pichler, Heran Zheng, Helene Muri, Jan Klenner, Diogo Kramel, Johannes Többen et al. "Estimating CO 2 Emissions for 108,000 European Cities." Earth System Science Data Discussions (2021): 1-23. Rausch, Sebastian, Gilbert E. Metcalf, and John M. Reilly. "Distributional impacts of carbon pricing: A general equilibrium approach with micro-data for households." Energy economics 33 (2011): S20-S33. Reiss, Peter C., and Matthew W. White. "Household electricity demand, revisited." The Review of Economic Studies 72, no. 3 (2005): 853-883. 19 Sallee, James M. Pigou creates losers: On the implausibility of achieving pareto improvements from efficiency-enhancing policies. No. w25831. National Bureau of Economic Research, 2019. Schmalensee, Richard, Thomas M. Stoker, and Ruth A. Judson. "World carbon dioxide emissions: 1950–2050." Review of Economics and Statistics 80, no. 1 (1998): 15-27. Severen, Christopher, and Arthur Van Benthem. Formative experiences and the price of gasoline. American economic Journal, forthcoming Sexton, Steven. "Automatic bill payment and salience effects: Evidence from electricity consumption." Review of Economics and Statistics 97, no. 2 (2015): 229-241. Stavins, Robert. The Relative Merits of Carbon Pricing Instruments: Taxes versus Trading . Review of Environmental Economics and Policy (2022); 16(1), 62-82. van Benthem, Arthur A. "Energy leapfrogging." Journal of the Association of Environmental and Resource Economists 2, no. 1 (2015): 93-132. Wackernagel, Mathis, Niels B. Schulz, Diana Deumling, Alejandro Callejas Linares, Martin Jenkins, Valerie Kapos, Chad Monfreda et al. "Tracking the ecological overshoot of the human economy." Proceedings of the national Academy of Sciences 99, no. 14 (2002): 9266-9271. Wagner, Gernot, and Martin L. Weitzman. Climate shock. Princeton University Press, 2016. Weitzman, Martin L. "Can negotiating a uniform carbon price help to internalize the global warming externality?." Journal of the Association of Environmental and Resource Economists 1, no. 1/2 (2014): 29-49. Weitzman, Martin L. "Voting on prices vs. voting on quantities in a World Climate Assembly." Research in Economics 71, no. 2 (2017): 199-211. 20 Table 1 Nations Included in the Engel Curve Regressions Nation Year Albania 2008 Angola 2008 Argentina 2003 Armenia 2005 Austria 2004 Azerbaijan 2002 Bangladesh 2000 Belarus 1998 Belgium 2004 Benin 2015 Bhutan 2003 Bolivia 2000 Bosnia and Herzegovina 2007 Botswana 2009 Brazil 2001 Bulgaria 2007 Burkina Faso 2009 Burundi 2006 Cabo Verde 2007 Cameroon 2007 Central African Republic 2008 Chad 2011 Chile 2000 Colombia 2001 Comoros 2004 Congo, Dem. Rep. 2012 Congo, Rep. 2011 Costa Rica 1989 Côte d'Ivoire 2008 Croatia 2009 Cyprus 2005 Czech Republic 2005 Denmark 2004 Djibouti 2012 Dominican Republic 2000 Ecuador 2003 21 Egypt, Arab Rep. 2010 El Salvador 2000 Estonia 2004 Eswatini 2009 Ethiopia 