A common theme in the vast literature on climate change is the estimation of models using historical data to make predictions many decades into the future. Although there is a large and growing number of these types of studies, researchers rarely return later to check the accuracy of their predictions. In this paper, we perform such an exercise. In Davis and Gertler (2015), we used household-level microdata from Mexico to predict future air conditioning adoption as a function of income and temperature. Revisiting these predictions with 12 years of additional data, we find that air conditioning in Mexico has accelerated, significantly exceeding our predictions. Neither errors in predicting income growth or rising temperatures, nor migration patterns, nor an overly restrictive model can explain the large prediction gap. Instead, our results point to the failure to account for falling electricity prices and technological changes in air conditioner efficiency as key drivers of the prediction gap. Institutional subscribers to the NBER working paper series, and residents of developing countries may download this paper without additional charge at www.nber.org.
We study how short-term political incentives affect firms’ compliance with environmental regulation. In China, city mayors face a sharp drop in promotion probability after age 57, creating a transitory enforcement window. Using a regression discontinuity design, we find that firms governed by promotion-eligible mayors reduce sulfur dioxide (SO2) emissions substantially more than under the same mayors after the threshold. These reductions are concentrated among firms without end-of-pipe abatement equipment and are driven by temporary cuts in operating hours and output rather than technological upgrades. We further document that government subsidies increase during the promotion window and are disproportionately directed to unequipped firms that curtail production, consistent with coordinated short-run compliance. A back-of-the-envelope calculation suggests that these transitory incentives account for 5.2% of the national SO2 decline during the Eleventh Five-Year Plan, but at a cost of 0.9% of total industrial output. We develop a dynamic model of firm compliance showing that such short-lived incentives induce inefficient output-based responses. Our findings highlight the limits of episodic regulatory pressure.
We study the interplay between individuals’ voluntary contributions to environmental protection and public environmental policy in a dynamic general-equilibrium model. To capture the observed discrepancy between the willingness of individuals to contribute privately versus publicly to environmental protection, we treat the quality of the environment as a public good and assume that voluntary contributions are motivated by a ‘warm glow’ of giving. We show that in the laissez-faire scenario without government intervention, the strength of warm glow negatively affects aggregate output and positively affects environmental quality in the long run. However, when the government intervenes and decides on environmental policy in the interests of individuals, the warm-glow strength has a negative long-run effect on both variables. Thus, our results suggest that advocating for private actions to ensure a better environment in the future may be counterproductive in the presence of public environmental policies.
We study the impact of dust-driven pollution on social conflict in West Africa, exploiting exogenous variation in PM2.5 concentrations caused by Saharan dust transported by Harmattan winds. Using a monthly panel covering 16 West African countries from 2003 to 2017, we combine satellite-derived PM2.5 data with geo-referenced conflict events. We instrument PM2.5 using dust emissions from the Bodélé Depression interacted with northeasterly Harmattan winds. We find that a one standard deviation increase in PM2.5 reduces the likelihood of conflict by 0.13 percentage points. The effects are concentrated in riots, one-sided violence, ethnic and religious conflict, and conflicts involving non-local actors. Similarly, we find that resource-related conflicts increase. The findings suggest that severe dust episodes constrain mobility, visibility, and social interaction, thereby reducing opportunities for violent confrontation, even as they intensify competition over scarce resources.
We analyze the potential for electricity storage to substitute for transmission capacity. We develop a spatial-dynamic model of the grid where a utility manager dispatches generators and batteries to serve load at the lowest possible cost. The addition of storage requires the manager to optimize charging and discharging across the day, making the problem dynamic. We identify general conditions under which transmission and storage can serve as substitutes. When transmission is constrained and the battery is being charged or discharged, storage and transmission are substitutes at that instant. The investment-relevant capacity value depends on the aggregation of these instantaneous scarcity values across all hours of the day: storage substitutes for transmission investment when additional storage capacity reduces the time-averaged shadow value of the transmission constraint. This suggests storage can serve as transmission in markets where transmission is frequently constrained and the battery rarely sits idle. We calibrate our model to the electric grid in Texas and show that transmission and storage are substitutes even when wind generation is curtailed. The ability to substitute storage for transmission enables utilities to forgo costly and contentious grid expansion projects.
