Working Papers

Watts and Bots: The Energy Implications of AI Adoption

With the rapid expansion of Artificial Intelligence, there are expectations for a proportional expansion of economic activity due to increased productivity, and with it energy consumption and its associated environmental consequences like carbon dioxide emissions. Here, we combine data on economic activity, with early estimates of likely adoption of AI across occupations and industries, to estimate the increase in energy use and carbon dioxide emissions at the industry level and in aggregate for the US economy. At the industry level, energy use can increase between 0 and 12 PJ per year, while emissions increase between 47 tCO2 and 272 ktCO2. Aggregating across industries in the US economy, this totals an increase in energy consumption of 28 PJ per year, or around 0.03% of energy use per year in the US. We find this translates to an increase in carbon dioxide emissions of 896 ktCO2 per year, or around 0.02% of the CO2 emissions per year in the US.

From Micro-level Weather Shocks to Macroeconomic Impacts

I develop a general equilibrium model of the macroeconomic impacts of microeconomic weather shocks accounting for production networks and non-linearities in production. In the model, weather shocks directly affect the productivity of local producers. I find a closed-form solution for the general equilibrium of the model and use it to draw insights. The first insight is that, though weather shocks are local, in the presence of networks and shared labor markets, direct effects from weather shocks can generate indirect effects that have consequences for producers throughout the economy. This suggests that empirical estimates that relate economic outcomes, such as income, to weather variations may be biased if they do not account for these spillovers. Second, I show that non-linearities in production from complementarities can generate non-linearities in the aggregation of local weather shock impacts and that this depends critically on variability in microeconomic impacts. Third, I show that the potential for labor reallocation can moderate the aggregate impacts of weather shocks, but again only if there is variability in microeconomic impacts. Using an empirical setting of 14 sectors across counties spanning the continental United States, I empirically constrain the importance of my theoretical findings. I find that, given inherent variability in weather shocks and in the response of industries to weather shocks, accounting for non-linearities in production increases the aggregate costs of weather impacts in the US economy by 33%. I find that free labor reallocation reduces these aggregate impacts by 10%.

Impact of solar geoengineering on temperature-attributable mortality (With D. Keith, W. Yang, G. Vecchi)

Temperature-attributable mortality is a major risk of climate change. Here we quantitatively analyze the capacity of solar geoengineering to reduce this risk. We model the climate response to solar geoengineering using the GFDL/FLOR model. We project temperature-attributable mortality using empirical estimates of the temperature-mortality relationship from Carleton et al. (2022). Globally, we find that solar geoengineering reduces temperature-attributable mortality. We further provide evidence that solar geoengineering may reduce temperature-attributable mortality more per degree of global cooling than emissions reductions due to differences in the climate response. These findings are robust to a variety of alternative assumptions about socioeconomics, adaptation, and solar geoengineering implementation.

Climate Damage in Convergence-Consistent Growth Projections (With S. Smulders, W. Rickels, M. Quaas, and J. Moreno-Cruz)

Climate-econometric extrapolations suggest that unmitigated warming could reduce average global incomes by over 20% towards the end of the century, significantly exceeding prior Integrated Assessment Model (IAMs) based damage projections of for example 7% as obtained with DICE (Burke et al., 2015). Here, we show that the discrepancy between the different projections can be resolved by accounting for growth convergence, an empirical approach that is consistent with the macroeconomic models underlying most IAMs. Our estimation suggests that an unmitigated warming scenario translates into global average income losses of 7%.

Foreign intervention in domestic climate policy (With J. Moreno-Cruz) (2022 CESifo Working Paper Version)

Given the global nature of the climate and the economy, climate policy that meaningfully reduces the impacts of climate change will require collaboration between sovereign nations. Characteristics of climate policy instruments, however, have made collaboration difficult. For example, emissions cuts suffer from a free-riding problem while solar geoengineering suffers from a free-driving problem. Foreign countries have at least three options to intervene in domestic climate policy in a Target country: i.) Agreement Interventions, such as international environmental agreements; ii.) Policy Interventions, such as rewards or sanctions; and iii.) Institutional Interventions, such as conflict. In this paper, we derive a theory of climate policy-motivated foreign intervention. We distill the fundamental properties of different climate policy instruments into a simple parameterization and examine the incentivizes and preferences of each foreign intervention option. Specifically, we consider intervention preferences for climate policies that suffer from free-riding or free-driving. We find that the ordering of foreign intervention preference depends critically on the parameters that determine if a policy suffers from free-riding or free-driving.