2010 Fiji 2002 Finland 2004 France 2004 Gabon 2005 Gambia, The 2010 Georgia 2002 Ghana 2005 Greece 2004 Guatemala 2000 Guinea 2012 Guinea-Bissau 2010 Haiti 2012 Honduras 2001 Hungary 2005 Iceland 2004 India 1993 Indonesia 2005 Iran, Islamic Rep. 2009 Iraq 2006 Ireland 2004 Italy 2004 Jordan 2006 Kazakhstan 2001 Kenya 2005 Kiribati 2006 Kosovo 2009 Kyrgyz Republic 2005 Lao PDR 2002 Latvia 2005 Lebanon 2011 Lesotho 2017 Liberia 2007 Lithuania 2005 Luxembourg 2004 Madagascar 2010 Malawi 2010 Malaysia 2016 Maldives 2002 Mali 2009 22 Malta 2007 Marshall Islands 2019 Mauritania 2008 Mauritius 2012 Mexico 1989 Micronesia, Fed. Sts. 2013 Moldova 2007 Mongolia 2010 Montenegro 2006 Morocco 2000 Mozambique 2008 Myanmar 2015 Namibia 2009 Nauru 2012 Nepal 2003 Netherlands 2005 Nicaragua 2001 Niger 2011 Nigeria 2018 North Macedonia 2004 Norway 2004 Pakistan 2001 Panama 2000 Papua New Guinea 2009 Paraguay 2001 Peru 1997 Philippines 2006 Poland 2005 Portugal 2004 Romania 2006 Russian Federation 2007 Rwanda 2010 Samoa 2008 São Tomé and Príncipe 2010 Senegal 2011 Serbia 2004 Seychelles 2006 Sierra Leone 2011 Slovak Republic 2005 Slovenia 2005 Solomon Islands 2012 Somalia 2017 South Africa 2010 South Sudan 2009 23 Spain 2004 Sri Lanka 2002 Sudan 2009 Sweden 2004 Switzerland 2007 Tajikistan 2009 Tanzania 2011 Thailand 2006 Timor-Leste 2001 Togo 2011 Tonga 2009 Tunisia 2005 Turkey 2004 Tuvalu 2010 Uganda 2012 Ukraine 2005 United Kingdom 2005 Uruguay 1992 Vanuatu 2010 Vietnam 2006 West Bank and Gaza 2004 Yemen, Rep. 2005 Zambia 2010 Zimbabwe 2017 24 Table 2 Durable Ownership Share Engel Curve Regressions (1) (2) (3) (4) Car Motorcycle Air Conditioner Refrigerator log(income) 0.139*** 0.0259 0.123*** 0.129*** (0.0213) (0.0204) (0.0203) (0.0365) Constant 0.162*** 0.318*** 0.0713*** 0.511*** (0.0188) (0.0175) (0.0147) (0.0316) Mean Y 0.3583 0.1608 0.14 0.6214 Nation/Year Fixed Effect Yes Yes Yes Yes Countries Included 84 79 70 84 Observations 775 595 540 680 R-squared 0.951 0.980 0.926 0.949 The unit of analysis is a nation/year/income quintile. The dependent variable is the share of households who own the durable. Robust standard errors in parentheses. The standard errors are clustered by country. *** p<0.01, ** p<0.05, * p<0.1 The regressions are weighted by the nation's population in the sample year. 25 Table 3 Durables Ownership Share Engel Curve Regressions (1) (2) (3) (4) Computer Television Washing Machine Cell Phone VARIABLES log(income) 0.180*** 0.133*** 0.0806*** 0.110*** (0.0120) (0.0336) (0.0265) (0.0153) Constant 0.0903*** 0.591*** 0.516*** 0.666*** (0.0104) (0.0295) (0.0276) (0.0130) Mean Y 0.294 0.7685 0.6395 0.7303 Nation/Year Fixed Effect Yes Yes Yes Yes Countries Included 116 89 74 113 Observations 3,050 780 730 2,720 R-squared 0.926 0.909 0.971 0.918 The unit of analysis is a nation/year/income quintile. The dependent variable is the share