Brazil's recent tax reform replaces several indirect taxes with a unified Value-Added Tax (VAT), introduces additional rates on goods harmful to health or the environment, exempts several food products, and provides cashback transfers to low-income households. We evaluate the environ mental and distributional effects of this reform by combining detailed household expenditure data with product-and state-level carbon footprint estimates that incorporate all emissions sources, including land-use change. The reform, as designed, is projected to increase consumption-based greenhouse gas (GHG) emissions by 1.7% (or 49 kilograms per capita), primarily because high-emission food products become tax-exempt. At the same time, lower-income households ex perience welfare gains through lower food prices and cashback transfers, revealing a trade-off between distributional and environmental objectives. We then evaluate counterfactual scenarios in which additional tax rates are extended to carbon-intensive goods and revenues are recy cled through targeted transfers and subsidies. These scenarios reverse the increase in emissions and achieve reductions with relatively limited welfare losses, particularly when revenues subsi dize lower-emission food consumption. Our findings show that targeted environmental taxation combined with compensatory transfers can better align tax policy with climate goals without sacrificing distributional objectives.
There are tipping points in the climate system that could lock the world into a higher-temperature regime. Many tipping points are characterised by Knightian uncertainty, that is, it is difficult to assign prior probabilities to their occurrence. This paper quantifies the economic cost of this uncertainty using a min-max regret robust policy in an integrated assessment model with temperature feedback effects. By not imposing prior probabilities over critical thresholds, the framework provides policy guidance when threshold forecasting and learning are limited. I compute the implied social cost of carbon under the robust policy and compare it to the complete information benchmark. The uncertainty premium is modest at the start of the transition, between 1% and 2.2% in 2020, but rises to approximately 20% by the end of the century as robust and complete information policy paths diverge. These results imply that tipping-point ambiguity is not a first-order driver of carbon pricing in the short run, but becomes significant over policy-relevant horizons, strengthening the case for precautionary abatement.
Wind power expansion is a central environmental policy objective, yet it may entail aesthetic dis amenities for nearby residents. Turbine visibility often transcends administrative borders, whereas political and economic processes remain more localized. We exploit this spatial asymmetry to iden tify the impact of cross-jurisdictional visual exposure on support for climate mitigation policies, proxied by pro-renewable voting in Germany between 1998 and 2021. We find no electoral back lash in earlier periods, but significant declines in support in more recent elections. These dynamics are consistent with shifts in debates over wind energy expansion and the growing salience of land scape concerns, with stronger effects in rural areas and in places with organized resistance, while localized economic benefits may partly offset adverse attitudes.
Over the past three decades, the number of U.S. landfills has fallen by 41%, while average capacity has risen by 130%. This paper studies the distributional consequences of this consolidation. We show these changes coincide with waste disposal shifting toward locations with lower population density and higher shares of low-income, Black, and Hispanic residents. Event studies of landfill openings rule out large effects on migration, though new landfills are sited in places with fast-changing demographics. Using a novel measure of regulation built from large-language-model classifications of state statutes, we show evidence that increased state regulations may contribute to increasing disparities.
Energy performance disclosure is increasingly paired with minimum energy performance standards to accelerate the capitalization of energy inefficiency into housing prices and strengthen homeowners’ incentives to renovate inefficient dwellings. We study the 2021 French policy implementing a progressive rental ban on the least energy-efficient dwellings (“energy sieves”) and examine whether it has affected house prices. Using more than 900,000 housing transactions over 2014–2024, we document that energy sieves sell at an additional discount of about 5 percent following the policy announcement. Consistent with expected renovation costs driving the post-ban-announcement discount, the discount is much larger for houses that are costlier to renovate. The increase in the supply of G-rated houses on the sales market is also concentrated among those properties. While the post-ban-announcement discount reduces the gap between expected renovation costs to achieve compliance and expected sale loss due to non-compliance, we find limited evidence of a renovation response.