Distribution of Physical Burden of Climate Policy (With J. Moreno-Cruz and K. Ricke)

By separating the source of emissions from the source of emissions reduction, negative emissions break away from the physical limits of national mitigation efforts. With negative emissions, countries can compensate for their current and past emissions, and even the past and future emissions of other countries. Thus, the possibility of negative emissions brings into sharp relief concerns about the fair and equitable distribution of the \emph{physical} burden of climate policy that move beyond financial aid or technological transfer commitments, but directly into physical emissions reductions. To think about the distribution of physical net emissions reduction burdens, we offer a simple framework that borrows from the mitigation wedges approach first introduced by Pacala and Socolow (2004). We adapt this framework to analyze the distribution of country-specific wedges with mitigation and negative emissions technologies, asking the questions: how much does any given country contribute to stabilizing the climate? And how does this change with different burden-sharing schemes?

Work in Progress

Country-level Incentives and Economic and Non-economic International Institutions Are Required for Effective Climate Agreements (With K. Ricke)

Reducing uncertainty around climate damages and benefits is important to identifying the magnitude and distribution of policy responses to climate change, but reduction of uncertainty has an unclear impact on incentives for cooperation. Here we quantify country-level non-cooperative behavior to estimate incentives for cooperation with precise country-level estimates of the social cost of carbon (SCC) and abatement costs across varying international institutional constraints. Environmental and economic outcomes improve with improved international institutions. However, regional information can also disincentivize broad cooperation between countries. We find that information can induce partial cooperation, but typically between just a few countries. Further we find that, even with very high climate damages, a grand coalition with global cooperation is unstable from a purely economic perspective. This highlights a need for exogenous (non-economic) motivations or institutional influences to achieve socially optimal climate outcomes.

“Bright Lights, Safe Nights? The Impact of Visibility on the Spatial Organization of Crime” (with Christopher Blackburn)

This paper exploits quasi-experimental variation from a street light retrofit program to test for the impact of improved visibility on urban crime rates. Our identification strategy leverages variation in the location of retrofits, the location of crime, and the time of day a crime occurs. Our main results suggest that improved visibility from street light upgrades had a negative effect on crime. We find the retrofit program reduced annual outdoor, nighttime crime by around 30% in retrofitted areas, corresponding to a 1% reduction in total annual crime. We find that property crimes are the most responsive to changes in visibility. This outcome is pertinent to policymakers as many cities upgrade or consider upgrading their street light inventories with LED technology. Without internalizing the savings from reduced criminal activity, cities would be undervaluing LED street lighting technology. 

“Coordination through Cooperation: Climate Clubs with Solar Geoengineering” (With Juan Moreno-Cruz and Sikina Jinnah)

Cooperation among countries to achieve abatement efforts capable of quelling climate change have continually faltered under the free-riding problem, prompting researchers to explore alternative strategies. Solar geoengineering (SRM) is a quick, cheap option that counteracts increasing global temperatures driven by anthropogenic climate change and may be able to overcome the free-riding problem. However, residual damages from climate change and side-effects from SRM may have differentially adverse effects. Researchers have called for the coordinated governance of SRM to prevent the creation of a free-driver effect where countries unilaterally deploy it to the detriment of others. To explore this twist on traditional climate policy, we build on a nascent literature examining strategic, international decision-making and optimal climate policy in the presence of SRM by exploring the role of cooperation. We analyze a two-stage game under an open-membership framework where countries are required to mitigate above a pre-determined level to join a club. Only members of the club are permitted to deploy SRM. We show that a third party must introduce and enforce this club for it to be stable. In the case of symmetric countries, outcomes show promising results, but the results are not strategically robust for large participation rates. However, using an illustrative example, we show how asymmetries between countries sensitivities to SRM can drive stability for higher participation rates.