of households who own the durable. Robust standard errors in parentheses. The standard errors are clustered by country. *** p<0.01, ** p<0.05, * p<0.1 The regressions are weighted by the nation's population in the sample year. 26 Table 4 Macroeconomic Statistics Sorted by Population Size 2014 Data Nation Population pm2.5 KWH Emissions Factor GDP Per-Capita Millions lbs/KWH 2017 $ PPP China 1364.270 59.767 3927.045 1.627 11917.340 India 1295.601 89.622 804.516 1.726 5116.629 United States 318.386 8.221 12993.966 1.141 57213.270 Indonesia 255.128 16.455 811.910 1.627 9801.166 Brazil 202.764 13.990 2619.961 0.353 15749.510 Pakistan 195.305 59.518 447.505 1.029 4171.292 Nigeria 176.405 48.633 144.525 0.750 5516.386 Bangladesh 154.517 68.395 320.210 1.103 3511.646 Russian Federation 143.820 16.582 6602.658 0.805 26057.160 Japan 127.276 12.587 7819.715 1.283 39255.280 Mexico 120.355 23.100 2157.324 1.001 18887.570 Philippines 100.513 20.260 696.347 1.323 6973.639 Ethiopia 98.094 35.044 69.199 0.001 1656.635 Vietnam 91.714 34.896 1423.700 0.826 6098.539 Egypt, Arab Rep. 90.425 76.560 1683.213 1.067 10353.670 Germany 80.983 12.751 7035.483 1.123 50770.610 Iran, Islamic Rep. 77.466 38.109 3022.122 1.115 13038.550 Turkey 77.229 42.596 2847.224 1.123 24881.730 Congo, Dem. Rep. 73.767 42.549 108.517 0.002 1029.791 Thailand 68.439 29.365 2538.796 1.128 15854.130 France 66.312 12.290 6939.944 0.077 43021.390 United Kingdom 64.602 10.801 5130.390 0.956 44154.110 Italy 60.789 17.769 5002.407 0.785 39898.530 South Africa 54.544 26.722 4198.046 2.059 12884.480 Myanmar 52.281 40.289 215.299 0.374 4020.037 Korea, Rep. 50.747 27.213 10496.514 1.223 37967.480 Tanzania 49.961 28.874 103.682 0.714 2284.962 Colombia 46.968 18.344 1312.200 0.375 13852.240 Kenya 46.700 28.949 164.325 0.394 3709.153 Spain 46.481 10.075 5355.987 0.630 35968.620 Ukraine 45.272 20.487 3418.569 0.922 12408.950 Argentina 42.670 14.467 3074.702 0.782 23550.100 Algeria 38.924 35.564 1362.872 0.919 11512.710 Poland 38.012 22.211 3971.800 1.886 26649.580 Sudan 37.978 48.142 256.756 0.461 4124.510 Small states 37.861 28.868 . 0.705 20369.980 Uganda 36.912 46.931 . . 2022.295 Canada 35.437 7.428 15588.487 0.333 47564.610 Iraq 34.412 56.567 1328.231 0.957 9991.616 27 Morocco 34.192 29.949 904.442 1.672 6912.178 Afghanistan 33.371 59.010 . . 2102.385 Saudi Arabia 30.917 75.608 9401.486 0.975 48209.140 Uzbekistan 30.758 28.900 1645.442 0.773 5764.493 Peru 30.090 27.583 1345.879 0.458 11877.080 Venezuela, RB 30.043 19.104 2719.138 0.460 . Malaysia 29.867 16.923 4651.951 1.343 23906.230 Other small states 28.364 32.466 . 0.714 22566.880 Ghana 27.224 25.890 351.301 0.166 4670.245 Angola 26.942 32.974 312.229 0.997 8239.831 Nepal 26.906 98.116 146.473 0.001 3217.448 Mozambique 26.286 22.614 478.921 0.080 1217.089 Yemen, Rep. 25.823 48.339 219.800 1.660 . Korea, Dem. People's Rep. 25.058 35.240 601.689 0.603 . Madagascar 23.590 23.518 . . 