Whereas hydrogen is currently a marginal energy carrier, the aim of the EU is to trigger low-carbon hydrogen production that covers around 10 percent of EU's energy needs by 2050. If this comes true, it will be an energy revolution. We study competition between green and blue hydrogen using an extended Salop model that encompasses the value chains for hydrogen and CCS. For blue hydrogen, we distinguish between exports of blue hydrogen—natural gas is converted into hydrogen close to the extraction site of natural gas and hydrogen is then transported over a long distance to the hydrogen consumers—and local production of blue hydrogen—natural gas is exported over a long distance and then converted into hydrogen near the hydrogen consumers. We find that mark-ups are highest in the case of exports of blue hydrogen production, which implies that total production of hydrogen, supply of blue hydrogen, and share of plants investing in carbon capture facilities are lower in the case of exports of blue hydrogen than in the case of local production of blue hydrogen. We provide numerical simulations of the alternative market outcomes and the corresponding first-best outcomes and identify policy instruments that should be used to sustain the first-best social outcomes in the cases of exports of blue hydrogen and local production of blue hydrogen.
: This study investigates convergent validity between elicitation formats within contingent valuation methodology, focusing on theory-driven formats designed to align with incentive compatibility. While earlier literature endorsed the single binary choice (SBC) format as the only incentive-compatible option, recent studies outline theoretical conditions under which payment card (PC) and open-ended (OE) formats may also encourage truthful preference revelation. We contribute to this debate by directly comparing these three theory-driven formats and assessing their robustness against bid-vector effects and other biases such as anchoring, yea-saying, and strategic misrepresentation. We conduct a large-scale, multi-country survey with 12,343 respondents across Poland, Sweden, Hungary, Spain, Germany, and the Netherlands, examining consumers’ willingness to pay (WTP) for an EU-wide biodiversity-enhancing program. The results indicate limited convergent validity across formats, with substantial discrepancies in WTP estimates. However, analysis of choice frequencies suggests that convergent validity holds for lower bid levels, whereas higher bids introduce differences. We argue that these differences stem from both behavioral biases and methodological factors, including bid vector design and econometric assumptions. Our results underscore the critical influence of bid structure on WTP estimates and highlight the need for further research to refine elicitation formats and econometric methods used in contingent valuation studies.
While the global space economy continues to rapidly grow, and public policy challenges related to satellites and debris in orbit emerge, there is surprisingly little evidence connecting theoretical economic frameworks with observed orbital usage. Using a state-of-the-science integrated assessment model combining economic modeling with physical models of the space environment, we find the observed distribution of satellite orbits is well-explained by profit-maximizing satellite operator behavior. Furthermore, we find strong evidence that collision risk plays a key deterrent role. Finally, the observed location of the largest satellite constellation, Starlink, is consistent with a “sole owner” over particularly economically desirable altitudes.
We analyze the distributional effects of climate policy by examining heterogeneity in households' carbon intensity of consumption. We construct a novel dataset that includes information on the carbon intensity of 1.7 million individual households from 88 countries. First, we show that hori zontal differences are generally larger than vertical differences. Then, we use supervised machine learning to analyze the non-linear contribution of household characteristics to the prediction of carbon intensity of consumption. Household income, proxied by total household expenditures, is usually an insufficient predictor for the additional costs of climate policy. Including household-level information beyond household income increases the accuracy of prediction. Our results highlight that, depending on the context, some compensation policies may be more effective in reducing overall heterogeneity than others.