1540.745 Australia 23.476 9.493 10071.399 1.594 47289.960 Cameroon 22.682 56.715 275.198 0.450 3362.665 Côte d'Ivoire 22.648 20.900 274.730 0.767 4161.940 Sri Lanka 20.778 26.923 531.091 1.316 11428.110 Romania 19.909 15.006 2584.412 0.733 23084.380 Niger 19.240 60.541 51.195 2.155 1127.614 Syrian Arab Republic 18.711 41.534 974.575 1.084 . Chile 17.759 22.868 3879.673 1.073 24173.060 Burkina Faso 17.586 33.969 . . 1907.949 Kazakhstan 17.288 13.750 5600.209 1.786 24355.760 Mali 16.934 33.916 . . 2075.401 Netherlands 16.865 12.829 6712.775 1.185 52187.000 Malawi 16.290 25.198 . . 1432.055 Ecuador 15.952 16.234 1376.394 0.920 12078.470 Zambia 15.400 28.493 717.347 0.060 3450.038 Guatemala 15.306 26.528 601.190 0.683 7939.375 Cambodia 15.275 28.583 271.367 0.850 3364.278 Senegal 14.175 36.827 229.352 1.819 2868.316 Chad 13.664 51.247 . . 1866.266 Zimbabwe 13.587 22.806 609.125 0.981 3195.768 Somalia 13.424 30.401 . . 873.494 Cuba 11.307 21.296 1450.883 1.110 . Belgium 11.209 13.234 7709.123 0.388 48747.680 Guinea 11.151 21.197 . . 2061.695 Rwanda 11.084 42.211 . . 1780.145 Tunisia 11.063 34.692 1454.643 0.893 10498.850 Greece 10.892 17.365 5062.606 1.487 28178.690 Bolivia 10.707 24.819 742.538 0.680 7730.638 South Sudan 10.555 41.901 43.582 2.121 . Haiti 10.549 16.262 39.056 1.944 2935.220 Czech Republic 10.525 17.391 6258.891 1.152 34386.700 Portugal 10.401 8.785 4662.601 0.684 30444.600 Benin 10.287 29.660 100.225 2.130 2975.855 Dominican Republic 10.165 14.188 1615.515 1.581 14499.630 28 Hungary 9.866 16.315 3965.958 0.596 26424.720 Burundi 9.844 39.388 . . 886.235 Sweden 9.696 6.959 13480.148 0.019 49259.000 Azerbaijan 9.535 20.710 2202.394 0.857 14875.780 Belarus 9.475 19.766 3679.979 0.915 19066.890 United Arab Emirates 9.214 37.983 11088.342 0.923 62378.610 Honduras 8.956 23.162 619.837 1.197 5177.415 Jordan 8.919 33.205 1864.932 2.036 10284.670 Austria 8.546 13.379 8355.842 0.277 52857.050 Tajikistan 8.253 46.481 1499.486 0.021 2884.200 Israel 8.216 22.059 6600.898 1.545 37474.850 Switzerland 8.189 11.389 7520.166 0.008 67682.690 Papua New Guinea 7.947 13.245 . . 3909.339 Hong Kong SAR, China 7.230 . 6083.270 1.905 56359.300 Bulgaria 7.224 20.744 4708.928 1.054 18747.370 Caribbean small states 7.173 19.967 3062.985 . 16095.280 Togo 7.138 28.936 154.665 0.294 1863.282 Serbia 7.131 26.492 4271.745 1.472 15226.320 Sierra Leone 7.017 17.698 . . 1997.509 Lao PDR 6.640 28.390 . . 6193.340 Paraguay 6.600 12.641 1552.385 0.000 11643.730 Libya 6.362 47.501 1811.055 1.475 12201.210 El Salvador 6.295 27.503 937.074 0.858 7990.459 Lebanon 6.261 30.445 2588.865 2.107 16970.440 Nicaragua 6.143 19.388 568.314 0.983 5443.278 Kyrgyz Republic 5.836 23.298 1941.222 0.182 4722.086 Denmark 5.643 11.067 5858.802 0.840 52048.340 Singapore 5.470 17.409 8844.688 0.906 87959.410 Turkmenistan 5.466 22.085 2678.766 0.910 12421.370 Finland 5.462 6.469 15249.989 0.354 44976.780 Slovak Republic 5.419 18.304 5137.074 0.351 27384.970 Norway 5.137 7.671 22999.935 0.020 62390.150 Costa Rica 4.795 17.544 1942.491 0.217 18669.100 Congo, Rep. 4.737 45.853 202.873 0.412 5556.705 Ireland 4.658 8.751 5672.064 0.806 58267.410 New Zealand 4.517 6.358 9012.731 0.247 40432.240 Central African Republic 4.464 49.356 . . 822.611 Liberia 4.360 14.942 . . 1621.345 Croatia 4.238 18.433 3714.383 0.478 23782.500 West Bank and Gaza 4.173 33.448 . . 5967.073 Oman 4.027 37.711 6445.581 0.942 30530.500 Mauritania 3.931 43.111 . . 5008.688 Panama 3.901 12.922 2064.178 0.947 27352.520 Georgia 3.719 22.967 2693.973 0.179 12254.650 Kuwait 3.691 55.624 15590.613 1.718 56647.440 Puerto Rico 3.535 9.089 . . 34070.250 Bosnia and Herzegovina 3.482 28.777 3446.765 1.396 12066.590 Uruguay 3.400 9.927 3085.190 0.197 22419.040 Mongolia 2.940 39.295 2006.387 2.137 10980.320 29 Lithuania 2.932 12.845 3821.145 0.521 29855.830 Armenia 2.912 33.086 1961.610 0.386 11019.840 Albania 2.889 18.884 2309.367 0.000 11586.860 Jamaica 2.875 14.074 1050.733 1.921 9438.477 Moldova 2.857 16.726 1725.617 0.857 10314.410 Qatar 2.459 78.833 14781.624 0.910 95578.260 Pacific island small states 2.324 12.005 . . 6753.211 Namibia 2.273 26.018 1652.572 0.018 10413.310 Botswana 2.089 24.151 1815.554 2.206 17264.400 North Macedonia 2.078 31.791 3496.520 1.629 14524.190 Slovenia 2.062 16.633 6727.999 0.509 33098.950 Lesotho 2.043 33.080 . . 2639.446 Gambia, The 2.024 30.622 . . 2038.802 Latvia 1.994 14.879 3507.405 0.414 25387.010 Gabon 1.884 41.320 1167.852 0.696 15437.260 Kosovo 1.813 . 2818.337 2.148 9214.406 Guinea-Bissau 1.692 26.941 . . 1740.897 Trinidad and Tobago 1.362 24.856 7092.959 0.913 28796.560 Bahrain 1.336 62.759 19596.983 0.910 48201.160 Estonia 1.315 7.706 6732.368 0.104 30602.600 Mauritius 1.261 15.293 2182.509 1.732 19240.220 Timor-Leste 1.174 20.450 . . 3264.563 Cyprus 1.152 18.294 3624.934 1.975 33207.770 Equatorial Guinea 1.122 44.302 . . 32436.550 Eswatini 1.095 17.979 . . 8192.528 Djibouti 0.899 41.260 . . 4238.944 Fiji 0.866 11.923 . . 12057.990 Guyana 0.763 23.648 . . 11244.450 Comoros 0.759 21.502 . . 2996.379 Bhutan 0.719 40.840 . . 9574.029 Montenegro 0.622 21.947 4612.341 0.990 17674.620 Macao SAR, China 0.590 . . . 155201.400 Solomon Islands 0.587 12.172 . . 2551.401 Luxembourg 0.556 10.858 13914.678 0.693 108414.800 Suriname 0.553 25.999 3596.745 0.802 20211.000 Cabo Verde 0.518 31.621 . . 6281.794 Maldives 0.435 9.921 . . 17568.410 Malta 0.435 14.435 4924.544 2.059 37230.280 Brunei Darussalam 0.410 6.181 10290.938 0.921 64190.820 Bahamas, The 0.371 18.495 . . 35659.930 Belize 0.353 24.421 . . 7323.145 Iceland 0.327 6.949 53832.479 0.000 50450.740 Barbados 0.285 25.183 . . 15041.550 French Polynesia 0.272 . . . . New Caledonia 0.268 . . . . Vanuatu 0.264 11.934 . . 2984.605 São Tomé and Príncipe 0.196 27.862 . . 