This paper examines the implications of different forms of corruption, i.e., grand corruption, petty corruption, and cut-money culture, on the formulation and enforcement of regulatory policies. Focusing on quota regulation in the context of natural resource extraction, it demonstrates the following. In the absence of cut-money culture, petty corruption never occurs in equilibrium, irrespective of whether the policymaker is honest or corrupt. However, when the policymaker is corrupt, the threat of petty corruption leaves no room for grand corruption unless net environmental damage due to extraction is sufficiently discounted, in which case grand corruption occurs. In contrast, the presence of cut-money culture induces corruption, either petty or both grand and petty, in equilibrium. Cut-money culture may reduce the equilibrium quota from the ‘no petty corruption enforcing quota’; however, whenever this occurs, total extraction and environmental damage rise, and welfare declines. Our results have important implications for designing corruption control mechanisms and governing natural resource extraction.
Large-scale experiments and surveys have become powerful tools for economists, measuring indi vidual economic preferences that help explain economic behavior. However, willingness-to-pay (WTP) for public goods, derived from best-practice stated preference methods, typically relies on response variations to estimate population average preferences. We develop a new WTP elicita tion method based on information theory and Bayesian updating that dynamically optimizes bids for a series of binary choices and independently elicits each participant's WTP while collecting responses. We conduct an online survey experiment with over 1400 adults aged 20 to 79 in Japan to explore the relationship between age and the value of a statistical life (VSL) in a pandemic con text at the individual level. First, Monte Carlo simulations demonstrate that our adaptive method can provide accurate individual-level estimates of WTP with fewer questions than traditional non-adaptive methods. These estimates are highly correlated with "true" values, even at the extremes of WTP. Second, the VSL estimate among adults does not decrease with age, as sometimes seen in the literature; instead, it hits its lowest point in the 30s and 40s, then increases with age, even after controlling for a comprehensive set of covariates. Third, economic preferences are strongly associated with VSL. Participants who are loss tolerant, present biased, and prosocial tend to have a higher VSL on the intensive margin. Risk-averse, impatient, and selfish participants are more likely to have zero WTP on the extensive margin.
Institutions are crucial for the allocation of natural resources. Surface water allocation in the west ern U.S. follows the prior appropriation doctrine, which distributes water based on the date when water was first appropriated for beneficial use. Under this system, during times of scarcity, water is allocated by the seniority date of water rights, with holders of more junior rights losing access to water first. Curtailments of water access can influence the profitability of irrigated production. This is critical in the arid West, where many croplands depend on irrigation. The Ricardian theory of rent suggests a premium associated with senior water rights, given that the net present value of long-term economic returns is higher due to more secure access to water. In this study, we use a hedonic pricing model to infer the economic premium of accessing surface water through senior ity of water rights. We do so by analyzing agricultural land sales as indicators of this premium in the absence of direct water rights sales data. In addition, we investigate the market valuation of storage rights and groundwater access. Specifically, we spatially link around 18,000 agricultural land sale transactions from 37 counties in Colorado with geo-coded data on their water rights. We find an average premium of access to surface water of about $6460 per acre, indicating a 64% higher value relative to comparable non-irrigated properties. Furthermore, we estimate the premium for water rights that are one year more senior in the portfolio to be about $46 per acre. Our results also indicate that properties with access to storage water rights and supplementary groundwater sell, on average, for 13% and 16% more, respectively, compared to irrigated prop erties with only surface water access. These findings highlight the importance of water security in shaping agricultural land markets and offer insights into how policy measures, such as expanding water markets or investing in storage reservoirs, can potentially improve allocation and reduce vulnerability for irrigated agriculture.
In the UK, aggregate emissions intensity has declined by about a factor of two over the last three decades. Prior research attributes most of this decline to reductions within industries rather than shifts in the composition of economic activity. This paper investigates whether such within-industry progress primarily reflects industry-specific factors or common forces operating across industries. Using a newly constructed panel of UK industry-level GHG emissions and gross value added for 1990-2022, we estimate a block-level dynamic factor model that decomposes changes in emissions intensity into global, block-level, and idiosyncratic components. We find that industry-specific factors account for the majority of variation in emissions intensity changes, though common shocks, either global or at the level of groups of industries, play a smaller but non-negligible role. We further show how patterns of co-movement partly reflect the way that emissions are recorded at the activity level and allocated to industries, a feature with implications for interpreting industry-level decarbonization dynamics.