3720.787 Samoa 0.192 12.644 . . 5785.488 St. Lucia 0.178 24.236 . . 14288.820 30 Channel Islands 0.164 . . . . Guam 0.161 12.023 . . . Curacao 0.156 . 4797.670 . 26826.590 Kiribati 0.109 11.394 . . 1992.234 Grenada 0.109 24.287 . . 14317.330 St. Vincent and the Grenadine 0.109 23.031 . . 11845.210 Virgin Islands (U.S.) 0.108 11.166 . . . Micronesia, Fed. Sts. 0.107 11.867 . . 3340.203 Aruba 0.104 . . . 35875.630 Tonga 0.101 12.058 . . 5861.464 Antigua and Barbuda 0.093 19.726 . . 18104.410 Seychelles 0.091 20.971 . . 24848.610 Isle of Man 0.083 . . . . Andorra 0.079 10.830 . . . Dominica 0.071 21.010 . . 12216.530 Bermuda 0.065 12.806 . . 77361.100 Cayman Islands 0.061 . . . 66326.890 Marshall Islands 0.057 11.070 . . 3545.207 Greenland 0.056 12.511 . . . American Samoa 0.056 13.317 . . . Northern Mariana Islands 0.055 10.313 . . . St. Kitts and Nevis 0.051 . . . 25521.200 Faroe Islands 0.048 . . . . Sint Maarten (Dutch part) 0.038 . . . 42946.680 Monaco 0.037 . . . . Liechtenstein 0.037 . . . . St. Martin (French part) 0.036 . . . . Turks and Caicos Islands 0.035 . . . 24392.020 Gibraltar 0.034 . 5692.937 2.130 . San Marino 0.033 . . . 57328.080 British Virgin Islands 0.029 . . . . Palau 0.018 . . . 17091.230 Tuvalu 0.011 . . . 3447.796 Nauru 0.010 . . . 13175.920 Eritrea . 44.375 96.635 2.119 . 31 Figure 1 4 2 Log(CO2/Pop) 0 -2 -4 6 8 10 12 Log(GDP/Pop) Cross-National Carbon Dioxide and GDP in 2018 1 Figure 2 3 % Change in CO2/Pop 2000 to 2018 2 1 0 -1 6 8 10 12 Log(GDP/Pop) in 2000 The Growth in Carbon Emissions Versus Log(GDP) in 2000 2 Table 5 The California Vehicle Stock’s Fuel Type Distribution and Age Distribution in 2018 and 2020 All Rich Non-Rich All Rich Non-Rich October 1st 2018 January 1st 2020 Fuel Type Battery Electric 0.74 1.79 0.37 1 2.47 0.49 Diesel and Diesel Hybrid 3.98 2.81 4.34 3.92 2.76 4.27 Flex-Fuel 4.06 3.13 4.38 3.89 2.9 4.23 Gasoline 86.91 85.34 87.5 86.62 84.65 87.34 Hybrid Gasoline 3.5 5.49 2.82 3.63 5.56 2.98 Hydrogen Fuel Cell 0.02 0.04 0.01 0.02 0.05 0.01 Natural Gas 0.11 0.07 0.12 0.09 0.05 0.1 Other 0.02 0.01 0.03 0.02 0.01 0.02 Plug-in Hybrid 0.67 1.32 0.44 0.81 1.54 0.56 Model Year % Model Year Before 2010 51.81 43.1 54.74 46.01 37.33 48.89 % Model Year Between 2010 and 2015 28.88 32.3 27.74 27.71 29.98 26.97 % Model Year After 2015 19.31 24.61 17.52 26.28 32.68 24.14 Each column for Fuel Type adds up to 100% Each column for Model Year adds up to 100%. Rich represents the set of zip codes whose percentage of tax filers with a reported income of over $200,000 in the year 2018 is greater than 10.6%. This represents the 75th percentile of the cross-zip code distribution. Non-Rich represents the remaining zip codes. 3 Table 6 Within European Nation Variation in Per-Capita Carbon Emissions Nation Count Mean CV Mean CV Austria 2107 12.372 1.419 10.442 1.890 Belgium 581 10.338 1.929 12.664 2.219 Bulgaria 270 5.328 4.014 4.923 2.993 Croatia 530 7.249 2.817 6.146 2.579 Cyprus 65 42.061 6.033 4.230 7.531 Czech-Republic 6778 12.976 13.437 10.813 10.089 Denmark 100 10.817 0.950 9.323 0.846 Estonia 4097 39.905 56.237 6.842 91.908 Finland 312 22.948 1.273 14.788 1.161 France 34882 9.729 6.077 6.098 6.090 Germany 11013 12.962 9.915 10.425 5.020 Great Britain 11522 26.145 7.919 7.359 8.244 Greece 330 9.473 2.542 6.307 2.784 Hungary 3156 9.734 7.993 6.077 7.565 Iceland 15 1.627 0.981 1.078 0.849 Ireland 103 9.146 1.116 8.161 0.986 Italy 7920 9.423 4.038 6.402 3.178 Latvia 119 12.017 1.189 7.334 1.384 Liechtenstein 11 3.972 0.458 3.933 0.539 Lithuania 517 11.720 1.635 6.942 2.854 Luxembourg 105 26.180 2.455 21.154 2.234 Malta 67 3.418 3.012 3.090 3.422 Netherlands 358 12.336 1.760 12.509 1.795 Norway 356 14.136 1.261 9.859 1.702 Poland 2478 10.313 11.286 8.588 8.470 Portugal 307 8.976 3.630 6.009 3.515 Romania 664 6.412 4.367 5.459 3.841 Slovak Republic 79 6.420 1.215 6.694 1.269 Slovenia 210 9.413 2.961 8.283 3.305 Spain 8156 19.408 9.367 6.687 4.347 Sweden 294 20.042 6.610 8.582 1.436 Switzerland 2214 9.119 2.972 4.885 2.505 Turkey 974 2.412 4.102 1.263 2.069 Ukraine 5007 29.431 5.376 8.546 1.743 Population Weighted Yes No 4 Table 7 Macroeconomic Statistics for Nations in the World Values Survey Sample 2014 Data WVS nations Population pm2.5 KWH Emissions Factor GDP Per-Capita Millions lbs/KWH 2017 $ PPP Andorra 0.08 10.83 Argentina 42.67 3074.7 14.47 0.78 23550.1 Australia 23.48 10071.4 9.49 1.59 47289.96 Bangladesh 154.52 320.21 68.4 1.1 3511.65 Bolivia 10.71 742.54 24.82 0.68 7730.64 Brazil 202.76 2619.96 13.99 0.35 15749.51 Chile 17.76 3879.67 22.87 1.07 24173.06 China 1364.27 3927.04 59.77 1.63 11917.34 Colombia 46.97 1312.2 18.34 0.37 13852.24 Cyprus 1.15 3624.93 18.29 1.97 33207.77 Germany 80.98 7035.48 12.75 1.12 50770.61 Ecuador 15.95 1376.39 16.23 0.92 12078.47 Egypt, Arab Rep. 90.42 1683.21 76.56 1.07 10353.67 Ethiopia 98.09 69.2 35.04 0 1656.63 Greece 10.89 5062.61 17.37 1.49 28178.69 Guatemala 15.31 601.19 26.53 0.68 7939.37 Hong Kong SAR, China 7.23 6083.27 1.9 56359.3 Indonesia 255.13 811.91 16.45 1.63 9801.17 Iran, Islamic Rep. 77.47 3022.12 38.11 1.11 13038.55 Iraq 34.41 1328.23 56.57 0.96 9991.62 Jordan 8.92 1864.93 33.21 2.04 10284.67 Japan 127.28 7819.71 12.59 1.28 39255.28 Kazakhstan 17.29 5600.21 13.75 1.79 24355.76 Kyrgyzstan 5.84 1941.22 23.3 0.18 4722.09 Korea, Rep. 50.75 10496.51 27.21 1.22 37967.48 Lebanon 6.26 2588.86 30.44 2.11 16970.44 Macao SAR, China 0.59 155201.41 Mexico 120.36 2157.32 23.1 1 18887.57 Myanmar 52.28 215.3 40.29 0.37 4020.04 Malaysia 29.87 4651.95 16.92 1.34 23906.23 Nigeria 176.4 144.53 48.63 0.75 5516.39 Nicaragua 6.14 568.31 19.39 0.98 5443.28 New Zealand 4.52 9012.73 6.36 0.25 40432.24 Pakistan 195.31 447.51 59.52 1.03 4171.29 Peru 30.09 1345.88 27.58 0.46 11877.08 Philippines 100.51 696.35 20.26 1.32 6973.64 Puerto Rico 3.53 9.09 34070.25 5 Romania 19.91 2584.41 15.01 0.73 23084.38 Russian Federation 143.82 6602.66 16.58 0.81 26057.16 Serbia 7.13 4271.74 26.49 1.47 15226.32 Thailand 68.44 2538.8 29.36 1.13 15854.13 Tajikistan 8.25 1499.49 46.48 0.02 2884.2 Tunisia 11.06 1454.64 34.69 0.89 10498.85 Turkey 77.23 2847.22 42.6 1.12 24881.73 Ukraine 45.27 3418.57 20.49 0.92 12408.95 United States 318.39 12993.97 8.22 1.14 57213.27 Vietnam 91.71 1423.7 34.9 0.83 6098.54 Zimbabwe 13.59 609.12 22.81 0.98 3195.77 6 Table 8 World Value Survey Regressions (1) (2) (3) Protect Environment Always Vote Right Wing VARIABLES Ideology Income -0.000152 0.000768*** 0.0135*** (0.000235) (0.000213) (0.00115) Primary Education 0.0106 0.0305*** 0.00879 (0.0124) (0.0112) (0.0608) Lower Secondary 0.0225* 0.0226* -0.0359 (0.0127) (0.0115) (0.0623) Upper Secondary 0.0515*** 0.0530*** -0.0985* (0.0121) (0.0110) (0.0594) Post-Secondary 0.0850*** 0.104*** -0.192*** (0.0143) (0.0130) (0.0703) Tertiary Education 0.0868*** 0.104*** -0.167** (0.0146) (0.0132) (0.0714) Bachelors 0.133*** 0.159*** -0.289*** (0.0131) (0.0119) (0.0642) Masters 0.131*** 0.138*** -0.266*** (0.0159) (0.0144) (0.0779) Doctoral Degree 0.127*** 0.190*** -0.662*** (0.0250) (0.0228) (0.123) Right Wing Ideology -0.0105*** 0.00340*** (0.000986) (0.000896) Female 0.00680 -0.0198*** -0.0508** (0.00470) (0.00427) (0.0231) Urban -0.0221*** -0.0199*** -0.199*** (0.00581) (0.00528) (0.0285) Age 41 to 60 -0.0168*** 0.153*** 0.112*** (0.00545) (0.00495) (0.0267) Age 60+ -0.0367*** 0.241*** 0.320*** (0.00726) (0.00660) (0.0356) Constant 0.580*** 0.482*** 5.729*** (0.0142) (0.0129) (0.0638) Nation Fixed Effect Yes Yes Yes Mean of Y 0.55 0.572 5.721 Observations 42,726 42,726 42,726 R-squared 0.054 0.146 0.078 Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 The Omitted Category are Males with early childhood education who are 40 years or younger. 7 For col 1, the survey asks preferences for protecting the environment vs. economic growth. “Here are two statements people sometimes make when discussing the environment and economic growth. Which of them comes closer to your own point of view? A. Protecting the environment should be given priority, even if it causes slower economic growth and some loss of jobs B. Economic growth and creating jobs should be the top priority, even if the environment suffers to some extent 1.- A: Protect environment 2.- B: Economic growth .” For col 2, the survey assesses voting in elections -- Vote in elections: National level 1.- Always 2.- Usually is coded as “voting” For col 3, the surveys assess people’s personal ideologies - Left-right political scale: In political matters, people talk of "the left" and "the right." How would you place your views on this scale, generally speaking? 1.- Left 2.- 2 3.- 3 4.- 4 5.- 5 6.- 6 7.- 7 8.- 8 9.- 9 10.- Right 8